'

sr>\;

«f*

;* ' ' ' -. !

' I ' 8

/ 'V".

'*■>.'■

'

Special Collections

Do not cW<~

HARVARD UNIVERSITY.

LIBRARY

OF THE

MUSEUM OF COMPARATIVE ZOOLOGY.

$tbre<\.

V^A^V^^^A^

BULLETIN

OF THE

MUSEUM OF COMPARATIVE ZOOLOGY

AT

HARVARD COLLEGE, IN CAMBRIDGE.

VOL. xxxv.

CAMBRIDGE, MASS., U. S. A.
1899-1900.

University Press:
John Wilson and Son, Cambridge, U.S.A.

CONTENTS.

Page
No. 1. — Reports on the Dredging Operations of the " Albatross " in 1891.
XXVII. Preliminary Account of Planktonemertes Agassizii, a new
Pelagic Nemertean. By W. McM. Woodworth. (1 Plate.) July,
1899 1

No. 2. — Contributions from the Zoological Laboratory. XCVL' The Anat-
omt and Physiology of the Mouth-Parts of the Collembolan, Orche-
sella cincta L. By Justus Watson Folsom. (4 Plates.) July, 1899 7

No. 3. — Studies from the Newport Marine Laboratory. XLII. Longitu-
dinal Fission in Metridium marginatum Milne-Edwards. By G. H.
Parker. (3 Plates.) October, 1899 43

No. 4. — Contributions from the Zoological Laboratory. XCVIII. Ovogen-
esis in Distaplia occidentalis Bitter (MS.), with Remarks on Other
Species. By Frank W. Bancroft. (6 Plates.) October, 1899 ... 59

No. 5. — Contributions from the Zoological Laboratory. XCIX. Observa-
tions on Non-sexual Reproduction in Dero vaga. By T. W. Gallo-
way. (5 Plates.) October, 1899 115

No. 6.— Contributions from the Zoological Laboratory. C. The Photo-
biechanical Changes in the Retinal Pigment of Gammarus. By G.
H. Parker. (1 Plate.) October, 1899 143

No. 7. — Contributions from the Zoological Laboratory. No. 105. The
Structure and Development of the Antennal Glands in Homarus
americanus Milne-Ed wards. By Frederick C. Watte. (6 Plates.)
December, 1899 151

No. 8. — Contributions from the Zoological Laboratory. No. 109. Matura-
tion and Fertilization in Pulmonate Gasteropods. By Henry R.
Linville. (4 Plates.) May, 1900 213

Bulletin of the Museum of Comparative Zoology

AT HARVARD COLLEGE.

Vol. XXXV. No. 1.

REPORTS ON THE DREDGING OPERATIONS OFF THE WEST COAST OF
CENTRAL AMERICA TO THE GALAPAGOS, TO THE WEST COAST
OF MEXICO, AND IN THE GULF OF CALIFORNIA, IN CHARGE OF
ALEXANDER AGASSIZ, CARRIED ON BY THE U. S. FISH COMMIS-
SION STEAMER "ALBATROSS," DURING 1891, LIEUT. COMMANDER
Z. L. TANNER, U. S. N., COMMANDING.

XXVII.

PRELIMINARY ACCOUNT OF PLANKTONEMERTES AGASSIZII,
A NEW PELAGIC NEMERTEAN. •

By W. McM. Woodworth.

[Published by Permission of Marshall McDonald and George M. Bowers,

U. S. Fish Commissioners]

With One Plate.

CAMBRIDGE, MASS., U. S. A. :

PRINTED FOR THE MUSEUM.
July, 1899.

Bulletin of the Museum of Comparative Zoology
AT HARVARD COLLEGE.

Vol. XXXV. No. 1.

REPORTS ON THE DREDGING OPERATIONS OFF THE WEST COAST OF
CENTRAL AMERICA TO THE GALAPAGOS, TO THE WEST COAST
OF MEXICO, AND IN THE GULF OF CALIFORNIA, IN CHARGE OF
ALEXANDER AGASSIZ, CARRIED ON BY THE U. S. FISH COMMIS-
SION STEAMER "ALBATROSS," DURING 1891, LIEUT. COMMANDER
Z. L. TANNER, U. S. N., COMMANDING.

XXVII.

PRELIMINARY ACCOUNT OF PLANKTONEMERTES AGASSIZIL
A NEW PELAGIC NEMERTEAN.

By W. McM. Woodworth.

[Published by Permission of Marshall McDonald and George M. Bowers,

U. S. Fish Commissioners.]

With One Plate.

CAMBRIDGE, MASS., U. S. A. :

PRINTED FOR THE MUSEUM.
Jdly, 1899.

No. 1. — Preliminary Account of Planktonemertes Agassizii,
a new Pelayic Nemertean. By W. McM. Woodwukth.

Since the final report on the Nemerteans taken by the " Albatross "
expedition cannot be ready for publication in the near future, owing to
absences of the writer from Cambridge, the following brief preliminary
account of a new pelagic Nemertean in the Albatross collections is now
presented. No free swimming Nemertean has been described since the
appearance of Moseley's 1 accounts of the two specimens of Pelago-
nemertes taken by the " Challenger." The Challenger material was
afterward reported on in detail by Hubrecht,2 and more recently Bur-
ger,3 in his great monograph, has made two distinct species out of the
two specimens taken by the Challenger merely upon an examination of
Moseley's sketches, basing his distinction upon the differences in the
number of lateral diverticula of the intestine in the two specimens.

The form to be considered here, like Pelagonemertes, was taken in
the Pacific Ocean, and in trawls from considerable depths. While the
Challenger specimens came from latitudes well outside of the tropics,
and from the western part of the Pacific, the specimens taken by the
Albatross came from near the equator in the eastern part of the ocean.
There are many points of resemblance between 'the two forms, resem-
blances iu form, color, and even finer structure, which will not be dis-
cussed here, but a careful comparison with the detailed description by
Hubrecht shows differences in structure of such fundamental importance
that it has been necessary to establish a new genus in Moseley's family
Pelagonemertidae.

"&v

1 Moseley, H. N. On Pelagonemertes Rollestoni, Ann. Mag. Nat. Hist., Vol.
XV. p. 165, 1875. A Young Specimen of Pelagonemertes Rollestoni. Ibid., Vol.
XVI p. 377, 1875.

2 Hubrecht, A. W. W. Report on the Ne'mertea collected by H. M. S. Chal-
lenger. Challenger Report, Zoology, Vol. XX., 1887.

B Burger, O. Nemertinen. Fauna u. Flora Golfes von Neapel, Monographie
XX., 1895.

VOL. XXXV. — NO. 1. 1

2 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY.

Family PELAGONEMERTID.E Moseley.

Pelagic Nemerteans with a broad, flattened, leaf-like, gelatinous, very
hyaline body. Rhynchocoelome extending nearly the entire length of
the body. Proboscis unarmed. No cephalic grooves or organs of special
sense. Intestinal tract dendroccelous.

Genus PELAGONEMERTES Moseley.

Mouth and proboscis openings separate and distinct. Supraceso-
phageal ganglia larger than suboesophageal. Median dorsal vessel
lacking. Lateral diverticula of the intestine comparatively few in
number.

Genus PLANKTONEMERTES, nov.

A common external opening for the mouth and proboscis. Supra-
cesophageal ganglia smaller than suboesophageal. Median dorsal vessel
present. Lateral diverticula of the intestine very numerous.

Planktonemertes agassizii sp. nov.

Five specimens were taken as follows : —

1. Station 3383, Lat. 7° 21' 0" N., Long. 79° 2' 0" W., 1832 fms.
6.51 A. M., March 8, 1891. Length 47 mm., greatest breadth 13.5 mm.,
greatest thickness 3 mm., color "orange." Figure 1.

2. Station 33G1, Lat. 6° 10' 0" N., Long. 83° 6' 0" W., 1471 fms.
7.33 A. M., February 25, 1891. Length of body 24 mm., length of
everted proboscis 28 mm., greatest breadth 9 mm., greatest thickness
2.5 mm., color "orange." Figure 2.

3. Station 3388, Lat. 7° G' 0" N., Long. 79° 48' 0" W., 1168 fms.
G. 41 A. M., Marcli 9, 1891. Length 14 mm., greatest breadth 5.5 mm.,
greatest thickness 1 mm., color "orange."

4. Same station as No. 3. Length 38 mm., greatest breadth 16 mm.,
greatest thickness 1 mm., color " orange." Figure 3.

5. Station 3406, Lat. 0° 16' 0" N., Long. 90° 21' 30" W., 551 fms.
6.47 A. M., April 3, 1891. Length 37 mm., greatest breadth 16 mm.,
greatest thickness 2 mm., color "pink." Figure 4.

WOODWORTH: PLANKTONEMERTES AGASSIZII. 3

The soundings given above indicate the depth at which the dredgings
were made with an open trawl. The colors given in quotation marks are
from notes taken by Alexander Agassiz on board the Albatross, and refer
to the living animal. A water color drawing made by Mr. Agassiz of the
living animal represented in Figure 4 shows the color as light brilliant
scarlet, the intestinal diverticula and proboscis showing as bands of
deeper color. The shape of the living animal as indicated by the sketch
was like that of many of the larger marine Turbellaria, with parallel
undular sides, and bluntly rounded at both ends.

All of the specimens except No. 2 (Fig. 2), which was not sectioned,
proved to be females. In specimen No. 4 (Fig. 3) the ovaries were
slightly developed, and could be seen ouly in sections. In the other
three specimens the ovaries were prominent even before the specimens
were subjected to a clearing reagent. Both specimens of Pelagonemertes
taken by the Challenger were also females.

The chief differences between Pelagonemertes and Planktonemertes
may be summarized as follows. In Pelagonemertes there are distinct
openings for the mouth and proboscis, the former being ventral, the
latter terminal ; the dorsal cerebral ganglia are much larger than the
ventral pair; the vascular system does not include a median dorsal
vessel ; the intestinal tract is comparatively simple, i. e. with few lat-
eral diverticula, five and thirteen in the two specimens so far known.
In Planktonemertes there is a common external opening for mouth and
proboscis, slightly subterminal in position ; the dorsal pair of brain gan-
glia are much smaller than the ventral pair; the vascular system in-
cludes a dorsal median vessel, which extends in the posterior portion
of the ventral wall of the rhynchoccelom, and unites with the lateral
vessels at the posterior end (this is seen in Figure 3) ; the intestine bears
a great many (moi'e than 50) lateral diverticula, and so profuse is the
dendritic branching of these that in a transverse section the body ap-
pears to be honeycombed with cavities of varying size and outline, the
gelatinous mesenchyma being but slightly developed.

A detailed study of this interesting form has already been made, but
will be delayed in publication until the completion of the report upon
the other Nemerteans taken by the expedition.

Museum of Comparative Zoology, Cambridge, Mass.,
May 10th, 1899.

BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY.

EXPLANATION OF PLATE.

AH Figures are of Planktonemcrtes agassizii.

The Figures were all of tliem drawn from unstained specimens cleared in oil of
cloves.

All Figures magnified about 1| times.

Albatross Ex. 1891.

Planktonemertes.

1.

#>*

/

m

\ *0

y*>

4?

F ..-7

a. * -

W. McM. W. Del.

Boston Eng. Co.

The following Reports have been published or are in prepa-
ration on the Dredging Operations off the West Coast of
Central America to the Galapagos, to the West Coast of
Mexico, and in the Gulf of California, in Charge of Alex-
ander Agassiz, carried on by the U. S. Fish Commission
Steamer " Albatross," during 1891, Lieut. Commander Z. L.
Tanner, U. S. X., Commanding.

A. AGASSIZ. II.1 General Sketch of the
Expedition of the "Albatross," from
February to May, 1891.

A. AGASSIZ. The Pelagic Fauna.

A. AGASSIZ. TheDjep-Sea Panamic Fauna.

A. AGASSIZ. I.2 On Calamocrinu?, a new
Stalked Crinoid from the Galapagos.

A. AGASSIZ. XXIII.23 The Echini.

JAS. E. BENEDICT. The Annelids.

R. BERGH. XIII.13 The Nudibranchs.

K. BRANDT. The Sagittaj.

K. BRANDT. The Thalassicolje.

C. CHUN. The Siphonophores.

C. CHUN". The Eyes of Deep-Sea Crustacea.

S. F. CLARKE. XI." The Hydroids.

W. H. DALL. The Mollusks.

W. FAXON. VI.3 XV.'« The Stalk-eyed
Crustacea.

S. GARMAN. The Fishes.

W. GIESBRECHT. XVL'5 TheCopepods.

A. GOES. Ill * XX.20 The Foraminifera.

H. J. HANSEN. XXII. 22 The Cirripeds

and Isopods.

C. HARTLAUB. XVIII. '« The Comatulfe.

W. A. HERDMAN. The Ascidians.

S. J. HICKSON. The Antipathids.

W. E HOYLE. The Cephalopods.

G. vox KOCH. The Deep-Sea Corals.

C. A. KOFOID. Solenogaster.

R. VON LENDENFELD. The Phosphores-
cent Organs of Fishes.

H. LUDWIG. IV.5 XII." The Holothu-

rians.
C. F. LUTKEN and TH. MORTENSEN.

The Ophiuridae.
OTTO MA AS. XXI.21 The Acalephs.
J. P. McMURRICH. The Actinarians.
E. L. MARK. XXIV.21 The Cerianthidse.
GEO P. MERRILL. V.6 The Rocks of the

Galapagos.
G. W. MULLER. XIX.19 The Ostracods.
JOHN MURRAY. The Bottom Specimens.
A. ORTMANN. XIV. " The Pelagic Schi-

zopods.
ROBERT RIDGWAY. The Alcoholic Birds.
P. SCHIEMENZ. The Pteropods and Het-

eropods.
W. SCHIMKEWITSCH. VIII.8 The Pyc-

nogonidse.
S. H. SCUDDER. VII.' The Orthoptera of

the Galapagos.
W. PERCY SLADEN. The Starfishes.
L. STEJNEGER. The Reptiles.
TH. STUDER. X.10 The Alcyonarians.
C. H. TOWNSEND. XVII.1' The Birds of

Cocos Island.
M. P. A. TRAUTSTEDT. The Salpidse and

Doliolidse.
E. P. VAN DUZEE. The Halobatidae.
H. B. WARD. The Sipunculoids.
H. V. WILSON. The Sponges.
W. McM. WOODWORTH. IX.9 The Pla-

narians and Nemerteans.
W. McM. WOODWORTH. XXVII."

Planktonemertes.

and Vol. XXIII., No 1, February, 1892,

» Bull M. C. Z., Vol. XXI., No. 4, June, 1891, 16 pp.
89 pp., 22 Plates

2 Mem M. C. Z., Vol. XVII., Vo. 2, January, 1892, 95 pp., 32 Plates,
s Bull. M. C. Z., Vol XXIV., N>. T, August. 1893, 72 pp.
* Bull. ft. C. Z., Vol. XXIII., No. 5, December, 1892, 4 pp., 1 Plate.
5 Bull. M. C. Z., Vol. XXIV.. No 4, June, 1893, 10 pp [Zool Auzeig., No. 420, 1893.]
« Bull M. C. Z., Vol. XVI., No. 13, July, 1893, 3 pp
7 Bull. M. C. Z , Vol. XXV , No 1 September, 1S93. 25 pp.
» Bull. M. O. Z , Vol XXV., No. 2, December, 1893, 17 pp. ,2 Plates.
9 Bull. M. 0. Z., Vol XXV., No. 4, January, 1894, 4 pp., 1 Plate.
i» Bull. M C. Z., Vol XXV., No. 5, February, 1894, 17 pp.

11 Bull. M. C Z . Vol XXV No. R, February, 1894, 7 pp., 5 Plates.

12 Bull. M. C. Z., Vol. XXV No 8, September, 1894, 13 pp., 1 Plate.

13 Bull. M. C. Z , Vol XXV X . 10, October, 1894, 109 pp , 12 Plates.
« Mem. M. C. Z., Vol XVII. No. 3, October, 1894, 183 pp.. 19 Plates.

15 Bull M. C Z , Vol XXV. No. 12. April, 1895, 20 pp., 4 Plates.

16 Mem M. C. Z.. Vol XVTTI , April. 1895. 292 pp , 67 Plates, 1 Chart
» Bull. M. C Z , Vol. XXVII , No. 3, July, 1895. 8 pp.. 2 Plates.

» Bull. M. C. Z , Vol XXVII . No 4, August, 1895. 26 pp., 3 Plates.
19 Bull. M C. Z , Vol. XXVII., No. 5, October. 1895, 14 pp., 3 Plates
™ Bull M. 0 Z., Vol XXIX , No 1. March, 1896, 103 pp., 9 Plates, 1 Chart.

21 Mem. M C. Z., Vol. XXTII , No. 1, September. 1897, 92 pp., 15 Plates.

22 Bull. M. i". Z , Vol XXXI , No 5, December. 1897. 37 pp., 6 Plates, 1 Chart.

23 Bull. M. C. Z., Vol XXKIL, No. 5, May, 1898'. 18 pp., 13 Plates, 1 Chart

24 Bull. M C. 7... Vol. XXXII.. No. S, August, 1*98, 8 pp.. 3 Plates.
« Bull. M. C. Z., Vol. XXXV.. No. 1. June, 1S99. 4 pp., 1 Plate

PUBLICATIONS

OF THE

MUSEUM OF COMPARATIVE ZOOLOGY

AT HARVARD COLLEGE.

There have been published of the Bulletins Vols. I. to XXXII. ,
of the Memoirs, Vols. I. to XXII.

Vols. XXXIV. and XXXV. of the Bulletin, and Vols. XXIII.
and XXIV. of the Memoirs, are now in course of publication.

The Bulletin ami Memoirs are devoted to the publication of
original work by the Professors and Assistants of the Museum, of
investigations carried 0:1 by students and others in the different
Laboratories of Natural History, and of work by specialists based
upon the Museum Collections.

The following publications are in preparation : —

Reports 011 the Results of Dredging Operations from 1877 to 1880, in charge of
Alexander Agassiz, by the U. S. Coast Survey Steamer " Blake," Lieut.
Commander C. P. Sigshee, U. S. N.,and Commander J. R. Bartlett, U.S. N.,
Commanding.

Reports on the Results of the Expedition of 1801 of the U. S. Fish Commission
Steamer " Alhatross," Lieut. Commander Z. L. Tanner, U. S. N., Com-
manding, in charge of Alexander Agassiz.

Contributions from the Zoological Laboratory, in charge of Professor E. L.
Mark.

Contributions from the Geological Laboratory, in charge of Professor N. S.
Shaler.

Studies from the Newport Marine Laboratory, communicated by Alexander
Agassiz.

Subscriptions for the publications of the Museum will be received
on the following terms : —

For the Bulletin, $4.00 per volume, payable in advance.
For the Memoirs, $8.00 " " "

These publications are issued in numbers at irregular inter-
vals; one volume of the Bulletin (8vo) and half a volume of the
Memoirs (4to) usually appear annually. Each number of the Bul-
letin and of the Memoirs is also sold separately. A price list
of the publications of the Museum will be sent on application
to the Librarian of the Museum of Comparative Zoology, Cam-
bridge, Mass.

Bulletin of the Museum of Comparative Zoology
AT HARVARD COLLEGE.

Vol. XXXV. No. 2.

THE ANATOMY AXD PHYSIOLOGY OF THE MOUTH-PARTS
OF THE COLLEMBOLAN, ORCIIESELLA C1XCTA L.

By Justus Watson Folsom.

With Fouk Platks.

CAMBRIDGE, MASS., U.S.A.:

PRINTED FOR THE MUSEUM.
Jilt, 1899.

Bulletin of the Museum of Comparative Zoology
AT HARVARD COLLEGE.

Vol. XXXV. No. 2.

THE ANATOMY AND PHYSIOLOGY OF THE MOUTH-PARTS
OF THE COLLEMBOLAN, ORCHESELLA CINCTA L.

By Justus Watson Folsom.

With Four Plates.

CAMBRIDGE, MASS., U.S.A.:

PRINTED FOR THE MUSEUM.
July, 1899.

No. 2. — The Anatomy and Physiology of the Mouth-Parts of
the Collembolan, Orchesella cincta L.1 By Justus Watson
Folsom.

INTRODUCTION.

The group Apterygogeuea (Apterygota) of Brauer comprises two
distinct groups, termed the orders Collembola (Lubbock) and Thysanura
(Latreille). The Collembola, to which Orchesella belongs, have been
comparatively little studied by naturalists, owing in part, no doubt, to
their being small, unobtrusive, and delicate insects. It is not surprising,
in view of the small size of the Collembola, that entomologists have
described their mouth-parts imperfectly and very diversely ; on the other
hand, it is remarkable that so much good work has already been accom-
plished by dissection, unaided by the improved modern methods of
technique.

An excellent resume of preceding descriptions of the mouth-parts has
been given by Lubbock ('73), to whose account, however, I shall make
certain additions.

Fabricius (1777) briefly mentions the palpi, mandibles, maxilla?, and
bifid labium of Podura, the genus in which all Collembola were origi-
nally placed. His description, especially inapplicable as regards the
palpi, has been commented upon by Latreille and Lubbock.

Latreille ('32), who noticed the labrum, was hardly more successful
than his predecessor. His account, quoted in full by Lubbock ('73), is
characterized by the latter as being "vague as well as inaccurate."

Bourlet ('39), in a brief paragraph, also quoted by Lubbock ('73),
adds " une languette large, saillante, ciliee, a deux divisions, chaque
division quadrifide." Like Fabricius, he erroneously supposed that there
were two pairs of palpi, to which he gave names.

Nicolet ('41), in his classical memoir, gives an extended and careful
description of the most salient characteristics of the mouth-parts, recog-

1 Contributions from the Zoological Laboratory of the Museum of Comparative
Zoology at Harvard College, under the Direction of E. L. Mark, No. XCVI.
vol. xxxv. — No. 2. 1

8 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY.

nizing the anomalous condition found in Achorutes. Although a few of
his interpretations are inaccurate and his drawings rather imaginative,
Nicolet decidedly improved upon preceding authors, and his work is of
great historical importance.

Lubbock ('62, '73) enumerates and figures the mouth-parts of Smyn-
thurus, Papirius, and Tomocerus, recognizing " second maxilla;," which
" are closely attached to the ligula." His clear descriptions are quite
accurate, as far as they go, and have never been disputed. His elabo-
rate elucidation of the body-muscles of Tomocerus and Smynthurus,
wbich no subsequent worker has attempted to repeat, leads one to wish
that Lubbock had extended his patient researches to the muscles of
the head.

De Olfers ('62) describes the coarse anatomy of the mouth-parts;
his account agrees in the main with Lubbock's. De Olfers notices
a " lingua " and " organa cochleariformia," the confluent margins of the
latter forming a tube, as expressed in the following passage : " Margines
organorum cochleariformium confluentes ut supra diximus tubulum
forman t, quern oesophagum appellamus." He is mistaken, however, in
terming this tube the oesophagus.

For the sake of completenesss, I allude to the work of Laboulbene
('65) upon the general anatomy of Anurida maritima, the mouth-parts
of which differ widely from the type prevailing among Collembola.

Meinert ('65), in his noted paper upon Campodea and Japyx, gives
important generalizations upon the mouth-parts of insects.

Packard (71) offers several suggestions upon the homologies of the
mouth-parts of Collembola.

Tullberg (72), in an especially important contribution to the study
of Collembola, describes and figures the skeletal portions of the mouth-
parts of Tomocerus vulgaris in some detail, showing the general struc-
ture and more evident relations of the parts, and adding much to the
meagre accounts of other authors. This author discovered and figured
salivary glands in Tomocerus flavescens.

Sommer ('84) describes the muscles of the pharynx and oesophagus.

Oudemans ('87), in a valuable anatomical work, gives a convenient
summary of the results of preceding investigators of the anatomy of
Collembola.

Nassonow ('87) figures the salivary glands of Lipura [Aphorura]
ambulans.

Von Stummer-Traunfels ('91), in the only article devoted exclusively
to this subject, discusses the mouth-parts of Collembola and Cinura with

folsom: mouth-parts of orchesella cincta. 9

reference rather to their systematic value than to their anatomical de-
tail. His accurate description, relating mainly to skeletal structures, is
of great value.

Fernald ('90, '90a) maintained that the so called " salivary glands " of
Anurida maritima opened into what Tullberg named the linea ventralis,
a median ventral groove terminating in the ventral tube.

Willem and Sabbe ('97) accept the results of Fernald, and infer that
the well known exudation from the ventral tube is secreted by the large
cephalic glands, which had been regarded as salivary in nature.

Upon the whole, our present knowledge of the mouth-parts of Collem-
bola is fragmentary and far from complete ; it is indeed insignificant,
compared with what is known of the mouth-parts of most other orders
of insects. The physiological side of the subject is practically untouched.

Methods.

The mouth-parts of Orchesella cincta1 were studied from dissections
supplemented by serial sections. Dissections were necessarily made
with needle-knives under dissecting or compound microscopes. First,
the chitinous skeletons were prepared from alcoholic or fresh material
by separating the mouth-parts in water and treating them with potassic
hydrate. The immediate muscular attachments were then studied from
dissections in water, weak alcohol, or glycerine, permanent preparations
being made in glycerine. The normal relations of the mouth-parts were
ascertained by rendering the entire head gradually transparent by weak
potassic hydrate. At a certain stage in the process, the mouth-parts
become prominent, owing to the pink or purple color they acquire from
the solution and diffusion of the ectodermal pigment they contain.
The stain thus obtained is permanent in glycerine or balsam dissolved
in xylol.

For killing and fixing, preparatory to sectioning, the principal agents
used were either hot water, hot alcohol, hot corrosive sublimate, or picro-
sulphuric acid. Hot water or hot alcohol was as good as anything, caus-
ing a well relaxed condition ; some of my best preparations were killed in
hot alcohol of seventy per cent. Embedding was most conveniently per-

1 The species used in this research has hitherto been called Orchesella flavo-
picta Pack. ; it is, however, synonymous with the European O. cincta L., as I have
ascertained by comparing Packard's types, which are preserved in the Museum of
Comparative Zoology, with examples of the latter species given to me by Dr. C.
Sehaffer of Hamburg, by whom they were identified.

10 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY.

formed in watch-glasses. Serial sections of the head, from 3 J /a to 10 fx
thick, were cut with a Minot-Zimmermann microtome, not only in the
three principal planes, but also obliquely. All sections were affixed to the
slide with Mayer's albumen mixture, stained on the slide and preserved
in xylol balsam. The most satisfactory of many staining methods tried
was Kleinenberg's hematoxylin followed by safranin. Victoria green,
like safranin, is a good stain for chitin. For differentiating nervous
structures, good results followed the use of Vom Path's picric-osmic-
acetic mixture. For elucidating the structure of the glossa and para-
glossse, the useful device of reconstruction in wax from sections was used.

Mouth, Pharynx, and (Esophagus.

The external appearance of the mouth has already been described by
previous authors, beginning with Nicolet; indeed, the figures of Tullberg
and von Stummer-Traunfels are so accurate that my own figure (Plate 1,
Fig. 1) is unnecessary, except for completeness. The mouth in repose
is tightly closed by the labrum, labium, and palpi, the palpi fitting
snugly into apertures on either side of the labrum. When the insect is
eating, the tips of the mandibles and maxillae may be seen projecting
a little from the mouth to grasp the food, but at other times they are
concealed within the capacious pharynx. In no other order of insects
is this curious protrusion and retraction of the jaws found.

The pharynx is evaginated to form four deep pockets, two on either
side. In the dorsal and ventral pair of pockets are situated respect-
ively the mandibles and maxilla) (Fig. 2). The glossa and paraglossa?
are median in relation to the other mouth-parts (Fig. 3). The lower
wall, or floor, of the pharynx is the dorsal surface of the labium, and its
upper or anterior wall is formed by the labrum.

The oesophagus ascends abruptly from the antero-dorsal part of the
pharynx (Figs. 2 and 3), hut quickly bends directly backward to the
middle of the mesothorax, where it terminates in a valve. The oesoph-
agus is a slender tube of uniform calibre, except that the lumen of its
anterior portion gradually enlarges from behind forwards.

The anatomy of the fore-gut has already been studied, especially by
Sommer ('85), whose excellent account of Tomocerus applies also to
Orchesella in most respects. There is a distinct intima of uniform
thickness (Fig. 4, ».), which I have shown by chemical tests to be chi-
tinous, and in some individuals this is furnished with numerous small
and irregular teeth, as mentioned by Sommer. In other specimens, how-

folsom: mouth-parts of orchesella cincta. 11

ever, no such structures are observed, and, if they are not artificially
produced, their presence may possibly depend upon the interval which
has elapsed since the last moult, the intima being shed and replaced at
every moult. The epithelial wall consists of a single, well developed
layer of pigmented cells (Figs. 3 and 4, eHh.). The boundaries of the
individual cells are not well indicated, but the arrangement of their
large round or oval nuclei is au indication of their size. These cells,
being doubtless of ectodermic origin, are, like the permanent "hypoder-
mal " cells, pigmented ; the pigmentation extends back through the
whole length of the oesophagus to the stomach. The epithelium
is thrown into four or five prominent longitudinal ridges (Fig. 4).
A delicate hyaline and homogeneous basement membrane (mb. ba.)
is distinguishable, surrounding which is a single layer^ of circular
or constricting muscles (Figs. 3 and 4, cstt.). Around the anterior
portion of the oesophagus, which Somnier designates as pharynx, the
muscle fibres are quite stout, but farther backward they gradually
decrease in size, as do their nuclei also, and finally disappear near the
base of the head. As there is but one nucleus to each circular fibre,
the statement of Somraer C&o, p. 690) that •" Jeder Eing entsprache
hiernach vielleicht je einer Muskelzelle " is unnecessarily cautious. The
nuclei of the circular muscles lie, as others have observed, in a single
(median dorsal) line (Fig. 3, nl.), and each nucleus is contained in a loose
sheath which is external to the muscle proper. The condition in
Orchesella is so nearly identical with that of Macrotoma [Tomocerus]
that I can apply Sommer's ('85, p. 691) description without change to
the former species : " Die Muskeln zeigen eine zarte feinkornige Aus-
senschicht, Perimysium, in welcher in ziemlicher Anzahl kleine runde
Kerne eingelagert sind. Dieses Perimysium ist besonders machtig an
den Muskelbiindeln des Kopfes entwickelt." There are no longitudinal
muscles enclosing the constrictors. All the muscles in the head are
conspicuously striated, and, in Sommer's ('85, p. 691) words, "Die im
Korper sich findenden Muskelbundel sind von verschiedener Dicke, die
einfachsten bestehen wohl aus einer einzigen Muskelfaser, wahrend die
starkeren aus einer Anzahl von Primitivbiindeln zusammengesetzt sind,
wie man an Querschnitten leicht sehen kaun. Die Insertion der Muskel-
bundel an die Chitincuticula erfolgt nicht unmittelbar, sondern mittelst
einer Sehne."

The pharynx and oesophagus are dilated by means of thirteen pairs
of muscles (Fig. 3, dil.), situated as follows. Four short muscles
(Fig. 3, dil. phy.) originate on the clypeus in paramedian positions, and

12 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY.

immediately above the depressors of the labrum. The more dorsal pair
pass directly backward, and are inserted side by side on the anterior
wall of the pharynx (Plate 4, Fig. 29, dil. phy.). The other pair are
similar, but longer, and are inserted considerably below the preceding.
These four muscles evidently pull upon the toothed epipharyns, and
serve to withdraw it from the teeth of the paraglossae. On the dorsal
side of the anterior part of the oesophagus are seven pairs of long
slender muscles which originate on the front, and are affixed to the
chitinous intima of the oesophagus (Fig. 3, dil. ce.). The members of
each pair are widely separated in origin, but converge as they approach
the cesoj)hagus, penetrate between the circular muscles and epithelial
cells, and are inserted dorso-laterally on the intima of the oesophagus
by means of short spreading tendons. The posterior three muscles of
either side unite to form a single head. Opposed to these dorsal muscles
are four pairs on the ventral side of the oesophagus, which have a
common tendinous origin on the anterior margin of the tentorium, and
run forward under the oesophagus, to which they are affixed ventro-
laterally in the same manner as the dorsal muscles. The function of
these dorsal and ventral muscles is manifestly to enlarge the gullet ;
thus they are antagonistic to the circular muscles previously de-
scribed.

Before describing the mouth-parts, it is best to consider an endo-
skeletal structure which is intimately concerned with them.

Tentorium.

The tentorium of Collembola has never been described ; to dissect it
out is extremely difficult ; in potash preparations it is partially de-
stroyed, and it is not easy, owing to its form, to make serial sections of
it which will permit of accurate reconstruction. The failure to recog-
nize the tentorium of Collembola as being the place of origin of the
principal cephalic muscles, and homologous with the same structure in
Orthoptera and other mandibulate insects, has led students to assign an
altogether undue importance to the " Stiitzapparat " of the ligula, which
has erroneously been regarded as a sort of substitute for a tentorium.
Partly as a result of this error, systematists have acquired an exagger-
ated opinion of the differences which separate Collembola and Thysanura
from insects of other orders.

The tentorium (Figs. 5 and 6) is a chitinized structure in the middle
of the head, underlying the oesophagus, extending upward on either side
of it, and held in place by three pairs of arms diverging from the median

FOLSOM: MOUTH-PARTS OF ORCHESELLA CINCTA. 13

plane. The ventral portion of the tentorium consists of a thin frontal
plate (Figs. 5 and 7, la./.), to the anterior margin of which are attached
the ventral dilators of the pharynx (Figs. 3, 6, and 7), and under which
may be seen certain of the muscles which adduct the mandibles (Fig. 7,
add. md.). From the frontal plate diverge two anterior arms (Figs. 5
and 6, Plate 2, Fig. 10, br. a.), which pass forward and downward, and
become united with the paraglossse (Plate 3, Fig. 22, br. a.). The
anterior arms bow outwards, and serve for the origin of the protrusors of
the mandibles (Plate 2, Fig. 14, prH.). A second, or dorsal, pair of arms
(Fig. 5, br. d.) diverge from the tentorium, and extend upwards on
either side of the supra-cesophageal ganglion to the skull. Each dorsal
arm is differentiated into two parts : a short proximal projection, which
is part of the tentorium proper, and a long distal strand, less chitinized
■ than the tentorium proper and distinctly fibrous in nature. As the
strand shows no trace of cross striation and is chitinous, it can hardly be
regarded as a muscle, but may be called a ligament. The third pair of
arras project behind the tentorium (Figs. 5 and 6, br. p.). Each pos-
terior arm curves downward, as well as outward,, and consists distally of
a ligament such as just described for the dorsal arms. The ligaments
not only become continuous with the body of the tentorium, but also
are securely attached to the heels (ex.) of the chitinous legs which
support the ligula (Fig. 6).

The responses of the tentorium to stains and to potash prove it to
possess three degrees of chitinization, the ligaments being least chitinous,
the anterior arms strongly so, and the body of the tentorium inter-
mediate in this respect. In preparations rendered transparent with
potassic hydrate, whether subsequently stained with safranin or not, no
trace of the tentorium is to be seen, except the anterior arms (Plate 2,
Fig. 10, br. a.) attached to the paraglossse. When the tentorium is
intact, the union of these arms with the rest of the endoskeleton is dis-
tinctly indicated by two curving sutures (Fig. 6, sut.).

The mass of muscles originating in the tentorium is at first bewilder-
ing ; it is, nevertheless, possible to trace each muscle to its insertion, or,
better, vice versa, and to infer its function. Having done this, I find
no muscles which might protrude or retract the tentorium as Meinert
('65) claims for Japyx. This author (Meinert, '67, p. 367) says,
"The opposite ends of the flexors of the mandibles, as well as of
their tensors, in Japyx are attached to a chitinous plate situated be-
tween the mandibles, and steadied by a double set of muscles." The
author figures the muscles referred to, and describes them as steadying,

14 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY.

protruding, and retracting the chitinous plate. The homologous muscles
of Oi'chesella, however, and probably of other Collembola, I believe
serve severally, not to move the tentorium, but to dilate the oesophagus
(Fig. 6, dil. ce.), to move the antennae, and to effect certain movements
of the entire head (Fig. 6, mu.). The tentorium appears to be immov-
ably fixed in place by means of the chitinous arms and ligaments
already described. As I shall show, contrary to the views of other
authors, the protrusion and retraction of the mandibles and maxillae are
accomplished, not by corresponding movements of either the tentorium
or the " Stlltzapparat," but in both cases by special muscles.

I shall now describe the mouth-parts in the order of their position,
passing from the dorsal toward the ventral side of the head.

Labrum.

The labrum is trapezoidal in external aspect (Plate 1, Fig. 1, Ibr.) and
wedge-shaped in sagittal section (Plate 1, Fig. 3). The external surface
bears three transverse rows of stout bristles. Between the labrum and
clypeus is a deep transverse suture, formed by the infolding of the chiti-
nous cuticula, which becomes thin to form a hinge (Fig. 3). At either
end of the hinge, however, the cuticula is swollen into a conspicuous
chitinous lobe, which projects into the pharynx to fit against a cor-
responding pi-ominence of the mandible. This relation between labrum
and mandibles was expressed by de Olfers ('62, p. 12) in the following
passage, which has been overlooked by subsequent writers : " Margo
posterior [labri] incrassatus et formam literse C imitans decrsum inflexus,
qua re fit, ut duo apices in cavum oris promineant, qui mandibulas sus-
tinent." The mandibles, when at rest, are held in place by these pro-
tuberances, and the surfaces against which the mandibles are applied
show stout pai*allel ridges, which perhaps hold the mandibles effectively.
Immediately behind the distal margin of the labrum are minute teeth
projecting into the mouth in a transverse row, which becomes inter-
rupted in the middle (Plate 2, Fig. 9, de.). Between these submarginal
teeth and the margin itself is a transverse groove, in which I have found
the apex of the glossa locked by means of a corresponding transverse
ridge (Plate 1, Fig. 3).

Tullberg ('72, p. 20) discovered an epipharynx in Tomocerus vulgaris,
and briefly described it in a sentence which I translate : " The pharynx
is bounded above [anteriorly] by a palate, or epipharynx, which consists
of a chitinous membrane furnished with several toothed elevations." The
same structure also occurs in Orchesella (Plate 2, Fig. 9, e'phy.). The

FOLSOM: MOUTH-PARTS OF ORCHESELLA CIXCTA. 15

teeth of the epipharynx are directed towards those of the paraglossae,

ia conjunction with which they appear to hold the food (Plate 4,
Fig. 30).

The only muscles within the labrum are dilators of the pharynx
(already described) and depressors of the labrum. The latter consist
of a pair of muscles (Plate 1, Fig. 3, dep.), which originate on the lower
margin of the clypeus in paramedian positions and converge downward
toward the place of insertion, which is a chitinous ledge or shelf project-
ing: inward from the anterior wall of the labrum. The contraction of
these muscles doubtless closes the upper lip. I find no muscles which
could conceivably elevate the labrum in opening the mouth. This being
the case, the most satisfactory alternative which suggests itself is to
assume that the external cuticula, which is bent like the letter S at the
clypeo-labral suture (Plate 1, Fig. 3), possesses an elasticity sufficient to
raise the labrum when the depressors are relaxed.

The labrum is supplied by a pair of short nerves, which originate from
the oesophageal commissures where the latter merge into the supra-
cesophageal ganglion. The nerves soon ramify and become distributed
between the hypodermal cells of the epipharynx and other regions. The
labrum is lined with a single layer of deeply pigmented hypodermis
cells with moderately large round nuclei (Plate 1, Fig. 8, h'drm.). Near
the base of the labrum and surrounding its central lumen, or body cavity,
are grouped large oval nuclei (Fig. 8) ; each nucleus occupies the base
of a filiform cell, which may often be traced directly to the base of one
of the stout seta? (set. sns.) which cover the exterior of the upper lip.
Sommer ('85, p. 703) regarded these in Tomocerus as sensory bristles,
and the large oval nuclei as belonging to ganglion cells, although he
gave little attention to the subject : " Was die Sinnesorgane betrifft, so
muss ich mich darauf beschrauken, dass sich eigenthumlich gestaltete
Borsten, welche ich als Sinnesborsten bezeichne, an den Beinen, den
Palpen, so wie der Ober- und Unterlippe vorfinden . . . sie stehen, wie
ich das an denjenigen der Oberlippe direkt nachweisen konnte, mit Ner-
venfiiden in Verbindung, welche aus einen Haufen von Nervenzellen
hervortreten."

The so called " neiwe-cells " have no direct connection with the central
nervous system, however ; they do adjoin a network of connective tissue
(Fig. 8). I am disposed to consider the filiform cells as glandular in
function, since they probably serve to produce the chitin of the sensory
bristles. Deeper than the zone of large oval nuclei may be seen true
ganglion cells, the nuclei of which are small and round, and in every

16 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY.

respect like those in the periphery of the brain, within the mandibles
(Plate 2, Fig. 17) and elsewhere. A delicate nerve fibre from each
ganglion cell penetrates between the glandular cells and appears to
accompany the filiform processes of the latter as far as the cuticula.

Mandibles.

In general form, the chitinous skeleton of either mandible (Plate 2,
Fig. 10, md.) is a modified, elongated hollow cone, a cross section of the
least modified part being almost circular (Plate 4, Fig. 31, md.). The
specialized regions consist of an anterior or dental portion, and a posterior
portion, named by Tullberg the fulcrum, for articulation and muscular
insertions. The apex of the mandible bears several sharp, incisive teeth
on its median side (Plate 2, Figs. 10, 11), invariably five on the right
mandible and four on the left in the many cases I have observed. Behind
the apex, also on the median side, is an extensive convex molar surface
(Figs. 10, 12) composed of minute raised teeth arranged in quincunx.
This denticulated molar surface is bounded ventrally by a row of several
large rounded conical teeth (de. v.). On the posterior end of the molar
face is a single blunt tooth at right angles to the median plane of the
head. At the base of the mandible is a conspicuous triangular medio-
ventral opening (Fig. 15, of.) through which the large adductor muscles
enter. Near the anterior angle of this aperture, on the median dorsal
side is a conical projection (Fig. 10, prJj. con.), serving for the insertion
of a rotating muscle. Other muscles are inserted on the mesal face of a
dorsal and oblique basal ridge (Fig. 10, crs. ba.). The extreme base of
the mandible, a prolongation of its dorsal wall, is formed into a blunt
pivot (Fig. 13, cdx.), upon which the mandible turns. This pivot is
peculiar in that it does not form part of an ordinary articulation ; it
simply rests freely in a chitinous pocket or stirrup (Fig. 13, sta.), from
which it is withdrawn when the mandible is protruded. The chitinous
stirrup is formed from the cuticular lining of the cavity in which the
mandible lies. The end of the stirrup against which the pivot bears
when in motion is thickened, thus offering better resistance (Fig. 13, cht.).
The pocket of the stirrup is fashioned from one extremity of an elongated,
trough-shaped strap, the other end of which passes through the hypo-
dermis and is continuous with the external cuticula of the head. I at
first thought that the stirrup was free to swing forward and backward,
carrying the mandible with it, but am now convinced that it is fixed
in place and supports the mandible only when the latter is retracted.

folsom: mouth-parts of orchesella cincta. 17

Immediately behind the stirrup is a gland (to be described later) which
may, as a secondary function, lubricate the pivot of the mandible.

The mandibles are situated in finger-like evaginations of the pharynx,
and, except for muscular and nervous attachments, are unconnected
with, the pockets in which they lie, as is easily demonstrated in trans-
verse sections of the mouth-parts. As von Stummer-Traunfels ('91,
p. 220) observes, " Die mandibeln sind ganz frei in der Kopfkapsel
gelegen, mit dem Stiitzapparate nur durch jener starken Muskel ver-
bunden, und verdanken die Stellung, die sie einnehmen, nur noch dem
Zuge der Kaumuskeln und einem Chitinvorsprunge an der Innenseite
der Kopfkapsel, auf dem sie mit ihren hinteren Enden pivotiren und
der diesem entgegenwirkt." It is evident that this arrangement facili-
tates the protrusion of the mandibles, which lie obliquely in the head
(Plate 1, Fig. 2, Plate 2, Fig. 10), their bases close beside the skull
on either side, while their apices converge, so that the opposing incisive
teeth and molar surfaces meet in the sagittal plane. On account of the
oblique position of the mandibles (Fig. 2), which in this agree with the
maxilla? and tongue, it happens that microtome sections frontal or trans-
verse in relation to the oesophagus are oblique in relation to these organs,
and vice versa. Throughout this paper, when I refer to frontal and
transverse sections, I shall use those terms with reference to the internal
mouth-parts, conceiving the axial line between the mouth-parts to be
their long axis, unless otherwise specified.

From the abundance of muscles in the head I have studied out the
surprising number of ten distinct pairs which are concerned in moving
the mandibles alone. This has been done by the study of serial sections
in different directions and of dissections. The diverse directions taken
by the muscles render them difficult to follow on sections in any single
direction. On the whole, however, most may be learned from sections
which are frontal, i. e. parallel with the plane in which the mandibles
lie, and I shall describe the muscles as studied in successive frontal
planes, beginning on the dorsal side. The order in which the muscles
are numbered is necessarily somewhat arbitrary, but is chosen as being
that in which they may with least difficulty be identified by any one who
may wish to study the subject hereafter.

1. Lateral Rotator. This muscle (Plate 2, Fig. 14, 1. rot. 1.) arises on
the skull at the side of the head, passes forward and downward, crossing
obliquely the dorsal surface of the mandible, and is inserted on the
conical, medio-dorsal projection of the mandible. The same muscle is
also represented as it appears in a transverse section of the mouth-parts

18 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY.

in Figure 16, 1. rot. I. Its function evidently is to rotate the mandible
in such a way as to raise the molar surface. The opposing muscle is
designated as 3. ret. rot.

2. Abductor. An extensive muscle (Fig. 14, 2. abd.), which arises at
the side of and anterior to the lateral rotator, passes obliquely forward
and downward, and is inserted obliquely on the lateral surface of the
mandible, which is excavated to receive the belly of the muscle. This
is the only muscle which can act in opposition to the adductor (9. add.),
and may also assist in the process of retraction.

3. Retractor and rotator. A slender muscle (Fig. 14, 3. ret. rot.),
which originates near the base of the skull in a dorso-lateral position,
passes forward and downward, and is attached to the inner side of the
basal ridge of the mandible described above.

4. Retractor. Although distinct from the preceding retractor, it
follows nearly the same course (Fig. 14, J^. ret.), but is more ventral ; it
originates nearer the median plane, and is inserted on the basal ridge
immediately behind its companion. Both these muscles are adapted to
withdraw the mandible into its socket after it has been protruded by
muscles 5 and G ; No. 3, however, appears in addition to be the only
muscle which is capable of rotating the mandible in opposition to Nos.
1, 7, 8, and 10, that is, so as to lower the molar surface.

5. Lateral Protrusor. The contraction of the outer of two slender
cylindrical muscles (Fig. 14, 5. pr't. I.) which originate on the anterior
arm of the tentorium results in protruding the mandibles. It passes
upward and backward from the tentorium along the mesal surface of
the mandible, and is inserted immediately under the insertion of No. 3.
Its function is doubtless to protrude the mandible by pulling its base
forward.

G. Mesal Protrusor. This muscle accompanies No. 5 (Fig. 14, 6. pr't.
ms.), at the side of which it originates, but its course is more ventral,
and its insertion is on the basal ridge just behind that of No. 5. The
last two muscles are distinct, but have the same function, that of pro-
truding the mandibles in opposition to muscles 3 and 4.

7. Rotator. A long stout muscle (Fig. 14, 7. rot.), which begins on
the side of a median dorsal chitinous projection near the base of the
head. The muscle runs forward, outward, and downward, and terminates
in a tapering pigmented tendon, which crosses under the base of the
mandible (Fig. 15, 7. rot.) and is inserted in the outer angle of the
large triangular opening.

8. Rotator. A long, powerful muscle, which originates near the base

folsom: mouth-parts of orchesella cincta. 19

of the skull, being behind the preceding muscle and crossing the median
plane (Fig. 14, 8. rot.). It also ends (Fig. 15, 8. rot.) in a tapering pig-
mented tendon, which is inserted close in front of the tendon of the last
described muscle. Both muscles must act as rotators, twisting the man-
dible so that its molar surface moves upward and outward.

9. Adductor. This muscle, the only one hitherto mentioned by writ-
ers, is the most powerful muscle in the head. It originates principally
on the tentorium (Fig. 14, 9. add.), passes directly outward, penetrates
the large triangular orifice of the base of the fulcrum (Fig. 15, 9. add.),
and is inserted on the inside of the lateral wall of the mandible. In
addition, several of its fibres pass under the tentorium (Plate 1, Fig. 7,
9. add. md.) and become continuous with similar fibres from the oppo-
site mandible. Thus, these fibres must pull against each other with the
effect of closing the jaws, but with this exception the adductors are at-
tached to the tentorium. These strong muscles are counteracted by
muscle No. 2.

10. Rotator. A long, slender muscle, beginning at the median dorsal
line, passing forward, outward, and downward, and inserted by a pig-
mented tendon along with Nos. 7 and 8. To avoid confusion I have
omitted this muscle from Figure 14, but the figures of No. 7 (Figs. 14
and 15, 7. rot.) will serve perfectly well for No. 10, if it be remembered
that the latter muscle lies under the former. It will be observed that
I have described as many as four muscles (Nos. 1, 7, 8, and 10) which
appear to rotate the mandible in the same direction ; I see no other
function for these muscles, however. The rotation in the opposite direc-
tion seems to be comparatively unimportant, being accomplished by a
single slender muscle (No. 3), the primary function of which is perhaps
retraction. Among the mass of antennal muscles originating on the
tentorium there is one which might easily be mistaken for a rotator of
the mandible, in function similar to No. 3. This muscle (Fig. 16, mu. at)
in some sections is moulded against the mesal surface of the conical
projection to which is inserted rotator No. 1. Iu other preparations,
however, in which the mandible happens to have been rotated so as to
remove the projection from its proximity to the antennal muscles, it
may be seen that the apparent rotator is really unattached to the
chitinous projection.

The nerves to the mandibles are the first pair of the infra-oesophageal
ganglion ; they arise from either side of the anterior part of the ganglion,
pass directly outwards, enter the mandible at the anterior angle of its
large lumen and extend the length of the mandible and into fine canals

20 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY.

in the chitinized apex. Each nerve fibre, soon after penetrating the
mandible, becomes enlarged to contain a small round nucleus (Figs. 16
and 17, cl.gn.). The mandible is lined with a thick layer of deeply
pigmented hypodermis (tidrm.), the confluent cells of which contain
large oval nuclei (shown in cross section in Fig. 16), strongly contrast-
ing in size with the nerve nuclei of the core of the mandible.

Maxilla.

The maxilla? (Plate 3, Fig. 18) are about as long as the mandibles,
and composed of two specialized regions : first, an anterior terminal
movable lobe (Fig. 18, cpt.), which is subdivided into several smaller
lobes and teeth ; secondly, a posterior framework bearing muscles and
supporting the terminal lobe and the palpus. The dorsal and outer
portion of the terminal lobe is wholly chitinous and bears three stout,
incurving claws (Fig. 19, get.). This portion de Olfers ('62) compares
with the galea of Orthoptera, and Packard ('71, p. 100), referring to
the mouth-parts of Tomocerus plumbeus, writes, "The middle lobe, or
galea, is nearly obsolete, though I think I have seen it in Smynthurus,
where it forms a lobe on the outside of the lacinia." Von Stummer-
Traunfels ('91, pp. 221, 223), however, says, " Es ist mir darum sehr
zweifelhaft, ob dieser Theil des Kieferapparates die Deutung als aussere
Lade wirklich verdient. . . . Bei Japyx noch zweifach gegliedert, ist
er bei Campodea schon mehr reduch't und fehlt bei den Collembolen
ganzlich." The careful comparative studies of the last mentioned author
give much weight to his opinion. Underlying the tridentate lobe, or
so called galea, are four chitinous lobes, or lamellae, each of which bears
on its inner margin a comb of fine teeth. Three of these lobes are
falcate in form, and the fourth or inmost lobe (Fig. 19, Icn.) bears a
prominent hook on its upper surface. These four fringed lobes probably
represent the lacinia, or inner lobe, of other insects. The seven lobes
and claws described appear to be firmly united basally, and, if so, cannot
move separately, but must all move together by means of the articula-
tion (Fig. 19, ate.) at the apex of the stipes. The movement is lateral
only, and the adduction is accomplished by muscles which terminate in
a slender chitinous rod (Figs. 18, 19, and 20, bac.) having the function
of a tendon, and so named by von Stummer-Trauufels. T^iis tendon is
attached to the base of the inmost lobe of the lacinia.

Considering now the framework of the maxilla, the stipes (Figs. 18-
20, stp.) is a stout chitinous structure, the form of which I may roughly

FOLSOM : MOUTH-PARTS OF ORCHESELLA CINCTA. 21

compare to a long, shallow boat, pointed at both ends and somewhat
crescentic in transverse section (Plate 4, Figs. 29-32, stp.). The two
thickened margins of the stipes are not parallel, however ; the anterior
portion of the dorsal margin is twisted toward the median plane of the
head, as shown in Fig. 18. The ventral margin, moreover, is sharply
incurved, as may be seen in transverse sections of the stipes (Plate 4,
Figs. 31, 32). The anterior end of the stipes is rounded where it artic-
ulates (Fig. 19, ate.) with the movable head (cpt.) of the maxilla. The
stipes is strengthened by a somewhat oblique cross rib connecting the
two margins ; the rib is prolonged beyond the dorsal margin as a free
projection (Fig. 18, prj.), the function of which I am unable to state;
I find no attachment of muscles or other structures upon it. The base
of the stipes is immovably fixed to the cardo, the junction being indi-
cated by a distinct suture (Fig. 18, sut.). The cardo (Fig. 18, car.) is
shaped like a shoe, the toe of which is attenuated to form a chitinous
ligament (Figs. 18 and 20, lig.'), which is continuous with a ligament
from the foot of the glossa, a suture showing that the now single liga-
ment originated from the union of two. By means of this peculiar
articulation — already noticed by de Olfers, Tullberg, and von Stummer-
Traunfels — and the ligament from the glossa to be next described are
permitted the movements of the maxilla as a whole. The toe of the
glossa, namely, is extended into a long flexible chitinous ligament (Fig.
10, lig.), which is fastened to the outside of the base of the stipes.
The length of this outer ligament evidently determines the extent to
which the maxilla may be protruded, the supporting stalks of the glossa
being stationary.

The rod of chitin previously mentioned as assisting in the adduction
of the head of the maxilla is crescentic in cross section (Plate 4, Figs.
29 and 30, bac). Articulating with the base of the rod is a chitinous
expansion (Fig. 18, exp.) for the insertion of four muscles. This ex-
pansion is nearly A-shaped in cross section (Plate 4, Fig. 31, exp.),
there being a dorsal longitudinal ridge with a sloping wing on either
side. In the angle between the ridge and the lateral wing is inserted
muscle No. 7 of the maxilla. The opposite or mesal wing is prolonged
backward in line with the rod (Figs. 18 and 20), and serves for the in-
sertion of muscles Nos. 2 and 3. The base of the rod itself is prolonged
into a short ligament (Fig. 18, lig.), by means of which the rod is con-
nected with the adjacent corner of the base of the paraglossa (Fig. 18,
pa'gls.). In order to show the ligament in Figure 18, I have represented
the maxilla as withdrawn from the paraglossa. Normally, however, the

22 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY.

rod (bac.) is situated in a lateral concavity of the ligula (Plate 4, Figs.
29, 30, gls.), and consequently underlies the lateral margin of the para-
glossa. The significance of the chitinized rod, or tendon, and its liga-
ment, I shall presently show when describing how the head of the
maxilla is abducted.

There are ten separate muscles which belong to each maxilla, exclud-
ing those of the palpi ; like the muscles of the mandibles, they are most
conveniently studied in frontal sections, although these must be supple-
mented by sections in other planes, as well as by dissections. The max-
ilke are more complicated than the mandibles, and are correspondingly
more difficult to understand ; after long study, however, I have but
little doubt as to their structure and movements.

1. Retractor and abductor. A long, slender muscle (Fig. 20, 1. ret.
aid.), which arises immediately beneath the origin of mandibular mus-
cle No. 7 (Fig. 14) on the same median dorsal projection, and passes
forward, outward, and downward to be inserted on the anterior con-
cavity of the cardo (cf. Fig. 18). This muscle appears to retract and
slightly abduct the entire maxilla by pulling the cardo backward. The
retraction must be of slight amount, however, as the two ligaments
which attach the cardo to the foot of the glossa prevent any extensive
retraction or protrusion.

2 and 3. Adductors. Two cylindrical muscles, distinct from each
other but lying side by side (Fig. 20, 2. 3. add., Plate 4, Fig. 32), which
arise on the dorso-lateral part of the skull, pass forward and downward
under the adductors of the mandible, and by means of several slender
tendons fuse with the posterior elongation of the chitinous expansion
just described (Fig. 20, exp.). I believe their function is, in co-operation
with muscles ISos. 4 and 7, to adduct the head of the maxilla. In Figure
20, muscles 2 and 3 are represented as interrupted, in order to show cer-
tain underlying muscles.

4. Adductor. A stout muscle (Fig. 20, 4- acid.), which arises on the
most anterior surface of the cardo, passes forward and is inserted on the
ventral surface of the chitinous expansion.

5. Protrusor and adductor. A slender muscle (Fig. 20, 5. pr't. add.),
which begins on the side of the tentorium, goes outward and somewhat
backward, and is attached to the anterior concavity of the cardo, just
beneath the insertion of muscle No. 1 .

6. Prot rusor and adductor. Similar to the last in origin and direction,
but more ventral and anterior in position (Fig. 20, 6. prt. add.) and in-
serted on the most anterior surface of the cardo, just under the insertion

folsom: mouth-parts of orchesella cincta. 23

of No. 4. Muscles Nos. 5 and 6 both appear to have the same function,
i. e. to protrude and adduct the entire maxilla by pulling upon the cardo.
Protrusion may evidently take place until the outer ligament (Plate 2,
Fig. 10, lig.) has become tense, whereupon additional contraction would
pi-oduce adduction of the maxilla.

7. Adductor. A stout mass of muscle fibres within the maxilla
(Fig. 20, 7. add.) arising along the base of the stipes and inserted ou
the lateral aspect of the chitinous expansion (Plate 4, Fig. 31, 7. add.).
This large muscle passes under the free dorsal projection of the stipes, to
which it does not appear to be fixed. The four muscles numbered 2, 3,
4, and 7, which are all attached to the chitinous expansion, which in turn
is articulated with the chitinous tendon (bac.) from the head of the
maxilla, probably adduct the head of the maxilla by retracting the rod,
so that the claws of the head are rotated in a frontal plane toward the
claws of the opposite maxilla, in order to meet them and grasp food ;
this function of the maxillae seems to be the most important one, judging
from the number and size of the muscles which close the claws.

8. Retractor and adductor. A short muscle on the ventral side of the
maxilla (Fig. 21,6". ret. add.), its course also shown faintly in Figure 20,
arising on the stalk of the glossa, passing obliquely forward and outward,
and inserted on the inflexed lower margin of the stipes. This muscle
must retract the stipes and draw it toward the median plane.

9. Protrusor and adductor. A broad, oblique muscle (Fig. 21, 9. pr't.
add.) beneath and running at right angles to the last, also arising on the
stalk of the glossa and inserted on the upturned border of the stipes.
By this muscle the stipes is protruded and also drawn towards the
glossa.

10. Adductor. A short, stout muscle (Fig. 21, 10. add., Plate 4,
Fig. 31) passing directly outward from the stalk of the glossa to the
inflexed margin of the stipes. The function of this muscle is clearly
to pull the stipes towards the glossa, i. e. to adduct it.

I believe that the three muscles last described have the important
function of abducting the head of the maxilla, or separating the claws of
the maxillae preparatory to grasping food. The head of the maxilla is
adducted by muscles numbered 2, 3, 4, and 7, which, by retracting the
rod, cause the head to rotate upon the end of the stipes. The reverse
movement, the opening, appears to be accomplished by the successive
retraction and adduction of the stipes by means of muscles Nos. 8, 1,
and 10, during which the head rotates upon the end of the chitinous rod
until in the position shown in Figures 18 and 19. In this condition the

24 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY.

rod has evidently been pulled back, as far as its ligament (Fig. 18 lig.)
allowed, by the action of the adductors of the head, and must manifestly
be thrust forward before it can again be withdrawn. This protrusion of
the rod appears to be a necessary result of the simple protrusion of the
stipes, by muscles Nos. 9, 5, and G ; the advancing stipes pushes forward
both the head and the attached chitinous rod until the ligament of the
latter has become tense in the opposite direction.

The above explanation of the movements of the terminal portion of
the maxilla is the only one I can offer, after long study. It appears
reasonable to me, and accounts for the presence of the peculiar chitinized
tendon or rod of the maxilla, as well as the unique ligamentous connec-
tion between the maxilla and paraglossa. I shall again refer to this
connection when describing the palpi.

The maxillae are supplied by the second pair of nerves from the infra-
cesophageal ganglion ; each nerve begins at the side of the ganglion, a
little behind the mandibular nerve, goes directly outwards for a short
distance and enters the maxilla at the posterior end of the chitinous ex-
pansion ; a branch is soon given off to the palpus. The anterior portion
of the maxilla is occupied by a core (Fig. 1 9, gl. et n.) of filamentous cells
which penetrate into the lobes of the laciuia through an orifice in the
base of the head of the maxilla; these filamentous cells are of two kinds
and precisely similar to those already described for the labrum, the base
of certain cells containing a large oval nucleus (Fig. 19), while inter-
vening cells are ganglionic, with small round nuclei. The maxilla is
lined with a single layer of confluent hypodermis cells, deeply pigmented
and containing round nuclei of moderate size.

Palpi.

The palpi, consisting of but one pair, are of special interest because von
Stummer-Trauufels considered them quite anomalous in position, having
described and figured them as separated from the maxilhe and joined to
the paraglossse. This author ('91, p. 226) emphasizes, "Die grosse
Unwahrscheinlichkeit, dass der sogenannte Maxillartaster der Collem-
bolen wirklich zur Maxille gehort, indem diese von jeuem vollstandig
getrennt ist und derselbe vielmehr in innigem Verbande mit der Para-
glossen steht." This view originated, however, with Tullberg, as his
description and figure show. I believe that this view is erroneous, and
that the palpi unquestionably belong to the maxilla?; I shall show how
the mistake mentioned might easily be made.

Each palpus (Fig. 18, pip.) is finger-shaped and composed of but a

folsom: mouth-pakts of orchesella cincta. 25

single segment. The extremity is provided with five bristles, each seated
upon a tubercle, the proximal bristle and its tubercle being much the
largest. The palpus lies dorsal to its maxilla and its base is attached to
the chitinous expansion of the maxilla, as represented in Figure 18. It
will be noticed that the palpus joins the expansion close to the ligament
which unites the chitinous rod (bac.) and the paraglossa (Fig. 18, pa'gls.),
so that the ligament might readily be mistaken for a prolongation of the
palpus. Careful study shows, however, that the ligament is directly con-
tinuous with the rod, and not with the palpus. If the palpus is con-
nected with the paraglossa at all, the attachment is only of the most
incidental nature.

The palpus contains at least two longitudinal muscles, which arise from
the chitinous supporting structure at its base and extend to its free ex-
tremity. The palpus is lined with confluent, deeply pigmented hypo-
dermis cells ; beneath the setse are filiform cells, each with an enlarged
base which contains a large oval nucleus ; a condition also found in the
labrum and labium.

Glossa and Paraglossa.

These structures have been briefly mentioned and differently named
by several authors. By de Olfers ('62) they were called respectively
lingua and organa cochleariformia ; by Lubbock ('62), ligula and second
maxillae ; by Meinert ('65), lingua and paraglossce ; by Tullberg ('72),
lamina hypopharyngis inferior and laminae hypopharyngis superiores ;
finally, by Grassi, Oudemans, and von Stummer-Traunfels, ligula and
paraglossce. Careful comparison has led me to believe that the ligula
and paraglossa? of Thysanura are the equivalents, respectively, of the
glossa and paraglossa? of other mandibulate insects; the Thysanura,
however, exhibit the more primitive or generalized condition of tongue,
which is not consolidated with the labium. The term ligula, at present,
is properly applied to the glossa and paraglossa? taken together, rather
than to the glossa alone.

The paraglossce (Plate 3, Fig. 22, pa'gls.) are two membranous,
transparent, chitinous appendages attached to the dorsal surface of the
glossa. The basal half of either appendage is firmly united to the
glossa, and therefore can have no power of independent movement.
The apical half is a free lobe, oval in cross section (Plate 4, Fig. 28,
pa'gls.). A paraglossa viewed from above presents somewhat the form of
an isosceles triangle with a rounded apex and nearly equalling the glossa
in length. The lateral surfaces are strengthened by being thicker and

26 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY.

more rigid (Fig. 22, cht.). The inner margins of the two appendages
approximate each other, leaving an elliptical aperture however (Fig.
22, lu.), beneath which may be seen a single median row of fine teeth
belonging to the dorsal surface of the glossa (Fig. 22). The inner edge
of either paraglossa is furnished with teeth along its posterior half.
Considering the teeth successively, the most anterior are blunt and
lie in a frontal plane ; as we pass backward the teeth are not only
longer and more slender, but, as they approach the median plane, also
become gradually erect, so that at length the posterior teeth project in
a parasagittal plane. This change in direction is brought about by a
curvature of the edge of the paraglossa, the dorsal surface of which, in
passing backward, gradually becomes more concave, causing the inner
margin to curve upward, or forward, as the parts naturally lie. The
concavity of the paraglossa matches the convexity of the molar surface
of the mandible on the same side of the head, and the sagittal teeth of
the paraglossse normally intervene between the grinding faces of the
mandibles (Plate 4, Fig. 30, ?nd.).

I find no well marked bundle of nerve fibres for the glossa and para-
glossa;, but many separate fibres, each from a ganglion cell, are given off
directly from the infra-oesophageal ganglion and penetrate between the
hypodermal cells of the tongue, which are for the most part attenuated
like the filamentous chitin-forming cells of the labrum.

The glossa (Plate 3, Fig. 23), situated under the paraglossse which it
bears, is an elongated, unpaired, chitinous organ, the median furrows of
which however probably indicate its derivation from a paired condition.
Three regions may be distinguished : a terminal, free portion (Plate 1,
Fig. 3, gls.), an intermediate region with which the paraglossa? are
fused, and a basal part, consisting of a pair of supporting stalks (pd.,
Plate 2, Fig. 10, Plate 3, Figs. 22, 23). The free portion is oval in
cross section (Plate 4, Fig. 28, gls.) ; viewed from above (Plate 3, Fig.
23) it terminates in front in an oval transparent lobe, across the end
of which is a subterminal dorsal fold or ridge (Fig. 23, pli.), which
interlocks with the labrum (Plate 1, Fig. 3). The upper surface of the
terminal lobe is provided on either side with a curving row of minute
teeth (Fig. 23, de.) borne upon a thickened chitinous ridge. A median
dorsal groove is present, in the course of which occurs what appears to
be an opening (Fig. 23, of.) into the interior of the glossa. The inter-
mediate region bears the paraglossae, the basal halves of which merge
with the glossa to form a single body ; the cavity of either paraglossa
also becomes confluent with that of the glossa (Plate 4, Fig. 29). The

folsom: mouth-parts of orchesella cincta. 27

lateral surfaces of the glossa are strongly concave (Plate 4, Fig. 29), and
receive the maxillae, which lie adjacent to it. The maxillae normally
approach each other in the space left free between the upper surface of
the end of the glossa and the under surfaces of the paraglossae (Plate 1,
Fig. 3). The glossa is prolonged behind into a pair of slender, diverg-
ing, chitinous stalks or legs, by von Stummer-Traunfels termed the
"Stiitzgerust " (Figs. 10, 22, 23, pdJ). The enlarged base of each bears
some resemblance to a human foot (pd.). The toe of the foot, which
underlies the cardo of the maxilla, is attenuated into a long ligament
which bends around the base of the cardo and attaches itself upon the
outside of the base of the stipes (Fig. 10, Eg.). The dorsal edge of the
stalk is extended and modified to form a second ligament (%.'), which,
as previously stated, unites with a ligament from the toe of the cardo, a
suture remaining to indicate where the two ligaments united.

The glossa is lined with a layer of well developed pigmented epi-
thelium, containing large oval nuclei (Plate 1, Fig. 3). A central
tubular cavity is left, however, which is a branch of the general body
cavity. The possible glandular nature of the tongue will be discussed
later.

Labium.

The labium, or lower lip (Plate 3, Fig. 24), is a single plate, formed
by the union of two lateral plates, as indicated by a conspicuous median
suture. Either half of the labium is divided by sutures into three dis-
tinct regions. The anterior region (j'^p.) bears five prominent tubercles,
from the apex and sides of which project several subulate bristles. The
intermediate region (?nen.) is of lunate form, thickly chitinized along the
mesal margin, hinged along the posterior curving suture, and beai'S a
few setae. The posterior and most extensive region (sb'men.) is almost
divided into two sclerites by an oblique suture, which however becomes
obsolete at its posterior end. This third region is laterally fixed to the
clypeus, is posteriorly distinguished from the gula by a deep oblique
suture, and bears a number of setae, many of which are barbellate.
The halves of the labium are firmly fused together, except anteriorly,
and are not easily separable ; but the fusion involves only the ventral
margins of the mesal faces of both plates, in such a way that a trough
or gutter is left within the buccal cavity (Fig. 25, and Plate 4, Figs. 28,
29, sid.). At the posterior end of the median longitudinal suture begins
the linea ventralis of Tullberg (Fig. 24, In. v.), an external median
ventral gutter formed by two thin longitudinal ridges of chitinous cu-

28 bulletin: museum of comparative zoology.

ticula, the margins of which approach each other without uniting, as is
shown by transverse sections (Figs. 26 and 27). This linea ventralis
may be traced hack along the median ventral line of the body as far as
the lobes of the ventral tube, which is a median appendage of the first
abdominal segment ; between the body segments, however, the liuea
ventralis suffers considerable interruption. I shall return to this
peculiar structure when considering the glands of the head.

The morphology of the labium of Collembola has never been eluci-
dated, and is difficult to understand, even after careful comparison with
other orders of insects. I am at present able to offer only a few sugges-
tions upon the subject. The glossa and paraglossse have no connection
with the labium, although fused with each other. Labial palpi have
been regarded as absent by other authors, yet the anterior regions of
the labium, which are doubtless tactile in function, may perhaps be
palpi in a morphological sense also. If a maxillary palpus of this same
insect with its setigerous tubercles be imagined as having become sessile,
through a shortening of the stalk, we have a counterpart of the terminal
lobe of the labium. I see no serious objection to considering the
two remaining sclerites of either side as mentum and submentum,
although the presence of a nearly obsolete suture, tending to divide
each submentum into two sclerites, certainly complicates the matter.

The movements of the labium are effected by five pairs of muscles, of
which three are elevators and two depressors. I shall describe these
muscles as they appear in successive sagittal sections, beginning at the
median plane.

1. Posterior depressor. This is the shortest of the labial muscles, and
is inserted on the ventral wall close to the median plane (Plate 1, Fig. 3,
and Plate 3, Fig. 25, 1. dep.p.). This muscle arises from the dorsal sur-
face of the thick chitinous wall of the salivary duct and runs forward,
downward, and inward to its insertion. It opens the mouth by pulling
upon the under surface of the lower lip. All the lateral muscles are seen
tranversely sectioned in Plate 4, Figures 28-31.

2. Mesal elevator. This muscle runs along the upper and inner margin
of each half of the labium (Figs. 3 and 25, 2. Ivt.ms.). The insertion is
at the anterior extremity, and the origin close beside that of the preced-
ing muscle, on the same chitinous duct. The function is to close the
mouth by diminishing the extent of the inner surface of the labium.

3. Anterior depressor (Fig. 3, 3. dep. a.). In origin, course, and
function this muscle is similar to the posterior depressor; it is more
lateral, however, and is inserted somewhat in advance of its companion.

FOLSOM : MOUTH-PARTS OF ORCHESELLA CINCTA. 29

4. Middle elevator. This muscle (Figs. 3 and 25, 4- Ivt. m.), which is
longer than any of the preceding, arises from the posterior arm of the
tentorium, bends downward, follows the inner wall of the labium directly
forward, aud terminates at the distal extremity of the labium. The func-
tion of the muscle is like that of the mesal elevator.

5. Lateral elevator. This is the longest and most lateral of the labial
muscles (Fig. 25, 5. Ivt. I.). As to insertion and course it is similar to
the middle elevator, but its origin is more posterior and dorsal, being
upon the dorsal chitinous surface of the salivary duct. The principal
dillei'ence between the three elevators described is that, by reason of
their different insertions, they may raise different portions of each
terminal movable lobe of the labium.

The labium is supplied by the third pair of nerves from, the infra-
oesophageal ganglion ; the nerves originate a little to one side of the
median plane from the ventral part of the ganglion, and pass forward,
downward, and outward into the dorsal portion of either half of the
lower lip. The labium is lined with a single layer of confluent, pig-
mented, hypodermal cells, with moderately large, round nuclei. Be-
neath the bristles of either half of the labium is a cluster of large oval
nuclei, which belong to filiform cells exactly like those of the labrum ;
ganglionic cells may also be distinguished, which bear the same relations
as those described for the upper lip.

The Cephalic Glands.

The glands in the head of Orchesella comprise two pairs, the principal
pair lying in the base of the head and occupying most of the region be-
hind the maxillae (Plate 1, Fig. 2, gl.). Each of the larger glauds consists
of a single tube, which is thrown into several longitudinal convolutions.
The deeper half of the tube is of larger caliber than the remainder, and
constitutes the secretory part (Plate 4, Fig. 32, gl.), while the more
superficial half is distinctly smaller in caliber and simply conductive in
function (Fig. 32, dt.) . The chitinous evacuating duct of either side
finally runs forward and downward on the mesal side of the gland, pass-
ing just beneath the maxillar evagination and trending toward the median
plane. The ducts open separately on either side of the extreme proximal
portion of the labial cleft (Plate 3, Fig. 25, dt.). As the mesal sur-
faces of the two halves of the labium are fused together along their
ventral edges only, a trough or gutter (sul.) is left above the line of
fusion, through which the secretion may run forward to the mouth
opening.

30 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY.

The epithelium of the glandular portion consists of large polygonal
cells (Plate 4, Fig. 33), containing large oval nuclei. The granular
cytoplasm is confined to the base of each cell, where it forms a dense,
deeply staining mass; the portion of the cell adjacent to the lumen of
the gland contains a clear non-staining substance, which sometimes ap-
pears to occupy large vacuoles and is probably secreted fluid. A delicate
chitinous intima may be distinguished as well as a thin basement mem-
brane. The dorsal wall of the gland proper is connected to the ventral
wall of the maxillary pocket by columnar cells (Fig. 33) with small oval
nuclei and fibrous cell body. These are probably modified hypodermis
cells ; between them may be seen unmodified cells, which are flattened
against the cuticula.

Transverse sections show the evacuating duct to be chitinous through-
out its whole length (Fig. 34), the component cells being indicated by
small, round nuclei only. The basement membrane is thin, but the
chitinous intima is thick.

In my account of the anatomy and function of the principal cephalic
glands, I differ widely from Fernald ('90) and also from Willem et
Sabbe ('97). According to Fernald ('90, p. 63), a curious anatomical
relation exists between the " salivary glands " and the ventral tube or
" vesicle " of Anurida maritima. I give the author's own words : " Pass-
ing forward [from the cleft of the abdominal vesicle] on the ventral
median line of the body to a median cleft in the lower lip is a small
tube, in the formation of which both hypodermis and cuticula take part.
In the posterior portion of the head are a pair of glands which resemble
salivary glands and which I regard as their homologues here. From
these glands a duct leads forward and soon fuses with its fellow, and the
median duct thus formed passes along the under surface of the buccal
cavity to a median cleft of the lower lip, where, instead of emptying into
the mouth, it turns downward and joins the ventral tube just described.
This remarkable relation of the parts concerned I am unable to explain,
although sure that no error of observation was made."

Willem et Sabbe ('97, pp. 131, 132) have recently offered an explana-
tion of the peculiar relation which Fernald described but left without an
interpretation ; they claim that the ventral tube of Smynthurus is an
adhesive organ, covered with a glutinous substance and add : " Ce liquide
est secrete, corarae l'a annonce Fernald pour Anurida maritima, par deux
glandes situees dans la tete, chez Sminthurus fuscus, elles sont logees dans
la region posterieure de la cavite" cephalique et occupent les protube-
rances verticales posterieures de la tete. Chacune d'elles est constitute

folsom: mouth-parts of orchesella cincta. 31

par un tube contourne qui, dans sa region glandulaire, se compose de
cellules plates a gros noyaux et a bordure striee ; le conduit excreteur,
plus etroit, debouche a l'extremite posterieure de la fente mediane de la
levre inferieure. Ces glandes, decouvertes chez Macrotoma flavescens par
Tullberg, qui les considerait comme glandes salivaires, ont ete figurees
ensuite cbez Lipura [Aphorura] ambulans par Nassonow qui les fait se
deverser dans la cavite buccale ; elles ont ete signalees chez Anurida
maritima par Fernald, qui a decouvert leur veritable role. De l'orifice
de la glande jusqu'au tube ventral, la secretion suit une gouttiere chi-
tineuse incomplete, qui court sur la ligne mediane de la face inferieure
de la tete et du thorax, en passant entre les pattes ; elle descend le long
du tube ventral pour aboutir au sillon qui separe les deux lobes de cet

organe."

I do not deny these observations upon Anurida and Smynthurus, espe-
cially in the face of Fernald's positive assertion, — I have not examined
those genera with reference to this question, as to do so would take me
beyond the scope of this paper, — but I have been unable to confirm the
observations in the case of Orchesella. In this species there is no com-
mon median duct, although the approximating sides of the labial cleft
might possibly be mistaken for such. I find no distinct opening through
the labium into the linea ventralis ; this structure is, as I have described,
and as Willem et Sabbe admit, " une gouttiere chitineuse incomplete,"
being always more or less open (Plate 3, Figs. 26, 27) throughout its
length, and moreover becoming more or less interrupted between the
body segments. In short, it is doubtful if it can have the function of
conveying even a viscid fluid. Furthermore, the exsertile processes of
the ventral tube are themselves well pi'ovided with unicellular glands
already described by Sommer ('85), but disregarded by Willem et
Sabbe ('97), sufficient to furnish the viscid secretions. It is my
opinion, then, that the larger cephalic glands of Orchesella are truly
salivary glands, as are those of Macrotoma [Tomocerus] and Lipura
[Aphorura] in the opinion of Tullberg and Nassonow.

The second pair of glands lie close to the skull on either side of the
head, between the bases of the mandible and maxilla. Each gland con-
sists of a somewhat conical mass of secreting cells converging downward
to a chitinous duct which follows the skull down, between the mandible
and maxilla, becoming triangular in cross section, and opens through
the lateral wall of the mandibular pocket, about half way down the
mandible.

The glandular cells (Plate 4, Fig. 35, gl) are polygonal, with large

32 BULLETIN: MUSEUM OF COMPAEATIVE ZOOLOGY.

oval nuclei ; the cytoplasm ftn*ms a close network, the interspaces of
which contain a clear substance.

The transition is abrupt between the gland and the duct, which ends
blindly (Fig. 35, dt.). The end of the duct, however, is provided with
many pores, which facilitate the passage of secreted fluid throuoh the
thick, chitinous wall of the duct. The mouth of the duct is distinctly
marked by ectodermal pigment, which is reflected from the wall of the
buccal cavity and lines the lumen of the duct for a considerable dis-
tance (Fig. 31, hi.).

The pivot of the mandible (cdx.), resting in its stirrup, abuts against
the gland, as shown in Figure 35. This leads me to think that the lubri-
cation of the pivot may possibly be an incidental function of the gland.

This pair of glands evidently corresponds to a pair described for Smyn-
thurus by Willem et Sabbe ('97, p. 132), who state, however, that " Les
conduits excreteurs des differentes cellules d'un meme cote se reunissent
en un canal collecteur qui, de la base de la mandibule, descend oblique-
ment pour se terminer dans la partie superieure de la cavite buccale,
dans Tangle forme par la mandibule et l'bypopharynx." In Qrchesella,
the gland cannot be said to open in the same place.

Regarding the possibility of the existence of a lingual gland, I may
say that the base of the glossa is lined with epithelial cells which are
unusually large and contain large oval nuclei (Plate 1, Fig. 3). There
is apparently a median opening, or at least a very thin place on the
upper wall of the glossa (Plate 3, Fig. 23, and Plate 4, Fig. 28, of.), but
I have not been able to trace its relation to underlying tissues on account
of the extreme delicacy and brittleness of the glossa, resulting in una-
voidable distortions in the process, of sectioning. On the other hand,
the central cavity of the glossa is undoubtedly a part of the general
body cavity, so that the evidence in favor of the tongue being glandular
is at most very slight.

The Physiology of the Mouth-Parts.

Almost complete ignorance of the physiology of the mouth-parts has
been but the natural consequence of an incomplete knowledge. of their
anatomy. Obviously, very little can be learned by direct observation
or experiment ; much may be inferred, however, from the structure and
relations of the organs. I have already described the action of the mus-
cles which are concerned, and may now briefly trace the history of the
food until the stomach is reached.

FOLSOM: MOUTH-PARTS OF OKCHESELLA CINCTA. 33

Orchesella cincta is a common species among decomposing leaves and
in moss ; it is most abundant among decaying pine needles and twigs,
upon which it feeds. The stomach usually contains minute irregular
fragments of wood, and the insect thrives when confined in a glass tube
with a moistened piece of decaying pine wood.

The observing Dr. Fitch ('63) is the only naturalist who has given
any account of the feeding habits of Collembola. I quote his observa-
tions ('63, p. 672) upon Smynthurus hortensis : "These Garden Fleas
are so minute that the human eye without the aid of glasses is wholly
unable to inspect their movements. The following observations will
therefore be the more interesting to the reader. It is some years since
that I noticed several of these insects on a piece of new pine board lying
in the garden. Wondering what they could find to attract them to
that situation, where I thought the odor of any turpentine in the wood
wrould rather make it repulsive to them, I was able to observe their
operations by approaching a magnifying glass to them gently, so as not
to alarm them and cause them to skip away, — ^he light colored surface
of the new wood enabling me to inspect their movements much more
accurately than could be done were they standing upon a darker colored
ground. Several of them were noticed, here and there, to have grasped
in their mouths what appeared to be an exceedingly minute flexible fibre
of the wood, fine as a fragment of a spider's web : and they were pulling
backward, at the same time shaking their heads slightly, evidently to
tear off these fibres. One of the fore legs was frequently used to crowd
this fibre more and more into the mouth, whenever it became peeled up
and too long to pull upon to advantage. Everything indicated that it
was for the purpose of food that they were thus tearing off this fine fuzz
from the surface of the new board. At one place was a small black spot
in the board, caused apparently by some old disease in the wood at this
point, which rendered it more soft and palatable to the insects, for two
of them were here busily occupied in gnawing the particles of matter
from the surface, as it seemed."

The stout setoe which project from the labrum, palpi, and labium, and
surround the mouth, are probably tactile in nature. It is possible that
the food is moistened with saliva before being taken into the mouth, as
the median trough of the labium is well adapted to convey saliva to the
border of the mouth.

In order to seize food, the mandibles leave their sockets and are pro-
truded a little from the mouth. The tips of the mandibles, by lateral
movements, grasp fibres of decaying wood, which are held between the

34 bulletin: museum of comparative zoology.

terminal incisive teeth, the four teeth of the left mandible interlocking
with the five of the right one. The food is pulled into the mouth bv
the retraction of the mandibles, assisted by the upper and lower lips,
then meets the secretions poured into the buccal cavity by the two pairs
of glands, and is grasped by the claws of the terminal maxillary lobes,
which move laterally ; the entire maxilla may also perform lateral move-
ments. The maxillae are situated on either side of the tongue, and their
tips interlock in the space left between the glossa and paraglossa?, so
that the retraction of the maxillae — which is slight, however, as con-
trasted with that of the mandibles — must pull the food along the
dorsal surface of the glossa, and through the space which intervenes be-
tween the paraglossae. In this operation, the lacunae brush along the
surface of the glossa, which is curved so as to conduct the food between
the paraglossae. On the concave dorsal surfaces of the paraglossae the
food meets the grinding faces of the mandibles, the upward rotary move-
ment of which may carry it to the projecting teeth of the paraglossae.
I have sometimes flmnd particles of wood held between the teeth of the
paraglossa; and those of the epipharynx, which is opposite. The coarse
ventral teeth of the mandibles crush the woody fibres preparatory to a
finer comminution by the denticulated molar surfaces. In the grinding
process, the powerful adductors play the principal part, supplemented
by rotary movements and possibly also by forward and backward rub-
bings. The downward rotary movement is much stronger than the
reverse, judging from the size and number of the muscles concerned in
the two acts. During mastication the mandibles are probably with-
drawn into their chitinous sockets, where the pivots encounter firm
resistance. The pivots are perhaps lubricated by glands already
described.

The comminuted food of the pharynx is sucked into the oesophagus.
This occurs by the constriction and subsequent dilatation of the fore gut.
Once within the oesophagus, the food may be forced back by peristaltic
action, resulting from the successive contraction of constricting muscles,
until the stomach is reached, where a valve prevents the return of the
food into the gullet.

Although Collembola are classed as mandibulate insects, it is evident
that they are also suctorial. The Collembola are closely related to
Campodea, a generalized type which is regarded as the representative of
a primitive form from which more specialized insects have been derived.
As already suggested by Lubbock, we may imagine the primitive insect
to have possessed mouth-parts resembling those I have described, ca-

folsom: mouth-pakts of orchesella cincta. 35

pable of further modification in either the rnandibulate or suctorial
direction. Indeed, the latter modification has already occurred in the
Collemholan genus Neanura.

In conclusiou, I desire to thank Mr. Samuel Henshaw, whose knowl-
edge of entomological literature has been of great service to me. I wish
to acknowledge my special indebtedness to Professor Mark ; for his careful
and kind supervision have been my greatest aid and encouragement.

36 bulletin: museum of comparative zoology.

BIBLIOGRAPHY OF THYSANURAN MOUTH-PARTS.1

Bourlet.

'39. Memoire sur les Podures. Mem. Soc. Spi. Agric Arts, Lille. Pt. 1,
pp. 347-417. 1 PI.

Bourlet.

'41-42. Memoire sur les Podurelles. Mem. Soc. Sci. Agric., etc. Nord.
Douai. Also separate, Douai, 1843, 78 pp, PI. 1.

Fabricius, J. C.
1777. Genera Iusectorum. [pp. 101-102.]

Fernald, H. T.

'90. Studies on Thysanuran Anatomy. (Prelim, coram.) Johns Hopk. Univ.
Circ. Vol. 9, pp. 62-63.

Fernald, H. T.
'90a. The Relationships of Arthropods. Studies Biol. Lab. Johns Hopk. Univ.
Vol. 4, pp. 431-513. Pis. XLVIII.-L.

Fitch, A.
'63. Eighth Report on the Noxious and other Insects of the State of New
York. Albany, N. Y.

Grassi, B.

'86. I progenitori degli Insetti e dei Miriapodi. L' Japyx e la Campodea.
Mcmoria II. Atti Accad. Gioen. Sci. Nat. Catania. Ser. 3, Tom. 19,
pp. 1-67. Tav. I.-V.

Grassi, B.

'86a. I progenitori dei Miriapodi e degli Insetti. Mcmoria III. Contribuzione
alio studio dell' Auatomia del genere Machilis. Atti Accad. Gioeu. Sci.
Nat. Catania. Ser. 3, T. 19, pp. 101-124. I Tav.

Grassi, B.
'86b. I progenitori dei Miriapodi e degli Insetti. Memoria IV. Cenni Ana-
tomici sul Genere Nicolatia. Bull. Soc. Entom. Ital. An. 18, pp. 173-
182. Tav. VII., VIII.

1 This Bibliography includes the works on Thysanura as well as Collembola,
and also the titles of articles relating to the cephalic glands of Collembola.

FOLSOM : MOUTH-PARTS OF ORCHESELLA CINCTA. 37

Grassi, B.

'87. Auatomia comparata dei Tisanuri e considerazioni general sull' organizza-
zione degli Insetti. Atti Accad. Lincei, Roma. Ser. 4. Mem. Vol. 4,
pp. 543-606. Tav. I.-V.

Grassi, B., e Rovelli, G.

'89-90. I progenitori dei Miriapodi e degli Insetti. Memoria VI. II sis-
tema dei Tisanuri fondato soprattutto sullo studio dei Tisanuri italiani.
Nat, Sicil. An. 9, pp. 25-41, 53-68, 77-87, 105-124. Tav. 1, 2.

Haliday, A. H.

'64. Japyx, a new Genus of Insects belonging to the Stirps Thysanura, in the
Order Neuroptera. Trans. Linn. Soc. Vol. 24, Pt. 3, pp. 441-447. PI. 44.

Laboulbene, A.

'65. Recherches sur l'Auurida maritima, insecte Thysanoure de la famille des
Podurides. Ann. Soc. Entom. France, Ser. 4, T. 4 (1864), pp. 705-720.
PI. 11.

Laboulbene, A.

'65. Description et anatomie d'un insecte maritime (Anurida maritima) qui
forme un genre nouveau dans l'ordre des Thysanoures et la famille des
Podurides. Compt. Rend. Seanc. Mem. Soc. Biol. Paris. Ser. 4, T. 1
(1864), Mem. pp. 1S9-206. PI. 1.

Latreille, P. A.

'32. De l'organisation exterieure et comparee des insectes de l'ordre des
Thysanoures. Nouv. ann. mas. hist. nat. Paris. Ser. 3, T. 1, pp. 161—
187.

Lubbock, J.

'62. Notes on the Thysanura. Pt. I. Smynthuridse. Trans. Linn. Soc, Vol.
23, Pt. 3, pp. 429-448. Pis. 45, 46.

Lubbock, J.

'73. Monograph of the Collembola and Thysanura. Rav Soc. 255 pp. 78
Pis.

Meinert, F.

'65. Campodese: en familie af Thysanurernes orden. Naturh. tidsskr. Raik.
3, Bd. 3, pp. 400-440. Tab. XIX.

Translation. On the Campodese, a Family of Thysanura. Ann. Mag.
Nat, Hist., Loud. Ser. 3, Vol. 20, pp. 361-378. Figs. 1867-
Nassonow, N.

'87. The Morphology of Insects of Primitive Organization. Studies Lab.
Zool. Mus. Moscow, pp. 15-S6. 2 Pis. 68 text figs. (In Russian.)

Nicolet, H.

'41. Recherches pour servir a l'histoire des Podurelles. Extr. Nouv. Mem.
Soc. Helv. Sci. Nat. Neuchatel. Vol. 6, 84 pp. 9 Pis. Also separate,
'43. Ibid. T. 4. 8S pp. 9 Pis.
vol. xxxv. — No. 2. 3

38 bulletin: museum of comparative zoology.

Olfers, E. de.

62. Annotationes ad Anatomiam Podurarum. Diss, inaug. Berolini. 34 pp.
4 Lith. Pis.

Oudemans, J. T.

'87. Bijdrage tot de Kennis der Thysauura und Collembola. Amsterdam.
104 pp. 3 Pis.

Translation. Beitrage zur Kenntniss der Thysauura uud Collembola.
pp. 147-226. Taf. I.-III. Amsterdam, 18S8.

Packard, A. S.

'71. Bristle-tails and Spring-tails. Amer. Nat. Vol. 5. pp. 91-107. Figs.

23-40. PI. 1.

Rovelli, G. [See Grassi, B., e Rovelli, G]

Sabbe, H. [See Willem, V., et Sabbe, H.]

Sommer, A.

'84. Ueber Macrotoma plumbea. Beitrage zur Anatomie der Poduriden.
Inaug. Diss. Gottingen. Also Zeitschr. wiss. Zool., Bd. 41, pp. 683-71S,
Taf. XXXIV., XXXV. 18S8.

Stummer-Traunfels, R. R. von.

'91. Vergleicheude Untersuchungen iiber die Mundwerkzeuge der Thysanuren
uud Collembolen. Sitzungsb. Akad. Wisseusch. Wien, Bd. 100, Abth. 1,
Heft 4, pp. 216-235. 2 Taf.

Tullberg, T.

'72. Sveriges Podurider. Kongl. Sveusk. Akad. Handlingar, Stockholm.
Ny Poljd. Bd. 10, No. 10. 70 pp. 12 Pis.

Willem, Y., et Sabbe, H.

'97. Le tube veutral et les glandes cephaliques des Smiuthures. Auu. Soc.
Eut. Belgique. T. 41. pp. 130-132.

folsom: mouth-parts of orchestslla cincta.

39

EXPLANATION OF PLATES.

[All figures are of Orchesella cincta L.]

ABBREVIATIONS.
Note. — Each muscle is numbered in agreement with the text.

a. Anterior.

abd. Abductor.

add. Adductor.

at. Antenna.

ate. Articulation.

bac. Rod.

6?-. Arm.

br. a. Anterior arm.

calc. Shoe.

car. Cardo.

cav. buc. Buccal cavity.

cdx. Pivot.

cht. Chitinous.

el. gn. Ganglion cell.

clyp. Clypeus.

cjit. Head.

era. Cranium.

crs. ba. Basal ridge.

c'stt. Constrictor.

eta. Cuticula.

ex. Heel.

d. Dorsal.

de. Teeth.

dep. Depressor.

dil. Dilator.

dt. Duct.

e'phy. Epipharynx.

e'th. Epithelium.

exp. Expansion.

/. Frontal.

fac. Surface.

ga. Galea ?

gl. Gland.

gls. Glossa.

gn. inf'ae. Infra-oesophageal gan-
glion.

gn. su'a. Supra-cesophageal gan-
glion.

gu. Gula.

h'drm. Hypodermis.

i. Intima.

i'cis. Incisive.

/. Lateral.

la. Plate.

lab. Labium.

Ibr.

Labrum.

len.

Lacinia.

lig.

Ligament.

In. v.

Linea ventralis.

lit.

Lumen.

Ivt.

Elevator.

m.

Middle, median.

m....m.

Median line.

mb. ba.

Basement membrane

md.

Mandible.

men.

Mentum ?

mol.

Molar.

ms.

• Mesal.

mn.

Muscle.

mx.

Maxilla.

n.

Nerve.

nat.

Natural size.

nl.

Nucleus.

oel.

Ocellus.

oe.

Oesophagus.

of.

Orifice.

or.

Mouth.

P-

Posterior.

pa' gls.

Paraglossa.

pd.

Foot.

pd'.

Footstalk.

phy.

Pharynx. ,

pig-

Pigment.

pi' my.

Perimysium.

pli.

Fold. "

pip.

Palpus.

}"'.h

Projection.

pr1). con.

Conical projection.

pr't.

Protrusor.

ret.

Retractor.

rot.

Rotator.

sb'men.

Submentum ?

set. sns.

Sensory bristle.

sta.

Stirrup.

stp.

Stipes.

sul.

Trough.

sut.

Suture.

bid.

Tendon.

tnt.

Tentorium.

v.

Ventral.

Folsoni. — Mouth-Parts, Orchesella cmcta.

PLATE 1.

Fig. 1. External aspect of the mouth. X 116.

Fig. 2. Diagram of the head, seen from the right side, to show the relations
between certain organs. X 55. The small figure in a circle represents
the natural size of the head.
Fig. 3. Eeconstruction from sagittal and parasagittal sections of the left half of
the head, imagined as seen from the right side. The numbers 28, 29,
30, and 31 refer to figures of Plate 4, bearing corresponding numbers,
which represent transverse sections made in the positions of the num-
bered lines of this figure. X 220.

Transverse section of the oesophagus. X 530.

The tentorium, viewed from the right side; reconstructed from serial
sagittal sections. X 220.

Dorsal aspect of the tentorium ; reconstructed from frontal sections.
X 220.

Sagittal section passing through the tentorium and adjacent structures.
X 440.

Parasagittal section of the labrum. X 220.

Fig.

4.

Fig.

5.

Fig.

6.

Fig.

7.

Fig.

8.

Folsom-Mouth Parts Orghe sell a

I

Ibt:

a,,/

),nl.'-

laf Moc

pi'rny

S.'add.md.

br.p.

fob.

Z.lvt.m.

©

X
h'l/rr,, .

pig

■'sit

set 'ins

,il,

Folsom. — Mouth-Parts, Orchesella ciucta.

PLATE 2.

Fig. 9. Internal aspect of the labrurn and clypeus, to show the epipharynx.

X 220.
Fig. 10. Dorsal aspect of the internal mouth-parts, in situ ; the paraglossae are

omitted. From dissections and potash preparations. X 110.
Fig. 11. Ventral aspect of the apex of the right mandible. X 440.
Fig. 12. Ventral aspect of the molar region of the right mandible. X 440.
Fig. 13. Dorsal aspect of the base of the left mandible, in situ; from a dissection.

X 440.
Fig. 14. Dorsal aspect of the left mandible with its muscles. Reconstructed from

serial sections, aided by dissections. X 116.
Fig. 15. Ventral aspect of the right mandible, to show certain muscles, and to

supplement Fig. 14. From a dissection. X 116.
Fig. 16. Posterior aspect of a transverse section of the right mandible, cut near

the anterior angle of the triangular orifice. X 220.
Fig. 17. Three ganglion cells from within a mandible. X 530.

"Fo lsom-Mo uth Parts Orche sella.

Plate 2.

md.

_,.-!"■/'

\
A-

'./,.

i ..car.

li9

12.

dr\

~~~^/acmol.

13.

26.

*tn_ (IL

■

md.

v.

Hdrm

sta

calc

iadd

Folsom. — Mouth- Parts, Orchesella eincta.

PLATE 3.

Fig. 18. Dorsal aspect of the framework of the left maxilla ; from a dissection.
X 116.

Fig. 19. Dorsal aspect of the head of the right maxilla ; from a dissection. X 530.

Fig. 20. Dorsal aspect of the left maxilla, with most of its muscles. Reconstructed
from serial sections, aided by dissections. X 116.

Fig. 21. Dorsal aspect of the more ventral portion of the left maxilla ; supplemen-
tary to Fig. 20. From serial sections and from dissections. X 116.

Fig. 22. Dorsal aspect of the right paraglossa, with underlying glossa ; from dis-
sections. X 440.

Fig. 23. Dorsal aspect of the glossa ; from dissections. X 440.

Fig. 24. Ventral aspect of the right half of the labium ; from a dissection. X 220.

Fig. 25. Dorsal view of the left half of the labium ; reconstructed. X 220.

Figs. 26 and 27. Transverse sections of the linea ventralis. X 530.

FOLSOMrMOUTH PaRTSOrCHESELLA.

.,.. pet1-

pd>.

car J,

21,

fa

/O rtt/t/

--— 9prl.adxt
■ arel./tdd

Uq.

;M1,'

i-lrt.m.

_.xul.

zlvtms

^dqxp

Folsom. — Mouth-Parts, Orchesella cincta.

PLATE 4.

Figs. 28-32. Sections nearly transverse to the mouth-parts, numbered in sequence,
beginning near the mouth, and cut approximately in directions
indicated by the numbered lines of Plate 1, Fig 3. Figs. 28-31,
X 220; Fig. 32, X 116.

Fig. 33. Transverse section through a secretory region of one of the larger
cephalic glands. X HO.

Fig. 34. Transverse section of an evacuating duct of the same gland. X 440.

Fig. 35. Sagittal section of one of the pair of smaller cephalic glands. X 220.

FOLSOMrMOUTH PaRTSOrCHE'..

4.

m. 28,

..-//>•

29.

MfTtv
S .. fir

Hit/

:'/ep.p

ocl.

J,9.

. >, ,

""7-

■iu.

J<r

■Si
d.

■ -

f/L...

P

met.

WE del.

The following Reports have been published on are in prepa-
ration on the Dredging Operations off the West Coast of
Central America to the Galapagos, to the West Coast of
Mexico, and in the Gulf of California, in Charge of Alex-
ander Agassiz, carried on by the U. S. Fish Commission
Steamer "Albatross," during 1891, Lieut. Commander Z. L.
Tanner, U. S. N., Commanding.

A. AGASSIZ. II.i General Sketch of the
Expedition of the "Albatross," from
February to May, 1891.

A. AGASSIZ. The Pelagic Fauna.

A. AGASSIZ. The Daep-Sea Panamic Fauna.

A. AGASSIZ. I.2 On Calamocrinus, a new
Stalked Crinoid from the Galapagos.

A. AGASSIZ. XXIII.23 The Echini.

JAS. E. BENEDICT. The Annelids.

R. BERGH. XIII. is The Nudibranchs.

K. BRANDT. The Sagittse.

K. BRANDT. The Thalassicolse.

C CHUN. The Siphonophores.

C CHUN. The Eyes of Deep-Sea Crustacea.

S. F. CLARKE. XI." The Hydroids.

W. H. DALL. The Mollusks.

W. FAXON. VT.s XV.!6 The Stalk-eyed
Crustacea.

S. GARMAN. The Fishes.

W. GIESBRECHT. XVI.'5 The Copepods.

A. GOES. Ill * XX.'-o The Foraminifera.

H. J. HANSEN. XXII.22 The Cirripeds

and Isopods.

C. HARTLAUB. XVIII. '» The Comatulas.

W. A. HERDMAN. The Ascidians.

S. J. HICKS0N. The Antipathids.

W. E HOYLE. The Cephalopods.

G. vos KOCH. The Deep-Sea Corals.

C. A. KOFOID. Solenogaster.

R. von LENDENFELD. The Phosphores-
cent Organs of Fishes.

H. LUDWIG. IV.5 XII." The Holothu-

rians.
C. F. LUTKEN and TH. MORTENSEN.

The Ophiuridse.
OTTO MAAS. XXI 21 The Acalephs.
J. P. McMURRICH. The Actinarians.
E. L. MARK. XXIV. « The Cerianthid*.
GEO. P. MERRILL. V.« The Rocks of the

Galapagos.
G. W. MULLER. XIX.19 The Ostracods.
JOHN MURRAY. The Bottom Specimens.
A. ORTMANN. XIV. w The Pelagic Schi-

zopods.
ROBERT RIDGWAY. The Alcoholic Birds.
P. SCHIEMENZ. The Pteropods and Het-

eropods.
W. SCHIMKEWITSCH. VIII.» The Pyc-

nogonidse.
S. H. SCUDDER. VIIJ The Orthoptera of

the Galapagos.
W. PERCY SLADEN. The Starfishes.
L. STEJNEGER. The Reptiles.
TH. STUDER. X.10 The Alcyonarians.
C. H. TOWNSEND. XVII." The Birds of

Cocos Island.
M. P. A. TRAUTSTEDT. The Salpidae and

Doliolidae.
E. P. VAN DUZEE. The Halobatidae.
H. B. WARD. The Sipunculoids.
H. V. WILSON. The Sponges.
W. McM. WOODWORTH. IX." The Pla-

narians and Nemerteans.
W. McM. WOODWORTH. XXVII."

Planktonemertes.

' Bull M.C. Z., Vol. XXI., No. 4, June. 1891, 16 pp. ; and Vol. XXIII., No 1, February, 1892,
89 pp., 22 Plates

2 Mem. M. C. Z., Vol. XVII., No. 2, January, 1892, 95 pp., 32 Plates,
s Bull. M. C. Z., Vol XXIV., N\ 7, August. 1893, 72 pp.

7 Bull.

M.

C.

1

Vol.

XXV

No.

1

» Bull.

M.

0.

z

Vol.

XXV.

No.

2

9 Bull

M.

0.

z.

Vol.

XXV.

, No.

i

io Bull.

M

c.

z.

, Vol

XXV.

, No.

5

» Bull.

M

c

z

Vol.

XXV

No.

6,

12 Bull.

M

c

z.

, Vol

XXV.

No.

8

is Bull.

M,

c.

z ,

Vol

XXV

No.

H

* Bull. M. C. Z., Vol. XXIII., No. 5, December, 1892, 4 pp. . 1 Plate,
s Bull. M. C. Z., Vol. XXIV.. No 4, June, 1893, 10 pp [Zool. Anzeig.
s Bull M. C. Z., Vol. XVI., No. 13, July, 1893, 3 pp

September, 1S93. 25 pp.
December, 1893, 17 pp., 2 Plates.
January, 1894, 4 pp., 1 Plate.
February, 1894, 17 pp.
February, 1*94, 7 pp., 5 Plates.
September. 1894, 13 pp., 1 Plate.
10, October, 1894, 109 pp , 12 Plates.
" Mem. M. 0. Z., Vol. XVIT. No. 3, October, 1894, 183 pp.. 19 Plates.
»5 Bull M. C. Z ., Vol. XXV. No. 12, April, 1895, 20 pp., 4 Plates.
is Mem. M. C. Z.. Vol. XVTTI., April. 1895. 292 pD , 67 Plates, 1 Chart
i' Bull. M. C Z , Vol. XXVII., No. 3, July, 1895. 8 pp.. 2 Plates.
is Bull. M. C. Z , Vol XXVII., No 4, August, 1895. 26 pp., 3 Plates.
19 Bull. M C. Z , Vol. XXVII., No. 5, October, 1895, 14 pp., 3 Plates
*• Bull. M. C. Z., Vol. XXIX., No 1, March, 1896, 103 pp., 9 Plates, 1 Chart.

21 Mem. M C. Z., Vol. XXIII., No. 1, September, 1897, 92 pp., 15 Plates.

22 Bull. M. O Z , Vol XXXI , No 5, December, 1897. 37 pp., 6 Plates, 1 Chart.

23 Bull. M. C. Z., Vol XXXIT., No. 5, May, 1898, 18 pp , 13 Plates, 1 Chart

24 Bull. M C. Z.. Vol XXXIT.. No. 8. August, 1898, 8 pp . 3 Plates.
2' Bull. M. C. Z., Vol. XXXV., No. 1, Juiy 1899, 4 pp., 1 Plate.

No. 420, 1893.]

PUBLICATIONS

OF THE

MUSEUM OF COMPARATIVE ZOOLOGY

AT HARVARD COLLEGE.

There have been published of the Bulletins Vols. I. to XXXI.,
and XXXIII. ; of the Memoirs, Vols. I. to XXII.

Vols. XXXIV. and XXXV. of the Bulletin, and Vols. XXIII.
and XXIV. of the Memoirs, are now in course of publication.

The Bulletin and Memoirs are devoted to the publication of
original work by the Professors and Assistants of the Museum, of
investigations carried on bv students and others in the different
Laboratories of Natural History, and of work by specialists based
upon the Museum Collections.

The following publications are in preparation: —

Reports on the Results of Dredging Operations from 1877 to 1880, in charge of
Alexander Agassiz, by the U. S. Coast Survey Steamer " Blake," Lieut.
Commander C. D. Sigsbee, U. S. N., and Commander J. R. Bartlett, U. S.N.,
Commanding.

Reports on the Results of the Expedition of 1891 of the U. S. Fish Commission
Steamer "Albatross," Lieut. Commander Z. L. Tanner, U. S. N., Com-
manding, in charge of Alexander Agassiz.

Contributions from the Zoological Laboratory, in charge of Professor E. L.
Mark.

Contributions from the Geological Laboratory, in charge of Professor N. S.
Shaler.

Studies from the Newport Marine Laboratory, communicated by Alexander
Agassiz.

Subscriptions for the publications of the Museum will be received
on the following terms : —

For the Bulletin, $4.00 per volume, payable in advance.
For the Memoirs, S-S.OO

These publications are issued in numbers at irregular inter-
vals; one volume of the Bulletin (8yo) and half a volume of the
Memoirs (4 to) usually appear annually. Each number of the Bul-
letin and of the Memoirs is also sold separately. A price list
of the publications of the Museum will be sent on application
to the Librarian of the Museum of Comparative Zoology, Cam-
bridge, Mass.

Bulletin of the Museum of Comparative Zoology
AT HARVARD COLLEGE.

Vol. XXXV. No. 3.

STUDIES FROM THE NEWPORT MARINE LABORATORY.
COMMUNICATED BY ALEXANDER AGASSIZ.

XLII.

LONGITUDINAL FISSION IN METRIDIUM MARGINATUM

MILNE-EDWARDS.

By G. H. Parker.

With Three Plates.

CAMBRIDGE, MASS., U.S.A.:

PRINTED FOR THE MUSEUM.

October, 1899.

Bulletin of the Museum of Comparative Zoology
AT HARVARD COLLEGE.

Vol. XXXV. No. 3.

STUDIES FROM THE NEWPORT MARINE LABORATORY
COMMUNICATED BY ALEXANDER AGASSIZ.

XLII.

LONGITUDINAL FISSION IN METRIDIUM MARGINATUM

MILNE-EDWARDS.

By G. II Parker.

With Thkke Platks.

CAMBRIDGE, MASS., U.S.A.:

PRINTED FOR THE MUSEUM.

October, 1899.

loaw

No. 3. — Studies from the Newport Marine Laboratory. Com-
municated by Alexander Agassiz.

XLII.

Longitudinal Fission in Metridium marginatum Milne-Edwards.

By G. H. Parker.

From early times double specimens of different species of Metridium
have been observed and recorded. Thus in the last century Dicque-
mare ('75, p. 229, Tab. VI. Fig. 2) described and figured a specimen of
M. diunthus with two complete oral disks. Similar specimens were ob-
served by Johnston ('47, p. 233), who called them monstrosities, and
interpreted them as cases of coalescence brought about by the gregarious
habit of the species. Thorell ('59, p. 10) and Gosse ('60, p. 20) like-
wise mentioned specimens with two disks as monstrosities, though Gosse
also expressed the opinion that they were due to a tendency to spon-
taneous division. Foot ('63, p. 64), who confirmed the observations of
his predecessors on the occurrence of animals with two oral disks, also
recorded the discovery of a specimen with two mouths on one disk and
further stated that these aberrations from the normal form are merelv
to be considered as monstrosities. G. Y. and A. F. Dixon ('91, p. 20),
in reporting the occurrence of specimens with two disks as well as those
with two mouths on one disk, mention them likewise as monstrosities.
Carlgren ('93, p. 109), after stating his belief that longitudinal division
is not uncommon in M. dianthus, adds : " Mehrere Forscher scheinen
Formen mit zwei Mundscheiben als monstrose ansehen zu wollen. Es
kann doch wohl als eine besonnene Langsteilung, die nicht zu Ende ge-
fiihrt ist, betrachtet werden."

Double specimens of M. marginatum were observed as early as 1847
by Professor Louis Agassiz, among whose unpublished drawings are three
figures, two • of which (Plate I. Figs. 2, 3) bear this date and probably
represent two views of the same animal ; the third figure (Fig. 1) bears
the date of 1 860. These figures were unaccompanied by any notes bearing

VOL. XXXV. — NO. 3. 1

44 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY.

ou the question of fission, but in the " Seaside Studies/' published in
1865 by Elizabeth C. Agassiz and Alexander Agassiz ('65, p. 11), longi-
tudinal fission is stated to occur in this species.

Finally, Torrey ('98, p. 347), who studied the Californian species
M. fimbriatiun, described double specimens which he believed to be in
process of fission.

The material upon which the present paper is based consisted of ten
specimens of Metridium marginatum, collected in part by myself and in
part by others. The sources of this material are acknowledged in the
account of the specimens given with the description of the figures. Here
each specimen has received a distinguishing letter. To the gentlemen
whose names are mentioned as having obtained certain specimens fur
me, I wish to express my indebtedness. I am also indebted to Dr.
H. C. Bnmpus, Director of the Laboratory at the United States Fish
Commission Station at Wood's Hole for many courtesies shown me
while working at the station, and I am under special obligations to
Mr. Alexander Agassiz for the privilege of working at the Newport Lab-
oratory, and for the use of unpublished drawings made for Professor Louis
Agassiz.

The animals collected were for %the most part stupefied by means of
magnesium sulphate in sea water, Tullberg's well known method, and
subsequently hardened in chromic acid and dissected by hand.

The specimens naturally fall into two groups : first, those with two
mouths on one disk, of which there were two examples, specimens A and
B (Plate II. Figs. G and 7) ; and secondly, those with two complete oral
disks, of which there were eight, C to J (Plate II. Figs. 4 and 5, Plate
III. Figs. 8-14).1 The two specimens (Plate I. Figs. 1-3) figured by
Professor Agassiz belong also to this group. These two groups were not
only distinguished by external anatomical differences, but also by certain
internal characteristics. In both representatives of the first group the
oesophageal tubes were Y-shaped, the single inner end opening into the
gastrovascular cavity and the two outer ends opening each through a
mouth. In specimen A (Plate II. Fig. 6) the bifurcation was close
to the oral disk and the oesophagus was for the greater part of its extent
a single flattened tube. In specimen B (Fig. 7) the oesophagus was
double excepting at its extreme inner end, so that its form is perhaps
more correctly described as Y-shaped. In all the specimens in the sec-
ond group the oesophageal tubes were entirely distinct from their snper-

1 One of these eight, specimen J, was sacrificed in an attempt to rear it; hence I
have been able to study the internal anatomy of only seven such specimens.

PARKER: LONGITUDINAL FISSION IN METRLDIUM MARGINATUM. 45

ficial to their deep ends (Plate III. Figs. 8—14). In the material at
hand, then, completely distinct disks were always associated with com-
pletely separate oesophageal tubes and single disks bearing two mouths
with only partially separate tubes.

Although the oesophageal tubes have been described as though they
were either entirely distinct or united only through their own substance,
'in two of the nine animals studied there were special membranes attach-
ing one tube to the other. In specimen C (Plate III. Fig. 8) the two
tubes were held together by a single membrane which extended from
their oral to their pedal ends, but which failed of connection with the
inner surfaces of both oral and pedal disks. In specimen B (Plate II.
Fig. 7) two such membranes were present and extended from the region
where the oesophageal tubes united almost to the oral disk. Both mem-
branes were closely attached to the tubes, and were deficient only next
the oral disk ; consequently the cavity which they enclosed was entirely
cut off from the gastrovascular cavity except near the oral disk, where it
opened freely over both membranes. Although the membranes in these
two specimens had the general appearance of mesenteries, they lacked
the characteristic muscle bands and mesenteric filaments. Membranes
much like these, but in one case bearing mesenteric filaments, have been
observed in M. fimbriatum by Torrey ('98, p. 347).

In all specimens excepting B, F, and H, each mouth was monoglyphic ;
in B (Plate II. Fig. 7) and F (Plate III. Fig. 11), each animal had a
monoglyphic and a diglyphic mouth ; in H (Plate III. Fig. 13) one
mouth was monoglyphic and the other aglyphic. In those cases where
the oesophageal tubes were united, A and B, the siphonoglyphs were, not-
withstanding this union, distinct throughout their whole lengths. The
siphonoglyphs of any double animal apparently do not occupy random
places, but are usually arranged symmetrically with reference to the
assumed plane of division. This is easily seen in Figures 8, 10, 1 2, and
14 (Plate III.). Its significance in connection with fission has been
pointed out by Torrey ('98).

Of the complete mesenteries the directives were always in pairs,
and always attached to the siphonoglyphs. Their arrangement conse-
quently corresponded to that of the siphonoglyphs, and hence need not
be described. The non-directives were also always in pairs, though not
infrequently one member of such a pair was incomplete. This occurred
in eleven cases in a total of 107, or in about 10% of the mesenteries in the
specimens examined. Including under the head of pairs of complete
mesenteries those cases in which only one member of a pair is complete,

46

bulletin: museum of comparative ZOOLOGY.

the number and general distribution of the non-directives is indicated in
the following tabulation.1

Specimen

Left Directive Pair

Intervening Non-directive Pairs
Right Directive Pair ....

Non-directive Pairs

Additional Directive Pair . .
Non-directive Pairs

1

16

G

1

3

1

13

H
1

1G
0
0

= 62

= 12

Total Non-directive Pairs

107

The numbers of non-directives in the animals with two mouths bear
an important relation to those of animals with single mouths. In an
enumeration of the mesenteries in the latter, made some time ago
(Parker, HJ7), I found that in 131 specimens the pairs of non-directives
were never fewer than three and never more than fourteen, and that the
average was about five and a half pairs (5.G+) for each individual. In
the nine double-mouthed specimens the non-directives were never fewer
than eight pairs, nor more than nineteen, and the average was a little
less than twelve (11.9 — ) for each individual. It thus appears that the
double-mouthed specimens have almost exactly twice as many non-direc-
tive mesenteries as the single-mouthed ones.

As can be seen from the tabulation already given, as well as from the
figures, the non-directives are unevenly but characteristically distributed,
there being usually a large group on one side and a small group on the
other side of the pairs of directives, called rights and lefts. Even in speci-
mens B, F, and H (Figs. 7, 11, and 13), where the directives are more
or fewer than two pairs, this same peculiarity in distribution may be said
to exist. In another respect the distribution of the non-directives is re-
markably regular. In specimen I (Plate III. Fig. 14) a pair of non-direc-

1 In this tahle the term left directive pair is applied to the pairs of directives
toward the left in the figures on Flate III., and to those toward the bottom in the
figures on Plate II. The corresponding pairs on the other oesophageal tubes are
called right directive pairs. Eight and left have in this connection only this simple
descriptive significance.

PARKER LONGITUDINAL FISSION IN METRIDIOI MARGINATUM. 47

tives is so placed that, while they start from nearly the same place on the
outer wall, they pass to different oesophageal tubes. The pair opposite
these has similar connections, so that at this stage the assumed plane of
division may he said to pass through primary entoccels on both sides.
The same is true of G (Fig. 12), and Torrey ('98, Plate XXI. Fig. 1) fig-
ures what seems to he a similar case, though in his description (p. 347)
he mentions only one pair of mesenteries as divided between the oesopha-
geal tubes. Apparently there is some discrepancy here, but whether it
is the description or the figure which is faulty, it is impossible to say.
In the remaining specimens examined by me, the division plane always
lay in primary ectocoels. It is noteworthy that in the nine cases
studied none showed the plane of division passing through a primary
ectoccel on one side and a primary entocoel on the other. In this re-
spect the specimens were always perfectly symmetrical, a condition
which may hold for other Actinians, as suggested by the symmetrical
division of Zoanthus thomensis as described by Koch ('86, p. 33).

The incomplete mesenteries formed what were usually irregular groups
in the primary ectocoels. Their arrangement seemed to bear so little
on the question of division that I have not attempted a description of
them. In the figures of transverse sections the incomplete mesenteries
are either represented in full in a given primary ectoccel or their pres-
ence is indicated by x. Primary ectocoels not marked in one or other
of these ways contained no incomplete mesenteries.

Of the double-mouthed specimens whose sexes could be determined,
four were females and three were males. Xo evidence of hermaphro-
ditism was observed, though the specimens were carefully scrutinized in
this respect. Obviously the double-mouthed condition is not peculiar to
either sex.

The interpretations that have been placed upon these double-mouthed
specimens have already been stated. Johnston's idea that they have
arisen by the fusion of two originally independent individuals seems to
me entirely unwarranted. M. marginatum is represented by individuals
showing extreme variations in color and markings, and yet the two mem-
bers of all double specimens seen by me have been strikingly similar.
If fusion were the means of forming double animals, particolored combi-
nations ought occasionally to occur, but such I have never seen.

There remain then the two suggestions of monstrosities and of stages
of fission. To test which of these was the correct interpretation, I at-
tempted to watch the process in what might be assumed to be a dividing
individual. This animal was found at Wood's Hole, August 6th, 1898,

48 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY.

and was kept under observation there during August and later at Bev-
erly Bridge till October 3d of the same year, a period of some eight
weeks. During this time the animal was kept in open sea water in a
marked locality, not in a laboratory aquarium, which my former expe-
rience had taught me might be unfavorable. When first seen its pedal
disk measured about three centimeters in diameter. It had two com-
plete oral disks, one of which had a monoglyphic and the other a di-
glyphic mouth. The cleft between the two oral disks was a deep one,
and I hoped soon to witness the separation of the two individuals.
When about eight weeks after capture it was last seen, it had increased
in size so that its pedal disk was nearly seven centimeters in diameter,
but the cleft between the two parts had in no wise increased. The speci-
men then unfortunately disappeared from its locality. The fact that it
grew considerably shows that it was under approximately normal condi-
tions, and yet so far as fission was concerned the animal was essentially
at the same stage at the end of the eight weeks as at the beginning. This
coincides with the experience of Torrey ('98, p. 351), who in describing
longitudinal fission in M. fimbriatum states that he has " not observed
a single instance of full severance of individuals, though a number of
dividing polyps have been kept in the laboratory for nine months."
Thus direct evidence of actual fission is wanting.

If now we turn to the specimens which have thus far been described,
it must be admitted that they can be arranged in a series passing from
less to more completely divided individuals. This in itself, however,
affords no more grounds for concluding that M. marginatum reproduces
by longitudinal fission than a similar series of partially double mammals
would establish longitudinal fission fur these animals. To make the
proof conclusive the final products of the process must be found.

As a rule, the larger specimens of M. marginatum are sessile animals.1
In large specimens that had undergone division the offspring ought
therefore to be found together. M. marginatum further shows great
variability in its colors and markings ; hence such natural pairs might
thus be distinguished from their neighbors, for the two descendants
would of course inherit directly the surface markings of their progenitor.

A search was made for natural pairs at Beverly Bridge, a locality

1 According to my own observation the locomotor activity of this species be-
comes rapidly less as the size increases. A specimen whose diameter was about
5 mm. moved as much as 9 cm. in twenty -four hours. A second one whose average
diameter was about 2 cm. never moved more than 1 cm. in a day, and a large one
whose diameter was about 8 cm. showed no perceptible movement in fifteen days.

PARKER : LONGITUDINAL FISSION IN METRIDIUM MARGINATUM. 49

where the species was represented by thousands of individuals in easily-
accessible positions. The search was facilitated by the fact that often a
given pile or plank was covered with individuals of almost exclusively
one type of coloration ; almost all in one restricted location were orange-
colored, or brown, or whitish, etc. This is doubtless due chiefly to re-
production by fragmentation from the edge of the pedal disk, a non-
sexual process which insures to the offspring the minute characteristics
of the parent. My object was to find pairs of animals individually sepa-
rate, but situated next each other and strikingly similar in color and
markings, though entirely unlike those surrounding them. Such in-
stances, if found, would afford strong evidence in favor of the actual
occurrence of longitudinal fission. A search of some three hours' dura-
tion, in which between two and three thousand specimens were in-
spected in their attached positions, resulted in the discovery of six such
natural pairs. These were pairs which fulfilled the requirements of the
case in every respect, in that they were composed each of two animals
strikingly similar in size and coloration, and entirely isolated from others
of a like kind by a surrounding group of individuals unlike them. "What
seemed to be natural pairs of this kind were often met with in other
places, but as it was important to accept only pairs that were well iso-
lated, and as these would naturally not occur frequently, it is not sur-
prising that only six such pairs were found.

The six pairs were placed each in a separate bottle and transferred to
the laboratory for further study. On examination all proved to be mono-
glyphic except one individual, which was diglyphic. The arrangement of
the complete mesenteries is indicated as follows : —

D-2-3-4-5-6-7.

First Pair D.2- 3-4-5-f -7-8.

(D-2-D-4-54-7.
Second Pair j D-2-3-4-5-6-7.

[D-2-3-4-5-6-7.
Third Pair D. 2-3-4-5- 6-7.

Fourth Pair
Fifth Pair
Sixth Pair

{S:

D-2-3-4-5-6-7-8-9.

D- 2-3-4-5-6.

2-3-4-5-6.

2-3-|-5-6.

D- 2 -3- 4-5.

D-2-3-4-5-6.

The first pair were found attached to a large mussel shell, and were
killed and hardened without being disturbed from their natural posi-

50 bulletin: museum of compaeative zoology.

tions. The arrangement of their siphonoglyphs and mesenteries is
shown in Figure 15 (Plate III.). So far as the anatomical peculiarities
of this and the five other pairs are concerned, they might well he con-
sidered as animals that had undergone division.

A further important feature of the six pairs was that in all cases mem-
bers of the same pair were of the same sex. Of the six pairs, three were
female and three.male. This uniformity is precisely what would be
expected, if these pairs arose by longitudinal fission, but quite the re-
verse of what we should anticipate had they been the result of accidental
juxtaposition. This evidence seems to me to favor strongly the view
that longitudinal fission is a regular method of increase in M. margina-
tum, though it does not exclude the possibility of some of the double-
mouthed forms being true monstrosities. It is, however, not my belief
that the specimen which was watched by me some two months and which
did not progress in division was a monstrosity. I am rather inclined to
Torrey's opinion, that the process of longitudinal fission is an extremely
slow one. This accords with what we know of the length of life of
Actinians, some having outlived human beings.

In all the double-mouthed specimens which I have examined the two
partial individuals were of about equal size, so that longitudinal fission
in M. marginatum may be justly described as equal. In no case have I
ever observed any evidence of distinctly unequal longitudinal fission,
though one of the specimens figured by Agassiz (Plate I. Figs. 2 and 3)
shows some considerable inequality. In this respect M. marginatum
seems to be strikingly different from M. fimbriatum, which according
to Torrey ('98) divides more usually by unequal than by equal longi-
tudinal fission.

In longitudinal fission, obviously, certain fundamental changes must
occur : new siphonoglyphs and new complete mesenteries, both directive
and non-directive, must be developed. In the material which I have
studied I have been able to discover no trace of the formation of new
siphonoglyphs. The usual symmetrical arrangement of two of the sipho-
noglyphs in each double animal suggests a process of division by which
one original siphonoglyph gives rise to two by longitudinal splitting, as
described by Torrey ('98). This is very likely one of the initial steps
in longitudinal fission, though it is not necessarily essential, as the con-
dition of specimen H (Plate III. Fig. 13) shows where the original sipho-
noglyph obviously did not divide.

Of the formation of new directives I have seen absolutely nothing.
The production of new complete non-directives ia possibly indicated in

PARKER : LONGITUDINAL FISSION IN METRIDIUM MARGINATUM. 51

several cases. In specimen C a curious condition was observed, which is
shown in the diagram (Plate III. Fig. 8). Two complete mesenteries
start from the column wall, but unite before they reach the oesophageal
wall. The cavity which they enclose may be called a primary entoccel
so far as their longitudinal muscles indicate, but it contains two incom-
plete mesenteries which pair off with the united complete ones. The
condition might be interpreted as due to the splitting of a complete me-
sentery and the supplementary development of incomplete ones, though
this interpretation is not without objections. I am more inclined to look
on the condition as purely abnormal. In another case, the mesentery
marked 1 in Figure 7 (Plate II.), whose mate is incomplete, is itself
complete only through a small portion of its oral length. This might be
a mesentery originally incomplete, which by active growth had completed
itself in the oral region. A second similar case was observed among
the six natural pairs already mentioned. Beyond these meagre facts, I
observed nothing suggestive of the methods by which new complete
mesenteries may be formed.

In all the specimens examined by me the fission was progressing from
the oral toward the pedal pole of the animal, never in the reverse direc-
tion, as described by McCrady ('59, p. 275) for Actinia cavernosa, by
('arlgren ('93, p. 31) for Protanthea simplex, and as may also occur in
Cjreactis as described by "Wilson ('89, p. 38), who however did not
place this interpretation on what he saw. The direction of fission in
M. marginatum corresponds to that generally found in M. fimbriatum
as described by Torrey ('98).

The kind of individuals produced by equal longitudinal fission in M.
marginatum can be easily inferred from the foregoing account. They
are usually monoglyphic, though occasionally diglyphic, possibly on rare
occasions aglyphic. A pair of directive mesenteries is regularly attached
to each siphonoglyph. The animals possess the usual number of com-
plete non-directive mesenteries. These, however, do not show the regu-
lar hexamerous arrangement often met with in this species, but are as
a rule very irregularly disposed. As I have elsewhere indicated
(compare Parker, '97, pp. 263, 264), M. marginatum is represented
by two chief types : a diglyphic one often characterized by a strictly
hexamerous arrangement of its mesenteries and a monoglyphic one
in which amongst other forms three sub-types, distinguished respectively
by having five, six, and seven pairs of non-directives, may be recognized.
The animals produced by longitudinal fission may come under any of
these heads except the regularly hexamerous diglyphic form, to which

52 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY.

about one fifth of the numbers of specimens ordinarily collected belong.
While I can confirm Torrey's ('98, p. 357) statement that diglyphic
and monoglyphic individuals may arise by the same mode of non-sexual
reproduction, I believe he has gone too far when he denies any possible
correlation between these structural types and the methods of repro-
duction. Since no mode of non-sexual reproduction has been shown to
give rise to a hexamerous diglyphic specimen, and since this type is rep-
resented by about one fifth of the number of individuals in the species,
it seems to me still possible that this may be the result of sexual i*epro-
duction.1 The value of this suggestion must, however, await further
investigation. Should it prove true, much of the irregularity in the
arrangement of the mesenteries in Metridium would be associated with
non-sexual reproduction.

Summary.

The double specimens of Metridium examined had either two mouths
on one oral disk or two complete oral disks. In the former cases the
oesophageal tubes were incompletely divided (Y- or V-shaped) ; in the
latter, there were two completely distinct tubes.

The mouths of the double specimens were usually monoglyphic, some-
times diglyphic, and in one instance aglyphic. Two siphonoglyphs were
usually placed symmetrically to the supposed plane of division.

A pair of directive mesenteries was always attached to each siphono-
glyph. There were about twice as many non-directive mesenteries in
double specimens as in single ones.

In any given case the assumed plane of division passed through either
two primary ectoccels or two primary entocoels, never through a primary
entocoel on one side and a primary ectocoel on the other.

The double specimens were either male or female, and showed no evi-
dence of hermaphroditism.

They are not due to fusion.

1 In briefly discussing this question, Torrey ('9S, p. 357) has quoted me in a
somewhat misleading way. It is true that I suggested that the monoglyphic and
diglyphic types might have the value of varieties, and also that they might be cor-
related with the methods of reproduction, but tins was not done in one breath, as
might be inferred from Torrey's statement. I mentioned the possible interpretation
as varieties on page 269 of my former paper (Parker, '97), and, after stating my dis-
inclination to accept this interpretation, I suggested on page 270 the possibility of
correlation with the methods of reproduction. Why Torrey should have com-
bined these two suggestions as one, and why he should have used quotation marks
for a piece of composition which is not mine, remain to be explained.

PAEKEE : LONGITUDINAL FISSION IN METEIDIUM MARGINATUM. 53

While some may be monstrosities, the occurrence of natural pairs
shows that longitudinal fission takes place.

This process is probably extremely slow.

While monoglyphic and irregular diglyphic specimens may be pro-
duced by this and other non-sexual processes, the production of regular
hexamerous diglyphic specimens by non-sexual methods has not been
observed. Such specimens number about one fifth of all collected, and
may be the products of sexual reproduction.

Cambridge, January 13, 1899.

54 BULLETIN: MUSEUM 01 COMPARATIVE ZOOLOGY.

PAPERS CITED.

Agassiz, E. C, and Agassiz, A.

'65. Seaside Studies in Natural History. Boston. Ticknor and Fields,
viii -f- 155 pp.
Carlgren, O.

'93. Studien iiber nordisclie Actinien. Kongl. svenska Vet-Akad. Hdlgr.
N. F., Bd. XXV. No. 10, pp. 1-118, Taf. I.-X.

Dicquemare, J. F.

'75. A Second Essay on the Natural History of the Sea Anemonies. Pliil.
Trans. Roy. Soc, London, Vol. LXV., 1775, pp. 207-248, Tab. VI.

Dixon, G. Y., and Dixon, A. F.

'91. Report on the Marine Invertebrate Fauna near Dublin. Proc. Roy. Irish
Acad., 3d Ser., Vol. II. pp. 19-33.
Foot, F. J.

'63. Notes on the Astrseacea of the Coast of Clare. Proceedings of the
Natural History Society of Dublin, for the Sessions 1859-1862
(inclusive). Vol. III. pp. 63-69.
Gosse, P. H.

'60. Actinologia Britannica. A History of the British Sea Anemones and
Corals. Loudon. Van Voorst, xl-f- 362 pp., 12 Pis.
Johnston, G.

'47. A History of the British Zoophytes. Second Edition. London. Van
Voorst, xvi + 488 pp., 74 Pis.
Koch, W.

'86. Ueber die von Herrn Prof. Dr. Greeff ira Golf von Guinea gesammelteu
Anthozoen. Inaugural-Dissertation. Bonn, Carl Georgi. Pp. 36,
5 Taf.

McCrady, J.

'59. Instance of incomplete longitudinal Fission in Actinia cavernosa Bosc.
Proceed. Elliott Soc. Nat. Hist. Charleston, S. C Vol. I. pp.
275-278.

Parker, G. H.

'97. The Mesenteries and Syphonoglyphs in Metridium marginatum Milne-
Edwards. Bull. Mus. Comp. Zool., Vol. XXX. pp. 259-273, 1 PI.

PARKER: LONGITUDINAL FISSION IN METRIDIUM MARGINATUM. 55

Thorell, T.

'59. Oni den inre byggnaden af Actinia plumosa Mull. Ofversigt af Kougl.
Vetenskaps-Akademiens Forhandliugar. Vol. XV. 1858, pp. 7—25,
Tab. I.

Torrey, H. B.

'98. Observations on Monogenesis in Metridium. Proceed. California Acad.
Sc, 3d Ser., Zoology, Vol. I. pp. 345-360, PL XXI.

Wilson, H. V.

'89. On the Occasional Presence of a Mouth and Anus iu the Actinozoa.
Johns Hopkins University Circulars. Vol. VIII. pp. 37, 3S.

3

<

56 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY.

EXPLANATION OF FIGURES.

All figures were taken from specimens of Metridium marginatum Milne-Edwards.
In the figures of transverse sections the dotted lines represent the axes of the mouths
projected upon the plane of section. All complete mesenteries are drawn. The
incomplete mesenteries usually form more or less irregular groups in the primary
eetoccelB. Each primary ectoccel containing incomplete mesenteries either has
these mesenteries represented in it or is marked with x ; any primary ectoccel not
designated in one or other of these ways contained no incomplete mesenteries.

Parker. — Metridium marginatum.

PLATE I.

Reproductions of figures of double specimens from the unpublished drawings of

Professor Louis Agassiz.

Fig. 1. Side view of a specimen with two oral disks. The original drawing is

marked, " Alive in Aquarial Gardens, May, 1860." Drawn by J.

Burkhardt.
Fig. 2. Oral view of a partly contracted specimen with two oral disks. The two

disks show a greater difference in size than is usual with this species.
Fig. 3. Oral view of what is probably the same animal as that shown in Figure 2,

but in a more contracted state.

The original drawings from which Figures 2 and 3 were taken were on the same
sheet of paper, which was marked, "Boston, Sept. 28, '47." Drawn by W. H.
Tappan.

Parker. - Metrtdium.

Pl. I.

\

Boston Sag. Co.

Paekeb. — Metridiuiu marginatum.

PLATE II.

Fig. 4. Lateral view of specimen D, with two complete oral disks, collected by the
writer and S. F. Williams at Beverly Bridge, Mass. The arrangement
of the internal parts of this specimen as seen in transverse section is
shown in Figure 9 (Plate III.). Female. Natural size.

Fig. 5. Oral view of the same specimen as that shown in Figure 4. Slightly
enlarged.

Fig. 6. Transverse section of specimen A, with one oral disk and two mouths.
The section is taken at a level at which the oesophagus is single.
Orally from this the oesophagus becomes double, each tube leading to
a mouth opening. Sex not determined. Collected by B. H. Van Vleck
at Cape Ann, Mass. X2.

Fig. 7. Transverse section of specimen B, with one oral disk and two mouths. The
cavities of the two oesophageal tubes were entirely distinct except near
the extreme aboral end, where the two tubes united to form a single
oesophagus. The oesophageal tubes are connected with each other by
a pair of membranes which resemble mesenteries without differen-
tiated muscle bands. These membranes extend from the region in
which the two oesophageal tubes separate almost to the oral disk. The
cavity between them opens into the gastrovascular cavity orally, since
both membranes are deficient in that region. Mesentery 1, though
complete orally, is incomplete through more than half of its pedal
length. Collected by the writer at Wood's Hole, Mass. Female. X3.

5ar ; •"

-.

Parker. — Metridium marginatum.

PLATE III.

All the specimens shown on this plate, except those represented in Figure 15, had
each two complete oral disks and two distinct oesophageal tubes,

Fig. 8. Transverse section of specimen C, in which the two oesophageal tubes are
connected by a single membrane which reached neither the oral nor
the pedal disks. Collected by the writer at Wood's Hole, Mass. Male.
X 1.6.

Fig. 9. Transverse section of specimen D, also shown in Fig. 4 (Plate II.). Fe-
male. X 2.

Fig. 10. Transverse section of specimen E, collected by B. II. Van Vleck at Cape
Ann, Mass. Sex not determined. X 2.

Fig. 11. Transverse section of specimen F, collected by C. W. Prentiss at Wood's
Hole, Mass. Male. X 3.

Fig. 12. Transverse section of specimen G, collected by J. I. Hamaker at New-
port, R. I. Femafe. X 2.

Fig. 13. Transverse section of specimen H, collected by R. H. Johnson at Wood's
Hole, Mass. Female. X 2.

Fig. 14. Transverse section of specimen I, collected by the writer at Wood's
Hole, Mass. Male. X 2.

[Specimen J, collected by the writer at Wood's Hole, Mass., was lost in an attempt
to rear it.]

Fig. 15. Transverse section of a natural pair of individuals, showing their mutual
relations so far as positions of siphonoglyphs, etc. are concerned. The
two specimens, which were entirely separate, were killed and preserved
without being removed from the shell on which they were found.

Boston

The following Rkports have been published or ark in prepa-
ration ox the Dredging Operations Off the West Coast of
Central America to the Galapagos, to the West Coast of
Mexico, and in the Gulf of California, in Charge of Alex-
ander Agassiz, carried on by the U. S. Fish Commission-
Steamer ".Albatross," during 1891, Lieut. Commander Z. L.
Tanner, U. S. X., Commanding.

A. AGASSIZ. II.i General Sketch of the
Expedition of the "Albatross," from
February to May, 1891.

A. AGASSIZ. The Pelagic Fauna.

A. AGASSIZ. The Deep-Sea Panamic Fauna.

A. AGASSIZ. I.2 On Calarnocrinns, a new

Stalked Crinoid from the Galapagos
A AGASSIZ. XXIII.2S The Echini.
JAS. E. BENEDICT. The Annelids.
R. BERGH. XIII. »s The Nudibranchs.
K. BRANDT. The Sagittaj.
K. BRANDT. The Thalassicolse.
C CHUN. The Siphonophores.
C. CHUN. The Eyes of Deep-Sea Crustacea.
S. F. CLARKE. XI." The Hydroids.
W. H. DALL. The Mollusks.

W. FAXON. VI.s XV. •« The Stalk-eyed
Crustacea.

S. GARMAN. The Fishes.

W. GIE3BRECHT. XVI.15 TheCopepods.

A. GOES. UI* XX.w The Foraminifera.

H. J. HANSEN. • XXII. 22 The Cirripeds

and Isopods.

C. HARTLAUB. XVIII. >« The Comatulas.

W. A. HERDMAN. The Ascidians.

S. J. HICKSON. The Antipathids.

W. E HOYLE. The Cephalopods.

G. vox KOCH. The Deep-Sea Corals.

C. A. KOFOID. Solenogaster.

R. von LENDENFELD. The Phosphores-
cent Organs of Fishes.

H. LUDWIG. IV. » XII." The Holothu-
rians.

C. F. LUTKEN and TH. MORTENSEN.

The Ophiuridae.
OTTO MAAS. XXI.-' The Acalephs.
J. P. McMURRICH. The Actinarians.
E. L. MARK. XXrV,2* The Cerianthidse.

GEO P. MERRILL. V.« The Rocks of the
Galapagos.

G. W. MULLER. XIX." The Ostracods.

JOHN MURRAY. The Bottom Specimens.

A. ORTMANN. XIV. w The Pelagic Schi-
zopods.

ROBERT RIDGWAY. The Alcoholic Birds.

P. SCHIEMENZ. The Pteropods and Het-
eropods.

W. SCHIMKEWITSCH. VIII. » The Pyc-

nogonidee.

S. H. SCUDDER. VII J The Orthoptera of

the Galapagos.

W. PERCY SLADEN. The Starfishes.

L. STEJNEGER. The Reptiles.

TH. STUDER. X.10 The Alcyonarians.

C. H. TOWNSEND. XVII17 The Birds of

Cocos Island.
M. P. A. TRAUTSTEDT. The Salpid» and

Doliolidae.
E. P. VAN DUZEE. The Halobatidse.
H. B. WARD. The Sipunculoids.
H. V. WILSON. The Sponges.
W. McM. WOODWORTH. IX.9 The Pla-

narians and Nemerteans.
W. McM. WOODWORTH. XXVII.**

Planktonemertes.

i

89 Pp

3

4
5
S

8
9
10
11
12
13
1*
15
16
17
19
lit
90
31
22
23
24

Bull M.C. Z., Vol. XXI., No. 4. June 1891. 16 pp. : and Vol. XX1IL, No. 1, February. 1892,

. 22 Plates

Mem. M. C. Z.. Vol. XVII., No. 2. January. 1892, 95 pp., 32 Plates.

Bull. M. C. Z., Vol XXIV.. Ni. 7, August. 1893, 72 pp.

Bull. M. C. Z .. Vol. XXIII., No. 5. December, 1892, 4 pp.. 1 Plate.

Bull. M. C, Z., Vol. XXIV.. No 4. June, 1893, 10 pp [Zool. Auzeig., No. 420, 1893.]

Bull M. C. Z., Vol. XVI.. No. 13, July, 1S;3, 3 pp

Bull. M. C. Z , Vol. XXV , No 1 September, 1893. 25 pp.

Bull. M. 0. Z . Vol. XXV., No. 2. Dei-ember, 1893, 17 pp., 2 Plates.

Bull. M. 0. Z., Vol. XXV., No. 4. January. 1894, 4 pp., 1 Plate.

Bull. M. C. Z.. Vol XXV.. No. 5. February, 1894, 17 pp.

Bull. M. 0 Z . Vol. XXV No h\ February, 1894, 7 pp.,5 Plates.

Bull. M. 0. Z., Vol. XXV. No 8, September. 1894, 13 pp., 1 Plate-

Bull. M. C. Z .. Vol XXV. No. 10, October, 1894, 109 pi> , 12 Plates.

Mem. M. 0. Z . Vol. XVII. No. 3, October, 1894, 183 pp. 19 Plates.

Bull M. C. Z , Vol. XXV No 12. April. 1895. 20 pp.. 4 Plate*

Mem. M. 0. Z . Vol. XVTII . April. 1895. 292 pp., 67 Plates, 1 Chart

Bull. M. C Z , Vol, XXVII., No. 3, July, 1895. 8 pp.. 2 Plates.

Bull. M. C. Z ., Vol XXVII., No 4, August. 18a",. 26 pp., 3 Plates.

Bull M. C. Z., Vol. XXVII., No. 5, October, 18il5, 14 pp., 3 Plates

Bull. M. C. Z.. Vol. XXIX., No 1. March, 1896, 103 pp., 9 Plates, 1 Chart.

Mem. M C. Z., Vol. XXITI., No. 1, September, 1897, 92 pp., 15 Plates.

Bull. M. 0. Z.. Vol. XXXI , No 5, December, 1897, 37 pp.. 6 Plates. 1 Chart.

Bull. M. 0. Z., Vol XXXII., No. 5, May, 1898, 18 pp., 13 Plates, 1 Chart.

Bull. M C. Z.. Vol. XXXII., No. 8, August, 1898. 8 pp., 3 Plates.

Bull. M. C. Z., Vol. XXXV., No. 1, July 1899, 4 pp., 1 Plate.

PUBLICATIONS

OF THE

MUSEUM OF COMPARATIVE ZOOLOGY

AT HARVARD COLLEGE.

There have been published of the Bulletins Vols. I. to XXXI.,
and XXXIII. ; of the Memoirs, Vols. I. to XXII.

Vols. XXXIV. and XXXV. of the Bulletin, and Vols. XXIII.
and XXIV. of the Memoirs, are now in course of publication.

The Bulletin and Memoirs are devoted to the publication of
original work by the Professors and Assistants of the Museum, of
investigations carried on by students and others in the different
Laboratories of Natural History, and of work by specialists based
upon the Museum Collections.

The following publications are in preparation : —

Reports oji the Wesults of Dredging Operations from 1877 to 1880, in charge of
Alexander Agassiz, by the U. S. Coast Survey Steamer " Blake," Lieut.
Commander C. D. Sigshee, U. S. N., and Commander J. R. Bartlett, U. S. N.,
Commanding.

Reports on the Results of the Expedition of 1891 of the U. S. Fish Commission
Steamer " Albatross," Lieut. Commander Z. L. Tanner, U. S. N., Com-
manding, in charge of Alexander Agassiz.

Contributions from the Zoological Laboratory, in charge of Professor E. L.
Mark.

Contributions from the Geological Laboratory, in charge of Professor N. S.
Shaler.

Studies from the Newport Marine Laboratory, communicated by Alexander
Agassiz.

Subscriptions for the publications of the Museum will be received
on the following terms: —

For the Bulletin, $4.00 per volume, payable in advance.
For the Memoirs, $8.00 " " "

These publications are issued in numbers at irregular inter-
vals ; one volume of the Bulletin (8vo) and half a volume of the
Memoirs (4to) usually appear annually. Each number of the Bul-
letin and of the Memoirs is also sold separately. A price list
of the publications of the Museum will be sent on application
to the Librarian of the Museum of Comparative Zoology, Cam-
bridge, Mass.

• OCT 24 1899

Bulletin of the Museum of Comparative Zoology
AT HARVARD COLLEGE.

Vol. XXXV. No. 4.

OVOGENESIS IN DISTAPLIA OCCIDENTALS RITTEK
(MS.), WITH REMARKS ON OTHER SPECIES.

By Frank W. Bancroft.

With Six Plates.

CAMBRIDGE, MASS., U. S. A. :

l'RINTED FOR THE MUSEUM.
October, 1899.

Bulletin of the Museum of Comparative Zoology
AT HARVARD COLLEGE.

Vol. XXXV. No. 4.

OVOGENESIS IN DISTAPLIA OCCIDENTALS RITTER
(MS.), WITH REMARKS ON OTHER SPECIES.

By Frank W. Bancroft.

With Six Plates.

CAMBRIDGE, MASS., U. S. A. :

PRINTED FOR THE MUSEUM.
October, 1899.

OCT 24- 1899

No, 4. — Ovogenesis in Distaplia occidentalis PaTTER (MS.), with
Remarks on Other Species.1 By Prank W. Bancroft.

CONTENTS.

PAGE

I. Introduction 59

II. Sexual Organs 62

1. Development in Distaplia oc-

cidentalis 62

Conclusions 66

2. Structure in the Adult ... 66

a. In Distaplia occidentalis . 68

b. Observations on Other Spe-

cies 73

c. General Considerations . . 74
.III. Incubatory Pouch 76

1. Historical Summary ... 76

2. Observations 77

IV. Envelopes of the Ovum ... 79

1. Primitive Follicular Epithe-

lium 79

2. Test Cells 80

a. Evidence for the Intraovu-

lar Origin of the Test Cells 81

PAGB

b. Evidence for the Follicular

Origin of the Test Cells . 82

c. Degeneration of the Test

Cells 84

d. Fate and Function of the

Test Cells 84

3. Secondary Follicular Epithe-

lium 88

4. Inner Follicular Epithelium . 88

5. Outer Follicular Epithelium

and Corpus Luteum ... 89

6. Observations on Styela mon-

tereyensis 92

V. Ovum 95

1. Cytoplasm 95

2. Germinative Vesicle ... 96

3. Nucleolus 102

4. Maturation 105

Bibliography 109

Explanation of Plates 112

I. Introduction.

The Distaplia material here described was collected by Professor
William E. Putter at Pacific Grove, California, during the summer of
1896, and is much of it in an excellent state of preservation. Davi-
dofF's ('89, p. 118) corrosive-acetic mixture, picro-acetic acid, picro-
sulphuric acid, and Perenyi's fluid were used for fixing, and of these the
first was by far the best, and yielded excellent results.

1 Contributions from the Zoological Laboratory of the Museum of Comparative
Zoology at Harvard College, under the direction of E. L. Mark, No. XCVIII.
vol. xxxv. — No. 4. 1

60 bulletin: museum of comparative zoology.

For the name of the species I am also indebted to Dr. Ritter, who
has kindly allowed me to quote his manuscript name, Distaplia occi-
dentalis. With the material at my disposal, which, for the most part?
consists of only pieces of colonies, and these from but one of the many
localities where the species has been encountered, it would be unwise to
attempt a complete diagnosis. I will, therefore, endeavor to give only
a few of what seem to be the distinctive features.

The colonies vary in shape, being both incrusting and pedunculate ;
in the latter case they are often mushroom-shaped. The color also
varies ; in preserved specimens it may be whitish, salmon-yellow, brown-
ish, or purple. The zobids are small, adult ones exceeding 3 mm. by a
few micra only, and they are always hermaphrodite, ovary and testis
being well developed at the same time. The ramifications of the pyloric
gland do not end in swollen ampullae. The stomach, though nearly
smooth on the outside, is thrown into irregular folds within, so that in
whole preparations it has a reticulated appearance. The test is com-
posed of the thin walls of numerous lacunae. It contains many ectoder-
mic vessels, and these do not anastomose. These characters easily
separate D. occidentalis from D. magnilarva (Delle Valle, '82, p. 200 ;
Lahille, '90, p. 157), D. lubrica (Drasche, '83, pp. 22-23), D. vallii1
(Herdman, '86, pp. 128-132), and D. livida (Sars, '50, p. 154; Huit-
feld-Kaas, '96, p. 11). It appears to be distinguished from D. clavata
(Sars, '50, p. 154; Huitfeld-Kaas, '96, p. 11) by the great length,
6 cm. of the narrowly clavate colony. This leaves D. rosea (Delle
Valle, '82, p. 202; Lahille, '90, pp. 174-175; Caullery, '95, p. 8) and
D. intermedia (Heiden, '93, pp. 348-349), with which our species
seems to be closely related. If D. intermedia is found to have unisexual
zooids, then this character will separate it from D. occidentalis. If, on
the other hand, the colonies examined by Heiden were not fully mature,
and the zooids become hermaphroditic later, then his species and D. rosea
must be united. Thus we get D. rosea as the closest, and a very close,
relative of D. occidentalis. In fact, the only differences seem to refer
to characters of minor importance, such as the range in shape and color
of the colonies, and the inequalities on the inner surface of the stomach.

1 If, as Herdman says (pp. 128, 130), D. vallii contains no bladder cells in the
test which "consists of a homogeneous matrix in which are scattered numerous
small test cells and larger pigment cells (Plate XVIII. Fig. 4)," then this species
must be considered rather distantly related to the othpr members of the genus, for
in all of these in which the structure of the test has been examined, it has been
found to be lacunar.

BANCROFT: OVOGENESIS IN DISTAPLIA OCCIDENTALIS. 61

But, as has already been said, I have not made the extended observa-
tions necessary to place beyond doubt the position of D. occidentalis.

The Styela and Chelyosoma material was mostly collected by myself,
at various places along the California coast, during the summers of 1894
and 1895, and studied about a year later. Four fixing media were used,
— Flemming's fluid, picro-sulphuric acid, glacial acetic acid, and Pere-
nyi's fluid, and of these the last gave the best results.

The investigation of the Styela and Chelyosoma material was done
for the most part at the University of California, and that of Distaplia
entirely at Harvard University during the college years of 1896-97 and
1897-98.

The stains used for the last species were mostly iron haernatoxylin
and a combination of methyl green and acid fuchsin, adapted from
Auerbach's ('96, p. 414) method B a. These two methods supplement
each other very nicely, and render the use of others almost superfluous ;
the haernatoxylin demonstrating most morphological features splendidly,
while the chemical constitution is brought out well by the methyl
green and fuchsin. The iron hematoxylin was employed in the usual
way ; fresh solutions of the stain were found to give decidedly the best
results, but for some things, such as nucleoli, which take a fresh solu-
tion too strongly, older ones are often to be preferred. In using the
double stain, sections were first treated in a mixture containing two
parts of a 0.1% aqueous acid fuchsin solution, to which a little acetic
acid had been added, and three parts of a 0.1 % solution of methyl
green. They were left in this mixture fur about fifteen minutes, and then
immersed in a 0.1% solution of methyl green for a few minutes longer.
From the stain the sections were transferred directly to 90%, and then
to absolute alcohol, xylol, and balsam. No especial care is necessary in
order to get a satisfactory differentiation. In addition to the two stains
mentioned, List's ('96, pp. 480-487) potassium ferro-cyanide methods
were also employed, but yielded no very satisfactory results. They will
be discussed later.

The sectioning of the younger ova presents no difficulties, but the
older ones, which, with the exception of the germinative vesicle, are pure
yolk, are exceedingly difficult to deal with, on account of the crumbling
of the yolk, and for them no satisfactory method was found. However,
all of the material that I had access to had been preserved in 80% or
90% alcohol for some time, so that I did not have an opportunity of
testing the effect of preservation in weaker grades of alcohol, formalin or
paraffine. As a rule, the whole colony was sectioned, but when dealing

62 BULLETIN: MUSEUM OF COMPAEATIVE ZOOLOGY.

with the maturation stages, the best results were obtained by dissecting
out the ova, .and then leaving them for a very short time in the higher
grades of alcohol, xylol, and paraffine. Four minutes in each of the
former, and three in the paraffine, melting at 50° Centigrade, secured a
good penetration, but did not entirely prevent crumbling. Painting
the sections with flexible collodion was resorted to, but while this held
together the section that was being cut, it also pulled part of the yolk
out of the next section, and imperfect series resulted. Double imbed-
ding with celloidin and paraffine was also tried, but this made the yolk
too brittle, so that I am still in quest of a satisfactory method for this
material. •

The sections were cut 6 /x thick, and fastened to the slide with tap
water, which was evaporated at about 36° after the sections had been
stretched by subjection to a slightly higher temperature. This method
gave perfectly satisfactory results, and possesses the immense advantage
of leaving no foreign substance on the slide.

I wish to express here my thanks to Dr. E. L. Mark, under whose
direction this research has been carried on, and to whose constant kind-
ness the credit of much that may be of value therein is due.

II. Sexual Organs.
1. Development in Distaplia occidentalis.

Old colonies which had already reached sexual maturity are the
only ones in which I have studied the development of the sexual or-
gans. In such colonies, all of the buds, even the smallest and least
differentiated, possess the fundament of the ovotestis. Figure 1 (Plate 1)
represents a section through such a bud. The large germ-cells, which
indicate the presence of the ovotestis, are situated entirely on one
side of the flattened inner vesicle, and all of the evidence indicates
that this side is dorsal ; for, as soon as the axes of the bud are estab-
lished by the development of the other organs, the ovotestis is found to
lie in the mid-dorsal line. The rather indiscriminate mingling of the
young oogonia and primordial follicle cells, and the superficial * position
of the oogonia as contrasted with the deeper position of the less differ-

1 Following Julin ('93, p. 95), I use the terms superficial and deep as relating to
the position of the structure in question with reference to the whole individual.
Peripheral and central are used only to describe position with reference to the organ
itself.

banckoft: ovogenesis in distaplia occidentalis. 63

entiated cells, which will go to make up the testis, are characteristic of
all the younger buds. On the other hand, the amount of separation of
the ovotestis from the inner vesicle is subject to considerable variation.
In the bud figured, the two apparently are fused in places, but in other
buds, which are entirely similar in all other respects, the line of demar-
cation of the ovotestis from the inner vesicle is most distinct. In these
cases it is seen that all the mesodermic elements on the dorsal side of
the inner vesicle, with the exception of a few scattered cells, are con-
tained in the fundament of the ovotestis. The conclusion, then, to be
drawn from the examination of the undifferentiated bud is that the
possibility of an endodermic origin for some of the unspecialized cells
of the ovotesticular fundament must be conceded ; but certainly the
greater part, and perhaps the whole of it, is derived from the mesoderm.

So far as the spaces between the cells of the ovotestis are concerned,
the figure represents a rather unusual condition, since in most cases the
fundament forms a compact mass. But on account of the unusual sepa-
ration, this bud demonstrates very clearly that in this stage Distaplia
has no " assise peripherique " such as Julin ('93, p. 97) and Floderus
('96, p. 175) found in the youngest stages examined.

In slightly older buds, where the fundaments of the intestine and
epicardium are present, the only change in the ovotestis is due to a few
of the ova having increased in size and assumed more distinctly the
characters of the oogonium. Its position is still dorsal, but it occupies
only the posterior end of the dorsal side of the bud. At still later
stages, even after the peribranchial sacs are quite large, the ovotestis
is apparently fused with the posterior end of the dorsal tube. This
connection is often very puzzling; in one series, for instance, the
lumen of the dorsal tube extends quite to the ovotestis, the latter
appearing as a thickening of the dorsal tube. In buds of these stages
there is considerable variation in the disposition of the organs, the
ovotestis being placed slightly to one side of the dorsal tube or directly
behind it, or even dorsal to its posterior end. Accordingly, I am
inclined to believe that the apparent union between these two organs
is not significant, but due rather to their close approximation and to
oblique sectioning. This position is further strengthened by the fact
that both before, immediately after, and even occasionally during,
this stage, the ovotestis is completely isolated from all other organs.
Whenever the ovotestis is thus isolated, it is produced anteriorly into a
genital strand, which occupies very nearly the mid-dorsal line. It
seems likely, then, in view of the very general isolation of the sexual

64 bulletin: museum of comparative zoology.

organs observed in the early stages, that the whole of these organs
is produced by the growth of the ovotesticular fundament of the un-
differentiated bud.

In older buds the genital organs are always distinctly separated from
the other organs, and the differentiation of ovary from testis is speedily
accomplished. Figures 2 and 3 (Plate l) represent two consecutive
cross-sections through a stage in which this separation is going on. In
Figure 2, the posterior section, there are two well-defined masses of
cells which in Figure 3 are united, the only indication of the future
separation being a slight notch on one side of the fundament. The
splitting takes place from behind forwards, and appears to be accom-
plished by a rearrangement of the cells in situ, and the secretion of a
structureless membrana propria between the two masses. Preceding
the actual separation, however, there is a differentiation consisting in
the fact that the smaller cells, which will produce the testis, are always
situated deeper than the others, and this, as Figure 1 shows, is charac-
teristic even of the earliest stages. Both ovary and testis are usually
solid masses of cells at this stage, and for some time later, but occa-
sionally one or both of them may develop a lumen (Plate 1, Fig. 3) ,
though in the case of the testis it is never well defined.

Figure 4 (Plate 1) gives a longitudinal section of a somewhat later
stage, the ovary and testis being completely separated from each other,
but both still attached to a common genital strand {fun. gen.). The
latter extends from the anterior ends of the genital organs to the left
peribranchial sac, with the walls of which it fuses. It is much thicker
and shorter in the earlier than in the later stages, when the distance
that it has to traverse becomes greater.

An examination of Figures 2, 3, and 4 shows that at this stage there
is a well-defined tendency for the long axis of the peripheral nuclei of
the testis or testicular part of the ovotestis to occupy a tangential
position. This marks the beginning of the peripheral epithelium that
bounds the testis at all later stages. In the ovary , however, such a
peripheral layer is absent in this stage, as in the earlier ones, so that
the young oogonia or their follicles extend to the surface of the organ.
This is exactly the condition found by van Beneden et Julin ('85, p.
334) for the same stage of these organs in Perophora. Their figures
(6 a and 6 b, Planche XYI.) show the peripheral layer well differenti-
ated in the testis, but not in the ovary, though the two organs are
still connected.

The fundaments of the oviduct and vas deferens are at first formed

BANCROFT : OVOGENESIS IN DISTAPLIA OCCIDENTALS, i 65

by the splitting of the genital strand, which is progressively converted
from behind forward into vas deferens and oviduct. But with the
growth of the bud, the undifferentiated portion of the strand is stretched
more and more, until it becomes a single row of cells attached end to
end, which connects the last differentiated portions of the two ducts
with the peribranchial epithelium.

From this time on cell-multiplication must take place pari passu
with the separation of the ducts. But as no mitoses were encountered,
it is useless to speculate as to whether this multiplication is accom-
plished in the region of the genital strand, the differentiated ducts, or
the mass of cells where the three join. Even before the ducts are
separated throughout their whole length a histological difference is
apparent, the nuclei of the oviduct cells being more flattened and tak-
ing a fainter stain than those of the vas deferens. Another point of
interest is that, contrary to what obtains in other species, both ducts
are solid strands, which do not acquire a lumen until much later.

While the ducts are being separated, the testis increases in size and
entirely covers the deep side of the ovary, so that in cross-section it
appears as a crescent, the tips of which abut against the ectoderm on
either side of the ovary. Later it becomes divided up into the eight
or more testicular pouches which are characteristic for the species.
Meantime the ovary increases in size and acquires a lumen, on the
deep side of which most of the ova are situated. With these last
changes all the essentials of the adult structure are reached, but the
position of the sexual organs with respect to the axes of the animal, and
the point of union of the genital strand with the peribranchial sacs,
still undergo some change.

In young buds with small peribranchial sacs, which have not yet
united to form the median cloaca, and without a trace of stigmata, the
sexual organs are situated in the mid-dorsal line, and also in the
intestinal loop which has developed around them. At this stage the
oesophagus opens into the right posterior corner of the branchial sac
and the anus is near the left posterior corner, so that the whole of the
post-pharyngeal digestive tract is in a frontal plane. The genital
strand extends forward in the median line and joins the tip of the left
peribranchial sac, where this comes nearest to the sagittal plane. Later
the point of union of the genital strand with the peribranchial sac
migrates still farther to the left, so that it comes to lie much nearer to
the anus. Now, as the bud grows and assumes the adult shape, the
whole abdomen is so twisted that right becomes ventral, left dorsal,

66 bulletin: museum of compaeative zoology.

and the loop of the digestive tract occupies the sagittal plane. By
means of this twisting, the anus becomes dorsal and the point of union
of the genital strand with the perihranchial sacs migrates from the
left through the cloaca to the right perihranchial sac; and thus the
adult relations of the organs are acquired.

Conclusions.

From this account it follows that the development of the sexual
organs in Distaplia occidentalis agrees with that of the three species
described by van Beneden et Julin ('85, pp. 328-349) and of Ciona
intestinalis described by Floderus ('96, pp. 173-181) in the following
important respects : —

1. The whole of the sexual organs of the adult is developed from a
single solid mass of cells of mesodermic origin, which during the later
period of its existence is connected with the cloaca by a genital strand.

2. The ovary is separated from the testis by a splitting of the primi-
tively single fundament which proceeds from behind forward. The
separation of the two ducts takes place in the same manner, though
here growth must intervene. The ovary and oviduct always lie on the
superficial side of the testis and vas deferens.

3. The cavity of the ovarial fundament is formed in such a way that
the germinative epithelium lies in its deeper wall.

Distaplia occidentalis differs from the species mentioned in that : —

1. The ovary and testis are usually solid, and the ducts always so,
at the time when they are separated from each other.

2. The genital strand is at first quite thick, and decreases in
diameter as the bud grows.

3. The fundament of the ovotestis is present in the youngest stages,
whereas in the other species it appears quite late in ontogeny.

With the development of these organs in Styelopsis grossularia
studied by Julin ('93), that of our species has less in common, because
in the two the ovaries and testes are of a different type. In Distaplia,
and the four species with which it has been compared, the ovary is
distinctly separated from the oviduct proper, in the walls of which no
ova arise, while in Styelopsis the deep wall is given over to the pro-
duction of ova throughout the whole extent of the oviduct. Further-
more, in Styelopsis the testis is divided into a number of lobes, each
one opening separately into the perihranchial cavity, whereas in the
other species it is a single organ with a single long duct. Accord-

BANCROFT: OVOGENESIS IN DISTAPLIA OCCIDENTALIS. 67

ingly, there is no genital strand at all in Styelopsis, and the separation
of ovary from testis takes place from the middle towards both ends ; in
the other respects mentioned, however, this genus agrees with Pero-
phora, Phallusia, Clavelina, and Ciona.

With respect to the presence of a peripheral epithelium surrounding
the developing ovotestis, there is more variation. In Perophora and
Distaplia this is present in the testis but not in the ovary, while in
Clavelina, according to the figures of van Beneden et Julin ('85, Planche
XV., Figs. 9, 10, and 13), in Styelopsis and in Ciona it is encountered
in both. This distribution is very suggestive, for the first two genera
have rather small zooids, and a simple structure, while the last three
are much larger, produce many more sex cells, and are more highly
specialized. It seems, then, that it is only in the larger and more
complex ascidians that this peripheral epithelium is formed around the
developing ovary.

For the explanation of the fact that the oogonia are present in the
undifferentiated Distaplia bud, we must undoubtedly look to the fact
that in this genus, as in Botryllus, the life of the individual zooid is
very short, so that apparently it would not have time to mature an
ovum if the latter were to start from a wholly undifferentiated cell, and
consequently it receives the ovum in a slightly elaborated form. In
these two genera, however, the structure and development of the sexual
organs seem to be entirely different, as according to Pizon ('92-'93,
pp. 257-271), the ovary and testis of Botryllus arise from separate fun-
daments, and no real ovary is formed at all, but merely a few follicles
attached by separate stalks to the peribranchial sacs.

Distaplia also agrees with Botryllus and disagrees with all other
species in which the matter has been studied in that the funda-
ments of the sexual organs and their ducts are usually solid. This
condition is probably the direct result of the early development of
these organs necessitating their being blocked out in solid masses
before enough cells to form vesicles were available.

On the whole, however, it is evident that the development of the
sexual organs in our genus conforms to the type first described by van
Beneden et Julin for Perophora, Clavelina, and Phallusia, and deviates
rom it only in minor respects due to the simpler structure and short
flife of the zooids in Distaplia.

68 bulletin: museum of comparative zoology.

2. Structure in the Adult.

a. In Distaplia occidentalis.

This species of Distaplia is always hermaphroditic, no unisexual
zooids such as have been described for D. magnilarva having been
encountered. Nor is there any provision for the prevention of self-
fertilization by means of protandry or protogyny. The testis is found
to be functionally active for a much longer period than the ovary, for
spermatozoa are always found in the vas deferens from the time the
oldest ova are about half grown until after the usual number of eggs,
two, has been laid.

The development of the genital organs in zooids of the same size
varies considerably according to the condition of the colony. If the
oldest zooids of a colony are laying eggs or have recently done so, then
all the younger ones and the older buds have their sexual organs more
fully developed than in individuals of the same size in colonies which
possess no mature sexual products. In the undifferentiated buds,
however, the same conditions obtained in all the colonies.

The position of these organs is on the right side of the intestinal
loop. When partially developed, they do not extend beyond the
intestine, either anteriorly or posteriorly, but when maturity is reached,
they project beyond it in both directions. In this condition the ovary
occupies the extreme hind end of the zooid and is entirely posterior
to the loop. The testis, which is much more voluminous, is also
mainly posterior to the loop, but its extreme anterior end is usually
within it.

Concerning the finer structure of the ovary, recent investigators are
not entirely in accord. Corresponding to the fact that in the develop-
ment of this organ a peripheral epithelium is described by Julin and
Floderus, which the earlier investigators had not mentioned, there is a
similar disagreement about the adult structure. But here the differ-
ence is of much more importance, for Julin, in tracing the peripheral
membrane into the adult ovary, arrives at an idea of the relation of
the germinative epithelium to the peripheral wall essentially different
from that of his predecessors. Van Beneden et Julin ('85, pp. 350,
358) and Maurice ('88, p. 456, 464) had found that the superficial
wall of the ovary passed insensibly into the germinative epithelium
occupying the lateral edges of the organ. The deeper1 wall, which is

1 There can be no doubt that in van Beneden et Julin's Figure 14 (Planche XV.)
the wall of the ovary of Clavelina rissoana that is placed uppermost and described

BANCROFT: OVOGENESIS IN DISTAPLIA OCCIDENTALS. 69

formed by a cylindrical epithelium in both Clavelina rissoana and
Fragaroides, and to which the stalks of the older follicles are attached,
is considered as having developed from the germinative epithelium.
According to this view, then, both the germinative epithelium and the
cylindrical epithelium connecting with and forming the stalks of the
follicles * are merely differentiations from the wall of the ovary and not
additional structures. In other words, the ovary is but a specialized
part of the oviduct.

On the other hand, Julin (p. 95) finds that in Styelopsis the super-
ficial wall of the ovary connects with a thin limiting epithelium and
not with the germinative epithelium. The latter represents an addi-
tional element lying within the cavity formed by the two peripheral
epithelia and not represented at all in the oviduct of such forms as
Clavelina and Fragaroides. As the ova increase in size, they push the
limiting epithelium before them, and retain their connection with the
rest of the ovary by means of it. Thus, according to this author, the
outermost layer of cells surrounding the ovum and continuous with
the stalk is not, as has been formerly supposed, derived from the
germinative epithelium, and is not part of the true follicular covering
of the egg.

Floderus, though he finds a limiting epithelium in the young
ascidian, does not agree with Julin concerning the structure of the
adult. He finds that the germinative epithelium continues into the
epithelium of the superficial wall (p. 187), that the epithelium which
he thinks corresponds to Julin's limiting epithelium does not connect
with the superficial wall (p. 185) , and that the stalks of the ova con-
nect with their follicles (p. 190).

But little evidence in favor of Julin's contention has been found in
the ovary of Distaplia occidentalis, which has essentially the structure

(p. 351) as "la voute de cette raeme cavite" is the deep or ventral wall. Maurice
(p. 459) takes it for granted that this wall is dorsal, and points out the difference
that would exist between Clavelina and Fragaroides on this point. Floderus
(p. 183-184) discusses the question fully and concludes that although van Beneden
et Julin do not state explicitly which side is dorsal, the description and the posi-
tion of the figure when compared with the other figures on the plate compel us to
believe that the authors consider the "voute" dorsal. However, leaving the ques-
tion of the intention of the authors aside, the ventral position of the germinative
epithelium in the fundament of the ovary in C. rissoana and of the ovarian follicles
in the adult ovary of C. lepadiformis, described by Floderus, point conclusively to
the fact that the same condition must obtain in the adult C. rissoana.

1 I propose in future to call the characteristic tissue constituting this epithelium
stalk tissue.

70 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY.

described for the same organ in Fragaroides. It consists of two sacs,
the first of which is an enlargement of the posterior end of the oviduct,
and whose cavity may be designated as the lumen of the ovary proper
(Plate 1 Pig. 5, lu. ov.). On the deeper wall of this sac is situated
the germinative epithelium, which occupies the greater part of its
extent. When viewed from within the animal, this epithelium
presents an oval outline with its long axis running in an antero-pos-
terior direction. The second sac, whose cavity is spoken of as the
lumen of the stalk tissue, is situated on the deep side of the first one,
and its lumen connects with that of the ovary proper by a narrow pas-
sage in the stalk, which is inserted in the deep wall of the ovary proper
at about the centre of the oval patch of germinative epithelium.
Figure 5 represents a cross-section through both these cavities and the
stalk connecting them.

In our species the adult ovary is composed of three tissues (Plate 1,
Pigs. 5, 6). First the thin pavement epithelium comprising the super-
ficiala wall of the organ (par/). Usually it forms also the lateral
edges and is often continued so as to constitute part of the deep * wall
(Plate 1, Figs. 5, 6, 7, par."). It is directly continuous with, and of
the same histological structure as the wall of the oviduct, both being
very thin and devoid of cilia.

Secondly, the germinative epithelium and the older ovarian follicles
developed from it. The former sometimes appears as directly con-
tinuous with the superficial wall, but more often, especially when the
ova are a little larger (Figs. 5, 6, 7), they seem to lie enclosed within
a peripheral epithelium which is itself continuous with the superficial
wall. The ova, which are the only indication of the germinative
epithelium, are very few, the total number in an adult zooid being
usually under twenty. Occasionally a very small oogonium, or an
ovum that may perhaps be considered undifferentiated (Fig. 5, ov'go.),
is found within the superficial Avail, but in the great majority of cases
they are confined to the deeper wall. Here they are usually so arranged
that the older the oogonium the nearer it is to the middle of the wall
to which the stalk tissue is attached (Fig. 5). There are very few or
no primordial ova in the epithelium, so that physiologically it is
really no germinative epithelium at all, but merely a repository for the
youngest oogonia. It is very interesting to note, however, that this
repository of oogonia takes on the same relations to the rest of the
ovary that the functional germinative epithelia of the more specialized

1 Used in the technical sense defined on page 62.

BANCROFT: OVOGENESIS IN DISTAPLIA OCCIDENTALIS. 71

genera- assume. In Distaplia occidentalis this arrangement of the
oogonia, and even their existence there, has reference to the phylogeny
of the genus rather than to present usefulness, for the usual number of
eggs laid hy a zooid is two, or at most three, and since in the adult
there is no provision for budding, all the younger oogonia must degen-
erate with the zooid and go to waste.

Thirdly, there is the stalk tissue (Figs. 5, 6, Us. pd.), which is
directly continuous with the gorminative epithelium or its peripheral
membrane, and also by means of the follicular stalks with the follicles
of all the older ova (Plate 3, Figs. 14, 15, Us. pd.). This, in its
typical condition (Figs. 5, 15), is a cylindrical epithelium with
rather irregular oval nuclei and a thick, structureless basement mem-
brane on its peripheral surface. The stalk tissue surrounds a cavity
(Fig. 5, lu. pd.) that is usually somewhat larger than the true cavity
of the ovary, with which it connects by a narrow stalk. This stalk
joins the rest of the ovary in the region of the germinative epithelium,
so that a series of cross-sections presents exactly the same appearance as
in Fragaroides. Posteriorly the deep wall is composed of a continuous
germinative epithelium; about the middle of the epithelium the cavity
of the stalk tissue connects with the cavity of the ovary proper and
divides the germinative epithelium so that in a cross-section of this
region it looks like a paired structure (Plate 1, Fig. 5). Anteriorly
the entire centre of the deep wall is made up of germinative epithelium
as it is posteriorly. As a result of this structure, it is seen that the
stalk tissue does not connect with the flat pavement epithelium com-
prising the wall of the oviduct. The cavity of the stalk tissue extends
posteriorly behind that of the ovary proper, and has the follicles
attached to it in such a way that the oldest ovum is usually the most
posterior one. From this account it follows that the ovary, though
not a paired organ, is a symmetrical one ; but even this is not always
the case. Occasionally the stalk tissue, instead of being inserted near
the centre of the deep wall, joins the latter near one side, so that it
connects with the superficial wall on one side and the germinative
epithelium on the other.

The question as to whether in Distaplia occidentalis there is a
limiting membrane which is distinct from the germinative epithelium,
seems to be answered negatively. The condition mentioned above,
and represented in Figures 5, 6, and 7, where the ova project strongly
into the cavity of the ovary, seems to indicate that the germinative
epithelium is bounded peripherally by a limiting membrane.

72 bulletin: museum of comparative zoology.

But it is impossible at this stage to tell whether the side of the
ovum next to the periphery of ovary is covered by but one cell layer,
the follicle, or by a limiting epithelium in addition to this. I have
never been able to differentiate a limiting epithelium in this position,
but the failure could not justify one in denying its existence. When
it comes to the formation of the follicular stalk, however, evidence is
obtained that is conclusive, for we get the external epithelium of the
ovary, which should extend around the ovum independently of the
follicle, connecting with it in the clearest possible manner (Plate 3,
Figs. 14, 15). Figure 14 shows part of a young ovum that is entirely
outside of the ovarian wall, but which is not yet attached to it by a
stalk. On the left the ovum is attached to the germinative epithelium,
as the tangential section of a much younger egg (ov'go.) shows. On
the right it joins stalk tissue which has not yet assumed its typical
appearance. Between the two are cells which make up both the fol-
licle and the wall of the ovary, and will probably differentiate into
follicle and stalk tissue. Figure 15 represents the same region in an
older ovum, where the stalk tissue presents its characteristic structure
and is seen to be directly continuous with the follicular epithelium.
From these facts the conclusion seems justified that in the adult
Distaplia ovary, just as in the developing one, we have no peripheral
limiting epithelium.

One feature that is not usually encountered in other species, but is
quite conspicuous in the ovary of this one, is the corpus luteum (Fig.
5, cp. hit.). This structure is merely the part of the follicular cover-
ing of the egg which remains behind when the latter is extruded from
the ovary. It will be described in detail after the origin of its con-
stituent elements has been considered.

The oviduct (Plate 2, Fig. 9, ov'dt.) is a narrow, thin-walled tube
extending up the right side of the abdomen and in immediate contact
with the ectoderm. Throughout its extent the vas deferens (va. df.)
is closely appressed to its deep wall. The most striking peculiarity
of the oviduct is that its diameter, even when distended by the passage
of an ovum, is very much less than the normal diameter of the ripe
egg. Accordingly, when the egg is passing through the duct, it is
greatly distorted, assuming the shape of a sausage (Plate 3, Fig. 17).
In this figure the ovum occupies a part of the oviduct, all of the
cavity proper of the ovary, and also the entire cavity enclosed by the
stalk tissue and its own follicle. The point where the stalk tissue
joins the germinative epithelium lining the deep wall of the ovary

BANCROFT: OVOGENESIS IN DISTAPLIA OCCIDENTALS. 73

proper is well indicated by the constriction at pd. Figure 16 shows
the normal size of the ovum before laying.

b. Observations on Other Species.

I have examined sections of the adult ovary in several other species,
and describe the condition in two of these, on account of the light it
throws on our conception of the fundamental structure of the ascidian
ovary.

The first of these species is Styela montereyensis,1 whose ovary
presents a structure similar to, but not identical with, that described
for Styela rustica by Floderus (p. 185-186). The ovaries are four
elongated tubes, two on each side, which open at their anterior ends
into the atrial cavity. The superficial wall is composed of a ciliated
epithelium, which connects with a germinative epithelium at the
lateral edges. The deep wall is much folded in an irregular manner,
and is composed of a thin pavement epithelium with occasional small
oogonia contained in its thickness. At times portions of the deep
wall, where these oogonia are crowded together, look a good deal like
germinative epithelium. I am convinced, however, that this is not the
case, for nowhere in the deep wall are oogonia of the smallest size
and undifferentiated ova, such as exist in the two lateral germinative
epithelia, encountered. The presence of these small oogonia outside
the germinative epithelium is due to the fact that the latter some-
times gives off its oldest oogonia after the younger ones, which are
thus carried out into the deeper wall before they are really ready, and
remain in it for a considerable period, while they are reaching the size
at which they project beyond the pavement epithelium. All the
older oogonia are situated on the deep side of this epithelium, presum-
ably attached to it by their follicular stalks, though these are but
seldom seen. In each of these four ovaries, then, we have essentially
the same structure that exists in Clavelina and Styelopsis, where the
ovary is single and situated in or near the sagittal plane. Further-
more, it is likely, though not proven, that these four ovaries are
developed from separate fundaments, for in a very much younger stage
(Plate 2, Fig. 12) , when the whole of the deep wall is made up of
the germinative epithelium, and the primordial ova are still dividing,

1 This is one of the commonest of our California ascidians. It was first de-
scribed by Dall (71, pp. 157, 158) as Cynthia montereyensis, and subsequently by
Fewkes ('89, pp. 134, 135) as Clavelinopsis rubra. Ritter ('93, p. 39) placed it in its
proper genus, Styela.

VOL. XXXV. — NO. 4. 2

74 bulletin: museum of compakative zoology.

and the vas deferens opens into the oviduct instead of the atrium,
these organs occupy the same relative position as in the adult.

The second species is Chelyosorna productum, Stimpson ('64, p.
161). Here the structure is similar, but not identical, to that
described by Floderus (p. 187) for the genera Ascidia, Ascidiella,
and Corella. The ovary is much branched, occupying the superficies
of the viscera, and also extending in between the stomach and intestine
and covering their deeper surface. In each lobe, however, the con-
ditions are different from those in the genera mentioned, in that the
germinative epithelia are in two strands, which are usually entirely
separated from each other except at the ends of the lobes, where they
fuse. In the great majority of cases, especially where the lobes are
superficial, they are so flattened that they are parallel to the surface.
One of the walls thus formed, which is usually the superficial one,
has its median part formed of a cubical epithelium with long cilia, and
on both sides of this are situated the germinative epithelia with the
largest ova away from the median line. Occasionally the germinative
epithelia are broad enough to reach the lateral edges, but these are
usually formed by fehe ova that are just large enough to extend beyond
the wall of the ovary. The other wall is made up of the same kind
of ciliated epithelium, and to it all the follicles of the older ova are
attached. Here again, then, we have the essential structure of
Clavelina and Styelopsis, but developed in each of the lobes of an
irregularly branching ovary. It must be said, however, that although
the structure of the two walls is quite constant, their position with
respect to the surface of the animal is somewhat irregular. The posi-
tion is most constant in the superficial lobes, but even here only about
76% have the wall containing the germinative epithelia superficial,
while in about 14% it occupies the deeper position; and in the other
10% the plane of the lobe is approximately at right angles to the
surface. In the lobes that extend between the stomach and intestine,
and over their deeper side, the variation in orientation is still greater,
depending more upon the adjacent organs than upon the surface of the
animal.

c. General Considerations.

Floderus (p. 250) established the law that in forms with bilateral
arrangement of the germinative epithelia the development within the
same takes place from without inward. In general this holds good
for the three species studied, though some exceptions are encountered.
It would be equally true to extend the law from the germinative

BANCROFT: OVOGENESIS IN DISTAPLIA OCCIDENTALIS. 75

epithelia to the whole ovary, but the case of the deeper lobes of the
Chelyosoma ovary shows that its applicability is limited to ovaries, or
their parts, that have a superficial position. Stated in its most general
form, then, the law of growth within the ascidian ovary of the regu-
lar type seems to be that:

In general, in ovaries or parts of ovaries having a superficial position
and bilateral symmetry, whether the germinative epithelium is double
or not, development takes place from the median line of the superficial
wall towards the deeper parts of the animal.

Another general question is whether the Clavelina type of ovary,
with two separate germinative epithelia, should be considered more
primitive than that of Distaplia and Fragaroides, where a single germi-
native epithelium occupies the greater part of the deep wall. It seems
to me that the condition in Distaplia is more primitive, for it occurs
in smaller and simpler species, and is in itself a simpler condition.
Furthermore, this contention is borne out by the young Styela ovary
(Plate 2, Fig. 12), where the germinative epithelium occupies the
whole deeper wall of the ovary, though it is double in the adult. It
is, however, possible that both the condition in Fragaroides and Styela
may be the result of a secondary fusion; and Julin's work on Stye-
lopsis, where the germinative epithelium is formed from two single
rows of cells extending along the lateral edges of the ovarian funda-
ment, and proliferating primordial ova and oogonia towards the middle
line, favor this view. In Styelopsis the two epithelia thus formed do
not come near enough to fuse, but it may be that they do in other
species. In Distaplia, however, there is certainly no fusion, for
the germinative epithelium is single from the first. There is about
the same number of oogonia in the undifferentiated bud that there is in
the adult, so that the whole ontogeny of these cells consists in rear-
rangement and growth. A few of these oogonia commence to increase
in size at a very early period, and, as these are usually near the middle
of the deeper wall, it is here that the first follicle and consequently
the stalk tissue is formed. But we should expect that occasionally an
ovum near the edge would commence to grow first, and accordingly we
actually do find that in a few cases the stalk tissue is attached near the
lateral edge of the ovary. In fact, from the study of the development
in Distaplia, one gets the impression that the bilateral symmetry is
accidental, depending upon the fact that usually the first ovum to grow
will be somewhere near the centre of the deep wall. But that this
cannot be true for other species, and may not be true in this one, is

76 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY.

shown by the fact that in Clavelina, Ciona, and Styelopsis the
doubling of the germinative epithelium occurs before the oogonia are
formed.

III. Incubatory Pouch.

1. Historical Summary.

Almost all compound and social Ascidians have some provision for
retaining their eggs within the colony until the larvae developed from
them are ready to enter upon their free swimming existence. The
organ where this is accomplished is usually called an incubatory pouch,
but its structure varies much in different groups. Usually the pouch
is merely a slight enlargement of one peribranchial sac, into the
bottom of which the eggs are laid, and from whose top the fully formed
larvae escape. Occasionally, however, as in Glossophorum sabulosum
described by Lahille ('90, p. 203), the pouch consists in the terminal
enlargement of the oviduct. Distaplia and Colella, and probably also
the allied genus Julinia, Caiman ('94), present an extreme develop-
ment of incubatory pouch, which consists of a large diverticulum con-
necting with the dorsal edge of the zooid by a narrow stalk. Within
the enlarged terminal part of the pouch the eggs and embryos are
always arranged so that the youngest are at the bottom; but how they
got there has been somewhat of a puzzle to ascidiologists. Delle Yalle
('81) published the first account of this structure; he supposed that the
ripe ova reached the peribranchial sac and were put into the pouch
together, and then fertilized. The retardation that the spermatozoa
would experience in travelling down the pouch and getting past the
first ova he thought accounted for the greater development of the
embryos near the mouth of the pouch. The next investigator of the
subject was Herdman ('86) , who studied this organ in Colella pedun-
culata and allied species. He describes the pouch (p. 89) as " merely
an enormous diverticulum of the peribranchial or atrial cavity " and
remarks that the neck is so narrow that ova can pass in but that the
larvae cannot escape. He explains the arrangement of the embryos
within the pouch by supposing that the ripe eggs all reach the peri-
branchial sac, are there fertilized and subsequently put into the pouch
in order, the one last fertilized going first. But he says that he has
no evidence for this, as eggs have never been found in the peribranchial
sac. Since then no investigator, so far as I know, has attempted to
account for the arrangement of embryos in incubatory pouchas of the

BANCROFT: OVOGENESIS IN DISTAPLIA OCCIDENTALS. 77

Distaplia type. Caiman ('94, p. 10), however, in describing the
genus Julinia, which is closely related to Distaplia, gives a description
of a spherical vesicle attached to the dorsal side of the thorax which
he believes to be an incubatory pouch, though no eggs or embryos were
found in it. The most significant thing about his account is that the
lumen of the oviduct seems to be continuous with that of the pouch.

2. Observations.

The incubatory pouch in Distaplia occidentalis is not so large as in
some other members of the genus. It usually contains two embryos,
and never more than three, according to my experience, while in D.
magnilarva, according to Delle Valle's figure ('82, Fig. 5), the pouch
may contain as many as eight. The conditions vary in different
colonies; in some the pouches contain but one or two embryos, while
in others three is the predominating number. The pouch is attached
to the posterior dorsal region of the thorax, a little to the right of the
median line, and when fully formed, though still attached to the zooid,
it extends posteriorly about as far into the colony as the zooid itself.

A careful examination of tbe structure of the pouch shows that it is
not merely a diverticulum from the peribranchial sac, but consists of
two parts which, for descriptive purposes, may be called the oviducal
and the peribranchial portions, though I do not know that they have
been developed from the oviduct and peribranchial sac respectively.
The oviducal part is a narrow tube, the anterior end of which connects
with the oviduct, and the posterior end with the bottom of the pouch.
Anteriorly the peribranchial portion is a narrow tube opening into the
posterior dorsal corner of the right peribranchial sac. Posteriorly,
however, it is enlarged to form the pouch proper, in which the develop-
ing embryos are lodged. Figure 8 (Plate 2) represents a young
pouch in which the relation of the oviducal (brs. ov'dt.) to the peri-
branchial portion (brs. pi'brn.) is shown. Both portions are of course
covered by the evaginated ectoderm (ec'drm.). The relations of the
stalk of the pouch to the zooid are best shown by means of cross-sections
(Plate 2, Figs. 9, 10, 11). The most posterior section (Fig. 9)
shows the stalk entirely separated from the zooid and imbedded in
the common test. Its oviducal portion (brs. ov'dt.) is seen to be nearest
to the part of the zooid containing the oviduct (ov'dt.). As the series
is followed anteriorly, the ectoderm of the stalk first joins that of the
zooid and then the wall of the oviducal portion becomes continuous

78 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY.

with that of the oviduct (Fig. 10, ov'dt.). Figure 11 is taken from
another zooid and shows the opening of the peribronchial portion of
the stalk into the right peribranchial sac. As the pouch is completely
separated from its zooid long before the larvae are mature, the only
function of this peribranchial orifice is to serve as a passage for the
spermatozoa. The vas deferens (ya.df.), which is closely applied to
the deep wall of the oviduct until the latter joins the oviducal part of
the pouch, extends much farther forward and opens into the cloaca near
the anus. It is thus seen that the oviducal portion of the pouch is a
continuation of the oviduct into the pouch, and that the egg never
reaches the peribranchial sac at all, but is conveyed directly to the
bottom of the pouch.

The lumen of the oviducal part of the pouch is even narrower than
that of the oviduct, so that here too the egg is greatly elongated in
its passage. Figure 18 (Plate 3) shows an egg the greater part of
which has just entered the pouch; but a small portion of it has. not
quite emerged from the oviducal tube, and can be s^ en as a small pro-
jection from the rest of the ovum. From the fact that there are no
muscles present in the pouch, and none but the heart muscles in the
abdomen, and because the ovum has shrunken away from the walls of
its conducting tube in preserved specimens, I believe that this com-
pression is not due to the killing reagent, but represents a normal
condition.

I have not been able to discover the pouch at the very beginning of
its development, and so cannot say whether it is formed as an evagina-
tion from the peribranchial sac, from the oviduct, or from both. It is
probable, however, that even if the first stages were found it would be
difficult to settle this matter, so that a comparative anatomical investi-
gation of the subject in Colella, for instance, where all grades of pouch
formation obtain, would be more profitable. I have, however, found
the fundament when it was only a short cylindrical outgrowth about
150 /x long. At this stage it presents all the essential features of the
adult except that the pouch proper is very little enlarged. In fact
Figures 9 to 11 (Plate 2) were taken from a stage but little older
than this. Zooids with pouches in this condition are about 2 mm.
long, that is about two-thirds grown. The vas deferens, however, is
filled with spermatozoa apparently ready to be discharged. Subsequent
development consists in a swelling of the pouch, and an elongation of
its stalk, so that the whole structure moves posteriorly into the colony at
some little distance from the zooid producing it. When the bottom of

BANCROFT: OVOGENESIS IN DISTAPLIA OCCIDENTALS. 79

the pouch is at about the level of the stomach, the first ovum enters it;
when the last ovum has entered, its end is about at the level of the
posterior extremity of tbe zooid.

None of the pouches containing the older embryos are ever found
connected to zooids, and as no adult zooids are ever encountered with-
out pouches, it is evident that the zooids from which the older pouches
arose have degenerated, and that those found in the colony along with
these older pouches belong to a subsequent generation. As a matter of
fact, degenerating zooids are not infrequently encountered, and in close
proximity to them are seen the pouches with which they were probably
connected. Most of the growth of the embryos takes place after the
degeneration of their mother; accordingly the pouch is much enlarged,
becoming even longer than the adult zooid was, and the narrow ante-
rior stalk apparently swells up to the size of the pouch proper. The
whole structure migrates through the test until its anterior end finally
comes into close connection with one of the common cloacal cavities,
through which the fully developed larvae reach the exterior.

IV. Envelopes of the Ovum.

So many summaries of this subject have been published, and such a
good one by its most recent investigator, Floderus, that a detailed
discussion of the literature would be superfluous. In fact the whole
subject has been left in a very satisfactory condition by this author,
who was the first to discuss the evidence for the formation of the follicle
cells and test cells, both from within the ovum and from sources which
are external to it. Accordingly, much that I have determined is con-
firmatory of his results, and of those of earlier investigators. How-
ever, as it was from the genus Distaplia that Davidoff ('89) drew what
is probably the strongest evidence for the intraovular origin of the test
cells, it seems worth while to discuss the conditions in this genus.

1. Primitive Follicular Epithelium.

Concerning the development of the primitive follicular epithelium
(van Beneden et Julin, '85, p. 357), I can confirm the account of
Floderus and all the other investigators of the subject since van Beneden
et Julin's time. The epithelium is formed from the primordial follicb
cells found in close proximity to the oogonia in the earliest stages

80 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY.

(Plate l, Figs. 3, 4; Plate 2, Fig. 12, cl. fol. pr.). Even in the
undifferentiated bud (Fig. 1) , there are found among the oogonia smaller
nuclei that will probably give rise to nuclei of follicle cells. As the
oogonium increases in size, these cells multiply and form a continuous
epithelium around it (Plate 1, Fig. 7; Plate 3, Figs. 21, 23, 13).
Julin ('93, pp. 106-109) states that the first follicle is composed of
three cells, one of which is the sister, and the others the products of
division of the cousin of the ovum; and also (p. 123) that the entire
follicle of later stages is derived from these three cells. In our species,
where the ovogenesis is not so rapid, and there is plenty of time for a
rearrangement of the cells, there is no evidence showing that the for-
mation of the follicle is so precise; and the fact that in one instance a
small ovum has been found in the follicle of an older one (Plate 3,
Fig. 26) seems to show that the follicle is made up from any cells that
happen to be in the vicinity of the growing oogonium. ,

The stalks connecting the younger follicles to the germinative epithe-
lium and the older ones to the stalk tissue are, as has already been
shown, differentiations from the germinative epithelium which connect
with the follicle, and are not derived from a limiting epithelium of
the ovary. However, even before the stalk of the follicle is com-
pletely established the differentiation of the primitive follicle into
secondary follicle cells and test cells has begun.

2. Test Cells.

Davidoff, in studying the origin of the test cells in D. magnilarva,
found evaginations of the wall of the germinative vesicle, and con-
cluded that these are constricted off to form the vesicular structures
present in the cytoplasm of the ovum. These vesicles, he says, some-
times appear empty and sometimes have a chromatic spot within them.
Their peripheral membrane stains very lightly, but as they move
towards the periphery of the ovum, they progressively take a deeper
stain and develop chromatic granules until they become the fully
developed nuclei of the test cells. Later their cytoplasm is formed
from portions of the cytoplasm of the ovum. Caullery ('94, p. 600),
who worked on the ovogenesis of D. rosea, does not confirm Davidoff's
results, but says that the test cells arise from the follicular cells by
mitosis. He does not, however, give the evidence upon which this
statement is based.

BANCROFT: OVOGENESIS IN DISTAPLIA OCCIDENTALIS. 81

a. Evidence for the Intraovular Origin of the Test Cells.

I have observed much of the evidence upon which Davidoff bases
his account, but do not consider it strong enough to establish his
contention. As a rule the wall of the young germinative vesicle is
perfectly smooth (Plate 3, Figs. 20-23, 13) , but in quite a number
of cases it is thrown into folds. A rather extreme instance of this
folding is shown in Figure 19 (Plate 3), which, however, represents
well the most characteristic feature, which is that the folds usually
occur at the ends of an oval germinative % vesicle. The long axis of the
germinative vesicle is often parallel to the edge of the knife used in
sectioning, so that in these cases the folding may be due to the shoving
of the section in cutting, but in other cases this cause cannot be in-
voked. Davidoff's figures represent the folding as being most pro-
nounced at the ends of the germinative vesicle, but not limited to that
region. In D. occidentalis, however, I found no cases of such exten-
sive folding as he has shown.

Other structures that probably influenced Davidoff's conclusions,
and are sometimes associated with the folding, are the cytoplasmic
vacuoles near the germinative vesicle (Plate 3, Fig. 19). Oblique
sectioning and a faint stain, such as the borax carmine used by Davidoff
is liable to give, might combine to obscure the conditions so that the
vacuole would appear to be part of the germinative vesicle. It may
be that some of the nuclear evaginations seen by Davidoff were formed
in this way; but, even if what he figures are actually evaginations
from the germinative vesicle, it would not follow that they represented
a normal condition. Thus the presence of the vacuoles (Fig. 19) just
opposite to the mfoldings of the membrane seem to show that both are
artifacts due to some shrinking process.

These vacuoles, however, are certainly some of the structures which
Davidoff has considered nuclear buds separated from the germinative
vesicle. In younger ova, where all of the cytoplasm takes a much
deeper stain, these vacuoles may have a stained periphery and thus
resemble a nucleus very closely (Plate 3, Fig. 22, vac).

The intravitelline bodies, from which Fol ('83), Sabatier ('84), and
Eoule ('85) have derived both follicle and test cells, and Pizon ('93,
pp. 284-290) the test cells only, and the true nature of which
Floderus first elucidated, are also present in Distaplia (Plate 1, Fig. 7;
Plate 3, Figs. 22, 23; Plate 4, Fig. 28, cp. ia'vt.) , though they are
much smaller and occur less frequently than in Ciona. These bodies

82 BULLETIN : MUSEUM OF COMPAEATIVE ZOOLOGY.

are often, though not always, surrounded by a clear area, probably due
to shrinkage, and especially when they are small, the combination
looks very much like the nuclei with pale membranes and central
chromatic spots figured by Davidoff ('89, Taf. 5, Figs. 6, 7). But
the irregular occurrence of the clear area, the lack of a chromatic mem-
brane in the later stages, and the absence of transitions between them
and the test cell nuclei, show that they cannot be nuclei, and have
nothing to do with the formation of the test cells.

b. Evidence for the Follicular Origin of the Test Cells.

The first indication of the formation of the test cells is found in
the form and position of certain nuclei in the primitive follicle.
Whereas at first all the nuclei of the primitive follicle are oval, when
viewed in section, later one is occasionally encountered which js nearly
spherical and projects some distance into the cytoplasm of the ovum
(Plate 3, Fig. 13, cl. fol.). The space between the nucleus and the
egg cytoplasm I do not suppose to be occupied by the hyaline cytoplasm
of the forming test cell, as Floderus does in similar instances (p. 234),
but consider it as due primarily to shrinkage. In some instances I
believe that clear spaces thus formed were partially occupied by cytoplasm
during life; but in these cases I think that there was nothing but the
thinnest possible layer of cytoplasm between the follicle nucleus and
the egg, for in my most satisfactory preparations of this stage, where
the follicular cytoplasm is stained, none is usually encountered in this
place (Plate III., Figs. 24, 25). Here it is seen that the stained
cytoplasm of the follicle is not quite half as thick as the clear spaces
in Figure 13, and that even here shrinkage vacuoles are occasionally
present. They also show that in Distaplia no structureless membrane,
or chorion, has as yet been secreted on the inner surface of the follicle,
the line of demarcation between this surface and the ovum being
merely the cell membrane of the follicle cells.

Figures 24 and 25 are from an ovum that is slightly older than that
shown in Figure 13, and in the height of test-cell production. About
half the follicle cells have rounded up, and either are on their way to
become test cells, or have actually become such. A majority of these
can be seen to be in cytoplasmic connection with the follicle. A
perfect series of the stages in the process can easily be made out, from
the enlarged follicle nucleus that is still in contact with the outer
membrane of that layer, through cases of progressive migration of the
nucleus and its accompanying cytoplasm into the ovum (Plate 3,

BANCROFT: OVOGENESIS IN DISTAPLIA OCCIDENTALS. 83

Fig 24, cl. tst.) to the condition where the test cell is entirely within
the latter, but still connected by a narrow strand with the follicle
(Plate 3, Fig. 24; Plate 4, Fig. 27, cl. tst.'), and finally to that where
it is entirely disconnected from the follicle, and the chorion passes
between the two (Fig. 27).

The way in which the test cells develop here is essentially that
given by Morgan ('90) and Floderus (pp. 234, 235), but the evidence in
the three cases differs slightly. Morgan has differentiated the bounda-
ries between the adjacent follicle cells, which neither Floderus nor
I have found, and has seen the cytoplasm of the forming test cell
extending between the follicle cells quite to the peripheral membrane
of the latter. Floderus finds both the follicle cell and the developing
test cell within a clear space, which he considers hyaline cytoplasm.
I find that in favorable instances the cytoplasm of the follicle and test
cells stains, and the two are seen to be in connection. When, however,
the cell boundary between them can be detected (Fig. 25) , the test cell
cannot be seen to extend quite to the periphery of the follicle, but
lies on its central surface. In this respect Distaplia differs from the
forms studied by Morgan, for the first step in the formation of the test
cell in Distaplia seems to be a displacement of the whole follicle cell
towards the inner surface, and only subsequently does its separation
take place.

Julin (p. 123), in describing the ovogenesis of Styelopsis, says that
there is an almost simultaneous mitotic division of all the follicle
cells, and when the products of this division come to rest, they are
arranged in two layers, the inner of which consists of test cells.
Caullery ('94, p. 600) agrees with Julin in deriving the test cells
from the follicle by mitosis. I have never seen any mitotic figures
in the follicle or test cells of Distaplia occidental is, but do not doubt
that they occur. It is likely, however, that in this species the process
is not accomplished as Julin describes it, for after the mitosis there
must be a differentiation taking place, by means of which the nuclei
of the prospective test cells become larger and more spherical than the
remaining follicle nuclei.

Shortly after the origin of most of the test cells, the chorion is
secreted between these cells and the follicle (Plate 4, Fig. 27); even
this, however, does not destroy the cytoplasmic connections of all the
test cells, and occasionally the union can be seen at much later stages.
With the formation of the chorion, the differentiation of the primitive
follicle into secondary follicle and test cells appears to be complete,

84 BULLETIN: MUSEUM OF COMPAEATIVE ZOOLOGY.

and the large subsequent increase in the numbers of the latter must be
due to their division. There is no reason to doubt that they divide
by mitosis, for at this stage they have a perfectly healthly appearance,
and in fact Davidoff (pp. 131-132) has described and figured mitoses of
both test and follicle cells in D. magnilarva.

c. Degeneration of the Test Cells.

The degeneration of the test cells in Distaplia is a much simpler
process than that occurring in some ascidians, consisting, as Davidoff
has shown, in the vacuolation of the cytoplasm. The cellular nature
of the product is always evident, as the distinctive character of the
nucleus is never lost, though this organ is forced intp a peripheral
position and changes its appearance slightly. This degenerative vacuo-
lation begins in our species at about the same time as the yolk for-
mation (Plate 4, Fig. 28) and is completed (Plate 5, Pig. 30) before
the last changes in the yolk occur. Figure 29 (Plate 4) shows a
few test cells in which the process is just beginning. At this stage the
test cells are often numerous enough to form a double row about the
periphery of the ovum (Fig. 28), but sometimes they form only a
single row. As the ovum passes through the oviduct into the pouch,
it seems to contract somewhat, forcing the test cells out, and the yolk
bodies between which they were formerly situated become arranged so
as to form a smooth surface at the periphery of the ovum (Plate 5,
Fig. 31). However, in spite of this tendency, there is occasionally
enough pressure exerted upon the test cells to force them partly into
the yolk again.

Up to this time the nucleus has changed little if any, but from now
on further changes in the test cells are almost entirely confined to this
organ. During the growth of the embryo, the nucleus loses most of its
chromatin, its membrane only staining, and becomes compressed be-
tween the cell membrane and adjacent vacuoles (Plate 5, Figs. 32, 34).

d. Fate and Function of the Test Cells.

Among the early investigators the belief that the test of the larval
ascidian was formed by the test cells was universal, and it was on this
account that Kupffer ('70, p. 122) applied this name to them. In 1872,
however, Herting ('72) showed that the test first appeared as a thin
layer next to the ectoderm of the larva and entirely within the so-called
test cells, which took no part in the process. This view has been gen-

BANCROFT: OVOGENESIS IN DISTAPLIA OCCIDENTALS. 85

erally accepted from that time ; but quite recently a vigorous supporter
of the older theory has been found in Salensky ('92, '94, pp. 441-449,
'95, pp. 618-621).

This author, finding that in Salpa and Pyrosoma cells which are de-
rivative of the follicle take a large part in the formation of the embryo,
turns to the compound ascidians expecting to find there the beginnings
of the same process. Accordingly, he makes out that in the compound
ascidians the test cells, or kalymmocytes, as he calls them, have a func-
tion to perform during the life of the embryo, though this differs in
different cases. In representative species of the genera Fragarium,
Amarcecium, and Circinalium he found a placenta attaching the embryo
to the wall of the cloaca, and the foetal part of this structure was de-
rived from transformed test cells. In Distaplia magnilarva, Diplosoma,
and Didemnum these cells have the function of forming the test of the
larva, the whole of it being derived from the test cells in Diplosoma, but
only a part of it in the other two genera.

In Distaplia, which is the only genus in which this process need
concern us, Salensky finds that at the time when the test is forming,
the test cells flatten out against the ectoderm of the larva so that
in places they form continuous layers. These he thinks secrete the
cellulose matrix ('92, pp. 113-114) and then metamorphose into the
cells lining the lacunar spaces with which the test is honey-combed.
The test, he thinks, continues to gi'ow in the same way, having layer
after layer of test cells added to its outer surface. Caullery ('94,
p. 600), the only other author who has studied this subject since the
publication of Salenky's first article ('92), comes to the conclusion that
the test cells take no part in the formation of the test, but does not
give any evidence.

In examining Distaplia occidentalis, an appearance was qnite fre-
quently noted that looked very much as if there were unmodified test
cells within the test itself. But when examined more carefully, the
substance within which the test cells were contained was found at stages
before there was any test being formed, and even before the egg had
segmented. It was also found within the cavities of some of the older
embryos, and was then concentrated on the same side of all the cavities.
If in these cases the substance was situated on the left side of the
cavities, then that on the outside of the embryo was always on the
left side and vice versa. Thus there was every reason to believe that
the substance surrounding the test cells was a coagulum that had
been thrown down by the fixing reagents, and had reached its character-

86 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY.

istic arrangement by means of gravity. From Salensky's ('92, p. 113)
statement that in an early stage the changing test cells are separated
from one another by a homogeneous substance that looks very much
like test matrix, I have thought that he might possibly have had some
more refined form of a similar coagulum before him ; but I cannot for a
moment suppose that the true nature of a state of affairs such as
exists in my preparations would have escaped him.

The flattening out of the test cells against the ectoderm and test
mentioned by Salensky, I find to be due to pressure from without.
Wherever the inner follicular epithelium is close to the embryo, as on
its sides, for instance, there the test cells are much flattened against the
surface (Plate 5, Figs. 32, 34, cl. tst.) ; but where the epithelium is
lifted from the surface, as in the triangular space it leaves when passing
over the tail, the great majority of the test cells are spherical. Occa-
sionally, however, oval ones are found here with their sides next to the
ectoderm or test, but in such cases the extreme flattening that occurs
on the sides has never been seen. Moreover, in reviewing quite a
number of series I have been unable to discover any transitions between
the test cells and cells within the test, or any cases in which the periph-
eral layer of the test matrix was not perfectly distinct from the
superimposed test cells. Thus no evidence at all has been found that
is favorable to Salensky's view of the test formation.

So much for the interpretation of Salensky's results. Concerning the
actual method of test development I must say that I think it can be
shown that the test in Distaplia is formed in such a way that the test
cells cannot be instrumental in producing it.

The first tunicin, or animal cellulose, that is laid down is that forming
the tail fin, which is quite well developed at a time when no other
tunicin can be detected either on the tail or the body of the embryo.
Now the test matrix of the fin has no cells within its substance, and so
must have been formed by the activity of cells that are outside of it.
As it is attached to the ectoderm on one of its three sides, and as the
test cells are in contact with only about half of the surface of the other
two, and are usually not pressed against it, and as these cells are in no
way different, from any of the other test cells, there is very little ground
for believing the test matrix of the tail to have been secreted by the
test cells.

On the body of the embryo an examination of the early stages shows
that here the whole thickness of the test is quite small, much less than
the thickness of a test cell, even though the latter is pressed against the

BANCROFT : OVOGENESIS IN DISTAPLIA OCCIDENTALTS. 87

test by the follicular epithelium (Plate 5, Fig. 32, tst.). In spite of this
fact, however, the outer layer of the test matrix is quite thin, and under
it the beginnings of the first lacunae (lac.) are seen. Occasionally we see
cells (Fig. 32, cl. ms'drm.) within the lacunae, but these have none of the
characters of the test cell, whereas they do resemble some of the mesen-
chyme elements very closely. It seems hardly possible, then, that the
test cells can have had anything to do with forming this structure.
Moreover, in following the series through ten sections, 6 /j. thick, not
another cell is found within the test of this region. It is seen therefore
that the first tunicin and some of the first lacunas are formed from the
ectoderm, not only without the intervention of the test cells, but also
without the aid of any cells lying within the test substance.

With further increase in the thickness of the test the lacunae increase
in size, and many mesoderm cells wander into them, and into the matrix
between them (Plate 5, Figs. 33, 34, cl. msdrm.). These cells often
flatten out against the lacunar wall, and line it for a variable distance,
but in the test that is about as thick as the test cells ; this condition is
rarer than at a later period, when there are several layers of lacunae.

During the whole of the development, up to the time when the larva
is extruded, the test cells are often closely pressed against the test ; but
there appears to be no marked decrease in the number of these ele-
ments, though, of course, on account of the greater size of the embryo,
in the later stages they become more widely separated. In this species,
then, the evidence points strongly to the conclusion that the test cells
have nothing whatever to do with the formation of the test, but that
the matrix of the latter, together with its lacunae, is formed from the
ectoderm, and that its cells are mesodermic elements that have subse-
quently wandered in.

What, then, is the function of the test cells in Distaplia, if they have
nothing to do with the formation of the test, and what is the function
of these cells in general] Many authors are not satisfied with their
role in nourishing the ovum, but think that they must be functionally
active as long as they remain near the growing embryo. Thus, Salensky
('92, pp. 109-110) says: " Es wurde vielmehr still schweigend ange-
nommen, die Testazellen hatten keine Bedeutung bei der Bildung des
Embryos der Ascidien. Ein solcher Schluss war fur mich von Anfang
an nicht ganz befriedigend." Floderus, after a careful discussion of the
literature, so far fails to realize the true nature of these cells that he
says (p. 244) : " Ich halte es fur wahrscheinlich, dass die Testazellen
eine Art von rudimentaren Bildungen sind, welche nunmehr eiue un-

88 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY.

bedeutende Kolle spielen, allein einstweilen dtirfte man die Frage nach
ihrer eigentlichea Fimction und ihrer richtigen Deutung gewissermassen
als eine offene bezeichnen konnen."

In considering the function of the test cells, it should always he borne
in mind that they are follicle cells derived from other follicle cells and
as such their mission is to convey nourishment to the ovum. The fact
that they lie under the chorion, and imbedded in the cytoplasm of the
ovum, does not remove them from this category, but merely puts them
in a more favorable position for conveying food to the egg cell. Start-
ing as normal and vigorous cells in the follicular epithelium, their ac-
tivity is so great that degenerative changes in the shape of vacuolation
appear in them, while the follicle cells proper still retain their normal
appearance. The early occurrence and complexity of these processes offer
the best possible evidence for the intensity of the activity. It is not sur-
prising, therefore, aud I think we have no right, a priori, to expect that

cells which have worked so hard that they have lost their vitality

cells in which degenerative changes have set in — should become further
involved in the developmental processes of the embryo.

3. Secondary Follicular Epithelium.

For some time after its complete differentiation this epithelium is
comparatively thin with flattened nuclei (Plate 4, Fig. 27). Soon,
however, as the cells multiply, the epithelium thickens, and the nuclei
assume a spherical shape (Plate 3, Figs. 26, 15, e'th. fol). The in-
crease in cells continues, so that soon they must arrange themselves
in two layers (Plate 4, Fig. 28). At about this stage morphological
differentiation sets in, and the inner nuclei fail to take the stain strongly,
thus becoming somewhat paler than before (Fig. 28, nl. e'th. fol. i.).
All cell multiplication now seems to cease, for with the subsequent
growth of the ovum the thickness of the follicle decreases, probably on
account of stretching (Plate 5, Fig. 30). But in spite of this thinnin^
out the follicle does not become one cell thick, as Davidoff maintains'
(p. 136). On the contrary, there are two quite distinct kinds of nuclei
(Fig. 30), and it is probable that at this stage the inner and outer follic-
ular epithelia are differentiated, but so closely pressed together that the
only indication of the differentiation consists in the nuclei.

4. Inner Follicular Epithelium.

It must be confessed that while the egg is still in the ovary the nuclei
of this layer are hard to find, and do not appear to be numerous. More-

BANCROFT: OVOGENESIS IN DISTAPLIA OCCIDENTALIS. 89

over, in the latest ovarian stages, where the follicle is even thinner than
in Figure 30, they are still less apparent. But later developments show
conclusively that we here have to do with the nuclei of the inner follic-
ular layer ; for when the egg has left the ovary it is covered by only a
thin pavement epithelium with pale nuclei at rather distant intervals
(Plate 5, Fig. 31, eth.fol. t\), while all the darker nuclei of the outer
follicle remain behind in the corpus luteum (Plate 1, Fig. 5 ; Plate 6,
Fig. 47). This inner epithelium persists, and covers the embryo during
the whole of its life within the colony, and becomes so stretched in the
later stages that even the nuclei are obliterated. It is thus seen that,
contrary to what Davidoff believed (p. 138), there is in Distaplia a
splitting of the secondary follicle into an outer and an inner epithelium,
which is actually accomplished only when the egg leaves the ovary,
though the two layers are probably, and their nuclei certainly, differ-
entiated some time before. Thus, in Distaplia, the results of van
Beneden et Julin on the origin and fate of these two epithelia are com-
pletely confirmed.

5. Outer Follicular Epithelium and Corpus Luteum.

As has already been said, one of the distinguishing features of this
layer is the comparatively deep stain that its nuclei take. Another
characteristic, which is just beginning to appear in Figure 30 (Pl»te 5),
but is much more distinct later, is that part of the cytoplasm also takes
a very deep stain. Tangential sections through the follicle of ova that
have reached the maximum size, about 300 p, show this stained cyto-
plasm situated in irregular patches around a central nucleus. In sec-
tions of the newly formed corpus luteum (Plate 6, Fig. 47) this
cytoplasm comes out very distinctly.

The corpora lutea (Fig. 5) are often conspicuous features of the
Distaplia ovary, as they are at first about 80 /x in diameter, and rem-
nants of them persist until an embryo of forty or fifty cells is developed
from the ovum formerly contained within them. When first formed,
this structure is a thick walled vesicle with a rather small cavity open-
ing into the lumen of the ovary. Almost the whole thickness of the
wall is composed of the cells of the outer follicle, which are pressed
together and elongated radially. Figure 47, which represents part of
the wall of a corpus luteum which the hind end of the ovum has just
left, shows most of the deeply stained cytoplasm of each cell located
between its nucleus and the periphery of the organ. Peripherally, this
vol. xxxv. — xo. 4. 3

90 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY.

cytoplasm seems to be attached to the hyaline membrane (inb. prp.)
surrounding the whole structure. This membrane is nothing but
a local differentiation of a similar covering of the ovary and all its
products, the consideration of which has been postponed until now.

Distaplia seems to have a special aptitude for the secretion of struct-
ureless membranes, for they cover the surfaces of nearly all the
visceral organs. In the case of the ovary and its products, the mem-
brane can usually be demonstrated with ease. Both the connective-
tissue cells and the epithelium of the ovary itself seem to participate
in the secretion. Origin from the first source is to be inferred both
from the frequency with which strands of connective tissue are seen to
join and extend along the wall of the ovary (Plate 1, Fig. 6, fun.
eorCt. tis.) ; and also from the fact that often the connective tissue cells
themselves are found on the wall of the ovary. That the membrane is
also in part secreted by the cells of the ovary is proved by its general
occurrence, but principally by the fact that on the periphery of the
stalk tissue, where Ave have the greatest number of cells per unit of
area, it is exceptionally developed, being much thicker than at other
places (Fig. 5, nib. prp.). It occurs also on the surface of the follicles,
varying much in thickness, but being in general best developed on
those that are about half grown. On the surface of the oldest follicles
are found here and there very flat nuclei that are probably derived from
wandering cells almost imbedded in this membrane. It is probably
nuclei of this kind, and the structureless membrane containing them,
that Julin considers to be the remnant of the limiting epithelium of
the ovary surrounding the follicle ; for he says that this limiting epithe-
lium has gradually become hyaline.

On the newly formed corpus luteum this basement membrane (Plate
6, Fig. 47, mb. prp.) is thicker and more conspicuous than on any
other part of the ovary, except some of the stalk tissue, and it is of
special importance in attempting to explain how the ovum is pressed
through the ovary into the oviduct. This pressure cannot he due to
muscles, as there are none in the abdomen except those of the heart.
Nor can an increase of blood pressure in the whole abdomen, due
either to long-continued beating of the heart in one direction, or to
the contraction of thoracic muscles, be invoked in explanation. For
while this might burst the follicle by pressing it against the ectoderm
and modifying its spherical shape, it would also tend to press the
oviduct out flat against the ectoderm, and so could hardly be effective
in moving the ovum into the oviduct. The only other available force

BANCROFT: OVOGENESIS IN DISTAPLIA OCCIDENTALS. 91

is one resident within the follicle of the ovum, and this might lie
either in the outer follicular epithelium or the memhrana propria. I
think that at first both of these structures are the active agents; but
the irregular crowding together of the follicle cells in the corpus
luteum seems to show that, during the later stages of the process, at any
rate, the cells have no active function. Finally, then, by this method
of exclusion, we come to the membrana propria as par excellence the
active agent in forcing the ovum into the oviduct, and find that there
is considerable direct evidence for this view. In the first place, both
follicular epithelium and membrana propria are presumably under
tension in the largest eggs, being formed by the stretching of structures
that were formerly thicker and of less area. Secondly, the membrana
propria of the corpus luteum is perfectly homogeneous and quite thick;
and it is very improbable that a membrane which was formerly very
much thinner, and covered an area about fourteen times its present
extent could assume this condition unless it had been under tension,
upon the removal of which it could contract into its present shape.
Thirdly, it has the appearance of elastic substance, being thrown into
the folds characteristic of such material. These folds, however, may
in part be due to the irregular pull exerted on the membrane by some
of the follicle cells attached to it. When once the ovum is started on
its course up the oviduct, the continuation of its progress seems to be
effected by the contractions of the oviducal epithelium, but the main
force in the initiation of the process seems to be the elastic tension of
the membrana propria of the follicle.

As soon as the ovum has entirely passed out of the ovary, the corpus
luteum begins to degenerate. The first change consists in the disso-
ciation of the deeply stained cytoplasm formerly associated with the
nuclei (Fig. 5). Strands of this material are still encountered, but
the characteristic arrangement is lost. At about the same time the
nuclei themselves appear paler, and the wandering mesoderm cells on
the surface of the structure are flattened against it, and together with
the membrana propria form an investing capsule (Fig. 5). Next
comes the constriction of the stalk of the corpus luteum, by means of
which it is separated from the stalk tissue. During this process the
lumen may or may not disappear, but after the distinct epithelial
connection with the stalk tissue is lost, the corpus luteum remains
connected therewith for some time by irregular strands of cytoplasm or
connective tissue.

Within the isolated corpus luteum many vesicles are formed contain-

92 bulletin: museum of comparative zoology.

ing nuclei in various stages of chromatolytic degeneration, while
outside the vesicles are other nuclei, which, though staining faintly,
appear to be in a healthy condition. It seems, then, that some of the
cells of the corpus luteum retain their vitality and digest their neigh-
bors, for the functional nuclei are so numerous as to make it improb-
able that they have all penetrated the thick membrana propria from
without.

Cases have been encountered where this cutting off from the ovary
has already taken place when the ovum that left the corpus luteum is
in the two or four cell stage. On the other hand, the same condition
has been found when the youngest embryo in the incubatory pouch
contained about sixteen cells. From this time on, the remnant of the
corpus luteum becomes smaller and smaller, still remaining a compact
mass; but finally, when the ovum has divided into about sixty cells,
it disintegrates and disappears.

6. Observations on Styela montekeyensis.

In this species the development of the follicular membranes pro-
gresses in much the same way as in Distaplia, but it is more difficult
to follow. The presence of a separate outer follicular epithelium,
however, is easily seen at a comparatively early stage; but, as is usual
with the simple ascidians, this epithelium is much thinner than the
inner follicle.

The most interesting events in the life of the follicle of this species
are the degenerative changes which take place in the test cells, and
those of the inner follicular epithelium. In the latter the process is
comparatively simple, and consists only in the development of a large
refractive body in the cytoplasm of the cell. It is first seen at a stage
when the yolk is fully formed, and the maximum size of the ovum is
nearly reached, and appears as a small speck in the cytoplasm, so close
to the nucleus that it may have been extruded from the latter. It
rapidly increases in size, until, just before the egg leaves the ovary,
it is nearly as large as the nucleus of the cell and is surrounded by a
lighter area, which may, however, be due to shrinkage. Together
with this area it is about the size of the nucleus, which now occupies
an excentric position. It does not take nuclear stains well, but when
a safranin, gentian violet, and orange triple stain is used, it is colored
an opaque brown, that makes it the most conspicuous object in the
ovary. This body can have no reference to any future function of the

banceoft: ovogenesis in distaplia occidentalis. 93

cell, for, very shortly after the egg enters the water, into which it is
extruded, all of the follicle cells are lost, and the egg is covered only
by the chorion, with a few cellular fragments adhering to its outer
surface. It seems rather to be some product of the intense metabolism
of the cell, Avhich the latter is attempting to get rid of, or to deposit in
some innocuous form.

The degenerative changes in the test cells are much more complex,
ending in a product that looks like Figure 46 (Plate 5) when stained
with a nuclear stain, and like Figures 42 and 43 (Plate 5) when
overstained in safranin. The latter stain brings out the essential
structure more clearly, and by means of it we see that the test cell has
resolved itself into a number of vesicles (vs.), each one of which con-
tains a central refractive, faintly staining corpuscle (cp. a). The num-
ber of vesicles varies from seven or eight to twice that number. They
are remarkably uniform in size, and when well stained it is only very
rarely that one is encountered which lacks the corpuscle, though it
often happens that nuclear stains do not bring out the corpuscles at all
satisfactorily. At the centre of the cell there is accumulated another
substance, which sends out thin lamellae between the vesicles. In
haeruatoxylin preparations (Plate 5, Figs. 44-46), the central sub-
stance appears homogeneous or finely granular, but safranin demon-
strates within it quite a number of very deeply staining bodies that
sometimes occupy almost the whole central space (Fig. 43). Of the
nucleus, not a trace can be seen, for, as will be shown presently, the
central bodies just mentioned are not composed of nuclear material.

The first step in the formation of these complex test cells begins early,
not very long after the formation of the test cells themselves, and
before the yolk granules appear. The first change noticed is an
irregular variolation of the cytoplasm (Plate 5, Figs. 35, 36). With-
in many of these vacuoles are seen bodies (cp. ia'vac.) which look
much like the central corpuscles of the vesicles of the final product, so
that it is most natural to suppose that the vesicles are developed from
the vacuoles of the earlier stage. But the conditions subsequently
encountered do not bear out this supposition.

The next step consists in further vacuolation, which is still more
irregular. Some of these vacuoles are much larger than any of the
vesicles of the final stage (Plate 5, Fig. 37) , and occasionally three-
quarters of the cell will be taken up by an immense vacuole crowding
the smaller ones and their central bodies to one end. Furthermore,
nearly half of the vacuoles have no stained bodies within them,

94 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY.

and these bodies themselves have increased in size, so that instead of
being comparable to the central corpuscle, they are nearer the size of
the entire vesicle of the final stage (Plate 5, Figs. 37, 41). At this
stage there are two kinds of stained bodies, — those within the vacuoles
(cp. ia'vac.) and those included in the lamellae between the vacuoles
(cp. ia'll.). Safranin (Fig. 37) does not allow one to distinguish
between these, but the triple stain of safranin, gentian violet, and
orange does, giving the intra vacuolar bodies a light yellow, and the
intralamellar ones a deep purple color. At a slightly earlier stage
the intravacuolar bodies also take the purple stain, so that it may be
that these two structures have a common origin. The nucleus takes
part in none of these changes. It is occasionally seen in safranin
preparations (Fig. 38, nl.) , and haematoxylin shows it to be practically
unchanged, and universally present at considerably later periods (Plate
5, Fig. 44).

The next and most important stage is that in which some of the
intravacuolar bodies have acquired a central more deeply stained
corpuscle. Their size is about the same as before, but they stain
more faintly (Fig. 38). In the next stage (Plate 5, Figs. 39, 40) all
of these bodies have developed a central corpuscle; they now take a
still fainter stain, and have approximated still more closely in size to
the vesicles of the final stage. In fact they resemble the latter so
closely that most of the transitional stages from one to the other are
found among the test cells of a single ovum. As these bodies acquire
their final dimensions, they assume a peripheral position, crowding the
remains of the intravacuolar cytoplasm towards the centre, in which
are contained the deeply stained intralamellar bodies that treatment
with safranin makes so prominent (Plate 5, Figs. 42, 43).

The nucleus, which from the start of the process has a peripheral
position, persists almost unchanged until the vesicles have become
thoroughly established (Fig. 44). It is still oval, but stains more
faintly than before. Later still, while the vesicles remain entirely
unchanged, the nucleus slowly degenerates, becoming at first paler,
and distorted by the pressure of the vesicles (Fig. 45), and then refus-
ing to take the stain altogether, but persisting as an irregular periph-
eral refractive patch (Fig. 42, nl.) , which finally seems to vanish
entirely (Fig. 46).

This process, like that in the follicle cells, has no prospective mean-
ing, for the test cells take no part in the development of the embryo.
In life they have a decided yellow color in this species, while the

BANCROFT : OVOGENESIS IN DISTAPLIA OCCIDENTALS. 95

yolk is green, so that, in the living embryo, they can be detected
easily, floating in the space between the embryo and the chorion, and
are seen to be left behind at the time of hatching. The process must
rather be considered as entirely degenerative, perhaps a futile attempt
to conform to the changed conditions and protect the cell from the
disastrous results of its active metabolism, by means of which the
ovum is nourished. It seems well to emphasize the intensity of this
metabolism.

V. Ovum.
1. Cytoplasm.

As is usual in ascidians, the cytoplasm of the young Distaplia ova is
filled with rather large granules that take a nuclear stain with avidity
(Plate 1, Fig. 4; Plate 3, Figs. 21, 22), but while the test cells are
being formed the stain becomes fainter and fainter, until it almost
vanishes. Together with the decrease in stainability, there is a
decrease in the size of the granules (Plate 3, Figs. 23, 13), until at
last they become so inconspicuous as to be negligible quantities
(Plate 4, Fig. 27). In later stages again they appear to become
more marked, but it is difficult to correlate the changes with the size
of the ova, for there is so much individual variation. It is clear,
however, that after the cytoplasm has ceased to stain deeply, good
preservation will often enable one to detect a cytoplasmic network,
the nodal points of which give the appearance of granules (Plate 3,
Fig. 26; Plate 4, Fig. 28). This condition comes out more distinctly
in the later stages figured, but indications of the same thing are seen
earlier, and the reticulum probably exists even where it is obscured
by the deep stain. The network persists up to the time of the forma-
tion of the yolk bodies, and is present in the central part of the ovum
when the periphery already contains much yolk (Fig. 28).

In Distaplia magnilarva DavidofF ('89, p. 155) did not find any
stages in the process of yolk formation, and hence concludes that the
yolk bodies are formed simultaneously throughout the whole extent of
the ovum. In our species, however, stages are often encountered
in which the yolk bodies are present at the periphery only (Fig. 28).
Here the largest yolk bodies are scattered amongst the test cells, and
as the germinative vesicle is approached they become progressively
smaller, until near the centre there is a gradual transition into the
cytoplasmic reticulum. However, where the yolk is forming the

96 bulletin: museum of comparative zoology.

whole of the cytoplasm seems to be converted into that substance, for
the intervals between the yolk bodies are not occupied by the cytoplasmic
reticulum, — which at this stage takes a distinct though faint stain, —
hut by a confused mass of substance, all of which has the characteristic
hyaline appearance of the yolk bodies, in which, however, no structures
of definite shape have been formed. After the yolk bodies are fully
established, the absence of undifferentiated cytoplasm between them is
still more marked, for, like Davidoff, I have been unable to discover
trace of any other substance between them (Plate 5, Figs. 30, 31;
Plate 6, Fig. 50). When fully formed, the yolk bodies are of very
large size for ascidian material, and of various shapes. Those appar-
ently first formed are oval or spherical, while those of later develop-
ment are angular and occupy the spaces left by the first ones.

Yolk formation begins when the ovum is about half grown (150 p
in diameter), and all of the cytoplasm is transformed into yolk some
time before the maximum size (about 300 /x) is reached. During the
growing period that follows the earliest stage at which all the
cytoplasm is converted into yolk, there is a little clump of yolk around
the germinative vesicle which has but indistinctly broken up into yolk
bodies (Plate 6, Fig. 49), and it may be from the periphery of this
clump that the new yolk bodies producing the increase in size are
formed. Ova possessing this clump of undivided yolk average about
200 fx, in diameter.

Later this clump becomes completely broken up into yolk bodies
(Fig. 50). But none of these are spherical, as is usually the case in
other regions, all being more or less elongated in a radial direction.
It is probable that growth continues even after this central yolk has
broken up, for the range in size of ova in this condition is from 230 //,
to 326 p. It is impossible to determine the diameter of the mature
ovum, because, as soon as it begins to leave the ovary, it is immediately
deformed, and even in the pouch it is oval instead of spherical. But
it is not probable that the mature ovum would exhibit so much
variability.

2. Germinative Vesicle.

The chromatin in the young germinative vesicle is usually situated
next to the membrane, while in the centre no structural elements of
any kind can be detected (Plate 1, Fig. 4, ov'go.). The chromatic
granules are sometimes of uniform size (Fig. 4, ov'go.), but usually
there is one which is much larger than the rest, and destined to become

BANCROFT: OVOGENESIS IN DISTAPLIA OCCIDENTALIS. 97

the nucleolus (Fig. 1, ov'go.). At this period all of the granules,
including the prospective nucleolus, are differentiated similarly — taking
a bluish tint — with the methyl green and acid i'uchsin double stain.
Soon achromatic fibres make their appearance. At first mainly periph-
eral, and inserted on the chromatic granules and the membrane,
they soon come to traverse the centre of the vesicle. It is rarely,
however, that chromatin is found upon them (Plate 1, Fig. 3, ov'go.).
At the same time that the achromatic network becomes differentiated,
the largest chromatin granule increases in size, and becomes more
nearly spherical and not so closely approximated to the membrane.
Although it still takes the same stain as the other granules, the
disparity in size and its regular shape justify us in calling it the
nucleolus. Its later development will be discussed in a separate
section.

The further changes that occur in the growth of the rest of the
germinative vesicle consist entirely in the elaboration of the achromatic
network and the chromatin granules upon it. The achromatic fibres
are not always seen, especially in the later stages, but staining in iron
haematoxylin brings out the chromatic elements very distinctly (Plate
3, Fig. 13; Plate 4, Fig. 27; Plate 6, Fig. 48). The latter forma
reticulum extending through all parts of the germinative vesicle, the
complexity of the structure increasing with the size of the vesicle, and
culminating in the condition shown in Figure 48 (Plate 6).

In addition to the reticulum, many isolated granules are found
throughout the vesicle and especially on its membrane. While all
these granules when stained with iron haematoxylin are very deeply
colored, and resemble closely the chromatic structures described for
other eggs, still, strictly speaking, we are not justified in calling them
chromatin; for chromatin is that substance which takes the chromatic
or basic aniline stain when treated with a combination such as methyl
green and acid fuchsin. Malfatti ('91), in testing the electivity of
this stain for known chemical substances, found that free nucleic acid
was stained pure green; nucleins, which contain less phosphorus, were
colored blue, while substances with still less phosphorus took the red
stain. Thus, as nucleic acid and the nucleins are derived from the
nucleus, it becomes highly probable that chromatin is made up of
these substances.

Now, when the germinative vesicle of this stage is treated with the
stain mentioned, the network that is so distinct in hsematoxylin prep-
arations is colored red, and the only substance staining green is con-

93 bulletin: museum of comparative zoology.

tained within the nucleolus, which will be considered later. During the
growth of the nuclear reticulum, then, there appears to be a concentra-
tion of nucleic acid within the nucleolus, while the reticulum itself
and all the other granules within the vesicle suffer a decrease in their
percentage of phosphorus.

The condition of the germinative vesicle shown in Figure 48
(Plate 6) is found in ova about 150 fx in diameter, in which the yolk
bodies are just beginning to be formed. During all of the subsequent
growth of the ovum, while it doubles its diameter, the germinative vesicle
decreases in size, usually diminishing to a little less than half its former
diameter (see Table, p. 100). During this shrinking, the vesicle exhibits
the wavy and stellate outline characteristic of the stages preceding the
maturation of ova (Plate 6, Figs. 49, 50) ; here, however, the shrivel-
ling is associated not with maturation but with yolk formation.

This shrinking of the germinative vesicle and the formation of the
maturation nucleus from it have been studied by Davidoff (pp. 156-159),
according to whom the process is very complex and quite unique. First,
the membrane disappears, and the major portion of the contents of the
germinative vesicle forms the " ergoplasma," or active cytoplasm, which
gradually becomes disseminated between the yolk bodies. The nucleo-
lus is the only structural element that remains behind, and it under-
goes the most complex modifications. It takes a deep stain, and becomes
shrivelled so that its outline is irregular and at times stellate. He says
he believes that there is an actual decrease in the volume of the nucleo-
lus accompanying these morphological changes, but his figures do not
bear him out in that statement. The next change is, that within the
shrivelled nucleolus small chromatic granules are differentiated, which
soon aggregate to form a compact chromatic body. This itself sub-
sequently becomes vesicular and contains a central granule. A few- of
the chromatic granules do not take part in the formation of the compact
chromatic body, but initiate the formation of a dense chromatic network,
with which the whole nucleolus becomes filled. At the same time a
membrane is distinctly differentiated about the periphery of the nucleo-
lus and we have developed from it a nucleus which is neither germinative
vesicle nor nucleolus, but the maturation nucleus or " Polkern."

In Distaplia occidentalis, most of the stages discussed by Davidoff are
abundantly represented; but I must differ from him in their interpreta-
tion and in the order of their sequence. I will first describe the process
as I conceive it to be, and then compare my results with those of
Davidoff.

BANCROFT: OVOGENESIS IN DISTAPLIA OCCIDENTALIS. 99

At the time when the first yolk is being formed, the germinative
vesicle has a full rounded outline and its maximum diameter of about
45 fji (Plate 6, Fig 48). As the formation of the yolk continues, and
the ovum grows, the germinative vesicle shrinks, its membrane becomes
wavy, and before all of the cytoplasmic reticulum has become trans-
formed into yolk, the vesicle has shrunk to a diameter of about 30 or
35 p. In the next stage (Fig. 49), the yolk extends up to the membrane
of the germinative vesicle, but the central yolk clump is not distinctly
separated into the yolk elements. In ova of this stage the germinative
vesicle has an average diameter of about 25 p. Finally (Plate 6,
Fig. 50), when the ovum is larger still and the central clump of un-
divided yolk has broken up, the germinative vesicle has shrivelled to a
diameter of about 20 jx (see Table, p. 100).

During the whole of this process, the methyl green and acid fuchsin
stain differentiates the nucleolus within the shrivelling germinative
vesicle. "With some other stains however it cannot be seen, for the ger-
minative vesicle is so deeply stained that no sti'ucture at all can be dis-
tinguished within it. Iron hematoxylin usually has this effect, and the
stain is so intense that excessive decoloration does not differentiate a
vesicle, but works from the periphery inward, removing all the stain from
around the edges, while in the centre it persists with undiminished in-
tensity. Occasionally, however, colonies are encountered in which the
stain does not work this way, but differentiates a vesicle, and in these
cases (Plate 6, Figs. 49, 50) the nucleolus and the remains of the retic-
ulum are plainly seen.

In order to make perfectly sure that the sequence noted above is cor-
rect, I established four classes, based upon the sequence mentioned, and
then measured all the eggs in these stages contained in several series,
placing each egg in the class to which it belonged according to the
structure of the yolk, and germinative vesicle. If my sequence is cor-
rect, by this means I should get a progressive increase in the size of the
ova in the successive classes, and when passing from one class to the
next, and corresponding to this a diminution in the size of the germina-
tive vesicle. This condition is actually what I did obtain.

The classes were : —

1. Yolk bodies formed at the periphery only. Germinative vesicle
with a full rounded outline.

2. Central part of the reticular cytoplasm not yet differentiated into
yolk. Germinative vesicle with a wavy outline.

3. Yolk extends from the periphery to the germinative vesicle, but

100

BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY.

the central clump has not yet broken up into yolk bodies. Germina-
tive vesicle stellate.

4. Central yolk clump has broken up into yolk bodies. Germinative
vesicle stellate.

Out of 48 eggs measured, the various sizes, expressed in micra (p),
were distributed as follows : —

Class.

No. of Ova

in
each Class.

. Smallest Ovum.

Largest Ovum.

Average.

Size of
Ovum.

Dimensions of
Germ. Ves.

Size of
Ovum.

Dimensions of
Germ. Ves.

Size of
Ovum.

Dimensions of
Germ. Ves.

1

2
3
4

10

6

15

17

133
134
170
230

56 X 50
46 X 26
31 X 30
28 X 25

173
191
243
326

52 X 42
33 X 25

28 X 15
20 X 19

151

161
205

287

47 X 37
36 X 28
28 X 20
23 X 17

Total

48

From this table, it is seen that the classes established on morphologi-
cal grounds contain groups of similar ova, and hence represent the actual
sequence of events. The column of averages especially shows well
how the germinative vesicle diminishes with the growth of the ovum.

Among the fifty-two ova examined, four exceptions were encountered
that would not fit well into any of the classes. Two of these were
small ova, about on the transition line between the second and third
classes, but in which the cytoplasmic conditions were very indistinct.
The other two belong to an entirely different category, and from their
rare occurrence in my early series, from which the above data were
taken, as well as in the later series, they must be considered abnormal.
In them the yolk bodies were completely differentiated in all regions,
but the contents of the germinative vesicle was finely granular and not
separated by a membrane from the yolk.

From this account it follows that the shrinking of the germinative
vesicle is a continuous process, and much simpler than Davidoff con-
sidered it. In attempting to interpret his results in the light of these
investigations, there is but one point that offers any obstacle. That is
his initial stage in the shrinking, in which the membrane of the
vesicle disappears and its contents penetrate between the neighboring

■

BANCROFT: OVOGENESIS IN DISTAPLIA OCCIDENTALS. 101

yolk bodies; while the nucleolus remains behind (Davidoff s Figures
18, 19, Taf. V.). It will be seen, however, that this condition is
exactly what I have found in the two abnormal ova mentioned above.
It may be that this appearance is more common in the species studied
by Davidoff than in the one with which I am familiar, and this stage
may possibly have some place in the normal series of events in that
species, but the close similarity of his other stages to mine does not
favor this view.

Davidoff's ergoplasm that in some stages surrounds what he takes to
be the nucleolus, but finally penetrates between all the yolk bodies, is,
I think, the central clump of yolk surrounding the germinative vesicle
in class three. When stained faintly, it sometimes looks slightly
granular. But so far as the penetration of the yolk by this substance
is concerned, I cannot confirm his results; for in the later stages, as in
the earlier, I find no interstitial substance whatever between the yolk
bodies.

The body that Davidoff takes to be the nucleolus in the later stages
is undoubtedly the shrunken germinative vesicle. In size, shape, and
in all other respects, the two objects are entirely similar. Curiously
enough Davidoff follows quite a number of these stages backwards, in
spite of the fact that he thinks that his " nucleolus " is shrinking, and
that his figures show the size of that structure to be increasing in the
successive stages. The central body which he finds developed within
his nucleolus is nothing but the true nucleolus of the vesicle, which I
have traced through all the stages, but which would often be obscured
by the stain he used. His main trouble consists in not having had
before him any of the first stages in the shrinking of the vesicle when
its membrane and nucleolus are easily seen. On this account also
he missed the entire process of the transformation of the cytoplasmic
reticulum into yolk, which occurs at the same period.

That this shrivelling of the germinative vesicle is closely associated
with the formation of the yolk, is suggested by the synchronism of the
two events, and it is interesting to note that the intense activity of
the nucleus, of which this shrinking is probably the result, begins
shortly after degenerative changes have commenced in the test cells.
It seems, then, that the test cells are particularly active in conveying
nourishment to the ovum in the early stages, whereas the nucleus
exerts its principal activity in the later stages in converting this
material into yolk.

102 BULLETIN : MUSEUM OF COMPAKATIVE ZOOLOGY.

3. Nucleolus.

The nucleolus was left at the time when it had just assumed a nearly
spherical shape, but was still attached to the membrane of the germinal
vesicle. It progressively assumes a more central position, but remains
attached to the membrane for a considerable period by means of a stalk
(Plate 3, Figs. 20-22). This stalk is at first quite large, but be-
comes smaller as the nucleolus increases in size, and finally disappears
altogether, never being encountered during the later stages. It cannot
be seen in all the young vesicles, and would of course be invisible except
in a profile view; but from the generality of its occurrence I believe
it is a normal structure serving to attach the nucleolus to the mem-
brane from which it has been derived. Floderus (pp. 214-215) has
described nucleoli with similar projections attached to them, and
believes that he has in them stages in the formation of " nebennucleoli,"
but the fact that the largest stalks are attached to the smallest nucleoli,
which is also brought out by his figures (Taf. X., Figs. 19, 20, 21),
has no significance according to this view.

When first differentiated, the nucleolus appears homogeneous with
all the stains used; soon, however, this condition changes and a cen-
tral medullary mass and an enclosing membrane can be distinguished.
In the most favorable haematoxylin preparations (Plate 6, Fig. 51) the
membrane takes a deep stain, the medulla (med.) appears granular and
more refractive, and outside of the membrane a faintly staining cortex
(ctx.) can be distinguished. With methyl green and acid fuchsin the
medulla is stained red, and is surrounded by a blue rim that is prob-
ably composed of both the membrane and the cortex seen in the
haematoxylin preparations. The next changes consist in simple
growth, the medulla, surrounded by its membrane, coming to occupy
an excentric position within the cortex, which is increasing more
rapidly than the medulla (Plate 6, Figs. 52, 53). Next we have
developed within the medulla, and often very close to its membrane, a
variable number of very highly refractive bodies (cp. ref.~). At the
same time the rest of the medulla becomes less refractive and more
finely granular (Plate 6, Figs. 54, 55). With the methyl green
and acid fuchsin combination, the cortex and medulla are with diffi-
culty distinguishable, both taking a pale bluish green stain; but the
refractive bodies are colored a very bright pure green. They are
therefore chromatin and apparently pure nucleic acid. At this stage,

BANCROFT: OVOGENESIS IN DISTAPLIA OCCIDENTALIS. 10

o

and for some time later, they are the only structures in the nucleus that
take a decided chromatic stain.

After further increase in size, hematoxylin stains the cortex deeper
and deeper, until it is harely possible to see the medulla and the
refractive bodies (Plate 6, Fig. 56). Later this deep staining is
still more pronounced, so that both medulla and refractive bodies are
entirely obscured and only a faint lighter central area perceived (Plate
6, Fig. 48). During the earlier stages also iron haematoxylin usually
stains the nucleolus so strongly that no structure can be detected within
it; and only exceptionally are the conditions illustrated in Figures 51
to 56 (Plate 6) to be made out. In such instances, however, they
may be brought out most distinctly. But with methyl green and acid
fuchsin, medulla and cortex are usually differentiated, though not so
clearly as with a favorable hematoxylin stain.

For the further history of the nucleolus I have depended almost
entirely upon the methyl green and acid fuchsin combination. All of
the structures increase in volume, the refractive bodies probably most
rapidly, until at the time when the yolk bodies are being formed some
are considerably over 1 fx in diameter. They still stain a bright green,
and are very refractive, but the largest are no longer homogeneous, ap-
pearing to be filled with many refractive granules (Plate 6, Fig. 58).
Though usually located in the medulla, some are occasionally found
within the cortex (Fig. 58). The cortex itself appears perfectly
homogeneous, while the medulla is finely granular, but both take the
same bluish tint. The medulla is no longer spherical, but flattened on
one side.

From now on, while the germinative vesicle is shrinking, and after
the yolk is formed, both cortex and medulla take a fainter and fainter
stain, until finally they can no longer be distinguished with the use
of this stain. During this process they usually are of a diffuse green
color, and sometimes the vicinity of the refractive bodies, which remain
with unimpaired distinctness, has a faint suffusion of green, which is
probably an indication of the presence of what is left of the nucleolus,
though its outlines cannot be seen. After the central yolk clump
has broken up, this stain has never shown anything of the nucleolus
except the refractive bodies; but I think that the other portions are
still there, for in the occasional well-stained haematoxylin preparations
of this stage it is still seen (Plate 6, Fig. 50), though in this case
too it is fainter than in the stages immediately preceding.

In addition to the stains already mentioned, I also tried List's

104 BULLETIN: MUSEUM OF COMPAEATIVE ZOOLOGY.

('96, pp. 480-487) potassium ferrocyanide methods in order to deter-
mine whether any of the structures in the nucleolus were composed of
the paranuclein demonstrated by these methods. List says (p. 488),
" Werden die Schnitte eine halbe Stunde lang in eine ganz schwache,
mit Salzsaure angesauerte Eisenchloridlo sung gestellt
(zu 50 ccm destillirtem Wasser gehe man 10 Tropfen einer 0,5% igen
Eisenchloridlosung und 5 resp. 15 Tropfen einer 1% igen Salzsaure),
und wird dann die Berlinerblaureaction ausgefiihrt, so bleiben
jSTucle'in und die verwandten Stoffe farhlos,die Suhstanz
des Nebennucleolus dagegen wird blau. Der name Paranu-
clein, gleichsam als Gegensatz zu Nuclein, ist chemisch daher vollkom-
men gerechtfertigt. " Both his first method, which consists in treating
the sections directly with two drops of 1.5% potassium ferrocyanide,
followed by one or two drops of hydrochloric acid, and his second
method, in which the sections are treated first with the mordant men-
tioned in the quotation above, were employed, and similar results
obtained. The refractive bodies were the only structures that took a
decided blue stain, and hence would be called paranuclein, according
to List. However, not all the refractive bodies in the same section
would be colored blue, but usually only the smaller homogeneous ones.
After treatment with potassium ferrocyanide, the sections were stained
in Mayer's hydrochloric acid carmine, and the principal result of the
cyanide method was to prevent the nucleoli taking the carmine stain.
The cortex was acted upon least, and usually took a light, but some-
times a rather deep, stain; but in the nucleoli in which medulla and
refractive bodies were differentiated, neither had a red color. They
were usually greenish yellow. In the younger ova, with nuclei about
20 fj. in diameter, the nucleoli were almost entirely colorless, having
but a faint bluish or greenish yellow coloration. The principal effect,
then, of List's method is to prevent subsequent coloration of the
nucleoli by carmine; but it also stains a few and sometimes all of the
refractive bodies blue. These bodies, then, are paranuclein, but we
have already seen that, with methyl green and acid fuchsin, they are
often the only structure in the nucleus that stains bright green, and
hence they are nuclein or nucleic acid. Which is correct? It may be
that the refractive bodies represent some chemical substance that has
not yet been tested by these reagents, and does not belong in the same
category as the nucleins and paranuclein. This, however, is hardly
probable, and I think that of the two methods the methyl green and
fuchsin is much more reliable on account of the chemical tests to which

BANCROFT: OVOGENESIS IN DISTAPLIA OCCIDENTALS. 105

it has been subjected; and believe that the refractive bodies are really
composed of nuclein or nucleic acid.

Of course the most interesting problem connected with the nucleolus
is whether the refractive bodies contain any of the chromatin that goes
to form the chromosomes of the maturation nucleus, but on account of
the small number of ova that I have found in the oviduct, and the
difficulty experienced in making perfect series of sections, I am unable
to offer any direct evidence on the subject. One series from an ovum
within the oviduct shoAved three of these bodies situated within the
remains of the disintegrated germinative vesicle, but no other distinct
structural elements were detected ; and when the ovum has reached the
pouch, the tetrads (Plate 6, Figs. 60, 61) are already formed, and no
trace of the nucleolus remains. On the whole, it seems rather improb-
able that the refractive bodies should form the chromosomes, and in
the latest ovarian stages there is other chromatic material in the ger-
minative vesicle. Occasionally the whole vesicle will be suffused with
a faint green coloration, and sometimes very minute green microsomes
are seen on the projections of the shrivelled vesicle. Accordingly, at
this stage, there is not the entire abs3nce of other chromatic material
that is encountered earlier.

4. Maturation.

The condition obtaining among the ova of Class 4, where the yolk
bodies are completely formed and the germinative vesicle much shriv-
elled, stellate, and about 20 /jl in diameter, is the last stage that I
have observed in the ovary. The next changes take place in the pas-
sage through the oviduct, during which the tetrads are formed.

A few observations have been made on these and subsequent stages,
but, owing to the difficulty of obtaining perfect series, they are so
scattered that they cannot be easily interpreted, and a discussion of
them would be unprofitable. Enough, however, has been seen to show
that the number of tetrads in the maturation nucleus is probably
twelve, and that two polar bodies are formed. No centrosome or
achromatic fibres of any kind have been detected either during matura-
tion or the first cleavage of the ovum, although I have several good
preparations of these stages. But, as before mentioned, the number of
figures is so small, and the variability in their appearance so great, that
as yet no connected history of the processes can be made out.

Cambridge, May, 1898.
VOL. XXXV. — NO. 4. 4

106 bulletin: museum of comparative zoology.

POSTSCRIPT.

Since the above was written, Professor Salensky has had the kindness
to afford me an opportunity of examining his sections of the embryos of
Distaplia magnilarva. As these preparations show a class of facts, not
observed in mine, which is of much importance in considering the origin
of the cells in the test, their consideration is imperative. But before I
proceed to discuss the differences, I will mention two points in which
our preparations agree.

1. What I have said about the flattening of the test cells against the
ectoderm (p. 86) applies equally well to those of Salensky's prepara-
tions that I have seen. The flattening appears to be entirely due to
the pressure of the embryonal membranes. Wherever the test cells
were flattened and the follicular membrane could be made out at all,
the membrane was seen to be pressed against the surface of the embryo ;
and, conversely, wherever the follicular membrane was seen to lie at
some distance from the ectoderm, as where it passes over the tail, the
test cells were never seen flattened against the ectoderm.

2. The granular substance sometimes found between the test cells in
D. occidentalis is also present in D. magnilarva. It seems, in this case
also, to be an artifact, being likewise found within the cavities of the
embryo, and principally on one side of the embryo. But it is more
finely granular and less one-sided in its distribution in D. magnilarva
than in D. occidentalis.

The diffei'ences between Salensky's preparations and mine are con-
nected with cells situated entirely outside the test-matrix, but having
a structure intermediate between that of the cells in the test and the
" test cells " (Salensky's kalymmocytes). Whereas, in my preparations,
I coidd find no intermediate stages, in Salensky's quite a perfect transi-
tion could be traced from cells that could not be distinguished with
certainty from typical kalymmocytes to those looking just like the
characteristic vacuolated cells of the test. But among the cells on the
outside of the test-matrix the series may be traced even farther, even
to the characteristic undifferentiated mesoderm cells with large nuclei
atid very scanty compact cytoplasm. These facts, and othei's to be
mentioned presently, have forced upon me the conclusion that in
D. magnilarva the mesoderm cells wander through the ectoderm, not
only into the test, but also on to the surface of the embryo, and there
undergo a degenerative vacuolation, the final result of which cannot be dis-

BANCROFT : OVOGENESIS IN DISTAPLIA OCCIDENTALS. 107

tinguished from a kalymmocyte. The facts for and against this interpre-
tation may be briefly stated as follows : —

1. The fact that, among the cells entirely outside the embryo, the
series may be traced not only to the typical vacuolated cell of the test,
but also to the undifferentiated mesoderm cell, tells against Salensky's
view that the kalymmocytes become rejuvenated and help to form the
test. For, if his view were correct, one would expect that the rejuve-
nation of the kalymmocytes would cease as soon as they had reached
the structure of the typical cell of the test, and that the concentration
of the cytoplasm would not go farther, and pi-oduce cells not to be dis-
tinguished from undifferentiated mesoderm cells. Undifferentiated meso-
derm cells are found in the test in comparatively small numbers ; and,
as it is certain that at least some of them wander through the ectoderm
in the undifferentiated condition, and later undergo vacuolation to form
the typical cells of the test, it seems probable that sooner or later all
the undiffei'entiated mesoderm cells within the test do the same. If
this is true, then the undifferentiated mesoderm cells found on the out-
side of the test are, according to Salensky's theory, destined to reverse
their development once more and become again moderately vacuolated,
after having already passed through that stage twice. Of course the
considerations adduced in this paragraph do not prove my point, for it
may be replied that, as the participation of the kalymmocytes in the
building up of the test implies a rejuvenation, it is not at all surprising
that some of the cells should become more rejuvenated than is neces-
sary. The above considerations, however, show that the presence of
undifferentiated cells on the outside of the test adds a further compli-
cation to Salensky's theory.

2. These undifferentiated mesoderm cells are found in such positions,
and associated with each other in such a way, that it is difficult to see
how they can have been formed from kalymmocytes. Thus, in one case,
in an embryo in which none of the cellulose test-matrix had been secreted,
an undifferentiated mesoderm cell was found within the cytoplasm of a
typical kalymmocyte, whose shrivelled nucleus was still distinctly visi-
ble. Near by, but still imbedded in the ectoderm, was another mesoderm
cell of exactly the same appearance, as if to show the course which the
first cell had traversed.

In several other places on the surface of the same embryo isolated
undifferentiated mesoderm cells were found. The majority, however, of
the mesoderm cells thus situated were aggregated into groups. In one
case a group of three was encountered, and in another a group of eight

108 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY.

or nine, a part of which was contained in each of two sections. If these
undifferentiated mesoderm cells had been developed from kalymmocytes,
we should not expect to find them thus massed together, but, instead,
distributed at intervals approximately corresponding to the distribution
of the kalymmocytes about the embryo. Nor should we expect to find
them penetrating other kalymmocytes. It must be admitted, however,
that neither of these facts is wholly inconsistent with Salensky's theory.

3. Undifferentiated mesoderm cells were found on the surface of
younger embryos, such as had not yet secreted any test-matrix, as well
as on the surface of older ones possessing a test of considerable thick-
ness. But while in these older embryos there were transitions to cells
that were highly vacuolated, like the kalymmocytes, in the younger ones
these transitions were entirely absent. This is exactly what should be
expected according to my interpretation, for the cells on the younger
embryos have just wandered out through the ectoderm, and have not
yet had time to become vacuolated, as they have in the older embryos.
I do not see how, according to Salensky's view, these facts can be ex-
plained, for if the undifferentiated mesoderm cells have been derived from
the kalymmocytes, the transition stages should be present in both cases.

4. The strongest objection to my theory, and one that will immedi-
ately present itself to every one, is the great improbability that cells
which have had such a different history as typical mesoderm cells and
kalymmocytes should both end in structures that cannot be distin-
guished from each other. It must be remembei'ed, however, that both
are of mesodermic origin, and that both are subjected to the same envi-
ronment. Improbable as such a convergence may seem, it does, how-
ever, occur in D. magnilarva ; for, in Professor Salensky's sections, I
have seen in the body cavity cells, certainly not displaced by the knife
in sectioning, which were vacuolated in such a way that only with diffi-
culty could they be distinguished from kalymmocytes, — one, indeed,
that had exactly the appearance of a kalymmocyte. The a priori im-
probability of such a convergence in development is thus shown to be
without weight. On the whole, then, I think it must be said that, in
spite of the series of transitions from kalymmocytes to cells resembling
those in the test (D. magnilarva), this series does not prove that the
kalymmocytes participate in the formation of the test.

Berlin, March, 1899.

BANCROFT: OVOGENESIS IN DISTAPLIA OCCIDENTALS. 109

BIBLIOGRAPHY.

*

Auerbach, L.

'96. Uutersuchungen iiber die Spermatogenese von Paludina vivipara. Jena.
Zeitschr. Naturwiss., Bd. 30, Heft 4, pp. 405-554, Taf. XXL, XXII.

Beneden, E. van, et C. Julin.

'85. Recherckes sur la morphologie des Tuniciers. Arch. Biol., Tom. 6,
Fasc. 2, pp. 237-476, PL VII.-XVI.

Caiman, W. T.

'94. On Julinia ; a new Genus of Compound Ascidians from the Antarctic
Ocean. Quart. Jour. Micr. Sci., Vol. 37, Pt. 1, pp. 1-17, PL L-III.

Caullery, M.

'94. Sur les Ascidies composees du genre Distaplia. Comp. Rend. Acad. Sci.
Paris, Tom. 118, No. 11, pp. 598-600.

Caullery, M.

'95. Contributions a. l'etude des Ascidies composees. Bull. Sci. Fr. Belg.,
Tom. 27, Fasc. 1, pp. 1-158, PL I.- VII.

Dall, W. H.

'71. Descriptions of sixty new forms of Mollusks from the West Coast of
North America and the North Pacific Ocean, with notes on others already
described. Amer. Jour. Couch., Vol. 7, pp. 93-160, PL XIII.-XVI.

Davidoff, M. von.

'89. Uutersuchungen zur Entwickluugsgeschichte der Distaplia magnilarva
Delle Valle, I. Die Reifung des Eies. Mitth. Zool. Stat. Neapel, Bd. 9,
Heft 1, pp. 113-178, Taf. V., VI.

Drasche, R. von.

'83. Die Synascidien der Bucht von Rovigno (Istrien). Eiu Beitrag zur
Fauna der Adria. Wieh, 1883.

Delle Valle, M. A.

'81. Nuove contribuzioni alia Storia naturale delle Ascidie composte del Golfo
di Napoli. Mem. Accad. Liucei, (3) Vol. 10, pp. 431-498, Tav. I.-X.

Delle Valle, M. A.
'82. Distaplia, nouveau genre de Synascidie. Arch. Ital. Biol., Tom. 1,
Fasc. 2, pp. 193-203, 1 PL

110 bulletin: museum of compaeative zoology.

Fewkes, J. W.

'89. New Invertebrata from the Coast of California. Bull. Essex lust.,
Vol. 21, Nos. 7-9, pp. 99-146, PI. I.-VII.

Floderus, M.

'96. Ueber die Bildung der Follikelhullen bei den Ascidien. Zeitschr. f. wiss.
Zool., Bd. 61, Heft 2, pp. 163-260, Taf. X.

Fol, H.

'83. Sur l'ceuf et ses enveloppes chez les Tuniciers. Bee. Zool. Suisse,
Tom. 1, Fasc. 1, pp. 91-160, PI. VII., VIII.

Heiden, H.

'93. Ascidise aggregatse uud Ascidiee compositae von der Iusel Menorca.
Zool. Jahrb., Abth. f. System., Bd. 7, Heft 3, pp. 311-361, Taf. XIII.

Herdman, W. A.

'86. Report on the Tunicata. II Ascidise Compositse. Challenger Rep., Zool.,
Vol. 14, pp. 1-432, PI. I.-XLIX.

Hertwig, O.

'72. Uiitersuchungen iiber den Bau und die Entwickelung des Cellulose-
mantels der Tunicaten. Jena. Zeitschr. Naturwiss. Bd. 7, pp. 46-73,
Taf. IV.-VI.

Huitfeld-Kaas, H.

'96. Synascidiee. Norwegian N.-Atl. Exped., Vol. 23, Zool., Tunicata, pp.
1-27, PI L, II.

Julin, C.

'93. Structure et developpement des glandes sexuelles ; ovo-genese, spermato-
genese, et fecondation chez Styelopsis grossularia. Bull. Sci. Fr. Belg.,
Tom. 25, Fasc. 1, pp. 93-154.

Kupffer, C.
'70. Die Stammverwandschaft zwischen Ascidien und Wirbelthieren. Arch.
mikr. Anat., Bd. 6, pp. 115-172, PL VIII.-X.

Lahille, F.

'90. Recherches sur les Tuniciers. Toulouse, pp. 1-130.

List, T.

'96. Beitrage zur Chemie der Zelle und Gewebe. I. Ueber die Farbung
thierischer Gewebe mit Berlinerblau. Mitt. Zool. Stat. Neapel, Bd. 12,
Heft 3, pp. 477-493, Taf. XXII.

Malfatti, H.

'91. Beitrage zur Kenntniss der Nucleine. Zeitschr. Physiol. Chem., Bd. 16,
Heft 1 u. 2, pp. 68-86.

Maurice, C.

'88. Etude monographique d'une espece d'ascidie composee (Fragaroides
.aurantiacum n. sp.). Arch. Biol., Tom. 8, Fasc. 2, pp. 205-495, PI. XVI.-
XIX.

bancboft: ovogenesis in distaplia occidentalis. Ill

Morgan, T. H.
'90. The Origin of the Test-cells of Ascidians. Jour. Morph., Vol. 4, No. 2,
pp. 195-204, PI. VIII.

Pizon, A.
'92-93. Histoire de la blasotogenese chez les Botryllides. Ann. Sci. Nat.
Zool., Tom. 14, Nos. 1-4, pp. 1-386, PI. I.-IX.

Ritter, W. E.
'93. Tunicata of the Pacific Coast of North America. I, Perophora annec-
tens n. sp. Proc Cal. Aead. Sci. (2), Vol. 4, pp. 37-85, PI. I.-III.

Roule, L.

'85. Sur le developpement des enveloppes ovulaires chez les Tuniciers. Rec.
Zool. Suisse, Tom. 2, Fasc. 1, pp. 195-202.

Sabatier, A.

'84. Sur les cellules du follicule et les cellules granuleuses chez les Tuniciers.
Rec. Zool. Suisse, Tom. 1, Fasc. 3, pp. 423-458, PI. XXII., XXIII.

Salensky, W.

'92. Ueber die Thatigkeit der Kalymmocyten (Testazellen) bei der Entwick-
luug einiger Synascidien. Festschrift f. Leuckart; pp. 109-120.

Salensky, W.

'94. Beitriige zur Eutwicklungsgeschichte der Synascidien. I. Ueber die
Entwicklung von Diplosoma Listen. Mitt. Zool. Stat. Neapel, Bd. 11,
Heft 3, pp. 36S-474, Taf. XVII.-XX.

Salensky, W.

'95. Beitrage zur Entwicklungsgeschichte der Synascidien. II. Ueber die
Entwicklung von Didemnum niveum. III. Allgemeiner Theil. Mitt. Zool.
Stat. Neapel, Bd. 11, Heft 4, pp. 4S8-630, PI. XXI.-XXIV.

Stimpson, W.

'64. Descriptions of new species of Marine Invertebrata from Puget Sound,
collected by the Naturalists of the Northwest Boundary Commission,
A. H. Campbell, Esq., Commissioner. Proc. Ac. Nat. Sci. Pliilad.,
Vol. 1864, pp. 153-161.

V

112

BULLETIN : MUSEUM OF COMPAKATIVE ZOOLOGY.

EXPLANATION OF PLATES.

All the drawings were outlined with a camera lucida; tube length, 160 mm. pro-
jection on the table. Zeiss lenses were used, except where especially specified.
Distaplia occidentalis is figured in all cases where the name of the species is not
mentioned. All the figures were drawn from sections, except where the contrary
is stated. The combination of lenses oftenest used was a y? immersion objective
and No. 2 ocular, and these, together with the camera, gave a magnification of
1300 diameters. Where no lenses and magnifications are mentioned this is the
combination used.

LIST OF ABBREVIATIONS.

a.

Space due to shrinkage.

lac.

Lacuna.

brs. ov'dt.

Oviducal part of incuba-

11.

Lamella.

batory pouch.

hi. oa.

Lumen of the ovary

brs. pi'brn.

Peribranchial part of in-

proper.

cubatory pouch.

hi. pd.

Lumen of the stalk tissue.

cl. con't. tis.

Connective-tissue cell.

nib. prp.

Mem bran a propria.

cl.fol.

Follicle cell.

vied.

Medulla.

cl.fol. pr.

Primordial follicle cell.

nl.

Nucleus.

cl. ms'drm.

Mesoderm cell.

nl. e'th.fol. i.

Nucleus of a cell in the

cl. sp. pr.

Primordial sperm cell.

inner follicular epithe-

cl. tst.

Test cell.

lium.

cp. c.

Central corpuscle.

nil.

Nucleolus.

cp. ia'll.

Intralamellar body.

oa.

Ovary.

cp. ia'vac.

Intravacuolar body.

oa-te.

Ovotestis.

C]). ia'vt.

Intravitelline body.

ov.

Ovum.

cp. hit.

Corpus luteum.

ov'dt.

Oviduct.

cp. ref.

Refractive body.

ov'go.

Oogonium.

cp. vt.

Yolk body.

pd.

Stalk or peduncle.

ctx.

Cortex.

par'.

Superficial wall of the

Cjlt'pl.

Cytoplasm.

ovary.

ec'drm.

Ectoderm.

par".

Deep wall of the ovary.

en'drm.

Endoderm.

sac. brn.

Branchial sac.

e'th.fol.

Follicular epithelium.

sac. pi'brn. d.

Right peribranchial sac.

e'thfol. ex.

Outer follicular epithe-

sac. pi'brn. s.

Left peribranchial sac.

lium.

stg.

Stigmata.

e'th.fol. i.

Inner follicular epithe-

te.

Testis.

lium.

tis. pd.

Stalk tissue.

e'th. g.

Germinative epithelium.

tst.

Test.

fun. con't. tis.

Connective tissue strand.

vac.

Vacuole.

fun. cyt'pl.

Cytoplasmic strand.

va. df

Vas deferens.

fun. gen.

Genital strand.

vs.

Vesicle.

gra. chr.

Chromatin granules.

vs. ex.

Outer vesicle.

in.

Intestine.

vs. i.

Inner vesicle.

Bancroft. — Distaplia occidentalis.

PLATE 1.

Fig. 1. Undifferentiated bud.

Fig. 2. Ovotestis of a young bud, showing ovary and testis separated, and the

beginnings of a peripheral epithelium about the latter.
Fig. 3. Section immediately anterior to Fig. 2. Ovary and testis are united.
Fig. 4. Sagittal section of a medium sized bud. Ovary and testis differentiated,

and both attached to the genital strand.
Fig. 5. Cross-section near the posterior end of the abdomen of an adult zooid,

showing the ovary with its lumen proper, and the lumen of the stalk

tissue, and a corpus luteum. D, 3; X 530.
Fig. 6. Same, from another zooid, showing the ovary. D, 3 ; x 530.
Fig. 7. Part of the germinative epithelium of an adult zooid, showing a young

oogonium projecting into the lumen of the ovary.

u

■

1 1 alpr.

Bancroft. — Distaplia occidentalia.

PLATE 2.

Fig. 8. Whole preparation of a young incubatory pouch, about \ adult size.

A, 2; X 91.
Figs. 9, 10. Cross-sections of the posterior end of the thorax of a zooid that had
nearly reached maturity, showing the region where the stalk of the
incubatory pouch joins the zooid. Fig. 10 is the more anterior sec-
tion. D, 2 ; X 380.
Fig. 11. From another zooid of the same age as that figured in Figs. 9, 10, show-
ing the opening of the pouch into the peribranchial sac. D, 2; X 380.
Fig. 12. Cross-section of the ovary and testis of a young Styda montereyensis.
Powell and Leland ^ immers. ; X 850.

Bancroft-Ovogenesis r ,

Plate 2.

ec'dnh.

brs

d/y.
pi '6m

YU lit"

ov'dt

1st #

'irs.pi.'hrn.

ec arm.
brs. 01

:

10.

n- ■/;,;,

//.

brn.cL.

• c

ectirrn

.

par

■ .

ft:

12.

t't It sac.pi 'wn

Bancroft. — Distaplia occidentals.

PLATE 3.

Fig. 13. Oogonium from an adult zooid. The test cells are just forming.

Fig. 14. Part of the wall of the ovary, showing an ovum projecting beyond the

ovary and just beginning to form a stalk.
Fig. 15. Same; ovum is much older, with a well -formed stalk containing differ-
entiated stalk tissue.
Fig. 16. Whole preparation of ovum, showing maximum size attained in the

ovary. Bausch & Lomb 2 in., Zeiss 2 ; X 40.
Fig. 17. Same; ovum partly extruded from ovary into oviduct. Bausch & Lomb

2 in., Zeiss 2 ; X 40.
Fig. 18. Same; ovum has not quite completely entered the pouch. Bausch &

Lomb 2 in., Zeiss 2 ; X 40.
Fig. 19. Section through a medium-sized ovum, showing bends in the nuclear

membrane and cytoplasmic vacuoles.
Fig. 20. Germinative vesicle of an ovum in a medium-sized bud.
Fig. 21. Ovum in a medium-sized bud.
Fig. 22. Young ovum from an old bud.
Fig. 23. Same ; ovum a little older.
Fig. 24. Follicle and periphery of an ovum a little larger than that in Fig. 13.

Test cells are being formed.
Fig. 25. Test cell and follicle cell from the same ovum.
Fig. 26. Follicle and test cells of a medium-sized ovum. A much smaller ovum

is enclosed within the thickness of the follicle.

Ban

13

/6.

"

■

K

ci/S/

/;

JK

vac

i- nth til.

V ~>

21.

w ^

J

»^

-

...

-.y

» •''''<"''

14-.

' *

el.ht..

I

f

- "■1T-V

. ■

«*&/

o

e*w

Ol\l/0

-js.

'o

■

•0

hsjnl

-*Q

■

te/>rf

hi in I .

* V

.-%•

n

-:

Bancroft. — Distaplia occidentalis.

PLATE 4.

Fig. 27. Ovum in which the test cells and secondary follicular epithelium have
just become differentiated. Reichert ^ immers., Zeiss 3; X 1240.

Fig. 28. Part of a half grown ovum (173 X 150/t), showing the formation of the
yolk, follicle, and test cells.

Fig. 29. Selected test cells from the same ovum.

Bancrdft.-Ovogenesis Distaplia.

4-.

27.

■

. «

3

tJ

*

* ' Sj

N ■

5

rffi*

S

•

>**»

, it

f$k

r >

" * %

..»

2

* ■

-.-

Q, 0

■fMi " *

0

kJt

(V

*

efafol

qua rt

^

1 •

I

*

_'<*

Bancroft. — Distaplia occidentalis.

PLATE 5.

Fig. 30. Periphery of an old ovum and its follicle.

Fig. 31. Periphery and follicular envelope of an ovum that has just reached the

pouch.
Fig. 32. Ectoderm and test at an early stage in the formation of the test.
Fig. 33. Section of the test when a little more developed.
Fig. 34. Test, still farther developed.
Figs. 35-46. Various stages in the degeneration of the test cells of Styela mon-

tereyensis.
Fig. 35. Ovum containing the test cell is 84 X 65 fi. Test cell shows vacuoles

and intravacuolar bodies.
Fig. 36. Ovum 92 X 69 p.
Fig. 37. Ovum 108X69^.
Fig. 38. Test cell from the same ovum.
Figs. 39, 40. Test cells, from another ovum, showing the formation of the central

corpuscle within the intravacuolar bodies.
Fig. 41. OvumgexSO.u.

Figs. 42, 43. From an ovum measuring 134 X 111 yu.
Fig. 44. Ovum 92 X 76 fi.
Fig. 45. Ovum 169 X 80^.
Fig. 46. Ovum, nearly mature, measuring 108 X 92 /u.

PLIA.

r,

cpcia

37. . ryi ia ; tw

38.

sOrvJ

epiaU.

id

cth.Mi

• i

nl.

30.

f'th fo/.ex .

5.

36.
cp.ia'vac

W.

M.

rpiarac.

•>'/ .(/ 1st

' Jcp.vf.

etAMi.

^

i4

f

■\

0

tcilrm.

ee'drm.

cl.ms'drnv.

F.W1-

Bancroft. — Distaplia occidentalia.

PLATE 6.

Fig. 47. Part of the wall of a corpus luteum from which the ovum has just passed

out.
Fig. 48. Germinative vesicle just as it is beginning to shrink. From the same

ovum as that in Fig. 28, measuring 173 X 150 ju. Keichert ^ imiuers.,

Zeiss 2 ; X 890.
Fig. 49. Germinative vesicle and yolk-clump in an ovum of 190 p. diameter.

Keichert ^ immers., Zeiss 2 ; X 890.
Fig. 50. Germinative vesicle and central yolk bodies at a later stage. Ovum is

254 jx in diameter. Zeiss ^ immers., 2 ; X 1300.

Figs. 51-59. Nucleoli in various stages of development. The measurements are

in fi.

Dimensions of Dimensions of the

the Ovum. Germinative Vesicle.

Fig. 51. indistinct 15 X 11

Fig. 52. 23 X 13 13 X 11

Fig. 53. indistinct 13 X 10

Fig. 54. 46 x 38 25 X 19

Fig. 55. 43 X 34 22 X 18

Fig. 56. 45 X 35 23 X 17

Fig. 57. 191 X •? 33 X 25

Fig. 58. 152 X 113 49 X 30

Fig. 59. 207 X 190 32 X 22

Figs. 60, 61. Two successive sections through the group of tetrads formed just
before the formation of the first polar cell. The ovum has just
reached the pouch.

Bancroft-Ovocenesi s Distaplia .

47

'ni.pm

V/V

cl.cm't.tis.

6

S6.

49.

<tr.

■57

5°

• o-

ified

ci/ref

68.

0 • .

60.

•^

(ft

JO.

ct.c

rill

5$

cprrf
med.

6/.

—*i^

cytpl

cp.vt...\

3 del

B.Meisc

The following Reports have been pibi.ished or are in prepa-
ration on the Dredging Operations off the West Coast of
Central America to the Galapagos, to the West Coast of
Mexico, and in the Gulf ok California, in Charge of Alex-
ander Agassiz, carried on by the U. S. Fish Commission
Steamer "Albatross," during 1891, Lieut. Commander Z. L.
Tanner, U. S. N., Commanding.

H. LUDWIG. IV. « XII." The Holothu-

rians.
C. F. LUTKEN and TH. MORTENSEN.

The Ophiuridse.
OTTO MAAS. XXI.21 The Acalephs.
J. P. McMURRICH. The Actinarians.
E. L. MARK. XXIV.24 The Cerianthidaj.
GEO. P. MERRILL. V." The Rocks of the

Galapagos.
G. W. MULLER. XIX.19 The Ostracods.
JOHN MURRAY. The Bottom Specimens.
A. ORTMANN. XIV." The Pelagic Schi-

zopods.
ROBERT RIDGWAY. The Alcoholic Birds.
P. SCHIEMENZ. The Pteropods and Het-

eropods.
W. SCHIMKEWITSCH. Vin.s The Pyc-

nogonidse.
S. H. SCUDDER. VTI.' The Orthoptera of

the Galapagos.
W. PERCY SLADEN. The Starfishes.
. L. STEJNEGER. The Reptiles.
TH. STUDER. X.10 The Alcyonarians.
C. H. TOWNSEND. XVII." The Birds of

Cocos Island.
M. P. A. TRAUTSTEDT. The Salpidee and

Doliolidfe.
E. P. VAN DUZEE. The Halobatidaj.
H. B. WARD. The Sipunculoids.
H. V. WILSON. The Sponges.
W. McM. WOODWORTH. IX.9 The Pla-

narians and Nemerteans.
W. McM. WOODWORTH. XXVII.2'
Planktonemertes

1 Bull M.C. Z., Vol. XXL, No. 4, June 1891, 16 pp. ; and Vol. XXIII., No 1, February. 1892,
89 pp., 22 1'

2 Mem. M. C.Z.", Vol. XVII., Xo. 2, January. 1^92, 95 pp., 32 Plares.
s Bull. M. 0. Z., Vol XXIV.. X >. 7, August, 1893, 72 pp.

*■ Bull. M. 0. Z., Vol XXIII., Xo. 5, December, 1892, 4 pp.. 1 Plate.
5 Bull. M. ('. Z., Vol XXIV.. Xo 4, June, 1893, 10 pp [Zool. Auzeig., No. 420, 1893.]
M. C. Z., Vol. XVI., Xo. 13, July, 1893, 3 pp

Vol. XXV , Xo 1 September, 1S93. 25 pp.
M. C. Z . Vol. XXV., Xo. 2. December, 1893, IT pp., 2 Plates.
Xo. 4. January, 1894, 4 pp., 1 Plate.
Xo. 5, February, 1894, 17 pp.
No. 6, February, 1894, 7 pp . 5 Plates.
Xo 8. September, 1*94. 13 pp., 1 Plate.
Vol XXV. Xo. 10. October, ls'.'l. 109 pp , 12 Plate?.
Vol. XVII Xo. 3. October. 1894, 183 PP- '9 Plates.
Vol. XXV. Xo 12. April. 1895, 2" pi' . 4 Plates
. Vol. XVIII., April. 1895, 2:'2 pp . 67 Plates, 1 Chart
Vol. XXVII . Xo. 3, July, 1895. 8 pp., 2 Plates.
Vol XXVII., Xo 4, August, 1895. 2fi pp . 3 Plates.
19 Bull. M C. Z , Vol XXVII., Xo. 5. October, 1895, 14 pp., 3 Plates
=" Bull M. 0. Z . Vol XXX , Xo 1. March, 1896, 103 pp.. 9 Plates, 1 Chart.

21 Mem. M C. Z., Vol. XXIII . Xo. 1, September, 1897, 92 pp., 15 Plates.

22 Bull. M. '' Z . Vol XXXI , Xo 5, December, 1897. 37 pp., 6 Plates. 1 Chart.
*> Bull. M. C. Z., Vol XXXII.. x,,. 5. May, 1898, 18 pp 13 Plates, 1 Chart
2< Bull. M C. 7... Voi xxx :.. Xo. S. August, 1898, 8 pp . 3 Plates.
2" Bull. M. C. Z., Vol. XXXV.. No. 1. July 1899, 4 pp., 1 Plate.

A. AGASSIZ. II.1 General Sketch of the
Expedition of the "Albatross," from
February to May, 1891.

A. AGASSIZ. The Pelagic Fauna.

A. AGASSIZ. The Daep-Sea Panamic Fauna.

A. AGASSIZ. I.2 On Calamocrinus, a new-
Stalked Crinoid from the Galapagos

A. AGASSIZ. XXIII.** The Echini.

JAS. E. BENEDICT. The Annelids.

R. BERGH. XIII.13 The Nudibranchs.

K. BRANDT. The Sagitt*.

K. BRANDT. The Thalassicolse.

C CHUN. The Siphonophores.

C. CHUN. The Eyes of Deep-Sea Crustacea.

S F. CLARKE. XI.11 The Hydroids.

W. H. DALL. The Mollusks.

W. FAXON. VI.3 XV.^ The Stalk-eyed
Crustacea.

S. GARMAN. The Fishes.

W. GIESBRECHT. XVI.1"' The Copepods.

A. GO as. Ill4 XX.» The Foraminifera.

H. J. HANSEN. XXII.22 The Cirripeds
and Isopods.

C. HARTLAUB. XVHT." The Comatula?.

W. A. HERDMAN. The Ascidians.

S. J. HICK30N. The Antipathids.

W. E HOYLE. The Cephalopods.

G. vox KOCH. The Deep-Sea Corals.

C. A. KOFOID. Solenogaster.

R. VON LENDENFELD. The Phosphores-
cent Organs of Fishes.

6 Bull

7 Bull. M
s Bui
'■' Bull M 0. Z., Vol. XXV.

w> Bull. M. C: Z., Vol XVV.

11 Bull. M. C Z . Vol XXV

12 Bull. \I. C. Z.

Vol. XXV.

13 Bull. M. C Z
11 Men. M. 0. Z .
>5 Bull M. C Z .
»6 Mem M. 0 Z
« Bull. M. C Z ,
» Bull. M. C. Z ,

PUBLICATIONS

OF THE

MUSEUM OF COMPARATIVE ZOOLOGY

AT HARVARD COLLEGE.

There have been published of the Bulletins Vols. I. to XXXI.,
and XXXIII. ; of the Memoirs, Vols. I. to XXII.

Vols. XXXIV. and XXXV. of the Bulletin, and Vols. XXIII.
and XXIV. of the Memoius, are now in course of publication.

The Bulletin and Memoirs are devoted to the publication of
original work by the Professors and Assistants of the Museum, of
investigations carried on by students and others in the different
Laboratories of Natural History, and of work by specialists based
upon the Museum Collections.

The following publications are in preparation: —

Reports on the Results of Dredging Operations from 1877 to 1880, in charge of
Alexander Agassiz, by the U. S. Coast Survey Steamer " Blake," Lieut.
Commander C. I). Sigsbee, U. S. N., and Commander J. U. Bar tie tt, U. S. N.,
Commanding.

Reports on the Pesults of the Expedition of 1801 of the U. S. Fish Commission
Steamer "Albatross," Lieut. Commander Z. L. Tanner, U. S. N., Com-
manding, in charge of Alexander Agassiz.

Contributions from the Zoological Laboratory, in charge of Professor E. L.
Mark.

Contributions from the Geological Laboratory, in charge of Professor N. S.
Slialer.

Studies from the Newport Marine Laboratory, communicated by Alexander
Agassiz.

Subscriptions for the publications of the Museum will be received
on the following terms : —

For the Bulletin, S4.00 per volume, payable in advance.
For the Memoius, $8.00

Those publications are issued in numbers at irregular inter-
vals; one volume of the Bulletin £8vo) and half a volume of the
Memoirs (4to) usually appear annually. Each number, of the Bul-
letin and of the Memoirs is also sold separately. A price list
of the publications of the Museum will be sent on application
to the Librarian of the Museum of Comparative Zoology, Cam-
bridge, Mass.

OCT 18 1899

Bulletin of the Museum of Comparative Zoology
AT HARVARD COLLEGE.

Vol. XXXV. No. 5.

OBSERVATIONS OX NON-SEXUAL REPRODUCTION IN

DERO VAGA.

By T. W. Galloway.

With Five Plates.

CAMBRIDGE, MASS , U. S. A. :

PRINTED FOR THE MUSE CM.

October, 1899.

Bulletin of the Museum of Comparative Zoology
AT HARVARD COLLEGE.

Vol. XXXV. No. 5.

OBSERVATIONS OX NON-SEXUAL REPRODUCTION IN

DERO VAGA.

Br T. W. Galloway.

With Five Plates.

CAMBRIDGE, MASS., U. S. A. :

PRINTED FOR THE MUSEUM.

October, 1899.

OCT 18 1899

No. 5. — Observations on Non-sexual Reproduction in Dero vaga.1

By T. W. Galloway.

CONTENTS.

1.

2.

3.

Introduction 115

Structure of the non-dividing
Animal and the Formation of
Segments 116

Position of the Bud-zone : Pri-
mary and Secondary . . . 117

Relation of Budding to Forma-
tion of Segments 119

Experiments on the Rate of
Budding 120

Methods 122

PAGE

7. Histological Features of the

Budding Process . . . . 122

a. Ectoderm 122

b. Entoderm 126

c. Mesoderm 129

d. Summary of Changes in the

Formation of the Mouth

and Pharynx 132

8. Conclusions 134

Bibliography 138

Explanation of Plates 140

1. Introduction.

The worm upon which the present study is made was originally de-
scribed by Leidy ('80) as Aulophorus vagus. A more complete account
of the anatomy and histology by Reighard ('84) makes it apparent that
its divergence from the species of Dero is not sufficient to warrant the
establishment of a new genus. The sexual features which I have
recently had the fortune to observe confirm this view.

The excellent work of von Bock ('97) on budding in Chsetogaster dia-
phanus, in which there is a review of the work of preceding authors upon
non-sexual reproduction in worms, renders it unnecessary for me to sum-
marize previous accounts here. His interpretations are, however, in
some respects so diverse from those of earlier authors as to make it
desirable that they should be tested in another group. Dero is one of
the most specialized of the Naidiform Oligochseta, and for this reason
presents certain interesting variations from the conditions encountered
in the forms that have been previously studied.

In the autumn, from September to November, Dero vaga is found
abundantly in tubes formed of Lemna leaves, or other small light objects,
at the surface of water in the ponds and ditches in the environs of

1 Contributions from the Zoological Laboratory of the Museum of Comparative
Zoology at Harvard College, under the direction of E. L. Mark, No. XCIX.
vol. xxxv. — no. 5. 1

116 bulletin: museum of comparative zoology.

Cambridge. At this time it usually occurs as single individuals or with
only two zodids in a chain. These, collected and placed in a suitable
aquarium, will thrive and multiply asexually throughout the winter.
The rate of division depends upon the food supply, by the regulation
of which certain interesting variants from the normal may be had.

2. Structure of the non-dividing Animal and the Formation

of Segments.

In individuals not sexually active and uot in process of budding, the
following regions may be distinguished: (1) prostomium ; (2) four
cephalic segments bearing ventral bristles only; (3) twenty to twenty-
five well developed body segments, with both dorsal and ventral bristle
bundles; (4) a region of incompletely formed segments passing pos-
teriorly into au undifferentiated region or zone, — all of which may
equal two or three adult body segments in length ; and (5) a completely
developed, complex structure, — representing probably at least one seg-
ment (anal segment), — which bears the anal orifice, the respiratory
lobes (pavilion), and the digitiform appendages.

The formation of new segments in such a worm is a process which
presents so many points of similarity to that of budding, that I wish to
give a brief description of it before proceeding to the consideration of the
latter phenomenon. New segments are invariably formed immediately
in front of the anal segment, and always from the anterior portion of the
undifferentiated zone. The length of this preanal zone is greater in well
nourished than in poorly nourished individuals. In this part of the body
all the structures which characterize segments in the more mature regions
are less and less differentiated as one proceeds posteriorly. Even the
organs which pass through this region in a functionally complete con-
dition, such as the intestine, the blood-vessels, and, in part, the nerve
cord, present simpler conditions than they do farther forward. The
nerve chain, for example, is represented chiefly by non-fibrous elements,
only a few fibres passing thixmgh to innervate the pavilion. The ventral
and lateral portions of the body cavity are shown, by transverse sections,
to be filled with a mass of indifferent tissue derived from cell multiplica-
tions in the ectoderm, the products of which break through the muscular
layers at definite places and then fuse together, much as in the budding
process (Plate 2, Figs. 8, 9, Plate 3, Fig. 15) described later. These
ectodermal ingrowths are arranged serially in the long axis of the worm,
and afford the earliest signs of segmentation.

galloway: non-sexual reproduction in dero vaga, 117

3. Position of the Bud zone : Primary and Secondary.

In such a worm, if the food supply be normal, a budding zone is
formed some distance anterior to the preanal or segment-forming zone.
Bourne ('91) has suggested that the segment at which bud formation
takes place is probably constant for each species. This is certainly not
the case here, at least not for the asexually produced zooids. I am not
yet able to say what may be true of the sexually derived individuals, for
I have not as yet discovered how, if at all, the sexually produced zooid
may be certainly distinguished from the non-sexually produced forms.
That probably can be settled satisfactorily only by rearing young worms
from fertilized eggs.

The cellular activity accompanying budding, it is to be observed, in-
volves only one segment, and the plane of separation is always midway
between dissepiments. The position of this budding segment in the
individuals which I have examined is variable within certain narrow
limits. It may occur as far forward as the 16th and as far backward
as the 21st setigerous segment. The following table, representing obser-
vations upon about three hundred individuals, gives in percentages the
frequency of occurrence in these two and the intervening segments.

TABLE I.

Setigerous Segment
Percentage of Occurrence

16th
7.6

17th
15 3

18th
38.2

19th
26.7

20th
7.6

21st
4.6

Table II. is a record of observations upon the position of the budding
segment in successive generations.

TABLE II.

Primary Division occurs in

Secondary Division occurs

Anterior Zooid, in

Posterior Zooid, in

Segment No.

No. of Cases.

Segment No.

No. of Cases.

Segment No.

No. of Cases.

20
19

18
16

1
4

4
1

20
. 19

18
16

1
4

4
1

20

(19
18

r 17

(19

118

-,17

1 16

17

1

2
1
1
1

1
1
1
1

113

bulletin: museum of comparative zoology.

It will be observed from this table that, when the anterior zooid again
buds, the zone of budding occurs at the same distance from the anterior
end as in the first instance, i. e. in the same body segment. This is not
true, however, for the posterior zooid. There may be either an increase

T&. "; 1. Cephalic segments (4) -f- pr'stm.

A< 2i

2. Body segments (11 to 16), unchanged-

>-_ — 3. Half-segment, producing the equivalent of 3 to 7..

■ 4 Half-segment, producing the equivalent of 1

■" 5. Body segments (5 to 10), unchanged

.'•■ 6. Preanal zone, producing

+ -i-±

'■•7. Anal segment, unchanged

--3
'4

••5

- 6
-7 J

>B

Fig. I.

Fig. II.

or a decrease in the number of segments in front of the budding zone
as compared with the condition in the parent. This variability is per-
haps correlated with the rate of growth as dependent upon food, etc.

It will be seen from the accompanying diagram (Figures I. and II.)
that the half-segment A, 3 (Fig. I.) produces A, 3-7 (Fig. II.), which

GALLOWAY: NON-SEXUAL REPRODUCTION IN DERO VAGA. 119

simulates the condition of A, 3-7 (Fig. I.), and consequently that the
second new individual, which is thus to be produced at the posterior
end of A (Fig. I.), arises from a region which was undifferentiated during
the first budding. The half-segment 4 (Fig. I.) produces a prostomium
and four cephalic segments, B, 1 (Fig. II.). At the same time, the indiffer-
ent preanal zone, 6 (Fig. I.) is adding segments to the region 5, which
before these additions embraces a small but variable (5-10) number of
body segments. Since the budding in no case takes place in a segment
so near the head as the tenth body segment, it follows that this must
occur in the undifferentiated region 6 (Fig. I.), and consequently it is
true for the posterior as well as the anterior zooid that, when a new
budding takes place, it occurs in a zone that was undifferentiated during
the preceding process of budding.

4. Relation of Budding to Formation of Segments.

If the worms are well nourished, the secondary divisious often com-
mence before the first is completed. A condition of this sort is shown
in Plate 1, Fig. 3. In such instances the secondary bud (z2) in each
zooid lies within the limits, either of the indifferent zone (anterior zooid),
or of still incompletely differentiated segments (a condition more usually
found in the posterior zooid).

There is thus shown to be a close relation between the process of
budding and that of forming new segments. For the material which
is destined to subserve the process of budding is at one time an indis-
guishable part of the zone (preanal) the posterior part of which is a con-
tinuous source of new segments. In every case the budding zone, on its
part, gives rise to two segment forming regions : a posterior, which is
normally limited to the production of four segments beside the cephalic
lobe ; and an anterior, which never loses the power of forming new seg-
ments, since some of the segments thus produced, or at least one of
them, retain the capacity of forming a new bud, by which the process is
repeated and the power perpetuated indefinitely. In other words, the re-
lation of the two processes in this worm is such as to lend support to the
view that here budding is a specialized form of normal segmentation.
The cycle of procedure occurs just as if, in the formation of segments
from the indifferent preanal zone, there were deposited at a certain stage,
in one or several segments, materials which are capable under appropriate
circumstances of giving rise to indifferent tissue and to new segments.

This fact allies itself with what is known of regeneration of lost parts.

120 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY.

As several authors have shown, the likelihood of the regeneration of a
given portion is decreased proportionally to the distance of the regenerat-
ing region from the part whose reproduction is expected. For example,
the head segments in the earthworm are much more likely to be regen-
erated from anterior segments than from middle or posterior ones. In
this case, too, the behavior suggests, one might say, a diminution of
head-producing material posteriorly.

In addition to this normally recurrent relation between these two
regenerative processes in Dero, I have found in two cases that, if a large
number of posterior segments ai'e removed from a worm in which the
budding zone has just begun, the process of budding continues, without
any effort being made to regenerate the lost tail piece. In such cases,
however, the budding departs from the normal course, since the prolifer-
ations in what should become the head of the posterior zooid are amor-
phous, none of the normal structures being produced ; but the anterior
half of the budding segment produces an indifferent zone and an anal
segment in the customary way, with, however, minor imperfections.
Here the behavior is as if the regenerative process had been transferred,
as a matter of economy, to the budding region instead of taking place
where the injury existed. In a case where only a small number of seg-
ments were removed, the cut being immediately in front of the indiffer-
ent preanal zone, regeneration occurred at the point of cutting, and the
ectodermal thickening of the budding zone disappeared. After the re-
generation of the tail piece was completed, the animal formed again the
budding zone in the original segment (in this instance the 18th setige-
rous) and ultimately divided.

It is a suggestive fact, that, furthermore, in the sexual individuals,
some of which are known to be asexually produced, I have found that
the testes are formed upon the posterior face of dissepiment iii/iv, i. e.
on a new dissepiment produced in the process of budding, in a region
which was of course located in the parent individual much posterior to
the usual place of production of gonads. The ovaries arise likewise
upon the posterior wall of the dissepiment between setigerous segments
iv/v. Although this is one of the original dissepiments, the new tissues
are closely related to it in their origin and development.

5. Experiments on the Rate of Budding.

In the effort to secure an abundance of budding material, there ap-
peared certain facts of practical and perhaps theoretical importance

GALLOWAY: NON-SEXUAL REPRODUCTION IN DERO VAGA. 121

relative to the rate of division. Two classes of culture solutions were
used in rearing the worms. Algae and other organic substances native
to pond water were supplied in the first ; the second consisted of an in-
fusion of boiled coru-meal in water in the proportion of 1 : 1000. Of the
first, three grades were roughly distinguished : (1) that in which the
organic matter was in such excess that striking evidence of decomposi-
tion was sure to occur ; (2) that in which the organic matter was present
in such quantities that no conspicuous impurity arose from its decom-
position; and (3) that in which the organic material was reduced to a
small per cent of what could have been employed without signs of de-
composition. In the first solution, bacteria, paramcecia, and stentors
appeared in great abundance, and it became necessary from time to time
to pour a part of this away and renew the water. In the other two
cases the water was not disturbed except as evaporation made renewal
necessary. In the corn-meal solution frequent changing was necessary
on account of rapid fermentation.

The following tabulated statement represents the results, but gives of
course only rough quantitative data as to the effect of the food supply in

10 Individuals
in each Culture.

Pond-water, with Algas and other organic Food Supply.

Corn-meal in
Water.

A. In Excess.

B. Medium.

C. Minimum.

D. 1 : 1000 parts.

15 days.

30 days.

15 days.

30 days.

15 days.

30 days.

10 days.

20 days.

Exper. 1
Exper. 2
Exper. 3
Average

18
22
21
20.3

23
19
22
213

15

18

14
15.6

28
30
26
28

13
12
15
13 3

15
14
16
15

18
20

19

25

22

23.5

1st 15
days.

2d 15

days.

1st 15
days.

2d 15
days.

1st 15
days.

2d 15
days.

1st 10
days.

2d 10
days.

Percentage )
Increase J

103%

5%

56%

80%

33%

13%

90%

23 5%

influencing growth, as measured by the rate of germination. In each
experiment, at the outset, ten individuals were taken approximately at
the same (an early) stage of division. The figures in the columns of
"days" indicate the number of separate worms at the end of the time

122 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY.

indicated. The "30 days" columns take no cognizance of the stages of
division of the individuals occurring at the end of the "15 days" periods,
and for this reason the two columns of each culture are perhaps not
strictly comparable. The horizontal line of percentages is reckoned
upon the number of individuals existing at the beginning of the period
in question.

It will be seen that during the first fifteen days solution "A " was pro-
ductive of the most rapid division, in some instances the second division
being accomplished ; but during the second period the worms scarcely
more than held their own. This is seemingly due to the prevalence of
bacteria. Culture " D " presents somewhat similar phenomena. In
" C " the falling off in the second period is probably caused by the dis-
appearance of food from the water and from the tissues of the body.
" B " is the control culture, and more nearly represents the ordinary rate
of reproduction under favorable conditions. These are winter cultures.
Propagation would doubtless be more rapid with fresh worms in summer
or fall.

6. Methods.

Dero does not present any special technical difficulties except that, in the
later stages of budding, separation of the zooids is likely to occur in the
process of killing, whereby control of the material is lost. I used as
killing and fixing reageuts with about equal success (1) a saturated
aqueous solution of corrosive sublimate plus 1% acetic acid (hot), and
(2) a solution of hot picro-sulphuric acid (Kleinenberg's). Staining of
excellent quality was secured by Heidenhaiu's iron-hrematoxylin method.
Sections were made from 6 to 12 /x in thickness.

7. Histological Features of the Budding Process.

In the discussion of the role played by the different embryonic layers
in the production of the new organs made necessary by budding, I shall
first treat in a general way of the changes occurring in each layer, and
the organs which are wholly developed therefrom. Afterward, at the
risk of some repetition, I shall correlate the share of each layer in the
formation of the mouth, and such other structui'es as involve more than
one germinal layer.

a. Ectoderm.

The first evidence of the bud-zone, as seen in optical sections of the living
worm, is a slight thickening of the ectodermal elements of the dermo-

galloway: non-sexual reproduction in deeo vaga. 123

muscular wall (Plate 1, Fig. 2, z' ; Plate 2, Figs. 11, 12). The thickening
is first manifest on the ventral aspect, extending ultimately as a girdle
around the segment. It is interseptal in position, and about equally
distant from the nearest dissepiments. That this thickening is a matter
of growth and not of local contraction is shown both by the increased
number of cells and the increased length of the segment. The peripheral
margins of the dissepiments bounding the bud-zone are crowded apart,
as indicated in Figure 2. The bristle bundles belonging to the involved
segment are forced backward into contact with the following dissepi-
ment. They become the bundles of the oth setigerous segment of the
posterior zooid. An external groove (Plate 1, Figs. 4, 5, sid.) is next
formed in the ectodermal thickening, and persists as the outward demar-
cation between the zooids, indicating the place of ultimate separation.
This groove deepens gradually but unequally at different points of its
circumference. The activity of the tissue is strictly confined to one seg-
ment, the segments in front of and behind this one undergoing no appre-
ciable changes. Hence, it is possible to affirm that the tissues of the
active segment posterior to the plane of the groove furnish the material
out of which are formed the prostomium, four new segments, and the
anterior portion of the fifth segment, with their contained structures.
Similarly, from the anterior half of the segment spring the whole set of
structures appropriate to the tail end of the worm. It is possible to
locate the fundaments of these organs from the beginning of the process.

The division of ectodermal cells may continue until a thickness of
three or four cells is reached in the dermal layer. In the mean time, as
may be seen in transverse sections, the proliferating ectodermal tissue
breaks through the muscular layer into the body cavity, especially in
the spaces between the longitudinal muscle bands (Figs. 8-10). These
invading elements on either side the median plane coalesce with each
other and with proliferations of the cells from the ventral nerve cord,
filling the ventral and latei-al portions of the body cavity in a very char-
acteristic way (Plate 1, Figs. 6, 7). In a sagittal section, a differentia-
tion of this internal cell-mass into an anterior and posterior portion is
apparent at a relatively early stage (Fig. 5). The cells on the con-
tiguous faces of these masses become arranged into bounding surfaces,
which are ultimately continuous with the external layer, and become, in
part, the boundaries of the deepening groove. Thus the latter is supple-
mented in its deeper portion by an internal delamination, and the two
together produce deep invaginations, especially from the latero-ventral
regions (Plate 3, Figs. 16, 17, sid.), which are paired and contribute to

124 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY.

the formation of the mouth. The apparent depth of the ectodermic
depression is increased by a pair of growths (directed backward and out-
ward) of the thickened ectoderm anterior to the groove in the latero-
ventral regions, to form the . digitiform appendages (Figs. 16, 17
pr'c. dg.).

The process of ectodermic ingi'owths between and through the muscle
bands, referred to above, is kept up both anterior and posterior to the
plane of separation of the zooids.

In the posterior zooid there are five distinct regions of cellular activity
in the ventral part of the body, namely : (1) the ectoderm in the spaces
between the lateral and ventral bands of longitudinal muscles (Plate 1,
Figs. 6, 7 ; Plate 2, Figs. 8, 9) ; (2) the ectoderm invading the ventral
muscle band and cutting off certain fibres (Plate 1, Figs. 6, 7, mu. v.
and mu.v'.) at either margin ; and (3) the cellular elements of the ven-
tral nerve cord itself. More dorsally, in the spaces between the dorsal
and lateral longitudinal muscles, there is likewise an ingrowth of ectoderm
on either side of the body, which in some measure supplements and fuses
with the masses already described. More dorsal still, approximately at a
level with the dorsal wall of the gut, is seen (Fig. 7) another irruption of
ectoderm so located as to cut off a strand from the ventral margin of
the dorsal muscle. This ingrowth differs from the former ones in the
very important fact that its extent in an antero-posterior direction is
extremely limited. The former proliferations occur one after another in
longitudinal series, one of each series to every segment. The ingrowth
through the dorsal muscle band, on the contrary, is limited to a single
pair of cell masses situated well Forward toward the plane of division.
The cells from this pair of ingrowths move dorsally and toward the
median plane until they rest upon the digestive tract, and finally meet
each other, forming the brain ganglia, whose elements later produce a
connective (Plate 1, Fig. 6, con't. crcce.). The detached portion of the
dorsal muscle band approximately equals in dimensions the whole lateral
band. The pair of irruptions through the marginal portions of the
ventral muscle, in conjunction with elements from the ventral nerve
cord itself, furnishes the material for the cord in the newly forming seg-
ments of the anterior zooid and the infra-pharyngeal enlargements of the
nerve cord in the posterior zooid. In the posterior worm there is a
degeneration of that part of the old nerve tract which lies in front of
the place to be occupied by the infra-pharyngeal ganglia, only a few
fibres remaining to preserve connection between the individuals until
separation is accomplished.

GALLOWAY: NON-SEXUAL REPRODUCTION IN DERO VAGA. 125

The ingrowths between the ventral and lateral muscle bands give rise,
in each of the four cephalic segments of the posterior individual, to a
pair of veutral bristle sacs, and to the ectodermal part of the nephridial
orgaus. There are no dorsal bristle bundles formed in the cephalic
segments of the posterior zooid.

The fate of the proliferation between the dorsal and lateral longitu-
dinal muscle bands is more obscure. The circumoesophageal connective,
in running from the brain ventrally, passes, as has been said, between the
main portion of the dorsal muscle baud and the lateral part of it, which
is thereby split off from the main band. • In its further course ventrad,
the connective again returns to a position inside the longitudinal mus-
cles. This it does by running inward between the split off portion of the
dorsal band and the lateral band of muscles, so that no part of the lateral
muscle band ever lies inside the connective. It is perhaps significant
that the cells of the lateral line make their appearance in the same
space between the dorsal and lateral muscle bands (Plate 1, Figs. 6, 7 ;
Plate 2, Fig. 10, prf. and cl. In. I.). The two sets of nervous structures,
— the circumoesophageal connective and the lateral line cells, — if the
latter are really nervous, are thus apparently in relationship by their
proximity. It is, furthermore, possible that the cells which push in
and go to form the brain do so here rather than at the more dorsal
position previously noted. The fact that the whole ectodermal arc periph-
eral to the detached portion of the dorsal band is very much thickened
would seem to support this view. In this event, we should be compelled
to admit that the future nervous elements produced at this place
grow dorsally in a position peripheral to the cut off portion of the dorsal
longitudinal muscle band, between which and the rest of the dorsal band
they pass into the body cavity. This would help to explain the course
taken by the fibres of the connective in uniting the brain and sub-
pharyngeal ganglion. It would also connect the brain in its origin with
the lateral line system, rather than with the ventral chain, the con-
nection with the latter being an altogether secondary one. However
this may be, the course of the nervous connective is associated with the
position of the lateral line cells in a very suggestive way. I am as yet
unable to state what, histologically, the origin of the connective fibres
may be. From analogy with other annelids we should expect them to
arise from the cells of the brain, or possibly from the subossophageal
ganglia, or both. The course of the connective is shown semi-dia-
grammatically in Figure 10, con't. crc'oe.

In the anterior zooid the course of development is somewhat different.

126 bulletin: museum of comparative zoology.

The ectodermic invasions occur in the same way as in the posterior zooid
and in corresponding positions, but the one which cuts oft' a lateral por-
tion of the dorsal muscle baud is ultimately repeated here for each new
segment, and occurs at the place where subsequently the dorsal bristle
bundles are found. In fact, these ingrowing elements constitute the
bristle forming organs. It is further to be noted that, at the time of
separation of the zooids, the bristle bundles of the anterior worm are in
a much less advanced state of development than those of the posterior
worm. Another point of difference is found in the fact that there is a
much more intimate fusion of the several ectodermic ingrowths of the
ventral and lateral regions with each other and with mesodermic ele-
ments in the anterior than in the posterior zooid. In this manner is
produced an indifferent zone, similar in all particulars to the undifferen-
tiated segment-forming zone previously mentioned as characteristic of
the preaual region. From the ectoderm comes, as in the posterior zooid,
the mass of tissue from which ventral nerve chain, ventral bristle sacs,
etc., are developed. All stages in the differentiation of these structures
may be found in a rapidly growing worm, as one passes forward. Im-
mediately in front of the anal segment, where the tissue is least differ-
entiated and the ventral nerve cord is to be developed almost de novo,
an ectodermic ingrowth in a mid-ventral position divides the ventral
longitudinal muscle and contributes to the ventral cells of the cord
(Plate 3, Fig. 15 ; Plate 4, Fig. 23). Proliferations which are lateral,
but still penetrate the ventral muscle, add cellular elements to the mar-
gins of the cord (Plate 2, Fig. 8; Plate 5, Fig. 24, gn.v.). Thus in the
formative region the cord can be resolved into a median and two lateral
constituents. It was the latter which Semper ('76) regarded as meso-
dermal. The fibres of the nerve cord naturally become less numerous
posteriorly. The more dorsal, i. e. the deeper fibres, are the first to ap-
pear ; this produces the condition figured in Plate 3, Fig. 15, in which
the fibrous tract appears to occupy a more and more dorsal position as
one proceeds toward the tail. The ectoderm on the ventral side of the
body in this region is especially thickened, being four or five layers of
cells deep (Figs. 15, 16, 22, 23, 21).

b. Entoderm.

It is to be borne in mind that before the beginning of division the di-
gestive tract in the region affected by that process is a simple tube with
interseptal enlargements. The wall of the intestine, reckoning from the

GALLOWAY: NON-SEXUAL EEPRODUCTION IN DERO VAGA. 127

lumen outward, here consists of (1) a layer of ciliated epithelium, in
which the free ends of the cells form a wavy contour ; (2) scattered sub-
epithelial cells lying in the basal portion of this epithelium (Plate 2, Fig.
13) ; and (3), surrounding all, a distinct basement membrane. Outside
the basement membrane is an investment of connective tissue, occasional
muscle fibres, and chlorogogue cells. The pharynx, on the contrary, is a
more highly specialized structure. Its wall is seen in cross section
(Plate 4, Fig. 18) to consist of a dorsal arched portion with a very thick
wall, a nearly flat ventral floor, and much thinner latero-ventral connect-
ing regions. The thin latero-ventral portions of the wall may be evagi-
nated to form a pair of longitudinal grooves (sul.). The flexibility of
this region allows considerable variability in the form of the tube, for
the floor may be infolded into the arched dorsal portion so as nearly to
obliterate the lumen, or it may be depressed until the lumen is nearly
circular in cross section. Strongly ciliated, long columnar epithelial cells
form the dorsal wall, aud extend ventrally somewhat more than half way
down the sides (Fig. 18, phy. d.). The ventral floor {phy. v.) is similarly
formed of cil'ated cells, but these are not so long as those of the dorsal
wall, while the wall of the groove is formed of cubical non-ciliated epithe-
lial cells. The new pharynx, formed as it is during budding, is derived
exclusively from material lying anterior to the dissepiment which origi-
nally marks the posterior boundary of the budding zone. This shows
that it must be either a modification of the posterior half of the intestine
of the segment which is involved in the budding process, or a new struc-
ture, the material for which is supplied from other growth centres within
the half-segment.

In a budding worm, while the body wall of the bud zone increases in
length by the rapid multiplication of the elements making up that wall,
there seems to be no corresponding growth on the part of the cells of the
intestine. The digestive tube is thus mechanically stretched ; this tends
to obliterate the inequalities of its calibre. (Compare Fig. 5 with
Fig. 16.) By the continuation of this stretching, the individual cells
are transformed from their original columnar character into relatively
thin pavement-like elements. I have no evidence that these ciliated
epithelial cells divide daring the budding process.

The first signs of cellular activity in the entoderm are seen in the
small, indifferent, sub-epithelial cells (Plate 2, Fig. 13). These become
more and more numerous, first in the ventral part of the gut, and later
in its lateral and dorsal walls ; they form at length a sheath of cells
around the epithelial layer which extends through the larger part of the

128 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY.

active segment, and is especially well marked in its posterior half (Plate
3, Fig. 1G). This cell multiplication may continue in certain places
until the layer en'drm.2 becomes five or six cells deep (Figs. 7-9).
The cytoplasm of the older layer (eri'drm.1) stains feebly and diffusely,
while that of the new layer (en'drm.2) takes a very deep stain. The
new entodermal layer attains a greater thickness in the posterior than
in the anterior zooid. The dorsal portion of the digestive tube of the
posterior zooid, beginning just behind the brain and extending to the
posterior limit of the budding zone, is the region of greatest thickening.
The thickness diminishes gradually from the dorsal to the ventral side
of the tube. Later a separation of the old from the new entoderm
occurs, and a cavity — crescent-shaped in cross section, the lumen of
the new pharynx — appears {phy. lu., Figs. 9, 1G). The new ento-
dermal cells next take on the typical columnar form and then become
ciliated (Plate 3, Fig. 17 ; Plate 4, Fig. 19). The action of the cilia is
manifest before the separation of the zooids, and before the rudimentary
lumen of the pharynx has any connection with either the old lumen or
the outside world. The old entodermal lining finally becomes detached
throughout the pharyngeal tract (Plate 2, Fig. 14, erfdrm}) and is swal-
lowed or ejected through the mouth.

It is important to note that the entodermal cells are even more com-
pletely separated into two functionally different regions, corresponding
to the two zooids, than are the cells of the ectoderm. There is a distinct
neutral zone between them, embracing the plane of division. This in-
terruption is most conspicuous dorsally and laterally, being only slight
ventrally (Plate 3, Fig. 16).

In the anterior zooid, immediately in front of the future plane of sep-
aration, two pairs of thickenings arise in a manner entirely analogous
to that just described for the pharynx of the posterior zooid (Plate 3,
Figs. 16, 17; Plate 4, Fig. 23; Plate 5, Figs. 24, 25, pav.), and in
these thickenings spaces are formed, as in the case of the pharynx. Ulti-
mately the dorsal and ventral cavities on the same side coalesce, and
likewise the right and left communicate across the median plane, both
dorsally and ventrally, thus forming a complete annular space of varying
calibre around the old tube. These new entodermal cells, like those of
the pharyngeal wall, bear cilia ; they constitute the lining of the pavilion.
The ectoderm has no part in forming the inner face of the pavilion, the
concrescence of the free margins of the two layers, ectoderm and ento-
derm, taking place at the margin of the pavilion. The digitiform
process, however, is wholly covered with ectoderm, the boundary be-

galloway: non-sexual reproduction in dero vaga. 129

tween ectoderm and entodewn being at the base of the inner side of
that appendage.

It is only in the region of the pavilion that the entoderm of the an-
terior zooid becomes especially active during budding, but anterior to
this the sub-epithelial cells form a more or less continuous layer beneath
the ciliated epithelium, in a maimer characteristic of the indifferent
preanal zone of the mature worm. As the digestive tube lengthens in
the formation of new segments, these sub-epithelial cells, in my opinion,
become interpolated between the ciliate cells, and thus help to form the
lining of the tube, for I find no evidence that the ciliate cells in this
region are undergoing division.

Summarizing the facts concerning the growth regions in the entoderm,
we may distinguish the following in the anterior zooid: (1) the ante-
pavilion region, where a single layer of sub-epithelial cells reinforces the
digestive cells, the latter not being sloughed ; (2) the pavilion thicken-
ing, from which the old entoderm is lost ; and in the posterior zooid,
(3) the pharyngeal region, of somewhat irregular shape, surrounding
the gut and extending posteriorly to the old dissepiment ; from this
region also the original entodermic lining is cast off. Between (2) and
(3) is a neutral zone, more extensive dorsally, narrowing below, in which
the old layer alone exists.

I shall deal more fully with this in discussing the mouth.

The old entoderm in the budding zone does not lose continuity until
the zooids separate ; thus there is no functional mouth nor anus until
separation. The cells of the original lining of the digestive tract show
signs of degeneration long before separation. The cell boundaries become
less distinct, the cytoplasm stains more diffusely and less intensely,
and vacuoles occur in the cytoplasm.

c. Mesoderm.

The general arrangement of muscle fibres is the same in Dero as in
other Oligochseta. The circular muscles lie beneath the dermis, and
beneath these the longitudinal fibres are grouped in four bands : (1) the
dorsal band occupies slightly more than one half the circumference of
the body ; (2) the narrow lateral band on either side is perhaps equal to
one eighth of the circumference, and is separated by a space from the
dorsal muscle ; and (3) the ventral band, which is somewhat broader
than the lateral bands, from which it is separated by a space.

In the formation of the budding zone the lateral bands are little in-
terfered with until the zooids separate, when their broken euds become
vol. xxxv. — No. 5. 2

130 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY.

spread out and attached to the lateral walls of the pavilion in the ante-
rior zooid, and to the prostomium in the posterior individual. The dorsal
and ventral muscle bands, on the contrary, are both split, it will be
remembered, by ectoderm ic invasions which penetrate between their
fibres. Two symmetrically located series of ingrowths penetrate the ven-
tral muscle, so that this band throughout the bud zone is made to consist
of a median (mu. v.) and two lateral portions (mu. v'.). During the in-
growth of the indifferent ectoderm tissue, the lateral parts are pushed in-
ward much more than the median, and in the posterior zooid they come
to lie inside the circle of the newly formed pharyngeal nerve ring (Plate
1, Fig. 6 ; Plate 4, Fig. 21, mu. v'.). These lateral fibres of the ventral
band, on account of their deeper position, are less interfered with by the
formation of the ventral groove than is the median strand. As a result,
they constitute the muscular connection on the ventral side of the zooids
which persists lougest. Thus the mid-ventral strand is sooner free to
form new attachments in connection with the mouth and head of the
new individual ; its fibres are in fact seen to be distributed to the
lower lip, to the sides of the mouth, and to the pharynx, persisting as
the principal ventral muscle. The lateral portion of the ventral band,
which passes within the nerve ring, becomes attached to the deeper por-
tion of the ectodermal constituent of the ventral surface of the mouth
(Plate 5, Fig. 2(5, mu. v'.).

In the budding zone of the anterior individual the old median band is
superficial even to the newly forming circular muscle fibres ; the baud
becomes broken up into smaller groups of fibres by the penetration of
ectodermal elements. The deeper lateral strands of the ventral muscle
band become more conspicuous, and a new median constituent is formed
between them ; in this way is reproduced the typical single ventral mus-
cle. Anterior and posterior to the region of ectodermal ingrowth the
strands merge into a normally continuous sheet.

The dorsal muscle band is likewise separated into a median and lateral
portions by the development of the nerve ring, which passes exterior to
the extreme lateral portions of the band (Plate 2, Fig. 10, mu. d'.),
but interior to the dorsal (median) part of it. Thus, while the zooids
remain connected, the brain is supported by a sling, as it were, com-
posed of the lateral strands of the dorsal muscle bands (Plate 2, Fig. 10 ;
Plate 3, Figs. 16, 17, mu d'.). In this position, some of the fibres become
attached to the brain capsule (Plate 5, Fig. 28) ; others remain attached
to the dermo-muscular wall at the anterior dorsal boundary of the pos-
terior zooid. The contraction of the latter fibres, when separation of the

galloway: non-sexual reproduction in dero vaga. 131

zooids takes place, in connection with the action of the circular muscles
to be described in detail later, pulls the ectodermal margin downward
and backward into connection with the median-dorsal entodermal and
dorso-lateral ectodermal elements previously mentioned. The median
dorsal fibres are distributed to the prostomium (Plate 5, Fig. 28, mu. d.).

New circular muscles are formed among the pre-existing fibres as the
bud-zone increases in length. The chief special modification occurs in
the immediate vicinity of the future separation. In connection with the
ectodermic ingrowths contributing to the formation of the mouth, cir-
cular fibres are carried inward in such a way as to constitute an invest-
ment of the new epithelium which forms the floor of the mouth (Plate 3,
Figs. 16, 17, mu. crc'oe., fibres cut crosswise).

Before separation of the zooids, these circular fibres (Plate 5, Fig. 25,
mu. crc'oe.) pass, like the circumcesophageal nerve ring, between the sep-
arated strands of the ventral muscle band, a lie for and part of their
course within the nerve ring. Another interesting fact concerning the
circular muscle fibres of this region is that they do not constitute a com-
plete circular band. The ends of the fibres, instead of meeting in the
mid-dorsal line, pass forward as well as upward, and are inserted into
the sides of the prostomium (Plate 5, Figs. 26, 29, mu. crc'oe.) ; they aid
in pulling the prostomium downward into its normal position when the
posterior individual becomes detached from the anterior one. The fibres
of this semicircular band (Figs. 16, 17, mu. crc'oe.) are coextensive, on
the ventral floor of the mouth, with the latero-ventral ectodermal invagi-
nations, and thus contribute to the buccal wall.

The radial fibres from the dermo-muscular sac to the pharynx repre-
sent longitudinal fibres, which have been diverted from their ordinary
course of growth, and have become attached to the wall of the newly
formed pharynx. Similarly, the fibres moving the bristle bundles are
apparently derived from the muscle bands in their immediate vicinity.
This is clearly true of those moving the setae in the plane parallel with
the axis of the body (i. e. the longitudinal fibres). There are also setal
muscles in the transverse plane, but I am not certain as to their origin.

The dissepiments in the undifferentiated region of the anterior zooid
seem to be formed in a somewhat mechanical way. The cell prolifera-
tions of the ectoderm are, from the beginning, discontinuous. They are
practically successive pairs of buds or ingrowths, which force before them
the portions of the peritomal lining and its connective tissue investment
which lie immediately opposite them. Between successive buds, i. e. in
the septal planes, the mesodermic elements remain attached to the per-

132 bulletin: museum of comparative zoology.

manent ectoderm. The ectodermic buds in their further growth continue
to carry the connective-tissue mesoderm before them until they come into
contact with the wall of the digestive tube. The connective-tissue par-
tition between the successive masses of ectodermic cells may be clearly
seen in longitudinal sections of stages as advanced as that of Figure 16
(Plate 3), the anterior masses being more clearly separated than the
posterior ones.

The blood-vessels undergo considerable branching and anastomosis in
the anterior and posterior ends of the mature animal. In the part of
the adult worm where budding is to take place, the dorsal and ventral
vessels are not connected by conspicuous circumintestinal loops ; but in
the budding zone, long before separation of the zooids, a complete vas-
cular loop is formed in each zooid close to the plane of division (Plate 3,
Fig. 17, vas. sng.crc). The complex anastomosis is not, so far as I can
determine, completed until the animals become independent. The loops
are of very small calibre until this occurs.

d. Summary of Changes in the Formation of the Mouth and

Pharynx.

In concluding this part of the subject, it is perhaps desirable to sum-
marize under a separate heading the chief features connected with the
rather complex processes involved in the formation of the structures
appropriate to the anterior end of the digestive tract. This is ren-
dered desirable from the fact that each of the germ layers contributes
to its production.

The wall of the new pharynx is formed around the old gut by a
proliferation of the cells occupying the base of the original epithelium,
and a delamination of the layer thus formed. The lumen of the pharynx
exists, as a lacuna dorsal to the gut, between the old and new entoderm
before the separation of the zooids. Dorsally, the new entodermal wall
does not reach forward beyond the position of the brain. Laterally, its
anterior margin passes downward and somewhat backward, but turns for-
ward again in the middle of the ventral side of the worm. To the curved
lateral margin of this new wall is applied a pair of ectodermal ingrowths,
whose history is as follows.

The groove marking the region of separation of the zooids, relatively
shallow in other regions, becomes specially deepened at two points, sym-
metrical to the median plane, and corresponding to the space between the
ventral and lateral longitudinal muscles. The axis of these depressions
lies in the plane of the groove, which inclines forward, and dorsally. The

galloway: non-sexual reproduction in dero vaga. 133

direction of the plane of the groove is shown in Figure 3. The depth of
the two infoldings is enhanced hy the outgrowing digitiform appendages
of the anterior zooid. When the depressions have reached a considerable
depth, they turn somewhat abruptly upward and backward on either side
of the digestive tract (Plate 3, Figs. 16, 17 ; Plate 4, Fig. 22, sul). The
ectodermal wall bounding this lumen applies itself to the lateral wall of
the old gut, but does not reach its mid-dorsal line. Posteriorly these in-
growths become continuous with the new pharyngeal wall described above,
and ultimately, after the separation of the zooids, their lumina, by the
breaking through of the tip of the invagination, are put in communication
with the lumen of the new pharynx (Fig. 19). This represents the con-
dition attained before the zooids separate ; the condition of the tissues
involved is shown in Figures 19-21. At the moment of separation, the
contraction of the semicircular muscles (Plate 3, Fig. 17; Plate 5, Fig.
26, mu. crc'ce.) and other muscles, both circular and longitudinal, in the
posterior zooid, pulls the whole prostomium downward, closing up the
communication of the body cavity with the outside world which other-
wise would result, bringing the aperture of the gut into a ventral, in-
stead of a terminal position. Simultaneously, the lateral fibres of the
dorsal muscle which lie inside the circumcesophageal commissure me-
chanically draw the ectoderm of the free anterior portion of the dorsal
wall inward and backward, under the brain, into the space left by the
failure of the lateral ectodermal ingrowths to meet in the mid-dorsal
line ; thus is formed the roof of the buccal cavity. A similar service in
the formation of the floor of the mouth is performed by the detached
constituents of the ventral muscle, in the mid-ventral line.

The buccal cavity is thus lined with ectoderm from two sources :
(1) the floor and roof are produced by a simple infolding of already formed
ectoderm, unaccompanied by a process of growth ; (2) the sides, which
extend farther posteriorly, by the latero-ventral ingrowths. The deep
groove in the ectoderm (Plate 3, Fig. 16, sul. ; Plate 4, Fig. 23)
persists as a lateral furrow at either side of the mouth (Plate 5, Fig.
26, ivag. os. I.).

When the zooids separate, and the lumina of the lateral ectodermal in-
vaginations become continuous with the lumen of the pharynx, there is
seen in this structure a pair of grooves (Plate 4, Fig. 18, sul.) which are
directly continuous with the ectodermal grooves of the mouth and buccal
cavity. By muscular contractions the old gut is loosened from its re-
maining points of attachment to the new, chiefly on the ventral aspect,
and the mouth and pharynx are functionally and structurally complete.

134 BULLETIN: MUSEUM OF COMPAKATIVE ZOOLOGY.

8. Conclusions.

1. Two normal regenerative pi^ocesses are to be distinguished in Dero
(1) that by which new segments are formed, limited to the region imme-
diately in front of the anal segment, and (2) the formation of the bud-
ding zone in segments which have been derived from (1). The first
process gives rise to undifferentiated materials from which the bud-zone
is produced ; the bud-zone, in turn, produces a segment-forming zone.
These two processes are so related in this species that budding may be
looked upon as a specialized method of segment formation.

2. The sexual gonads in Dero vaga arise upon the posterior walls of dis-
sepiments iii/iv and iv/v, in individuals known to be produced asexually.
The dissepiment iii/iv, on which the testes occur, is produced de novo
in the budding process, and dissepiment iv/v is the bounding dis-
sepiment of the budding segment. These facts show a close relation
betweeu the sexual elements and the new structures formed from the
indifferent cell masses during the budding process, and suggest that
both are referable to the preanal segment-forming region as their
source.

3. The budding zone in Dero is formed, and division takes place mid-
way between dissepiments, i. e. in the middle of the segment, if the dis-
sepiments mark the boundaries of segments. Semper ('76) states that
in Nais barbata and Xais proboscidea budding occurs between two old
segments. Von Kennel ('82) says that the bud is formed immediately
behind a dissepiment in Ctenodrilus pardalis. Von Bock ('97) describes
the bud as forming between segments in Chsetogaster. His Fig. 14, Taf.
VI. illustrates this and shows the dissepiment splitting. It appears to
me, however, desirable that the method of the formation of new segments
in Chpetogaster be more fully made out. It seems to me that the split-
ting of the dissepiment may possibly be a part of the process of the
formation of new segments rather than of the immediate production of
the more specialized bud-zone, and that the separation occurs, after all,
in the middle of this newly formed segment. The appearance of the gut
in Figure 14 would suggest this interpretation. We know that new
segments are formed in these regions in Chaetogaster, and it is not
clear how much of this precedes and how much follows the budding
process.

galloway: non-sexual reproduction in dero vaga. 135

4. The histological steps of bud formation in Dero agree for the most
part with those described for Chsetogaster by von Bock. This is true of
the orio-in of the brain from paired ectodermic ingrowths, the reinforce-
ment of the ventral nerve cord by serial ectodermic invasions between
the longitudinal muscle bauds, the ectodermic origin of the bristle sacs
in the newly formed segments, the origin of the pharynx from the ento-
dermal cells of the old gut, the paired ectodermic invaginations con-
cerned in the formation of the mouth, and the absence of a proctodseum
in the anterior individual. The chief points of difference which should
be noted are as follows.

(a.) In Chsetogaster the old entoderm and the wall of the body cavity
unite directly in the anal segment. In Dero there is an outgrowth of
newly formed ciliated entoderm which unites the old entoderm with the
body wall. This new entoderm lines the pavilion. The anus is at the
place of union of the old and new entoderm. According to von Kennel
(82), there is a proctodeum in Ctenodrilus pardalis.

(b.) While the mouth and pharynx of Dero and Chsetogaster present
certain common features in their mode of origin, there are important dif-
ferences between the two worms in each region. In Chsetogaster the paired
ectodermal invaginations ultimately run together across the median plane
to the mouth, and they secure union with the new pharynx before the
separation of the zodids. The ectodermic invaginations in Dero, on the
other hand, never meet, producing merely the buccal sinus at either
lateral angle of the mouth cavity. The floor and roof of the mouth are
formed by mechanical infolding of the appropriate portions of the anterior
margin of the body wall around the old gut. This old ectoderm is brought
into contact with the new entoderm and with the new ectoderm of the
invaginations mentioned above, after the separation of the zodids. The
pharynx is formed in Chsetogaster by an outgrowth of the entodermal cells
on the floor of the old gut. This cell mass comes to be occupied by a lumen,
single behind but continued forward as a pair of curved branches which
come to communicate with the ectodermal invaginations which are the
beginnings of the mouth. The old gut becomes constricted, and, together
with the nerve cord and the ventral body wall lying between the funda-
ments of the mouth, is resorbed. In Dero the pharynx is formed around
the old gut, especially well developed on the dorsal side, and the cavity of
the gut becomes the lumen of the pharynx, the old wall being split away
from the new entoderm, but not losing its continuity until the zodids
separate, whereupon the old wall is entirely broken down and lost. The
differences cited above are perhaps to be correlated in part with the fact

136 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY.

that the zooids of Dero are much less inclined than those of Cheetogas-
ter to remain associated after the development of the bud zone has gone
so far as to render them capable of independent existence.

(c.) For Chaetogaster von Bock ('97, Taf. VI. Figs. 9, 13) figures the
connective as passing superficial to the lateral longitudinal muscle bands,
and describes the brain as arising by the fusion of the ectodermic con- .
tributions from the spaces between the dorsal and lateral (" obere Mus-
kelliicken"), and from those between the lateral and ventral muscle
bands (" untere Muskelliicken "). These cell masses grow dorsad, over-
arching the gut, and fuse with the corresponding masses of the other
side to form the brain. According to my observations on Dero, the
brain is formed wholly by cells which break through the dorsal longitudi-
nal muscle band. When the connective is formed, it lies superficial to
a detached lateral portion of the dorsal muscle and returns to the body
cavity between this and the lateral muscle (the region at which the cells
of the lateral line occur). The regions of ectodermal activity respon-
sible for the elements of the brain and connective are thus seen to lie
immediately superficial to the lower margins of the dorsal muscle band.
From this place there is an upward growth to form the brain, and a
ventral growth which marks the course of the connective as it passes
toward the ventral cord. The position of the mature connective ex-
ternal to certain longitudinal muscle fibres is thus rendered intelligible.
In a similar way the region from which nervous cells spring on the
ventral side, being superficial to the lateral fibres of the ventral muscle,
gives rise to elements, a part of which penetrate the ventral muscle band
(cutting off a lateral portion of it) to fuse with the ventral nerve
chain, while others pass into the body cavity between the ventral and
lateral muscles. The position of the connective in its course from the
brain to the ventral ganglia is thus determined, and, as in the dorsal
region, is superficial to the lateral portion of the ventral longitudinal
muscle.

While it is entirely possible that the differences in the families studied
may account for the apparent differences in the origin of the nervous ele-
ments, the course of the connective as figured by von Bock would suggest
that the region superficial to the lateral muscle band is the place at
which the nervous elements in this case arise, and that the more dorsal
ingrowth is chiefly, if not wholly, the source of the brain, — the one
between the lateral and ventral bands marking the course of the connect-
ive, rather than contributing directly to form the brain. Otherwise, it
is very difficult to see why the connective should not follow the line of

GALLOWAY: NON-SEXUAL REPRODUCTION IN DERO VAGA. 137

least resistance along the inside of the body wall, rather than penetrate
the body wall, as its course actually indicates. I am aware that this
suggestion is based upon general considerations and upon analogy merely,
while the care shown throughout von Bock's paper gives abundant ground
for the trustworthiness of his results.

I have great pleasure, finally, in expressing the sense of obligation I
feel to Dr. E. L. Mark, under whose direction this work has been done,
for his advice and assistance throughout its prosecution.

138 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY.

BIBLIOGKAPHY.

Beddard, F. E. *

'95. A Monograph of the Order Oligocheeta. Oxford, xii + 769 pp.,
5 Plates.

Bock, M. von.

'97. Ueber die Knospung von Chaetogaster diaphanus. Jena. Zeitschr., Bd.
31, Heft 2, pp. 105-552, Taf. 6-8.

Bourne, A. G.

'86. On Budding in Oligochseta. Report of Brit. Assoc., pp. 1096, 1097-
Bourne, A. G.
'91. Notes on the Naidiform Oligochseta. Quart. Jour. Micr. Sci., Vol. 32.
pp. 335-356, Plates 26, 27.
Bulow, C.
'83. Die Keimscliichten des wachsenden Schwanzendes von Lumbriculus
variegatus, nebst Beitragen zur Anatomie und Histologie dieses Wurmes,
Zeitschr. f. wiss. Zool., Bd. 39, pp. 64-96, Taf. 5.

Bulow, C.

'83a. Ueber Theilungs- u. Regenerations vorgange bei "Wiirmern (Lumbriculus
variegatus, Gr.). Arch. f. Naturg., Jahrg. 49, Bd. 1, pp. 1-96.

Hatschek, B.

'58. Studien iiber Entwickelungsgeschichte der Anneliden. Arbeit, a. d. Zool.
Inst. Wien, pp. 277-404, Taf. 23-30.

Hepke, P.
'96. Zur Regeneration der Naiden. Zool. Anzeiger, Bd. 19, No. 520, p. 514.

Hepke, P.
'97. Ueber histo - und organogenetiscBe Vorgange bei Regenerations-processen
der Naiden. Zeitschr. f. wiss. Zool., Bd. 63, pp. 263-291, Taf. 14 und 15.

Hescheler, K.

'96. Ueber Regenerations Vorgange bei Lumbriciden. Jena. Zeitschr.,
Bd. 30, pp. 177-290 ; Bd. 31, pp. 521-604, Taf. 21-26.

Keller, J.
'94. Die Ungeschlechtliche Portpflanzung der Susswasserturbellarien.
Jena. Zeitschr., Bd. 28, pp. 370-407-

Kennel, J. von.
'82. Ueber Ctenodrilus pardalis. Arbeit, a. d. Zool. Inst. Wiirzburg, Bd. 5,
pp. 373-429, Taf. 16.

GALLOWAY: NON-SEXUAL REPRODUCTION IN DERO VAGA. 139

Leidy, J.
'80. Notice of some Aquatic Worms of the Family Naides. Amer. Nat,, Vol.
14, pp. 421-425.
Leuckart, F. S.
'51. Ueber die ungeschlechtliclie "Vermehrung bei Nais proboscidea. Arch,
f. Naturg., Jahrg. 17, Bd. 1, pp. 134-138, Taf. 2.

Perrier, E.

'70. Sur la reproduction scissipare des Naidiues. Compt. Rend., Tom. 70,

p. 1304..
Randolph, H.
'92. Regeneration of the Tail in Lumbriculus. Journ. Morph., Vol. 7,
pp. 317-344, Plate 19.
Reighard, J.
'84. On the Anatomy and Histology of Aulophorus vagus. Proc. Araer.
Acad. Arts and Sci., Vol. 20, pp. 88-106, Plates 1-3.

Rievel, H. von.

'96. Ueber die Regeneration des Vorderdarmes und Enddarmes bei einigen
Anneliden. Zeitschr. f. wiss. Zool., Bd. 62, pp. 2S9-341, Taf. 12-14.

Schultze, M.
'49. Ueber die Fortpflanzung durch Theiluug bei Nais proboscidea. Arch,
f. Naturg., Jahrg. 15, Bd. 1, pp. 293-304.

Schultze, M.

'52. Noch eiu Wort iiber die ungeschlechtliclie Vermehrung bei Nais probos-
cidea. Arch. f. Naturg., Jahrg. 18, Bd. 1, pp. 3-7, Taf. 1.

Seeliger, O.
'96. Natur u. allgemeiue Auffassung der Knospenfortpflanzuug der Metazoen.
Verhandl. der Deutsch. Zool. Gesellsch., pp. 25-58.

Semper, C.

'76. Ueber die Verwandtschaftsbeziehungen der gegliederten Thieve. Arbeit,
a. d. Zool. Inst. Wiirzburg , Bd. 3, pp. 115-404, Taf. 5-15.
Vejdowsky, F.

'84. System und Morphologie dev Oligochaeten. Pvag. 166 pp., 16 Taf.

Wagner, F. von.

'90. Zuv Kenntniss der ungeschlechtlichen Fortpflanzung von Microstoma.
Zool. Jahrb., Abth. f. Anat. u. Ontog., Bd. 4, pp. 349-423, Taf. 22-25.
Wagner, F. von.

'93. Einige Bemerkungen iiber das Verhaltniss von Ontogenie und Regenera-
tion (Vorlaufiger Bericht). Biol. Centralb., Bd. 13, pp. 2S7-296.
Wagner, F. von.

'97. Zwei Wort zuv Kenntniss dev Regenevation des Vovdevdavmes bei Lum-
briculus. Zool. Anzeiger, Jahrg. 20, No. 526, p. 69.

Zeppelin, Graf M.

'83. Ueber den Bau und die Teilungsvorgange des Ctenodrilus monostylus,
nov. sp. Zeitschr. f. wiss. Zool., Bd. 39, pp. 615-652, Taf. 36, 37.

140

BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY.

EXPLANATION OF PLATES.

ABBREVIATIONS.

a.

Anterior.

mu. r.

Radial muscles.

an.

Anus.

mu. v.

Ventral muscles (longitu-

cl. gn.

Ganglion cells.

dinal).

cl. In. 1.

Cells of the lateral line.

mu. v'.

Portion of ventral muscle

con't. crc'oe.

Circumoesophageal con-

detached by ectodermal

nective.

ingrowth.

con't. tis.

Connective tissue.

nph.

Nephridium.

d.

Dorsal.

a.

OSsophagus.

di'sep.

Dissepiment.

OS.

Mouth.

ec'drm.1

Old ectoderm.

P-

Posterior.

ec'drm.'2

New ectoderm.

pav.

Pavilion.

en'drm.1

Old entoderm.

pav. lu.

Lumen of pavilion.

en'drm.2

New entoderm.

phy.

Pharynx.

e'th.

Epithelium.

phy. d.

Dorsal wall of pharynx.

f'br. n.

Fibrous elements of ner-

phy. ec'drm.

Ectoderm of pharynx.

vous system.

phy. en'drm.

Entoderm of pharynx.

gn. su'ce.

Brain.

]>hy. lu.

Lumen of pharynx.

gn.v.

Ventral nerve ganglion.

phy. v.

Ventrai wall of pharynx.

in.

Intestine.

pr'c. dg.

Digitiform process.

i'vag.

Invagination.

I"'/' '

Proliferation.

i'cag. os. 1.

Lateral groove of mouth.

pr'stm.

Prostomium.

lu.i

Old lumen.

sac. drm-mu

. Dermo-muscular sac.

lu.2

New lumen.

sb-e'th.

Sub-epithelium.

mb. ba.

Basement membrane.

set. d.

Dorsal bristle sacs.

ms'drm.

Mesoderm.

set. v.

Ventral bristle sacs.

mu.

Muscles.

sul.

Groove.

mu. crc.

Circular muscles of body-

v.

Ventral.

wall.

vas. sng. d.

Dorsal blood-vessel.

mu. crc'oe.

Circumoesophageal mus-

vas. sng. crc

Vascular loops.

cles.

vas. sng. v.

Ventral blood-vessel.

mu.d.

Dorsal muscles (longitu-

z.

Undifferentiated preanal

dinal).

zone.

mu.d'.

Portion of dorsal muscle

2.1

Primary bud zone.

detached by ectodermal

Z.2

Secondary bud zone.

ingrowth.

zo'd. a.

Anterior zooid.

mu. 1.

Lateral muscles (longitu-
dinal).

zo'd. p.

Posterior zouid.

Galloway. — Dero vaga.

PLATE 1.

Fig. 1. Optical sagittal section of posterior portion of a worm not in process of
budding, showing undifferentiated preanal zone (z.). X 30.

Fig. 2. Similar view of portion of an individual in which the bud zone (z.1) is
already distinguishable. X 30.

Fig. 3. Budding worm in which the secondary bud zone (z.-) is recognizable.
The Roman numerals indicate the number of the segment behind the
prostomium of the original worm; the Arabic, those of the zooids.
The accented Arabic numerals, T-4', denote the zone of undifferen-
tiated materials out of which are to be formed the cephalic segments
of the individual produced by the second budding. X 100.

Fig. 4. Parasagittal section of a budding segment showing ectodermal thickening
and ingrowth on ventral surface only, at a fairly advanced stage. Sec-
tion considerably to one side of the median plane. X 655.

Fig. 5. Parasagittal section of the whole segment at a later stage. X 218.

Fig. 6. Cross section of a stage much more advanced than the preceding, through
the posterior zooid, showing the position of the brain (gn. su'oe.). X 218.

Fig. 7. Transverse section of a similar stage, posterior to position of the brain,
showing the regions where the ectoderm grows in between the muscle
bands. X 218.

Galloway- Dero.

Plate 1.

.pay

z.a 3.

20_ 2/ -

I

prt.

s.

en "drm '

on su oe

el in '

vas.snq.'i.

eri'drm!

prt.

rn'Hn

. IHU.L

ffnv:

/"'/'

!

Galloway. — Dero vaga.

PLATE 2.

Fig. 8. Transverse section of the same worm as that from which Figure 7 is
drawn, but anterior to the plane of separation. X 218.

Fig. 9. Transverse section of posterior zooid ill the region of the pharynx (phy.).
The stage is later than that represented in Figures 6-8. The lumen ap-
pears as a cavity formed by the splitting of the old entoderm (en'drm.1)
from the new (en'drm.2).

Fig. 10. Semi-diagrammatic representation of the course of the circumcesophageal
connective (con't. crc'oe.) in its relation to the ganglia, the muscles, and
the lateral line cells (cl. In. /.). X 200.

Fig. 11. Parasagittal section showing the ectodermal thickening on the ventral
surface of the budding segment, at a very early stage. X 500.

Fig. 12. Similar section at a later stage. X 500. (Figures 11, 12, are to be com-
pared witli Figures 4 and 5.)

Fig. 13. Part of longitudinal section of normal entodermal layer of intestine, show-
ing sub-epithelial cells {sb-e'tk.). X 800.

Fig. 14. Sagittal section of head of posterior zooid which has been separated from
anterior zooid one hour, showing the extent of ectoderm and entoderm
in the formation of the mouth and pharynx, in the median plane
Ec'dnn:1 marks the boundary between new ectoderm and new ento-
derm ; ec'dnn.1, the boundary between the newly formed ectoderm and
the old ectoderm of the mechanically infolded edge of the prostomium,
formerly attached to the anterior zooid. X 218.

Galloway-Dero.

T E 2 .

\O/LStC0€.

- . trtrf I

set, v. _

rtucv/

in.t'r/rm

sul.

rin.su.oe

■

mi. ha.

fe -

— et ■/'■/,, .'

■

aihtrrn?

T.W.G. del.

B.Meise.

Galloway — Dero vaga.

PLATE 3.

Fig. 15. Sagittal section showing development of the ventral nerve chain in the
preanal zone of a recently separated zobid. X 312.

Fig. 16. Parasagittal section, considerably to one side of median plane ; stage
about the same as that represented in Fig. 9. X 218.

Fig. 17. Similar section, later stage. X 375.

VLLOW/

rrui.ei:

ma,d ffn.sii.of.

r '- 3.

phy. hi . pky.enMrm

-

16.

sheA.

pa I :

Hi t/nii_ .

' rjrm .

" • ■irt.v.fl-*-)

pr'cdf.

par,.

■

■■■'• rrr- sndrm..

1.5.

miLcrc. ■ '

I?.

p.

ifn.r

ras.snff.cra.;

B.Meise.

Galloway. — Dero vaga.

PLATE 4.

Fig. 18. Cross section of the pharynx of a mature worm. X 312.

Figs. 19-24. Posterior face of cross sections from a single series through a bud-
zone in which separation is imminent. X 312. The higher the num-
ber, the farther forward the section is in the series. Figures 19-21
belong entirely to the posterior zooid. Figures 22-24 show parts of both
zooids. Sections 10 /x thick.

Fig. 19. Section through pharynx of posterior zooid; dorsal lumen conspicuous.

Fig. 20. Next section in front of that of Figure 19 showing anterior portion of dor-
sal pharyngeal lumen.

Fig. 21. Same series, two sections anterior to the preceding.

Fig. 22. Section (3rd in front of that of Figure 21) passing through ectodermal
invagination (sh/.) between the zooids. The entodermal wall of pavilion
(en'drm.) is seen below the gut.

Fig. 23. The section immediately in front of the last, showing that on the left side
the plane of section is farther cephalad than on the right side.

Galloway-Dero

Plate 4-.

1$.

_ phy </

75

/•«.-

,-.- ■ -n'r/rm.

21.

gn.su oe.

pkv estfrrri.

'fj

S3.

conlcrc'oi.

' ■ii'tlrn.*

23.

ert 'drnt f

p;\ i

ec'drm

20.

r*t, r//7'l

'III

Galloway. — Dero vaga.

PLATE 5.

Fig. 24. Only the margin of the prostomium of posterior zobid appears in this sec-
tion, which is three sections anterior to that of Fig. 23. The ventral
part of the section passes through the undifferentiated segment-form-
ing zone of the anterior individual. (Compare Fig. 15, Plate 3, which
represents a more advanced stage in longitudinal section.)

Fig. 25. Frontal section through stage of about the same age as that shown in
Fig. 16, Plate 3. X 218.

Fig. 26. Frontal section through head of posterior zouid which has been separated
from the anterior zooid two hours. X 166.

Fig. 27. Section of same series, four sections ventral to the preceding. The lateral
grooves of the mouth are shown (i'vag. os. I.). X 166.

Fig. 28. Parasagittal section, nearly median, through the head of a posterior
zooid, showing distribution of muscles. Ventral nerve cord is not
shown. X 166.

Fig. 29. Parasagittal, but more lateral, section from the same series. X 166.

tn'drin'

Dl '.sijn.

5.

26.

ivay.os.l_..

mu.v

■roe. i

pav.

.■'■•■■■

■

0\ m-i

en'drm2

rri'drm .'

27.

pr'stnt.
fvag.os.l

.eridrm'

en (Inn

■pr'stm

■

-

1

J'f

set.v.

VII

'///

., mit.oL'

■

, gn.su.oe.

gnv

29.

■

The following Reports have been published or are in prepa-
ration on the Dredging Operations off the West Coast of
Central America to the Galapagos, to the West Coast of
Mexico, and in the Gulf of California, in Charge of Alex-
ander Agassiz, carried on by the U. S. Fish Commission
Steamer " Albatross," during 1891, Lieut. Commander Z. L.
Tanner, U. S. N., Commanding.

A. AGASSIZ. II.i General Sketch of the
Expedition of the "Albatross," from
February to May, 1S91.

A. AGASSIZ. The Pelagic Fauna.

A. AGASSIZ. The Deep-Sea Panamic Fauna.

A. AGASSIZ. I.2 On Calamocrinus, a new

Stalked Crinoid from the Galapagos
A AGASSIZ. XXIII.23 The Echini.
JAS. E. BENEDICT. The Annelids.
R. BERGH. XIII. i3 The Nudibranchs.
K. BRANDT. The Sagittaa.
K. BRANDT. The Thalassicolas.
C. CHUN. The Siphonophores.
C. CHUN. The Eyes of Deep-Sea Crustacea.
S. F. CLARKE. XI.u The Hydroids.
W. H. DALL. The Mollusks.

W. FAXON. VI.3 XV.w The Stalk-eyed
Crustacea.

S. GARMAN. The Fishes.

W. GIESBRECHT. XVL>5 The Copepods.

A. GOES. Ill* XX.m The Foraminifera.

H. J. HANSEN. XXII. 22 The Cirripeds

and Isopods.

C. HARTLAUB. XVIII." The Comatula?.

W. A. HERDMAN. The Ascidians.

S. J. HICKS0N The Antipathids.

W. E HOYLE. The Cephalopods.

G. von KOCH. The Deep-Sea Corals.

C. A. KOFOID. Solenogaster.

R von LENDENFELD. The Phosphores-

H. LUDWIG.

rians.
C

IV.5 XII." The Holothu-

cent Organs of Fishes.

F. LUTKEN and TH. MORTENSEN.

The Ophiuridse.
OTTO MA AS. XXI.21 The Acalephs.
J. P. McMURRICH. The Actinarians.
E. L. MARK. XXIV.2* The Cerianthidae.

GEO P. MERRILL. V.« The Rocks of the

Galapagos.
G.#W. MULLER. XIX.19 The Ostracods.
JOHN MURRAY. The Bottom Specimens.
A. ORTMANN. XIV.12 The Pelagic Schi-

zopods.
ROBERT RIDGWAY. The Alcoholic Birds.
P. SCHIEMENZ. The Pteropods and Het-

eropods.
W. SCHIMKEWITSCH. VIII.8 The Pyc-

nogonidse.
S. H. SCUDDER. VII.' The Orthoptera of

the Galapagos.
W. PERCY SLADEN. The Starfishes.
L. STEJNEGER. The Reptiles.
TH. STUDER. X.10 The Alcyonarians.
C. H. TOWNSEND. XVII." The Birds of

Cocos Island.
M. P. A. TRAUTSTEDT. The Salpidaa and

Doliolidae.
E. P. VAN DUZEE. The Halobatidse.
H. B. WARD. The Sipunculoids.
H. V. WILSON. The Sponges.
W. McM. WOODWORTH. IX.» The Pla-

narians and Nemerteans.
W. McM. WOODWORTH. XXVII."

Planktonemertes.

89 pp

Bull M.O. Z., Vol. XXI., No. 4, June 1891, 16 pp. ; and Vol. XXIII., No 1, February. 1892,

, 22 Plates

Mem. M. C. Z., Vol. XVII., No. 2, January, 1892, 95 pp., 32 Plates.

Bull. M. C. Z., Vol XXIV., N,o. 7, August, 1893, 72 pp.

Bull. M. C. Z., Vol XXIIIy No. 5, December, 1892, 4 pp.. 1 Plate.

Bull. M. C, Z., Vol. XXIV.. No 4, June, 1893, 10 pp. [Zool. Anzeig., No. 420, 1893.]

Bull M. 0. Z., Vol. XVI., No. 13, July, 1893, 3 pp

Bull. M. C. Z , Vol. XXV , No 1 September, 1893, 25 pp.

Bull. M. C. Z . Vol. XXV.. No. 2. December, 1893, 17 pp., 2 Plates.

Bull. M. 0. Z., Vol. XXV., No. 4. January, 1894, 4 pp., 1 Plate.

Bull. M. C. Z., Vol. XXV., No. 5, February, 1894, 17 pp.

Bull. M. C Z , Vol. XXV No. 6, February, 1894, 7 pp., 5 Plates.

Bull. M. C. Z., Vol. XXV. No. 8. September. 1894; 13 pp., 1 Plate.

Bull. M. C Z . Vol XXV. No. 10, October, 1894, 109 pp , 12 Plates.

Mem. M. 0. Z . Vo'. XVII. No. 3, October, 1894, 183 pp.. 19 Plates.

Bull M. C. Z , Vol XXV. No 12. April. 1895, 20 pp.. 4 Plates.

.Mem M. 0 Z Vol XVIII.. April. 1895. 292 pp., 67 Plates, 1 Chart

Bull. M. C Z , Vol. XXVII . No. 3, July, 1895. 8 pp.. 2 Plates.

Bull. M. C. Z , Vol XXVII , No 4, August, 1895. 26 pp , 3 Plates.

Bull. M. C. Z , Vol. XXVII., No. 5, October, 1895, 14 pp., 3 Plates

Bull M. C. Z.. Vol XXX , No 1, March, 1896, 103 pp., 9 Plates, 1 Chart.

Mem M C. Z., Vol. XXIII , No. 1, September, 1897, 92 pp., 15 Plates.

Bull. M. U. Z , Vol XXXI , No 5, December, 1897. 37 pp., 6 Plates, 1 Chart.

Bull. M G. Z., Vol XXXII., No. 5, May, 1898. 18 pp , 13 Plates. 1 Chart.

Bull. M C. Z..Vol XXXi I.. No. 8, August, 1S98, 8 pp.. 3 Plates.

Bull. M. C. Z., Vol. XXXV., No. 1, July 1899, 4 pp., 1 Plate.

PUBLICATIONS

OF THE

MUSEUM OF COMPARATIVE ZOOLOGY

AT HARVARD COLLEGE.

There have been published of the Bulletins Vols. I. to XXXI.,
and XXXIII. ; of the Memoirs, Vols. I. to XXII.

Vols. XXXIV. and XXXV. of the Bulletin, and Vols. XXIII.
and XXIV. of the Memoirs, are now in course of publication,.

The Bulletin and Memoirs are devoted to the publication of
original work by the Professors and Assistants of the Museum, of
investigations carried on by students and others in the different
Laboratories of Natural History, and of work by specialists based
upon the Museum Collections.

The following publications are in preparation : —

Reports on the Results of Dredging Operations from 1877 to 1880, in charge of
Alexander Agassiz, by the U. S. Coast Survey Steamer " Blake," Lieut.
Commander C. I). Sigsbee, U. S. N., and Commander J. K. Bartlett, IT. S. N.,
Commanding,

Reports on the Results of the Expedition of 1801 of the U. S. Fish Commission
Steamer "Albatross," Lieut. Commander Z. L. Tanner, U. S. N., Com-
manding, in charge of Alexander Agassiz.

Contributions from the Zoological Laboratory, in charge of Professor E. L.
Mark.

Contributions from the (Geological Laboratory, in charge of Professor N. S.
Slialer.

Studies from the Newport Marine Laboratory, communicated by Alexander
Agassiz.

Subscriptions for the publications of the Museum will lie received
on the following terms : —

For the Bulletin, $4.00 per volume, payable in advance.
For the Memoirs, $8.00 " "

These publications are issued in numbers at irregular inter-
vals ; one volume of the Bulletin (8va) and half a volume of the ■
Memoirs (4to) usually appear annually. Each number of the Bul-
letin and of the Memoirs is also sold separately. A price list
of the publications of the Museum will be sent on application
to the. Librarian of the Museum of Comparative Zoology, Cam-
bridge, Mass.

OCT 18 1«M

Bulletin of the Museum of Comparative Zoology
AT HARVARD COLLEGE.

Vol. XXXV. No. 6.

THE PHOTOMECHANICAL CHANGES IN THE RETINAL
PIGMENT OF GAMMARUS.

By G. II. Parker.

With < ixi; Plate.

CAMBRIDGE. MASS., U.S.A.:

PRINTED FOR THE MUSEUM.

October, 1899.

Bulletin of the Museum of Comparative Zoology
AT HARVARD COLLEGE.

Vol. XXXV. No. 6.

THE PHOTOMECHANICAL CHANGES IN THE RETINAL
PIGMENT OF GAMMARUS.

By G. H. Parker.

With One Plate.

CAMBRIDGE, MASS., U. S. A. :

PRINTED FOR THE MUSEUM.

October, 1899.

OCT 18 1899

No. 6. — The Photomechanical Changes in the Retinal Pigment of
Gammarus.1 By G. H. Pakker.

The following studies were made with the intention of comparing the
photomechanical changes in the retina of one of the simpler crustaceans
with those already known in the decapods. In this respect, very few of
the simpler forms have been studied, and almost all that has been done
is contained in a single paper by Szczawinska ('91). This paper appeared
shortly before Exner's ('91) well known monograph, in which the pho-
tomechanical changes of the compound eye first received a consistent
physiological interpretation, and consequently it does not touch several
important questions raised by Exner's work.

The lower crustaceans studied by Szczawinska were Gammarus, Phro-
nima, and Branchipus, and of these, judging from the figures given,
Gammarus has the most pronounced photomechanical changes. I have
therefore studied the common American form, Gammarus ornatus Milne-
Edwards. Vigorous individuals of this species were kept, some in the
dark and some in the light, for six hours, and were then killed by being
momentarily immersed in hot water (85° C). They were afterwards cut
in sections, and their eyes studied and compared.

The structure of the eye in Gammarus ornatus has already been de-
scribed (cf. Parker, '91; p. 68), and agrees almost exactly with that of
G. pulex as given by Carriere ('85, p. 156). Szczawinska's ('91, p. 534)
description of the eye in G. roeselii is so different from the accounts
given for the other two species that I am persuaded it must be in part
erroneous.

In G. ornatus the corneal cuticula (Fig. 1, crn.), which is not facetted,
is covered internally by a corneal hypodermis (cl. crn.) the cells of which
are not regularly grouped with reference to the ommatidia. The axis of
each ommatidiurh is occupied distally by a two-celled cone (con.) and
proximally by a slender rhabdome (rhb.). These two structures are
entirely distinct from each other and not continuous, as described by

1 Contributions from the Zoological Laboratory of the Museum of Comparative
Zoology at Harvard College. E. L. Mark, Director. No. C.
vol. xxxv. — no. G. *

144 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY.

Szczawinska ('91, p. 536). The transparent axis thus formed is sheathed
laterally by live elongated pigment cells, the retinular cells (cl. rtn.').
Each retinular cell consists of three parts : a flattened distal portion
applied to the side of the cone ; a middle rod-like portion (Fig. 2, cl. rtn.')
lying against the rabdome ; and an enlarged basal portion proximal to
the basement membrane (mb. ba.) and containing the nucleus. The
proximal end of this last part becomes attenuated and forms a retinal
nerve fibre (fbr. r.) which passes to the optic ganglion. All parts of
these cells except the nucleus may contain more or less blackish pig-
ment. The space between the ommatidia is filled with a coarse granular
pigment, whitish by reflected light and containing nuclei (nl. sn.). This
material is made up of the accessory pigment cells, the boundaries of
which are not easilv discernible.

In Szczawinska's description of the eye in Gammarus roeselii, the
three parts of the retinular cells described above are stated to be each
a separate cell containing its nucleus. She correctly identified the
nucleus in the proximal portion. What she believed to be the nuclei of
the middle parts (Planche XVI. Fig. 4, n.^pg.) are without question the
nuclei of the accessory pigment cells, which she failed to recognize as
such. In the distal portions of the retinular cells of Gammarus ornatus
nothing resembling nuclei could be discovered, and I am of opinion that
she was mistaken in attributing such bodies to the coi-responding parts
in G. roeselii. The continuity of the three parts of the retinular cells
in G. ornatus can be clearly demonstrated in serial transverse sections,
in longitudinal sections, and in isolation preparations, and I therefore
believe that these three portions are only parts of one cell. If this is
true of G. ornatus, it is probably also true of G. roeselii, and I am
strengthened in this belief as Szczawinska's own figures (Planche XVI.
Figs. 4, 5) admit more readily of this interpretation than they do of her
own. In these respects, then, her account is probably at fault.

In an eye of G. ornatus that had been subjected to light for some six
hours (Fig. 1), a considerable amount of black pigment was found uni-
formly distributed through the distal and middle portions of the retinular
cells, thus sheathing the cone and rhabdome laterally (Fig. 2). The
proximal portion of each cell contained a few irregularly scattered pig-
ment granules except near the nucleus, where the pigment was more
abundant.

In an eye from an animal kept some six hours in the dark (Fig. 3),
the pigment in the distal portions of the retinular cells presented the
same condition as in the eyes exposed to light. The middle portions,

PARKER: CHANGES IN RETINAL PIGMENT OF GAMMARUS. 145

however, were almost entirely devoid of pigment, while the proximal
portions were as densely filled with pigment as the distal portions.

Obviously the changes induced by the presence or absence of light
affect the pigment of only the middle and proximal parts. When an
animal that has been kept in the dark is exposed to the light, the pig-
ment that is massed in the proximal parts of the retinular cells (Fig. 3)
migrates distally and fills the middle portions, without however entirely
abandoning the proximal parts, especially around the nucleus (Fig. 1).
When an animal that has been kept in the light is placed in the dark,
the pigment in the middle portions (Fig. 1) migrates into the proximal
parts till almost no pigment is left in the middle portions. In other
words, the presence of light induces a distal migration of much of the
pigment from the proximal parts and the absence of light brings about a
proximal migration of almost all the pigment in the middle parts.

In none of my observations was there any evidence of photomechanical
changes in the accessoi-y pigment cells.

Aside from the dispai'ity due to the different anatomical descriptions
of the eyes, the physiological results given in this paper confirm in the
main those given by Szczawinska ('91, p. 548). In one respect only is
there a significant difference. Szczawinska claims that in G. roeselii the
pigment in what I have called the distal parts of the retinular cells shows
photomechanical changes. In G. ornatus no evidence of such changes
could be found, and since in G. roeselii, according to the figures given
by Szczawinska (Planche XVI. Figs. 1, 2), the supposed evidence of
these changes may be entirely the result of a slight difference in the
planes at which the sections have been cut, it may fairly be doubted if
these changes occur at all.

The relations that the photomechanical changes, described above, bear
to the physiology of the eye in G. ornatus are not far to seek. Light pass-
ing through the axis of any cone in the eye of this animal would be con-
ducted directly to the rhabdome under the given cone. Light entering
a cone obliquely to its axis would fall upon one of its pigmented sides,
where, if not absorbed, the light would suffer reflection. As the sides of
the cone are not parallel but approach proximally, the light would not
undergo simple internal reflection as in a cylinder, but would be so
turned at each reflection that it would eventually be discharged from
the end of the cone at which it entei*ed. Thus oblique light would not
reach the underlying rhabdome at all. This action, by which the axial
light of the cone is conducted to the rhabdome and the oblique light is
discharged, has been called by Exuer ('91, p. 59) the catoptric action of

146 bulletin: museum of comparative zoology.

the cone, and has been fully described by him. In accordance with this
relation, then, the rhabdome of each omniatidium would receive light
from an external region corresponding to the outward projection of the
axis of the cone distal to it.

As the pigment which surrounds the cone is merely concerned with
the absorption of the lateral rays, and as these rays would be equally
disturbing whether the animal were in bright light or dim light, it fol-
lows that no photomechanical changes in correspondence with changes in
the intensity of the light should be expected in this pigment, and as a
matter of fact in G. ornatus no such changes have been observed.

The axial light, which according to the foregoing account finds its way
into the rhabdome, must have a certain degree of intensity in order to
stimulate that organ. Ordiuary daylight is presumably more than suffi-
cient to call forth this stimulation, and such superfluous light as may
pass to the edges of the rhabdome or through it is probably absorbed by
the black pigment that in bright light (Figs. 1, 2) surrounds that body.
In dim light, however, there must be times when the light which en-
ters the rhabdome is scarcely intense enough to stimulate that organ.
Under such circumstances the more oblique rays, which ordinarily would
be absorbed by the black pigment on the sides of the rhabdome, would
materially aid in stimulating it if they were turned back into the rhab-
dome. That these rays are probably thus turned back is shown by the
fact that in dim light the black pigment is removed from the rhabdome
and the surrounding whitish reflecting pigment of the accessory pigment
cells is exposed (compare Figs. 2, 4).

Changes exactly comparable with these have been described in the
proximal retinular cells of the higher mistaceans. The changes in the
distal retinular cells, the iris pigment of Exner, which are connected
with the formation of superposition images, find no representatives in
the eyes of Gammarus ornatus. As this eye is in many respects primi-
tive, it is likely that the ancestral crustacean ommatidium possessed a
catoptric cone and a retinula provided with a reflecting apparatus for
use in dim light. In the differentiation of the higher type of crustacean
ommatidium the reflecting mechanism was retained, but the catoptric
cone gave way to a second type of cone provided with special pigment
cells whose movements were associated with the superposition images
formed by this type of cone.

PARKER: CHANGES IN RETINAL PIGMENT OF GAMMARUS. 147

Summary.

1. In Gammarus ornatus photomechanical changes in the retinal pig-
ment are not observable in the accessory pigment cells or in the distal
parts of the retinular cells, but are limited to the black pigment in the
middle and proximal portions of the retinular cells.

2. In the light the middle portion of each retinular cell is well filled
with pigment, thus enclosing the rhabdome in a black sheath ; the
proximal portion contains scattered grains except near the nucleus,
where the pigment is more massed.

3. In the dark the middle portion of each cell is almost free from
pigment, which now fills the proximal part.

4. In the dark the removal of the black pigment from around the
rhabdome exposes the accessory pigment cells, which probably act as
reflecting organs, and in very dim light turn such rays as have escaped
laterally from the rhabdome back into that structure, thus aiding in
an effective stimulation of this organ.

Cambridge, December 15, 1898.

$

148 bulletin: museum of comparative zoology.

PAPEES CITED.

Carriere, J.
'85. Die Sehorgane der Thiere vergleichend-anatomisch dargestellt. R. Olden-
bourg. Miinchen uud Leipzig. 6 + 205 pp., 147 Abbildmigen und 1
Tafel.

Exner, S.

'91. Die Physiologie der facettirten Augen vou Krebseu und Insecten.
Deuticke, Leipzig und Wieu. vi -f- 206 pp., 7 Taf.

Parker, G. H.

'91. The Compound Eyes in Crustaceans. Bull. Mus. Comp. Zool. Harvard
Coll., Vol. XXI. pp. 45-140, Plates I.-X.

Szczawinska, W.

'9SL Contribution a l'etude des yeux de quelques Crustaces et Recberches
experimentales sur les mouvements du pigment granuleux et des cellules
pigmentaires sous l'influence de la lumiere et de l'obscurite dans les yeux
des Crustades et des Arachnides. Arch, de Biol., Tom. X. pp. 523-568,
Planches XVI., XVII.

Parker. — Retinal Pigment of Gammarus.

EXPLANATION OF THE PLATE.

All the figures were taken from preparations of the eyes of Gammarus ornatus
M.-Edw. They were drawn with the aid of an Abbe' camera, and are magnified
570 diameters.

ABBREVIATIONS.

cl. cm. Corneal hypodermis. fbr. r. Retinal nerve fibre.

cl. rtn.' Retinular cell. mb. ba. Basement membrane.

cl. sn. Accessory pigment cell. nl. rtn.' Nucleus of retinular cell.

con.
cm.

Cone. nl. sn. Nucleus of accessory pigment cell.

Corneal cuticula. rhb. Rhabdome.

Fig. 1. Longitudinal section of an ommatidium, showing the arrangement of pig-
ment due to exposure to bright light.

Fig. 2. Transverse section of a retinula from such a preparation as that shown in
Figure 1, and taken near the level marked rhb. in that figure.

Fig. 3. Longitudinal section of an ommatidium showing the arrangement of pig-
ment due to the absence of light.

Fig. 4. Transverse section of a retinula from such a preparation as that shown in
Figure 3, and taken near the level marked rhb. in Figure 1.

I Parker-Photc

LIGHT.

DARK.

clrbil

cLsn

nl sn.

mb 6a.

v,..m mm

'&-:'■ $\i''< 'X&-A &<'•"$

&•■; md ~Mk mi

>»£.-'■ *£*5*": .-,?*akv -SawR'

..nl rlJi'

rhh.

■I rtii

...

4:

The following Reports have been published or are in prepa-
ration on the Dredging Operations off the West Coast of
Central America to the Galapagos, to the West Coast of
Mexico, and in the Gulf of California, in Charge of Alex-
ander Agassiz, carried on by ths U. S. Fish Commission
Steamer " Albatross," during 1891, Lieut. Commander Z. L.
Tanner, U. S. N., Commanding.

A. AGASSIZ. II.i General Sketch of the
Expedition of the "Albatross," from
February to May, 1891.

A. AGASSIZ. The Pelagic Fauna.

A. AGASSIZ. The Deep-Sea Panamic Fauna.

A. AGASSIZ. I.2 On Calamocrinus, a new-
Stalked Crinoid from the Galapagos
A. AGASSIZ. XXIII.23 The Echini.
JAS. E. BENEDICT. The Annelids.
R. BERGH. XIII. is The Nudibranchs.
K. BRANDT. The Sagitta;.
K. BRANDT. The Thalassicolfe.
C CHUN. The Siphonophores.
C. CHUN. The Eyes of Deep-Sea Crustacea.
S. F. CLARKE. XI." The Hydroids.
W. H. DALL. The Mollusks.

W. FAXON. VI.3 XV.i-s The Stalk-eyed
Crustacea.

S. GARMAN. The Fishes.

W. GIESBRECHT. XVI.is The Copepods.

A. GOES. Ill" XX.» The Foraminifera.

H. J. HANSEN. XXII. 22 The Cirripeds

and Isopods.

C. HARTLAUB. XVIII.'s The Comatulse.

W. A. HERDMAN. The Ascidians.

S. J. HICKSON. The Antipathids.

W. E HOYLE. The Cephalopods.

G. vo.v KOCH. The Deep-Sea Corals.

C. A. KOFOID. Solenogaster.

R. VON LENDENFELD. The Phosphores-
cent Organs of Fishes.

H. LUDWIG. IV.5 XII. » The Holothu-

rians.
C. F. LUTKEN and TH. MORTENSEN.

The Ophiurida?.
OTTO MAAS. XXI.21 The Acalephs.
J. P. McMURRICH. The Actinarians.
E. L. MARK. XXIV.24 The Cerianthidaa.
GEO. P. MERRILL. V.« The Rocks of the

Galapagos.
G. W. MULLER. XIX." The Ostracods.
JOHN MURRAY. The Bottom Specimens.
A. ORTMANN. XIV. « The Pelagic Schi-

zopods.
ROBERT RIDGWAY. The Alcoholic Birds.

P. gCHIEMENZ. The Pteropods and Het-

eropods.
W. SCHIMKEWITSCH. VIII.s The Pyc-

nogonida?.
S. H. SCUDDER. VII.7 The Orthoptera of

the Galapagos.

"W. PERCY SLADEN. The Starfishes.
L. STEJNEGER. The Reptiles.
TH. STUDER. X.10 The Alcyonarians.
C. H. TOWNSEND. XVII." The Birds of

Cocos Island.
M. P. A. TRAUTSTEDT. The Salpidse and

Doliolida?.
E. P. VAN DUZEE. The Halobatida?.
H. B. WARD. The Sipunculoids.
*H. V. WILSON. The Sponges.
W. McM. WOODWORTH. IX.» The Pla-

narians and Nemerteans.
W. McM. WOODWORTH. XXVII."

Planktonemertes.

89 pp

Bull M. C. Z., Vol. XXI., No. 4, June 1891. 16 pp. ; and Vol. XXIII., No 1, February. 1892,

, 22 Plates

Mem. M. C. Z., Vol. XVII., No. 2, January, 1892, 95 pp., 32 Plates.

Bull. M. C. Z., Vol XXIV., No. 7, August, 1893, 72 pp.

Bull. M. C. Z., Vol XXIII., No. 5, December, 1892, 4 pp., 1 Plate.

Bull. M. C. Z., Vol. XXIV.. No 4, June, 1893, 10 pp [Zool. Anzeig., No. 420, 1898.]

Bull M. C. Z., Vol. XVI., No. 13, July, 1893, 3 pp

Bull. M. C. Z , Vol. XXV , No 1 September, 1893, 25 pp.

Bull. M. C. Z , Vol. XXV., No. 2. December, 1893, 17 pp., 2 Plates.

Bull. Mi 0. Z., Vol. XXV., No. 4. January, 1894, 4 pp.. 1 Plate.

Bull. M. C. Z., Vol XXV., No. 5, February, 1894, 17 pp.

Bull. M. C Z , Vol. XXV No. 6, February, 1894, 7 pp., 5 Plates.

Bull. M. C. Z., Vol. XXV. No 8, September. 1894, 13 pp., 1 Plate.

Bull. M. C Z . Vol XXV N... 10, October, 1894, 109 pp , 12 Plates.

Mem. M. 0. Z . Vo\ XVlf. No. 3, October, 1894, 183 pp.. 19 Plates.

Bull M. C. Z , Vol. XXV. No 12, April, 1895, 20 pp.. 4 Plates.

Mem M. C. Z . Vol. XVI1L, April. 1895. 292 pp., 67 Plates, 1 Chart

Bull. M. C Z , Vol. XXVII., No. 3, July, 1895. 8 pp.. 2 Plates.

Bull. M. C. Z , Vol XXVII , No 4, August. 1895. 26 pp., 3 Plates.

Bull M C. Z , Vol. XXVII., No. 5, October, 1895. 14 pp., 3 Plates

Bull. M. C. Z.. Vol XXTX , No 1, March, 1896, 103 pp., 9 Plates, 1 Chart.

Mem M C. Z., Vol. XXIII , No. 1, September. 1897, 92 pp., 15 Plates.

Bull. M. iv Z , Vol XXXI , No 5, December, 1897. 37 pp., 6 Plates, 1 Chart.

Bull. M. C. Z., Vol XXXIT., No. 5, May, 1898. 18 pp , 13 Plates, 1 Chart.

Bull. M C. Z.. Vol XXX M.. No. 8, August, 1898, 8 pp.. 3 Plates.

Bull. M. C. Z., Vol. XXXV.. No. 1, July 1899, 4 pp., \ Plate

PUBLICATIONS

OF THE

MUSEUM OF COMPARATIVE ZOOLOGY

AT HARVARD COLLEGE. •

There have been published of the Bulletins Vols. I. to XXXI.,
and XXXIII. ; of the Memoirs, Vols. I. to XXII.

Vols. XXXIV. and XXXV. of the Bulletin, and Vols. XXIII.
and XXIV. of the Memoirs, are now in course of publication.

The Bulletin and Memoirs are devoted to the publication of
original work by the Professors and Assistants of the Museum, of
investigations carried on by students and others in the different
Laboratories of Natural History, and of work by specialists based
upon the Museum Collections.

The following publications are in preparation : —

Reports on the Results of Dredging Operations from 1877 to 1880, in charge of
Alexander Agassiz, by the U. S. Coast Survey Steamer " Blake," Lieut.
Commander C. I). Sigshee, U. S. N.,and Commander J. R. Bartlett, V. S. X.,
Commanding.

Reports on the Kesults of the Expedition of 1891 of the U. S. Fish Commission
Steamer " Albatross," Lieut. Commander Z. L. Tanner, U. S. N., Com-
manding, in charge of Alexander Agassiz.

Contributions from the Zoological Laboratory, in charge of Professor K. L.
Mark.

Contributions from the Geological Laboratory, in charge of Professor N. S.
Slialer.

Studies from the Newport Marine Lai oratory, communicated by Alexander
Agassiz.

Subscriptions for the publications of the Museum will be received
on the following terms : —

For the Bulletin, $4.00 her volume, payable in advance.

For the Memoirs, $8.00 lt tk ''

These publications are issued in numbers at irregular inter-
vals; one volume of the Bulletin (8vo) and half a volume of the
Memoirs (4to) usually appear annually. Each number of the Bul-
letin and of the Memoirs is also sold separately. A price list
of the publications of the Museum will be sent on application
to the Librarian of the Museum of Comparative Zoology, Cam-
bridge, Mass.

DEC 22 1899

Bulletin of the Museum of Comparative Zoology
AT HARVARD COLLEGE.

Vol. XXXV. No. 7.

THE STRUCTURE AND DEVELOPMENT OF THE ANTENNAL
GLANDS IN HOMARUS AMERICANUS MILNE-EDWARDS.

By Fredeuick C. Waite.

With Six Plates.

CAMBRIDGE, MASS , U. S. A. :

PRINTED FOR THE MUSEUM.

December, 1899.

Bulletin of the Museum of Comparative Zoology
AT HARVARD COLLEGE.

Vol. XXXV. No. 7.

THE STRUCTURE AND DEVELOPMENT OF THE AXTEXXAL
GLAXDS IX HOMAEUS AMERICAXUS M1LXE-EDWARDS.

Br Frederick C. Waite.

With Six Plates.

CAMBRIDGE, MASS., U. S. A. :

PRINTED FOR THE MUSEUM.

December, 1899.

DfcC 83 1899

No. 7. — The Structure and Development of the Antennal Glands
in Homarus americanus Milne-Edwards} By Fkedekick C.
Waite.

CONTENTS.

Introduction
Methods . .

II

Structure in the Adult

A. Gross Anatomy

B. Finer Anatomy
Development . . .

A. Historical Summary

B. Development in the Embryo

C. Development in the Larvae
(a) In the First Larva . .

1. General Structure . .

2. Histology

PAGE

151
154
157
157
162
168
168
172
182
182
182
185

II. Development, continued.

(b) In Older Larvse ... 188
III. Theoretical Considerations . 191

A. Homology of the Antennal

Glands with the Ne-
phridia of Annelids . . 191

B. The Number of Metameric

Organs of this Nature in

Crustacea 196

Summary ■ 201

Bibliography 205

Explanation of Plates 210

Introduction.

The antennal glands of Crustacea have been the subject of much dis-
cussion ever since the first discovery cf them in Astacus by Rosel von
Rosenhof (1755, p. 321). The true relation of the parts of the organ
was for a long time overlooked, and the gland proper was considered as
distinct from the storage reservoir and duct to the exterior. The former
was supposed to be connected with the digestive tract, either as a sali-
vary gland or as the source of the gastroliths, while the latter was taken
for a sense organ, first as auditory, later as olfactory. Brandt ('33,
p. 64) showed that the parts which had been considered as two organs
of different functions really constituted a single structure. To this was
ascribed the function of hearing.

All observations were confined to Astacus until Semper ('61) extended
his work to Lucifer. Shortly afterward Claus ('63) added observations

1 Contributions from the Zoological Laboratory of the Museum of Comparative
Zoology at Harvard College, under the direction of E. L. Mark, No. 105.

152 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY.

upon a few other forms. The first extensive comparative work was that
of Grohben ('80), who studied these organs in a number of Entomostraca
and Malacostraca, and sought to prove them homologous in the two
groups.

The earliest evidence of the excretory function of these organs was
afforded by Will und Gorup-Besanez ('48), who made the following
statement (column 828) : "... in der That haben wir im sogenannten
griinen Organ des Flusskrebses (Astacus fliwiatilis) und im Bojanus'-
schen Organ des Teichmuschel ( Anodonta) einen Stoff aufgefunden, der
Reaktionserscheinungen zeigte, die mit der grossten Wahrscheinlichkeit
auf Guanin hinweisen, doch gebrach es uns bisher an dem nothigen
Material, urn entscheidende Versuche damit anzustellen." Similar chem-
ical experiments were carried on by several later investigators, but the
most conclusive evidence of the excretory activity of these organs rests
upon extensive experiments in intra vitam injection and feeding which
were inaugurated by Kowalevsky ('89), and extended in subsequent
papers by him and by other authors, especially by Cuenot ('94).

By far the most comprehensive work on excretion in Crustacea is that
of Marchal ('92). Until the publication of this work, the major part of
the literature, especially as regards histology, was on Astacus, with addi-
tions on some of the Carididse (Grohben '80, Kowalevsky '89, and
Weldon '89 and '91). Marchal in his memoir gives the results of a
careful reseai'ch on the morphology, histology, and physiology of a long
series of species representing all the typical genera of the Decapoda.
The work was carried on by the methods of modern technique, including
the intra vitam method of Kowalevsky ('89).

This is the only paper that deals with the antennal gland in Homarus.
The species studied by Marchal (H. vulgaris), however, was not the
same as that of the present paper. In H. vulgaris, according to Mar-
chal (pp. 156-163), the gland presents an entirely different appearance
from that seen in Astacus, although the main divisions into gland proper,
vesicle, and duct are found in both. In H. vulgaris the gland as seen
from above is triangular in shape, the antero-posterior axis being the
longest. The posterior tip is curved mediad, and the periphery is more or
less notched. The dorsal face is slightly concave, the ventral markedly
convex.

The structure is much less complex than in Astacus, and there are
two distinct regions besides the duct, — the saccule and the labyrinth.
The former is thin and flattened out over the dorsal face of the gland,
covering it except for a narrow projecting margin. The saccule has a

waite: antennal glands in homarus americanus. 153

large central cavity, from which radiate ramifying culs-de-sac. The
labyrinth is three times as thick as the saccule and forms the greater
bulk of the gland. It is divided by an antero-posterior fissure on the
ventral face, so as to be U-shaped, the arms of the U pointing anteriorly.
The connection with the overlying saccule is at the end of the lateral
arm, that with the excretory duct at the end of the median arm. In
the latter case, this is by means of several parallel canaliculi, which open
as pores in the wall of the duct. The vesicle is a dorsal outpocketing
from the duct, and is not in direct connection with the gland. The
labyrinth consists of a spongy tissue formed by the walls of innumerable
canals, which anastomose with one another without any determinable
system.

The cells of the saccule are narrow, crowded, with rounded free ends,
and without cuticula, while in the labyrinth the cells are markedly
striate, and surmounted often by clear vesicles. A cuticuloid covering
is sometimes present.

The present state of opinion as to the function of the antennal glands
in Decapoda is not one of complete accord. It needs but a superficial
examination of the epithelium in these organs to convince one of their
glandular nature. The intra vitam experiments of Kowalevsky ('89) and
Cuenot ('94) show conclusively that the cells take up grains of coloring
matter introduced into the circulation and eliminate them. Moreover,
with the exception of the ductless branchial glands and certain cells of
the liver, this function is confined, as far as we at present know, to
these antennal glands. The glandular secreting cells of the intestine
perform no such function. Conclusions like those of Will und Gorup-
Besanez ('48), who asserted the presence of guanin, have been reached
by H. Dohrn ('61),1 by Kirch ('86), and especially by Griffiths ('85).
Griffiths got definite tests for uric acid, and he carefully describes the
chemical analysis by the murexid method ; further, he obtained small
traces of the base guanin. Szigethy ('84) also found what he believed
to be uric acid crystals. In each of these cases the analysis was upon
the gland in Astacus.

Marchal ('92, pp. 237-245), on the other hand, made analyses of the
excremental fluid of Maia squinado, and could get no traces of urea or
uric acid, but did obtain nitrogen. He however obtained an acid some-
what analogous to uric acid, which he terms carcinuric acid. This be-

1 I quote Dohrn and Kirch on the authority of Gerstaecker ('95, p. 1009), as the
original papers are inaccessible to me.

154 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY.

longs to the series of the carbopyridic acids. There is also found a base,
leucomaine, which is analogous to the vegetable alkaloids. He denies
that the gland in Astacus contains uric acid.

It is difficult to reconcile these diverse results. It is undoubtedly
true that the antennal glands eliminate certain ammonia compounds,
and it is also shown by Marchal ('92, p. 243), that one of their func-
tions is to rid the organism of the excess of mineral salts present in the
blood, and Kirch ('86) has found that in Astacus these organs contain
glycogen. Gerstaecker ('95, pp. 1009, 1010) concludes that, in view of
all these results, we must consider the antennal glands of Decapoda not
as closely analogous to a kidney, hut as excretory organs important in
the general metabolism ; and, by reason of recent work, that we must
concede the physiology of the antennal glands to be in many respects an
open question, the solution of which will require much research, and
that it is not wholly improbable that these glands, like the liver, are
subject to temporarily changing functions.

This seems a fair and conservative statement, which expresses the pres-
ent state of our knowledge in regard to the function of these organs.

The work represented by this paper was done during the years
1896-97 and 1897-98 in the Zoological Laboratory of Harvard Univer-
sity. During the year 1896-97 my work was aided by privileges af-
forded by a Morgan Fellowship in that University. I wish here to
express my indebtedness to Prof. E. L. Mark, Director of the Labora-
tory, for his kindness in many ways, and for his supervision and able
criticism of the work during its progress. I am also indebted to Mr.
Alexander Agassiz for the opportunity of working at his Newport Lab-
oratory during part of the summer of 1897, and for appointment to a
table at the United States Fish Commission Laboratoi-y at Wood's Hole
during the summer of 1896 ; to Hon. J. J. Brice, formerly United States
Commissioner of Fish and Fisheries, for courtesies extended at various
times at the Wood's Hole and Gloucester stations of the United States
Fish Commission ; and to Prof. F. H. Herrick and Dr. G. H. Parker for
the use of embryonic material of certain stages.

Methods.

The methods employed were different in the embryonic, larval, and
adult stages.

In studying the finer anatomy of the adult organ, I have had re-
course to maceration, teasing, and serial sections. For sectioning, the

WAITE : ANTENNAL GLANDS IN HOMARUS AMERICANUS. 155

most serviceable fixing reagents were, in the order named, the chrom-
osmo-picro-acetic mixture of vom Eath followed by pyroligneous acid,
Perenyi's fluid, and corrosive sublimate. For staining the material
fixed in Perenyi's fluid and corrosive sublimate, I have used Heidenhain's
iron hematoxylin, Delafield's hematoxylin, and Ehrlich's acetic acid-alum
hematoxylin. For a double stain orange G of Griibler after either of
the last two is of value. In general the carmines are not serviceable
in this connection, but picro-lithium carminate has afforded very good
results on the wall of the vesicle.

Both surface preparations and serial sections were used in studying
embryos. The yolk was sometimes dissected away before sectioning, but
at other times it was sectioned with the embryo. The selection of the
killing agent to be used depends in a measure upon which of these two
methods is to be followed in sectioning. The specimens to be used for
examination by transmitted light must necessarily be separated from the
yolk. This was accomplished with fine needles and a small stream of
fluid from a pipette, the work being done under a dissecting microscope
and in alcohol of a lower grade than that in which the embryos had been
kept. This change in the grade of alcohol distends the egg membrane
so that it may easily be shelled off. The difficulty of the separation of
the yolk depends upon its friability and the age of the embryo.

I have used for killing the following : (1) water, heated to boiling and
then quickly poured over the embryos ; (2) saturated aqueous solution
of corrosive sublimate, used either hot or cold ; (3) Perenyi's fluid ;
(4) Kleinenberg's picro-sulphuric mixture ; (5) Flemming's fluid, —
weaker mixture. Of these the hot water and the corrosive sublimate
treatment render the yolk granular and friable, so that it is easily sep-
arated from the embryo ; at the same time the yolk particles are so
loosely united that it is impracticable to section embryos with the yolk
attached. On the other hand, treatment with Perenyi's fluid makes the
yolk so firm and tough that it is almost impossible to dissect off the em-
bryo without serious injury to it. Treatment with picro-sulphuric mix-
ture, if of short duration (15 to 30 minutes), allows the yolk to be
removed in large masses, and completely, without injury. Longer treat-
ment (60 to 90 minutes) interferes with such removal, but solidifies the
yolk so that it can be easily sectioned with the embryo. The weaker
Flemming's fluid when used for a short time (20 to 30 minutes) gives
a tenacious rind enclosing the embryo, which is easily separable from
the coarsely granular central part of the yolk. A longer treatment (45
to 90 minutes) gives the best condition I have obtained for sectioning

156 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY.

entire the yolk and embryo. The time of exposure to any killing fluid
should vary with the age of the embryo.

I have tried a variety of stains. Of the many carmine combinations
used, I have found Grenadier's aqueous alum carmine the most service-
able. The best results are gained from material killed with hot water.
Of the lnematoxylins, I have found Ehrlich's acetic acid-alum mixture
as good as any, and the most reliable. It stains material fixed by any of
the methods mentioned, the best results being obtained after fixing with
Perenyi's fluid or with Flemming's weaker mixture. I have found, how-
ever, that the staining is much better if the hamiatoxylin is followed by
an aniline dye. Of several tried I have found orange G of Griibler
much the best ; consequently I have used this excellent combination in
most of my work. The results are precise and clear, and the manipula-
tion of staining is very simple. Sections 6§, 10, or 13J micra in thickness
are stained on the slide in Ehrlich's mixture from 20 to 30 minutes, then
washed in acidulated alcohol until they are very pale, next neutralized
in a two percent solution of bicarbonate of soda, then thoroughly washed
in distilled water and put into a saturated aqueous solution of orange G.
This is allowed to act from 10 to 20 minutes, the time being determined
by watching the action of the stain. The slide, after draining, is trans-
ferred directly to absolute alcohol, which sets the stain and at the same
time washes out the excess. After clearing in xylol, the sections are
mounted in xylol balsam, the stain being thoroughly permanent.

In studying the larval organ I have depended entirely upon serial
sections, for the gland is too small to permit satisfactory dissections, and
the shell of the appendage does not clear readily enough to permit study
of the entire organ in situ. The first larva was obtainable in abundance,
so that various methods could be tried, but the older larvae are more dif-
ficult to get, and as a matter of fact all of the older larvae which I had
were already preserved when they came into my hands.

The most successful results with the first larva were obtained with
the weaker Flemming's mixture, Perenyi's fluid, corrosive sublimate, or
Kleinenberg's picro-sulphuric mixture. I give these reagents preference
in the order named ; the last, however, is not available in the older larval
or the adolescent stages, when the shell comes to contain considerable
quantities of lime salts.

For this stage several stains have been used. I finally settled upon
Ehrlich's acetic acid-alum hematoxylin, however, as the most reliable
stain for the gland. For a double stain I have used this followed by acid
fuchsin or orange G, the latter being, in my opinion, preferable.

WAITE: ANTENNAL GLANDS IN IIOMARUS AMERICANUS. 157

The paraffin method of embedding has been used exclusively. The
cuticular shell in the larval stages is rather resistant, and soon dulls the
knife, hut with a sharp knife good series 6§ micra in thickness were
obtained.

I. Structure in the Adult.

The structure of the adult antennal gland in the genus Homarus has
been described by only one writer, Marchal ('92, pp. 15G-1 G3, Plate VII.
Fig. 1), who gives an account of the organ in H. vulgaris. There is no
published account of this gland in H. americanus except a short abstract
of this paper (Waite, '98).

A. Gross Anatomy.

The gland, with its accessory structures, is situated at the base of the
second antenna, and the most of it is within the cephalothorax. It
occupies the greater part of the space beneath, and on the side of the
masticatory stomach, anterior to the voluminous hepato-pancreas. The
organ as a whole may be divided roughly into three parts, which are
easily recognized and sharply separated, namely : (1) the gland proper,
often referred to hereafter simply as the gland ; (2) the overlying
vesicle, or stoi-age reservoir; and (3) the duct leading to the exterior,
together with the tubercle upon which it opens.

The gland proper lies almost entirely on the ventral floor of the cephalo-
thorax, but its anterior median lobe extends into the base of the antenna.
It lies close to the sagittal plane, the median edges of the two glands
being but about five millimetres apart.1 ]N"one of the principal axes of
the gland proper are parallel to the principal axes of the body. The
dorsal face is so inclined to the frontal plane of the animal that its
anterior edge is 20° or 30° more dorsal than its posterior edge, and its
lateral border 25° or 30° more dorsal than its median. The chief axis
of the gland as viewed from above runs from the anterior notch or hilus
to the most distant point of the posterior edge (Plate 1, Fig. 1). This
axis makes an angle of about 30° with the sagittal plane of the animal,
its anterior end being nearer that plane.

The general shape of the gland proper as seen from above resembles
that of a cordate leaf with a blunt point, the notch being anterior (Plate 1,
Fig. 1). The dorsal face is irregularly concave. This concavity is filled

1 All measurements given for the adult gland are for an average adult ninp.
inches (= 23 cm.) in length from tip of rostrum to tip of telson.

158 BULLETIN: MUSEUM OF COMPAEATIVE ZOOLOGY.

by the ventral part of the overlying vesicle, the wall of which is closely
applied to the smooth dorsal surface (if the gland (Plate 1, Fig. 2, par. vs.).
The ventral face of the gland is in general convex, but has in it depres-
sions which conform to the underlying hard parts of the endophragmal
system and shell, and to the muscles upon which the gland rests. The
ventral surface differs much in appearance from the smooth dorsal surface.
It is uneven, being thrown into numerous ridges and furrows, resembling
to some extent the gyri and sulci of the cerebral hemispheres in higher
mammals. All these furrows radiate in a general way from a region of
the gland lying near the bottom of the hilus, but the individual furrows
are comparatively short, and separated by ridges from other furrows in
the same radius. The furrow lying in the main axis of the gland is the
most prominent and deepest one. It is a direct continuation of the hilus.
The dorso-venti'al diameter varies, but is greatest in the centre of the
organ, and thins out to a translucent edge on the borders. The outline
of the border of the gland is irregular. In the middle of the anterior
edge there is a sharp V-shaped hilus extending posteriorly 2 to 3 milli-
metres (Plate 1, Fig. 1, hi.). The bottom of the hilus marks very nearly
the centre of the gland. The median edge of the gland presents a sim-
ple curve convex toward the median plane of the body, and having a
radius of about a centimetre. The lateral edge is S-shaped, the poste-
rior bend of the S giving the gland the appearance of being notched to
accommodate the flexor antennarius muscle, on which it does in part
rest. This notch is 2.5 or 3 mm. deep and 5 or 6 mm. broad. The
edge of the gland for the most of its extent is bluntly toothed or lobed,
so that the outline is more or less sinuous.

For purposes of description the gland proper may be said to present
three lobes : a median anterior (lob. m-a., Plate 1, Fig. 1), a lateral
anterior (Job. l-a.), and a posterior one (lob. p.).

The gland measures in its longest diameter, from the edge of the
median anterior lobe to the posterior apex, 13 to 15 mm. ; from the
bottom of the lateral notch to the middle of the median border, 9 to
10 mm. Its greatest dorso-ventral thickness is about 3 mm. It is thus
evident that the gland is mnch flattened.

In color the gland is a light olive, the ventral face being of a darker
shade than the dorsal ; the latter in some individuals approaches yel-
lowish brown. The light dorsal surface is surrounded by a narrow
border of the darker shade from 1 to 2 mm. in width (Plate 1, Fig. 1,
marg.), which is left without tint in the figure. This border widens to
quite a broad band on the median anterior lobe. The area of lighter

WAITE : ANTENNAL GLANDS IN HOMAEUS AMEPJCANUS. 159

color corresponds with the extent of the endsac ; that of the darker
border, with that part of the dorsal face of the labyrinth which is not
covered by the endsac, or which shows through the attenuated edge of
the endsac.

From the median anterior lobe there is given off a smaller lobe 2.5 to
3 mm. in diameter and 5 to 6 mm. in length. This is directed caudad
and ventrad, and tapers to a point. Its general shape is conical. In
color it is so distinct from the other parts of the gland that Marchal
('92, p. 159) has called it the white lobe. The tapering point opens into
the excretory duct in a region ventral to the centre of the gland. This
opening is not single, but consists of several minute pores, to each of
which there leads one of a series of converging tubules which are in turn
formed by the confluence of other smaller tubules in the basal part of
the white lobe. These pores are on the dorso-median wall of the duct,
2 to 3 mm. from its external orifice. By these the products of the gland
are discharged into the duct. This is the only communication which the
gland proper has with its accessory structures, there being no direct com-
munication between the gland and the vesicle, as there is in Astacus.
Therefore the vesicle must be filled by the backing up of the secreted
products in the duct from the point where the pores open, a process
which is presumably effected by closure of the external opening of
the duct.

The vesicle is situated in the cephalothorax at the side of the masti-
catory stomach. Its anterior end lies against the muscles of the body
wall. On its median face it presses against the stomach, while poste-
riorly it reaches the hepato-pancreas and extends between its lobes.
Ventrally, it is in part closely and firmly attached to the dorsal face of
the gland (Plate 1, Figs. 1, 2, par. vs.). Posterior to the gland it lies
free upon the floor of the cephalothorax. Its shape is roughly oval, its
long axis lying in a parasagittal plane. Usually the posterior end is
slightly divided into two blunt lobes. Since the thin elastic walls easily
conform to the surrounding organs, its actual shape in detail is irregular,
depending upon the state of its own distention and that of the stomach,
hepato-pancreas, and reproductive organs, and the contraction or relax-
ation of the muscles in this region. In some individuals it seems to
fill every niche of unoccupied space ; in others there are considerable
spaces between it and the body wall.

I have found no ligamentous attachments for holding the vesicle in
place other than that to the dorsal surface of the gland, already noted,
and those to the blood-vessels and nerves which come to its surface.

160 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY.

The wall of the vesicle is uniformly very thin and transparent in all
parts except immediately around the orifice of the duct, where it is
somewhat thicker; but there is no muscular sphincter at this point.
The histological structure of the wall is described in another connec-
tion (pp. 166, 167).

At the anterior end of the vesicle, at a point on a line with the me-
dian border of the hilus and a little in front of the anterior edge of the
gland, there is an orifice which opens into the duct leading to the exte-
rior (Plate 1, Fig. 1, of. i.). This orifice is circular, or slightly oval, and
1 to 1.5 mm. in diameter. It is noticeable that it retains its shape when
the walls of the vesicle are collapsed.

The duct is a thin-walled tube, oval in cross section, — being slightly
flattened dorso-ventrally, — and lies wholly within the antenna. From
the orifice in the vesicle it extends downward and slightly forward and
outward for a distance of 5 to 7 mm. Its average diameter through
this region is 2 mm. At the level of the coxo-basal joint the duct turns
abruptly through an arc of about 45° in a parasagittal plane of the body,
and runs downward and backward to the tubercle on the ventral face of
the coxopodite, through which it opens to the exterior. Its diameter
through this second part of its course is somewhat greater than in the
first part, but diminishes noticeably as it enters the tubercle. There is
a slight dilatation at the point where the tubules from the white lobe
enter. The entire length of the duct is not over 1.5 cm., and its aver-
age diameter is perhaps 2.5 mm.

I find no sphincters or other muscular structures in or around the
wall of the duct, other than the muscles of the antenna, among which it
runs. It seems, then, that the ejaculatory process, by which the secre-
tion of the gland is sometimes forced out of the external orifice, cannot
have its origin in any intrinsic action of the duct. It may arise from
concerted action of the muscles of the antenna, but since no especially
active movement of the antenna has been noted in connection with this
function, it is probable that the force producing such ejaculation has its
seat in the muscular walls of the vesicle, the duct being passive.

The external orifice of the duct is situated at the apex of a truncate
conical tubercle, which rises from the general ventral surface of the cox-
opodite of the second antenna. This tubercle is nearly in the midventral
line of the appendage. It does not stand perpendicular to the general
surface, but its axis is inclined slightly forward and medianward, so that
if the antennae were at rest in the horizontal position, the axes of the
two tubercles prolonged would intersect at a point in the sagittal plane

WAITE : ANTENNAL GLANDS IN HOMARUS AMERICANUS. 161

about 2.5 cm. in front of, and 4 cm. ventral to, the bases of the first
antennae. The posterior and median sides of the tubercle are more ele-
vated from the general surface than are the anterior and lateral sides.
The truncated face of the tubercle is a circle with a diameter of about
1 mm., and is covered with a membrane which has concentric markings.
The actual orifice is situated in the centre of this membranous area, and
is less than 0.1 mm. in diameter. The opening into the tubercle from
the main cavity of the appendage, as may be seen by inspecting the
iuside of a moulted shell, is semicircular, the anterior edge being straight ;
from this edge a toothed process of the shell, serving for the attachment
of muscles, projects backward.

The blood supply of the antennal gland is mainly from two soui'ces :
(a) the antennal artery (arteria lateralis of Gerstaecker) ; (b) the ster-
nal artei-y. The antennal artery arises directly from the heart, and
passes obliquely forward and downward close to the dorsal body wall.
In that part of its course where it runs near the vesicle, it gives off
several small branches, which are distributed to the lateral wall of this
organ. This artery sends branches into the gland proper at three re-
gions : (1.) As the artery passes ventrad along the anterior face of the
antennal flexor muscle, a large branch is given off, which goes on the
lateral and ventral side of this muscle (i. e. outside of the muscle), and
enters the lateral region of the posterior lobe of the gland on its ventral
face, to which region it is distributed. (2.) As the antennal artery
passes ventral to the lateral anterior lobe of the gland it sends several
small branches into the ventral face of this lobe ; and one large branch
to the bottom of the hilus, whence it passes dorsad to be distributed to
the endsac (Plate 1, Fig. 1, art. sac). Tin's branch Marchal ('92,
p. 1G2) terms the " art ere sacculaire." (3.) From the dorsal branch of
the antenna] artery a small branch is given off, which runs ventrad and
mediad to enter the median anterior lobe of the gland. It enters on
the ventral face close to the anterior edge. The antennal artery after
giving off the artery to the endsac passes iuto the second antenna. In
the region of the coxo-basal joint of this appendage two small branches
are given off to the walls of the duct and to the tissues at the base of
the tubercle.

The second source of blood supply is the sternal artery. A branch
from this artery comes to the median surface of the posterior lobe, where
it divides. The larger of the resulting branches enters the extreme pos-
terior region of the lobe on the ventral face close to the edge : the
smaller passes forward along the edge of the gland and enters the ven-

162 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY.

tral face of the median anterior lobe. Other smaller vessels enter the
ventral face of the gland at several points, coming from the trunks sup-
plying the tissues upon which the gland rests.

To summarize, the lateral part of the labyrinth, together with all the
endsac, is supplied by the antenna! artery, while the median part —
except the tip of the median anterior lobe — is supplied from the sternal
artery.

The nerve supply of the gland is from the nerve trunk running to the
second antenna. Immediately after this trunk emerges from the super-
cesophageal ganglion, it gives off a branch which passes outward and
backward to enter the median anterior lobe of the gland. The point of
entrance is on the ventral face near the median edge. The nerve trunk
to the second antenna also sends several small branches into the edge of
the median anterior lobe ; these arise at points more distal than that at
which the main branch arises. I have not found any nerve branches
entering at the hilus, nor in the posterior region in a position correspond-
ing to the blood supply from the sternal artery.

B. Finer Anatomt.

The gland proper consists of two distinct parts. The relation of the
two parts — the endsac (sac. trm.) and the labyrinth (Iby.) — is seen in
Figure 2 (Plate 1), which is a transverse section of the gland proper
about half a millimetre posterior to the bottom of the hilus. The end-
sac is dorsal to and covers the whole of the labyrinth except its extreme
edges. Its average dorso-ventral depth at this region is about one
fourth that of the labyrinth. This ratio is greater than the average for
the whole gland.

The endsac shows a considerable regularity in the arrangement of its
cavities. At the bottom of the hilus thei-e is a single large chamber of
irregular shape, from which outpocketings radiate. These are separated
from one another by folds of the wall of the endsac, which form complete
septa. These outpocketings branch dichotomously many times, and thus
produce a system of compartments which become smaller and smaller
until the edge of the endsac is reached. In these compartments there
are secondary foldings of the wall forming incomplete septa, which are
more frequent on the ventral than on the dorsal wall of the compart-
ments. In Figure 2 (Plate 1) the compai-tments near the edge of the
endsac are cut crosswise ; those near the centre, lengthwise or diagonally.

The dorsal face of the endsac has attached to it the closely adhering
ventral wall (par. vs., Figure 2, Plate l) of the overlying vesicle. This

WAITE: ANTENNAL GLANDS IN HOMABDS AMEEICANUS. 163

consists, however, of only the glandular epithelium of the vesicle wall, the
muscular layer — to be described later (p. 167) — being in this region
absent. Between the basement membrane of the epithelium of the
vesicle and the wall of the endsac there is a rich plexus of blood-ves-
sels, — branches of the sacculary artery, — some of which are shown
cut across in Figure 2 (i-as. sng.).

The septa of the endsac, both complete and incomplete, consist each of
a fold of the epithelial lining of the sac (Plate 1, Figure 3, cl. sac. trm.),
embracing a sheet of connective tissue (tis. con't.) between its two lay-
ers. This connective tissue is highly vascular, being principally formed
of the walls of blood vessels and blood lacunae. The glandular epi-
thelium of the endsac forms of course a continuous layer, lining all com-
partments and investing all septa. It is everywhere a single layer thick,
and its cells have certain distinctive characters. In shape they vary
both according to location and to condition of activity. They are most
elongated on the dorsal wall (Figure 4) and at the place of junction
with the cells of the labyrinth ; most rotund on the ventral floor and on
the septa which rise from it (Figure 3, cl. sac. trm.).

The difference of shape due to the state of activity is more striking.
Cells lying side by side (Figure 4) vary much in this respect. Among
shorter cells of nearly uniform diameter there are many elongated cells,
the free ends of which are expanded into globules often in diameter two
or three times that of the basal portion of the cell. The contents of
these swollen ends are less dense and stain less deeply than do the basal
portions of the cells. The globules become detached and pass to the ex-
terior, the lumen of the endsac being more or less filled with various
sized globules of this nature (Figure 4, gib.). Each, when detached, —
if free from mechanical pressure, — is spherical and contains a granular
clot nearly filling it. This clot is not homogenous, but made up of
granules of various sizes and degrees of opacity. Various progressive
stages in the constricting off of the globules from the cell are seen, so
that there can be no doubt that they arise primarily from the cells.
These globules seem to be composed of a mass of secretion products en-
closed in a capsule, which is part of the cell wall. They are not detached
cells, for they never contain a nucleus, nor, so far as I have seen, any
chromatin particles. Further, there are no evidences of nuclear division
in the glandular epithelium, as would probably be the case if these
globules were degenerate cells ; for if some cells became degenerate and
passed off, division would be necessary in the normal cells to make up
the loss.

164 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY.

These glandular cells are crowded with a mass of spherical vacuoles,
which are smaller in the attached end of the cell than in its free end
(Figure 4), and sometimes are not found at all in its basal part (Figure
5). The cytoplasm also contains granules of various sizes and shapes,
which differ widely in refractive powers. These are found throughout
the cell, often in compact opaque masses. Such masses are also common
in the globules (Figure 4).

In sections prepared by certain methods, namely, in material fixed in
saturated aqueous-corrosive sublimate, or in vom Path's platinic-aceto-
osmic mixture, acicular crystals are found in many cells, though not
in all. Such crystals are not found in material treated with Perenyi's
fluid or with Flemming's weaker mixture, nor in any sections which
have been decolorized with alcohol acidulated with 1% HC1. These
crystals are all fine and needle-like, but differ in length (Fig. 5, cry., and
F'ig. 6). It is noticeable that they are confined to the cells of the end-
sac, none being found either in the lumen of the endsac or in the cells or
lumina of labyrinth or vesicle.

Szigethy ('85, p. 109) found in the gland of Astacus crystals which he
regarded as uric acid. I have not seen his figures in the original paper,
and so cannot compare them with those I have described. In the present
case, I think it probable that the crystals are artifacts ; at any rate,
more detailed investigation is necessary before they can be considered of
physiological importance. I have simply made a record of the condition
noticed.

The nuclei of the cells of the endsac are oval and usually situated in
the basal half of the cell ; but in some cases they are crowded toward
the free end (Figure 4). The cells rest upon a membrane (tub. ba.}
Figs. 4 and 5) of appreciable thickness, but without nuclei.

There is but one region of communication betiveen endsac and labyrinth.
This is in the main axis of the gland immediately posterior to the bottom
of the hilus. The section from which Figure 2 is drawn passes through
this region. The communication is between the central lumen of the
endsac and the lateral anterior lobe of the labyrinth. The median an-
terior lobe has no direct communication with the endsac, but it will be
remembered that it is from the median anterior lobe that the white lobe
— which is the direct passage to the exterior — arises (p. 159). There-
fore the products of the endsac must pass through the lateral-anterior,
the median-anterior, and the white lobes in the order named before reach-
ing the duct to the exterior. The communication between the endsac
and the lateral-anterior lobe is about 0.4 millimetre in diameter, and

waite: antennal glands in homarus americanus. 1G5

runs obliquely ventrad and laterad. It is short, and soon branches to
become continuous with the lumina of the various labyrinth tubules.

At the border of this short channel, as indicated at a in Figure 2,
there is a sudden transition from tbe cells of the endsac to those of the
labyrinth, there being no cells of intermediate character. The boundary
of this opening is the only place where the cells of endsac and labyrinth
come into contact. At all other places where endsac and labyrinth come
together, the basement membranes of the two are face to face.

The labyrinth is not a single pocket with evaginations like the endsac,
but is composed of a system of branching tubules ; these are not, how-
ever, simply coiled tubules, for they undergo anastomosis at frequent
intervals, and it is therefore impossible to trace any single tubule for a
great distance. Moreover, the tubules are not of uniform calibre, but
have numerous outpocketings, which still further complicate the appear-
ances presented in sections. This communicating system of lumina is so
complicated that it is impossible to say that it presents a definite plan.
It is certain, however, that it does not represent one or a few coiled tu-
bules, as has been described for Astacus. The whole system of laby-
rinthine passages is lined by a continuous layer of epithelium. The
basement membrane (Figs. 7 and 8, mb. ba.) of this epithelium is
somewhat thicker than that of the epithelium of the endsac. The spaces
between the epithelial lining of adjacent tubules is occupied by a vascu-
lar connective tissue (Fig. 7, tis. con't.), like that described in the septa
of the endsac (p. 163). The blood-vessels and lacunae found in this tissue
contain blood corpuscles (Fig. 8, vas. sng.).

The epithelial cells in different regions of the labyrinth do not show
differences of form, but they are very unlike those of the endsac. The
cytoplasm is denser and stains more deeply, and the vacuolated condition
so characteristic of the cells of the endsac is not seen. Instead, the
cytoplasm contains large numbers of granules of very uniform size and
power of refraction, and exhibits a striation which is perpendicular to
the basement membrane. The oval nuclei lie near the middle of the
cell, and are somewhat larger than those of the epithelium of the endsac.
But the distinguishing feature of the epithelium of the labyrinth is its
free border. This differs in different glands, and in different parts of the
same gland ; the extreme conditions may, indeed, be seen even in neigh-
boring regions of the same tubule. In Figures 7 and 8 are shown the
extremes of the conditions seen. The free border of the cells shown in
Figure 8 appears striate, and there is formed a layer which, though it is
not a true cuticula, resembles one. The parallel striae represent partitions
vol. xxxv. — no. 7. 2

166 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY.

between adjacent vacuoles which form a single superficial layer and are
by pressure elongated in the long axis of the cell. In Figure 7 some
cells exhibit large globular protrusions from their free ends. The cyto-
plasm of these protrusions is less dense than that at the base of the cell,
and they bear some resemblance to the globules found on the cells of the
endsac. Other cells in Figure 7 show two or three small globules, and
still other cells might have been shown having a larger number of pro-
portionally smaller globules. I believe that the large globules arise from
cells in the condition seen in Figure 8, by the progressive fusion of neigh-
boring small vacuoles at the free ends of the cells, accompanied by an
expansion and protrusion of the wall of the free end of the cell. As far
as my observations go, I can agree with the main points of Marchal's
('92, pp. 228-230) description of this process, except that I do not find
secondary smaller globules within the larger ones.

The large terminal globules evidently become detached from the cells,
since great numbers of such globules (Fig. 7, gib.) of various sizes ap-
pear free in the lumina of the tubules. They resemble the globules from
the endsac in being spherical, but their contents are much more homo-
geneous, the granules being nearly alike in size and refractive properties.
This production of globular bodies with granular contents characterizes
the cells of both endsac and labyrinth, but the manner of their formation
differs somewhat in the two cases. There is, then, a typical true secre-
tion by these cells, not merely a filtration of substance through them.

The vesicle has a wall without folds or protuberances. It is com-
posed of three layers; the inner one epithelial (Plate 1, Figs. 10 and
11, la. e'th.), the outer one muscular (la. mil.), and between these a third
layer containing a system of blood vessels (Figure 11).

The epithelium is throughout one layer deep. The cells in surface
view (Fig. 9) are seen to be irregularly polygonal. They are often elon-
gated parallel to the axis of an underlying blood vessel. The resemblance
to the cells of the labyrinth is close, as is to be expected from their sim-
ilarity of origin (p. 189). The cytoplasm is granular and striate, but
shows more vacuoles (Figs. 9, 10, and 11) than occur in the cells of the
labyrinth. The nuclei are oval and situated near the middle of the cell.
There is a basement membrane of appreciable thickness, which may be
separated from the cells by maceration and teasing. The free border of
the epithelium (Figs. 10 and 11) presents the same condition as that seen
in Figure 8 from the labyrinth; but I do not find the other extreme condi-
tion (Fig. 7), with large globules attached to cells, although some of the
earlier intermediate conditions between these two extremes are found.

waite: antennal glands in homakus amekicanus. 167

The muscular layer is more variable in thickness than is the epithelial
layer. It is composed of long spindle-shaped cells (Fig. 12) of different
lengths, containing centrally located nuclei. The cells and nuclei are
much flattened in one plane. The edge of the cell is frequently crinkled
so as to produce a sinuous outline (Fig. 13). The cell contents are
fibrillar, as is well shown when a cell is frayed out in teasing. The
fibrillae are likewise crinkled. There is usually present an axial bundle
(Fig. 13, fas. ax.) of fibrillae, which seem to be more closely bound to-
gether than are the others ; these stain with hematoxylin more deeply
than the rest of the cell. The thickness of this muscular layer varies.
It may be several cells (Fig. 10, la. mu.) or a single cell deep (Fig. 11,
la. mu.), or it may disappear altogether, as it does where the wall of the
vesicle is attached to the dorsal face of the endsac (Fig. 2, par. vs.). In
any given place in the muscular sheath the cells lie parallel with one
another, there being no crossing or formation of meshwork.

Between the basement membrane of the epithelium and the muscular
sheath there is a layer of connective tissue containing a plexus of blood-
vessels. One of these is seen cut crosswise in Figure 11 (vas. sng.).
These vessels present the ordinary structure seen in the vessels of other
parts of the body, and are readily distinguishable from the two other
layers of the wall of the vesicle.

The position and external appearance of the tubercle upon which
the duct opens has already been described (p. 160). In Figure 14
(Plate 1) is shown a section in the long axis of the appendage through
this tubercle. It is seen that the base is much constricted by an infold-
ing of the integument on the anterior side. This is the toothed process
referred to on page 161. As the duct passes through this narrow region,
it is constricted, but within the tubercle again enlarges. It opens
through the middle of the integumental membrane which covers the
face of the tubercle. This opening varies in different specimens from 25
to 40 micra in diameter (Figure 14, of. ex.). The tegumental membrane
is composed of three layers, which, passing from within outward, are,
(1) the thin non-calcified layer of the shell ; (2) the non-pigmented cal-
cified layer (Figure 14, la. ex.) ; (3) the cuticula or enamel layer. The
pigmented calcified layer (Figure 14, la. ex'.) stops abruptly at the edge
of the truncate face of the tubercle. The three layers of the membrane
are traceable into the duct on the anterior side, but not on the posterior.
However, even on the anterior side these rapidly thin out and disappear
entirely at a short distance from the outer face of the membrane. This
continuation of the integument into the anterior floor of the duct forms

168 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY.

a sort of operculum (Figure 14, op.). The epidermal cells upon which
this rests, and by which of course it has been produced, are much elon-
gated, and their deep ends are attached to a ligament (Figure 14, lig.).
To this ligament are attached, in turn, other elongated epidermal cells,
the basal ends of which are attached to the integument of the " toothed
process." I have no evidence that either set of these epidermal cells is
contractile, but they are both peculiarly elongated, and if they are
contractile we have here a mechanism for opening the external orifice
by depressing the operculum. This would allow the escape of the
excretory products, they being forced out by contractions of the
muscular layer of the wall of the vesicle.

The space between the duct and the wall of the tubercle is filled by
the dermis (Figure 14, drm.), in which are found tegumental glands
(gl. e'drm.). At its base are striated muscles (mu.) belonging to the
antenna, but having no connection with the operculum. On the tubercle,
especially around the edge of its truncate face, there are numerous
sensory hairs (set. sns.).

II. Development.

A. Historical Summary.

The published work on the development of the antennal glands in
decapod Crustacea is meagre and lacks both detail and completeness.
In most cases, it is only a by-product of researches along some other
line, or part of a general consideration of the life history of some
species ; this may account for the contradictory statements on the
subject.

Rathke ('29) in a work much in advance of his time describes (p. 51)
for the crayfish embryo, at the time when pigmentation of the optic
lobes is well begun, the formation of a slender narrow plate, whose great-
est diameter is directed from dorsal-posterior to ventral-anterior. This
plate presents a rounded outline dorsally and ventrally ; its posterior
edge is concave ; its anterior convex. The outer surface is in contact
with the integument, the inner with the yolk, from which it has its
origin "through deposit of plastic material." The duct to the exterior
was not made out. These " Speicheldriisen," while remaining in con-
nection with the yolk, increase in size and become more spherical
(p. 60). Just before hatching, there may be seen in these organs small
round green spots, the occurrence of which may be assumed to indicate
the beginning of secretory activity. When the yolk disappears, they

waite: antennal glands in homarus americanus. 169

have a discoidal form, are green in color, and lie against the wall of the
stomach (p. 64).

This early account is nearly correct as far as it goes, hut it deals with
only those parts resulting from the ectodermic invagination ; the author
failed, moreover, to see the connection with the exterior, which would
have shown him the relations of the organ and would have prevented
the use of the term " Speicheldriisen " in relation to them.

The next reference to the embryology of the organ is by Lereboullet
('62), who describes (pp. 760, 761), also in the crayfish, the formation
of certain glandular bodies of unknown function at the stage when the
stomach grows forward to produce the lateral pouches. These bodies
are formed of a simple transparent tube, which in one part of its course
is coiled, and is directed toward the base of the corresponding external
antenna. The green glandular body also formed by coiling of the same
tube is not seen until later.

It is difficult to reconcile this last statement with the preceding, but
it seems to indicate that the author believed that there were two glands
of different kinds arising in this region, but not simultaneously. The
part first referred to is probably the duct to the exterior.

A. Dohi-n ('70, pp. 253, 254) describes in the basal segment of the
antenna, in the early embryonic stages of Scyllarus arctus, a round cell-
mass surrounded by a membrane, and outside this by "einer starken
Hypodermisschicht." In Palinurus vulgaris, according to Dohrn, the
antennal glands lie in the basal segment of the antenna (p. 261). They
are surrounded by the hypodermis, but are free from it. The central
part is a cavity in which lie rather close together large free cells, while
the wall of the gland is continuous by means of the duct with the ven-
tral and inner wall of the base of the antenna. The free cells within the
cavity are probably of the same origin as those of the wall, and ultimately
all are incorporated into the wall of the gland. I think it probable that
the author erred at this point, for the " free cells " are probably of differ-
ent origin from the others, i. e. of mesodermic origin, and are the Anlage
of the end sac.

Reichenbach ('77, p. 191) finds that the green glands arise by invagi-
nation of the ectoderm at the stage when the maxillipeds begin to appear,
but he gives no figures of them.

Grobben ('80, p. 104) says the antennal glands, as well as the shell
glands in Moina, arise from the middle germ layer, but he fails to give
arguments, descriptive figures, or authority to warrant this conclusion
concerning the antennal glands, or even to say in what species it

170 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY.

occurs. I cannot therefore agree with Kingsley ('89, p. 30) in saying
that Grobben " showed that the gland belonged to the mesodermal

tissues."

Ishikawa ('35, p. 422, PI. XXVIII. Figs. 68, 90-94) finds in Atye-
phira, in the base of the second antenna at the time of the appearance
of the first pair of maxillipeds, a circular group of about eight granular
ectodermic cells. In these is formed a cup-shaped depression from the
exterior, which has a large mouth and grows deeper while the opening
narrows. " The cells which are concerned in the formation of it [the
antennal gland] are all ectodermic." By the time of hatching, the de-
pression has become a canal having three or four convolutions and filled
with granular fluid.

Iieichenbach ('86, pp. 97, 98) sees in sections of the embryo of Astacus
with well outlined maxillipeds (Stage G) the Anlagen of the glands ap-
pearing in the basal segment of the second antennae as recently formed
plugs of ectoderm (Taf. X. Fig. 125 u. 126, g. D.). The arrangement
of cells shows that a sac is to appear, although no lumen is as yet present.
At a slightly older stage (Stage H), with pereiopods marked off, surface
views show a crescentic arrangement of cells bordering an involution.
The opening of the ectodermic invagination is the permanent mouth of
the gland. This invagination grows anteriorly, making a slight turn.
The Anlage remains in connection with the ectoderm, and at first does
not even approach mesodermic tissue. At later stages (Stage J, with
distinct pleopods, and Stage K, with strongly developed eye pigment)
flat connective-tissue cells begin to surround and soon envelop the gland
(Taf. XIII. Fig. 205).

Thus far, with the exception of the unsupported statement of Grobben
('80), there is a consensus of opinion as to the ectodermal origin of the
glands.

With this opinion Kingsley disagrees. He found ('89, p. 29, Plate
II. Fig. 49) in Crangon vulgaris a patch of mesoderm stretching into
the base of the second antenna as a solid plug at a stage when the
eye pigment existed as a crescentic line (Stage G). In a somewhat
older embryo (Stage H) "a cavity (Plate II. Fig. 61) appears in this
tissue, and the cells lining it take a well marked epithelial character,
their boundaries being distinct." At no " stage does the green gland
have any connection with any other cavity within the body. . . . The ex-
ternal opening to the gland is not formed until after hatching." By this,
I infer, is meant the confluence of the lumina of the endsac and the duct.
The mesodermal origin of the antennal gland in Crangon agreed with

WAITE : ANTENNAL GLANDS IN HOMARUS AMERICANUS. 171

that for the shell gland of Phyllopods, as already worked out by Grobben
('79, p. 23), and later confirmed by Lebedinsky ('91, p. 152).

Lebedinsky ('89, p. 197, see also '90, p. 184) has found that the
" Segmentalorgan " in the first maxilliped of Eriphia is formed from
both ectoderm and mesoderm, and from this he concludes ('92, p. 233)
that the antennal glands, the shell glands, and " Segmentalorgane " in
Crustacea are serially hom6logous organs.

Allen ('93, p. 338) says that in a larva of Palsemonetes a few days
old the green gland consists of an endsac, which communicates by means
of a U-shaped tube with a very short ureter opening at the base of the
antenna, and that the distal portion of the tube is slightly enlarged and
may be termed a bladder. At hatching, the deeper part of the gland is
a mass of cells without lumen, but ureter and external opening are pres-
ent. Very early in larval life the cells of this deeper portion separate
from one another, giving rise to the lumen, and the gland probably be-
comes functional at that time. The development of the organ in the
larva consists chiefly in the enlargement of the bladder, which, arising
at the elbow of the U-shaped tube, grows mediad and dorsad to a great
extent. The shell gland, opening at the base of the second maxilla, is
the functional kidney during embryonic life, but it is not found in young
adults.

In his later paper Allen ('93a) gives figures (Plate XXXVI. Figs. 1
and 2) showing that in the young larva the eudsac differs histologically
from the tubule and bladder, the two last being lined with an epithelium
composed of cells which are striated and capped with a cuticular struc-
ture. The plexus of tubules between the endsac and the bladder arises
from the origiual tubule in that region "obviously . . . by the splitting
up of the single tubule" (p. 406).

Boutchinsky ('95, pp. 78, 79, Tab. II. Fig. 51) found that in Iphinoe
the ectodermal part of the gland arises as a true invagination from the
exterior, with a lumen at all times connected with the outside world.
The deep end of this invagination is surrounded by a mass of meso-
dermic cells. In Gebia littoralis, at a stage (Tab. IV. Fig. 85) when
the telson has grown forward so that it is even with the mandibles, he saw
(pp. 169, 170) the first appearance of the antennal gland as a compact
ball of mesodermic cells (Tab. VI. Fig. 143) with distinct walls and
large granular nuclei. The ectodermic invagination occurs at a later
stage (Fig. 153), a lumen meanwhile having formed in the mesodermic
part of the gland. The mesodermic part soon elongates and becomes
horseshoe-shaped (Fig. 154). He reaches two general conclusions in

172 bulletin: museum of comparative zoology.

regard to the embryology of the gland : — (1) that it is formed from an
inner mesodermic and an outer ectodermic part ; (2) that the formation
occurs at a time when the mesoderm shows no regularity in the distri-
bution of its cells, these being scattered without visible ordei\

I have published (Waite, '98) a short abstract of some of the chief
points in the development more fully described hereafter.

B. Development in the Embryo.

In the following account of the development of the antennal glands
in Homarus, it should be stated at the outset that it is very difficult to
fix definitely the time intervals between successive stages in the em-
bryonic development. The determination of age in hours or even in
days is not of much importance. Since the rate of development is
known to depend largely upon temperature, the conditions of embryos
of the same age must vary with every station, with eggs extruded at
different seasons of the year, and also with eggs extruded at the same
date in a given locality in different years. Since, then, rate of develop-
ment is dependent upon many variables, the known intervals between
the killing of different embryos in the same series are of only relative
value, not being accurately comparable with the intervals of any other
series. Further, even in the same series, in which apparently the same
conditions have prevailed, there is individual variation.

Notwithstanding these sources of uncertainty, I shall state the ap-
proximate age at which the several stages appear, this approximation
being arrived at by comparing the several series of my material with
one another and with the published figures of Bumpus ('91) and Her-
rick ('95). I shall give in each case the date of killing, which will in-
dicate the seasonal influences to which the development was subjected.

The earliest embryo which it is necessary to consider in dealing with
the development of the antennal gland was killed on August 29, and
had reached the late nauplius stage (Plate 2, Fig. 15). This embryo
has the cephalic lobes well marked off, and the first (at. 1) and second
antennas (at. 2) and the mandibles (md.) well defined. The second an-
tenna; are biramous and reach nearly to the posterior limit of the pleonic
flexure. Of the post-oral appendages, the buds of the first maxillae have
just begun to appear.1

1 This stage is a little further developed than that figured by Herrick ('95) in
Cut 32, Plate I., which he estimates by comparison with embryos raised from seg-

waite: antennal glands in homarus americanus. 173

In an embryo of this age I find in sections that the basal region of
the second antenna is completely filled with cells having certain distinct-
ive characters (Plate 2, Fig. 18, ms'drm.), which serve to differentiate
them from the outer wall (ecdrm.) of the appendage. These axial
cells are oval or slightly elongate, with rounded outlines, and less uni-
form in size and less densely granular than the rectangular ectoderm
cells (ecdnn.), which form the outer wall of the appendage. The nuclei
of the axial cells are more oval than the nearly spherical nuclei of the
ectoderm cells. In some sections it is to be seen that this plug of axial
cells is connected with similar plugs of cells in the base of the first an-
tenna and of the mandible by means of a single layer of cells immedi-
ately beneath the ectoderm of this region.

According to recent works on the formation of the germ layers and of
the appendages in Decapods, — I refer especially to Reicheubach ('86, p.
24, Taf. IX. Fig. 84, and Taf. X. Fig. 121), Bumpus ('91, p. 245, Plate
XVIII. Fig. 6), and Herrick ('95, p. 209), — the cells which form the
axis of the appendage are clearly mesoderm, in the generally accepted
meaning of the word.

In embryos one day older (Plate 2, Fig. 16), in which the first max-
illae (mx. 1) are marked off, a small number of these axial mesodermic
cells in the base of the second antennae have become differentiated from
the rest. They are destined to become the deeper part of the antennal
gland. These cells are nearly on a level with the general surface of the
body. They are larger than the remaining mesodermal cells, and have
the cytoplasm less densely granular, especially toward the centre of the
cell, and their nuclei also are larger and stain much more deeply with
haematoxylin. The nuclei are likely to be eccentric, and in many cases
they are surrounded by a clear zone of non-granular or very sparsely
granular cytoplasm (Plate 2, Figs. 19, 20, 21, 22, sac. trm.).

I believe that there is at first only one of these differentiated cells in
each antenna, although I have not found this condition in any of my
sections. My reason for thinking so is that in several cases in which
only two nuclei are present I have found these close together, and with
their adjacent sides somewhat flattened (Plate 2, Fig 19, sac. trm.), as
if recently divided. I have not found, however, any mitotic phenomena
which might serve as proof of a recent division. On the other hand, it

mentation stages (see his Tables 17 and 18, p. 56) as 14 to 16 days old. With this
as a basis, the embryo of my Figure 15 would probably be between 15 and 17 days
old. It falls between the Stages N and O of Bumpus ('91, p. 248, Plate XIV.),
and is a little further developed than Reichenbach's ('86, p. 47, Taf. III. Fig. 10)
Stage G of Astacus.

17-4 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY.

is possible that there are two cells differentiated at the same time.
Sections preceding and following the one seen in Figure 19 show that
only two of these differentiated cells are present in this appendage
at this stage. In three other specimens also I have found only two
such cells, so that there is certainly a two-cell stage in the development
of this organ. These nuclei divide, but divisions are not necessarily
simultaneous, for I have found cases in which there are three, four
(Fig. 21), five, six, seven, eight, nine, and twelve nuclei present. In a
few cases mitotic phenomena are found in some of the nuclei, but in
most cases, even where an odd number of nuclei is present, there are no
mitotic figures. These nuclei are distinctly different from those of the
ectoderm and of the surrounding mesoderm in the appendage. There
can be no doubt about these differentiated cells being of mesodermic
origin. In another particular this region differs from the surrounding
mesoderm; although the nuclei increase in number, there is not a cor-
responding increase of distinctly separated cells. There are partial cell
walls (Plate 2, Figs. 19, 21, par. cl.), but the cell areas are not fully or
sharply defined by cell membranes. This condition is found in indi-
viduals which have been treated by one or other of several different
methods of killing, hardening, and staining ; it does not appear there-
fore to be due to the influence of particular reagents.1 The result is a
partial syncytium, or cell complex, in which the nuclei are irregularly
distributed, but often show a tendency to take a peripheral position
(Plate 2, Figs. 23, 24).

Bergh ('88, p. 228) has found a parallel condition in the nephridia of
Criodrilus. My results do not, however, agree with the condition found
by Boutchinsky ('95, p. 169, Tab. VI. Fig. 143) in Gebia, in which, as
he says, the cells at a similar stage are separated by distinct walls.

In Homarus this syncytial mass has an even oval outline, the longer
axis of the oval lying nearly in the transverse plane of the body, but
directed obliquely laterad and ventrad at an angle of 20° or 30° with
the frontal plane. The evenly rounded outline of the mass is preserved
even where it is separated from the mesenchyme by only a thin layer of
mesodermic cells. This oval syncytium with its contained nuclei becomes
the endsac of the antenna! gland, and in the further description will be
so designated.

1 It should, however, be stated that my material had been in alcohol for some
time, which may have caused disintegration of cell walls in this particular region.
The cell walls of neighboring regions, and of corresponding positions in other ap-
pendages, are however firm and distinct.

waite: antennal glands in homaeus ameeicanus. 175

Iu an embryo about 22 to 24 days old (September 5th),1 represented
by Figure 17 (Plate 2), the spheroidal form of the end sac is well marked
(Plate 2, Fig. 23, sac. trm.). At this stage, however, when there are
twelve nuclei, it is not visible from the ventral surface of the appendage,
nor is there any circular or crescentic arrangement of nuclei and cells at
the base of the appendage, such as several authors figure as representing
the first trace of the antennal gland. Such an appearance accompanies
the ectodermic invagination, of which there is as yet no trace, either in
surface view or in sections.

In an embryo from the same brood two days older (September 5th to
7th) there appears in sections a definite vacuole (Plate 2, Fig. 24, vac.)
situated centrally in the endsac, which here has only nine nuclei. This
vacuole has a smooth distinct outline, which is evidence that it is not an
artifact. It is clearly not intercellular, for there are no cell membranes
near it. From this stage on there are seen in the endsac of each indi-
vidual a few well marked vacuoles of various sizes (Plate 2, Fig. 22, Plate
3, Fig. 26).

It seems probable that the lumen of the endsac originates by the con-
fluence of the vacuoles of the syncytium. I can therefore agree with
Kingsley ('89, p. 29) that the lumen of the endsac is neither an enclos-
ure of part of the body cavity nor in any way connected with it. A pro-
cess parallel to this is described by Bergh ('88, p. 228) in the formation
of the nephridia of Criodrilus.

It is interesting to notice that the formation of the lumen of the end-
sac begins before there is any trace of the ectodermic invagination, and
only eight days (August 30th to September 7th) after the earliest differ-
entiation of the endsac is seen, whereas in Paleemonetes (Allen, '93,
p. 338) the lumen of the endsac does not appear until early in larval
life and is there intercellular.

The next stage to be noticed is from another series. It is slightlv
older than that figured by Herrick ('95) in Cut 34, Plate I., and between
Stages 0 and P of Bumpus ('91, PI. XIV.). The tip of the telson has
grown cephalad until it is somewhat in advance of the base of the second
antenna ; there is no pigment in the eye. There have elapsed since fer-
tilization, to judge from a comparison with Herrick's figures, 23 to 25
days, but the actual interval is greater, — probably 28 to 30 days, — as

1 This embryo was from the same brood as that from which Figure 16 was
drawn, showing the earliest recognized condition of the endsac. Since the younger
was killed on August 30th, and the older on September 5th, there is an interval of
six davs between the two.

176 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY.

these embryos were killed in October. In most individuals of this age
there is no sign of ectodermic invagination in the region of the endsac.
The endsac (Plate 2, Fig. 25, sac. trm.) is oval and contains in different
individuals from 18 to 26 nuclei, which are often only partially separated
by cell walls, these being evident in the peripheral portions of the organ
only. The central area of the endsac is a granular mass with one or
more large irregular vacuoles, or several small ones, while the nuclei
lie peripherally in compartments between the incomplete cell walls. In
some cases the cell walls are completed, and thus a definite boundary to
the lumen is formed. This is the condition of the endsac when invagi-
nation of the ectoderm begins, but in some individuals, even after in-
vagination is well started, the cell walls of the endsac are incomplete and
its interior is still a granular mass with several vacuoles (Plate 3, Fig.

26, sac. trm., vac).

In some precocious individuals of this batch of embryos I find evi-
dence of the beginning of the invagination, — a slight depression of ecto-
derm cells. In Figure 25 (Plate 2) at the median ventral border of the
endsac appears a thickening of the ectoderm (i'vag. ec'drm.). This rep-
resents the beginning of the ectodermic ingrowth.

In sections of the stages succeeding this it is seen that the ectodermic
growth extends proximad and dorsad.1 The axis of this line of growth
is not parallel with the antero-posterior axis of the body of the embryo,
but in passing proximad trends slightly laterad, so that true parasagittal
sections of the embryo cut the invagination obliquely. The ingrowth
consists of a plug, or perhaps better, a sheet of cells, the width of the
sheet being however small. This sheet grows up around the end-
sac and lies against its distal and dorsal faces (Plate 3, Figs. 26,

27, i'vag. ec'drm.).

From all the evidence afforded by my sections I believe that this sheet
of cells arises by a process of growth which is most active at its free
(deep) end. The conclusive evidence must rest upon signs of mitosis in
numerous series of stages covering the period of this growth. I have not
sufficient material at command to settle the point definitely.

However this may be, this ectodermic ingrowth occupies a narrow
space between the distal and dorsal faces of the endsac and the wall of

1 In referring to the embryonic organ I must use designations of direction which
seem contradictory. For that face of the second antenna which lies nearest the
mandible and is now the dorsal face is ultimately the ventral, and must be so des-
ignated. It assumes its normal position at the time of hatching by revolving about
its base as a centre through an arc of about 135° in a parasagittal plane.

WAITE : ANTENNAL GLANDS IN HOMAPJJS AMEKICANUS. 177

the appendage. As a result, in a parasagittal section — at this stage a
proximo-distal section of the appendage — there is to be seen a single
row of cells (Plate 3, Fig. 26, i'vag. ec'drm.). Preceding or following
sections of the series usually show that there is an accompanying row by
the side of this, i. e. the sheet is two cells wide, and one cell deep dorso-
ventrally. In a few cases there is evidence in transverse sections that
there are three rows through at least part of its course, these being so
arranged that the cross section is triangular. These rows are not dis-
tinctly separated but interdigitate, so that the dividing line is anything
but straight and sharp, as would be the case were this formation an ac-
tual invagination. The line — there is no lumen — between these rows
may represent a potential lumen derived from the outside world, but I
can trace no actual connection. There is, to be sure, in the region of the
connection with the ectodermic wall, a slight depression (Plate 3, Fig.
26, fos.) in the surface of the ectoderm, but the cellular arrangement
does not warrant the conclusion that this is a real invagination. I be-
lieve therefore that the ectodermic ingrowth is a solid plug, and not an
invaginated sac.

As this plug advances along the dorsal (now ventral) face of the endsac
and approaches the proximal border, it is subjected to less stress, its
proximo-dorsal border being in contact with only the mesenchyme and
the fluids of the body cavity. The plug here enlarges ; at first there
are in this region one or two extra rows of cells (Plate 3, Fig. 26, i'vag.
ec'drm.), — cross sections often show four rows, — a little later the plug
enlarges into a knob-like body in which an intercellular lumen appears
(Plate 3, Fig. 27, i'vag. ec'drm., hi.). There is no reason to believe
that the lumen thus ai'ising is an enclosed part of the body cavity, and
there is no evidence to show that it is actually a part of the outside
world. I believe that it is a lumen of independent origin, just as is that
of the endsac ; but in the case of the ectodermic ingrowth it is inter-
cellular, while in the endsac it begins as an tWra-cellular space.

This ingrowth of the ectodermic plug and its expansion into a knob
containing a lumen is a rapid process. I have found all the stages hith-
erto noted in the same batch of eggs from a single female, and all killed
at the same moment. In September and October this condition is at-
tained in about 30 days, probably in somewhat less time in eggs extruded
in the middle of summer.

The condition shown in Figure 27 (Plate 3) progresses distad and
ventrad along the ectodermic plug towards the region of its connection
with the ectodermic wall of the appendage. The axial line separating

178 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY.

the cell rows is now distinct and even, and quite different from the line
separating the two interdigitating rows of earlier stages. This axial line
represents a potential lumen, and by separation of the apposed walls the
lumen is formed here at later stages. Figure 28 (Plate 3) is a parasa-
gittal section from an embryo about 35 days old. The endsac (sac. trm.)
has its cell walls well marked and the lumen (lu. ) is sharply bounded.
The ectodermic ingrowth (ivag. ec'drm.) shows in longitudinal section
two rows (many rows in all) to a point near the distal end of the endsac.
The cells are more columnar than in earlier stages. Some of the ecto-
dermic cells of the wall of the appendage are much elongated. They go
to form the ligaments seen in this region in later stages.

Figure 29 (Plate 3) is a transverse section at a little later stage (about
40 days, September 10th to October 20th). Here the lumen of the ecto-
dermic plug (i'vag. ec'drm.) is evident and can be traced to the exterior
(of. ex.). This duct has arisen as the outward extension of an intercel-
lular lumen, first appearing in the region of the deep proximo-dorsal end
of the ectodermic plug.

There is thus formed in the embryo of about six weeks an endsac with
definite lumen, and an ectodermic ingrowth with lumen, and with a duct
opening to the outside world, but these two lumina are not in connec-
tion. The duct of the ectodermic ingrowth persists as the permanent
duct of the adult organ. However, its lumen is not always visible, espe-
cially in the deeper regions, for the growth and resulting pressure at
times apparently obliterate it by causing a close apposition of the walls,
which, however, do not actually fuse, for at all subsequent stages when a
section in the region is torn the lumen is reopened.

There are still two important points to be determined in the embry-
ology of the antennal gland. When, and at what point, does the lumen
of the endsac become continuous with that of the ectodermic sac? In a
series of embryos taken from a female kept in confinement during the
winter (see Herrick, '95, No. (3) 18, Table 18, p. 56) I find that in those
killed April 1st (estimated age 273 days) the lumina are separate, but
in embryos of the same series killed May 1st (estimated age 303 days)
the lumina are continuous. The details of the union of the two lumina
I am unable to describe, as I have no intermediate stages.

For some time before this last stage is reached, the lumen of each part
is seen in sections, but the two lumina are separated by the closely ap-
posed walls of the two sacs (Plate 3, Fig. 31 ) . The break through the two
walls occurs on the dorsal face of the proximal region of the endsac and
at the terminal end of the ventral part of the forked proximal end of the

WAITE: ANTENNAL GLANDS IN HOMARUS AMERICANUS. 179

ectodermic sac (cf. Fig. 30, Plate 3). After communication is established,
an abrupt transition in the character of the wall is still noticeable at the
point of union. This level may be termed the mes-ectal line, since it is
the line of junction of the mesodermic with the ectodermic part of the
organ. There is a valve-like flap of cells projecting from the wall of the
endsac into the orifice connecting the two cavities. This condition is
still evideut in the first larva (Plate 6, Fig. 51, vlv.).

The brood of this female kept in confinement during the winter was
hatched 33 days after the killing of the specimens which showed con-
necting lumina (about 330 days after fertilization), so that when the end-
sac comes into communication with the exterior the embryo has already
passed through ten elevenths of its embryonic period. It is possible that
the establishment of the communication between the two sacs occurs in
the earlier part of the period embraced between the ages of 273 and 303
days, but even if it does not occur until the latter part of the period,
the lumen of the endsac is in direct communication with the exterior at
least a month before the end of embryonic life. The only records that I
find of the time when this connection is effected in other Crustacea is
that made by Allen ('93, p. 338) for Pakemonetes, and by Kingsley ('89,
p. 30) for Crangon. Kingsley states that " the external opening to the
gland is not formed until after hatching." Allen says that the forma-
tion of the lumen of the endsac and its communication with the exterior
occur early in larval life. In Homarus, however, the lumen of the end-
sac communicates with the outside world late in embryonic life. If, as
Allen thinks, the establishment of this condition marks the beginning of
functional activity, then the antennal gland in the lobster is functional
as an embryonic organ. But I cannot agree with Allen. On a priori
grounds we may expect an excretory organ in the lobster embryo. Ex-
cretion takes place in many yolk-bearing embryos, and special embryonic
organs having this function occur in many cases, such as the embryonic
nephridia of Annelids and pulmonate Gasteropods, and the shell gland in
the embryo of many Crustacea. In some decapod Crustacea the shell
gland is functional in the embryo, but atrophies soon after hatching.
Whether it is present in Homarus I am unable to say ; but I believe it
probable that excretion in the embryo is performed by some such special
organ rather than by the antennal gland, for in the latter I find no direct
evidence of excretory activity during embryonic life. The cavity of the
gland contains neither a granular clot nor excretory globules, such as are
found in the larval and adult stages, nor have the cells of the wall of the
ectodermic sac the marked striation which is seen later when glandular

180 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY.

activity certainly exists. It may well be that excretion in the embryo
is less abundant than in the predatory larva, because of the less active
metabolism and the more perfect food supply of the yolk. The presence
or absence of excretion products in the organ during embryonic life
might be demonstrated on fresh material by micro-chemical tests, and
the question thus definitely settled ; but my material is not so preserved
as to permit the application of such tests.

After the formation of the ectodermic ingrowth, as shown in Figures
26 and 27. (Plate 3), the further embryonic development consists, then,
in four chief phases : — (a.) The line representing its lumen extends to
the outside world and so establishes a potential duct. This is a rapid
process and is completed shortly after the stage shown in Figure 28
(Plate 3). The condition previously noted at the deep ind of the ecto-
dermic ingrowth — several rows of cells — progresses and finally reaches
the exterior near the point where the ectodermic ingrowth first appeared
(Fig. 25). This potential duct is open in many cases (Fig. 29). The
opening of the duct is on the anterior side of the appendage, somewhat
toward its ventral face. The ectodermic duct and sac extend from this
external orifice dorsad and proximad, covering the distal and dorsal faces
of the endsac, but not until a later stage does it pass ventral to the end-
sac, nor ever much beyond its proximal border.

(b.) The ectodermic sac becomes extended and complicated through-
out embryonic life by the formation of evaginations, which begin to ap-
pear as soon as the lumen of the duct has been extended to the external
world. These outgrowths are flattened, probably by mechanical pressure.
There are three regions at which principally these evaginations occur.
One of these is proximal, i. e. at the deep end of the ectodermic ingrowth.
Here the outgrowth soon becomes forked, there being a ventral portion
nearer the endsac and a proximal portion more nearly in line with the
axis of the appendage. Figure 30 (Plate 3) shows these forks, but the
proximal one is seen only in cross section. It extends at right angles to
the main axis of the appendage and to the plane of the section, approx-
imately parallel to the frontal plane of the embryo. The notch separat-
ing these forks never becomes very deep, but it is recognizable from the
second month of embryonic life until the late larval stages. On the
dorsal face of the ectodermic sac in the region of its sharpest curvature
is a second area of evagination. This outgrowth, which is dorsad, does
not exist at the stage shown in Figure 30, but it appears slightly later.
It has a considerable horizontal extent (Figure 31, Plate 3), being all
that part of the ectodermic sac which later lies distal to the duct. The

WAITE: ANTENNAL GLANDS IX HOMARUS AMERICANUS. 181

third region is along the entire length of the ectodermic sac in the dorso-
inedian area. An idea of the position of this region may be gained from
Figure 31 (Plate 3), which is a transverse section. It is there repre-
sented by the long dorsal tract of the ectodermic sac, extending mediad
near the letters "sac. trm."

The complications in the shape of the ectodermic sac arise from the
differential growth in these three regions. The ventral fork (Figure 30)
grows ventrad around the proximal face of the endsac, and then distad
between it and the wall of the appendage, meanwhile extending laterad
and mediad in the frontal plane. It is this extension that finally con-
nects with the endsac. The proximal fork has grown at the same time,
but extends more nearly in the axis of the appendage. Meanwhile the
dorsal extension has grown considerably dorsad and distad, and the dorso-
median evagination has extended horizontally. As a result of all these
growths the parts of the ectodermic sac have come to embrace the end-
sac on all sides, though not completely, for it is seen in the larval stages
that through a part of its extent the endsac lies against the ventral and
lateral walls of the appendage.

All these growths and extensions of the ectodermic sac are evagina-
tions, and carry with them part of the original lumen, but they are in no
respect convolutions. The complicated lumen thus arising is suppressed,
as far as actual cavities are concerned, until the latter half of embryonic
life. In an embryo of 152 days (July 1st to December 1st) no actual cav-
ity is observable, only the line marking the meeting of apposing walls,
but in an embryo of 303 days (July 1st to May 1st) the lumen is an actual
cavity at almost all points. In stages intermediate between these the
lumen becomes more and more open as age advances. We thus see that
the distention of the lumen occurs in large part prior to the junction
(see p. 179) of the lumina of the endsac and ectodermic sac.

(c.) The endsac does not suffer much complication during its embry-
onic growth, but it becomes much flattened and compressed between the
extensions of the ectodermic sac. The chief growth is laterad and dorsad,
and gives rise to the condition found in the first larva, where the endsac
lies in large part lateral to the ectodermic sac (Figure 40, Plate 5).

(d.) Finally the lumina of endsac and ectodermic sac become contin-
uous (see p. 178), and the organ has then practically reached the
larval condition.

vol. xxxv. — no. 7. 3

182 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY.

C. Development in the Larvae.

(a.) In the First Larva.

The stages at which the various phases of the development of the an-
tennal gland in the embryo appear are not precisely fixed by time intervals
because of seasonal variation in the rate of development. They may be
referred to particular conditions of prominent external organs, such as
the pigmentation of the eyes or the position of the tip of the telson, with
much greater precision, but even this allows of no close comparisons with
parallel stages in other genera or families. The stage at which the em-
bryo hatches, marking the transition from the passive embryonic to
the active larval life, is sharply defined in most free-living Crustacea,
and it would seem that this is a time at which comparisons may
be made most safely. Moreover, the transition to the predatory life
of the larva must have a marked influence upon metabolism, and
therefore upon the excretory organ and its functions. It has there-
fore seemed well to consider this stage in some detail.

My material was all killed within twelve hours after hatching, and
hence represents the earlier part of the first larval period, the extent of
which may be as much as five days (Herrick, '95, p. 172).

1. General Structure.

With the hatching of the embryo a considerable alternation occurs in
the topography of the antennal gland, but this is due to mechanical
causes rather than to growth. The second antenna, which has been
folded back against the embryo so that its distal end is posterior to its
proximal end, is released in hatching and swings — approximately in a
parasagittal plane — through about 135 degrees, to a position in which
the distal end is anterior to the proximal. Thus in the basal region of
the appendage the dorsal face is shortened and thereby relieved of the
stress which has existed at that place during embryonic life. This
removal of pressure from the extensions and evaginations of the ecto-
dermic sac brings about a change in its form, which passes from the
flattened constrained shape to a more rounded one. This permits the
separation of the apposing walls and the establishment of a large lumen.
Likewise the endsac, which has been compressed between the dorsal and
ventral portions of the ectodermic sac, is relieved from this pressure and
expands as much as the larger space allows.

In Figures 32 to 38 (Plate 4) are shown the second, third, fifth,
sixth, eighth, eleventh, and twelfth sections respectively from a series

waite: antennal glands in iiomarus americanus. 183

of tliirteen sections 10 micra thick through the left antennal gland of
the first larva. The sections are cut parallel to the long axis of the
antenna and perpendicular to the frontal plane of the larva, and are
viewed from the lateral face, Figure 32 being the most lateral one
shown. The members of the series reproduced do not extend near
enough to the median plane to include the opening of the duct to the
exterior, but the relations of this duct to the ectodermic sac and to the
external orifice may be inferred from Figure 38..

The ectodermic sac — the walls of which are represented in a lighter
shade than those of the endsac — is oval in shape, flattened vcntrally,
and elongated in the axis of the antenna (Figures 32 to 38, sac. ee'drm.).
The endsac (Figures 33 to 35, sac. trm.) lies in a depression of the ven-
trolateral face of the ectodermic sac, as is readily to be understood from
its position in the sections and in the series as a whole. The three
evaginations of the ectodermic sac — proximal, ventral, and dorsal —
are still indicated, though less prominently than in the embryo. The
passage from the lumen of the endsac to that of the ectodermic sac —
partly closed by the valvular Hap — is shown in Figure 33.

The endsac (Figures 33, 34, sac. trm.) now lies closely applied to the
ventral wall of the ectodermic sac, a relation which is changed in later
stages (see p. 190). A ventral recess of the ectodermic sac shown in
Figures 37 and 38 leads distad and mediad to the duct communicating
with the external world. This is better shown in Figure 39, a section
of the left gland from another series, but viewed from the median face.
This being cut in a more favorable plane, at a slight angle with the long
axis of the appendage, shows that the ventral recess continues as a duct
to the external opening (of. ex.) at the summit of a papilla on the ven-
tral wall of the appendage. It is clear from this figure that the duct is
connected with the ventral face of the ectodermic sac. The position of
the duct marks the original course of the ectodermic ingrowth in the
embryo.

The conclusions reached by the examination of such parasagittal
sections are supplemented by the examination of transverse sections.
Figure 40 (Plate 5) is the anterior face of a transverse section through
the first larva in the region of the base of the second antennse, and
shows the general position of the antennal glands, which lie almost
wholly within the appendage. Figures 41 to 48 (Plate 5) exhibit the
anterior faces of the third, fifth, seventh, ninth, tenth, twelfth, four-
teenth, and seventeenth sections of a series through the right gland,
Figure 41 being the most anterior. The sections, each 10 micra thick,

184 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY.

are cut perpendicular to the long axis of the antenna. No section plane
passes through the external orifice of the duct, but the orifice is near the
plane of the section shown in Figure 46. The opening to the exterior,
it will be observed, is not at the middle of the ventral face of the ap-
pendage, but more toward the median side. The communication be-
tween the lumina of the endsac and ectodermic sac does not appear, as
it is small and its axis is inclined to the plane of the section (cf. Figure
33), but it lies between Figure 47 and the next following section. Its
position is indicated in Figure 47 by a. It will be noticed that
the endsac lies partly ventral and partly lateral to the ectodermic sac
(Figures 44 to 48), and that its free margin extends dorsally, while
the ectodermic sac extends ventrally. This is the beginning of a pro-
cess of growth that is destined to reverse the relative positions of the
two parts as found in the embryonic stages, where the endsac was ven-
tral and the ectodermic sac dorsal (cf. Figs. 26, 27, 28, Plate 3).

Finally, Figure 49 (Plate 6) represents the dorsal face of a frontal
section of the right gland, — the tenth, counting from the dorsal face of
the gland, in a series of eighteen, each 10 micra thick, — in which is
seen the lateral position of the endsac and the communication between
the endsac and the ectodermic sac at a. The passage is situated at
the lateral and proximal part of the endsac.

The evidence from sections in the three planes shows that the endsac
lies ventral and lateral to the ectodermic sac ; that the ectodermic sac
is elongated in the axis of the antenna, and gives off the duct to the ex-
terior from its ventral face well over on its median border; and that the
gland lies almost wholly within the appendage and not in the cavity of
the body proper.

The greatest diameter of the gland measured in each of the three axes
gave as a result of the average * of several cases the following : —

Maximum proximo-distal axis, 0.3 mm.
" dorso-ventral axis, 0.2 "

" latero-median axis, 0.2 "

The average length of the first larva is 7.8 mm. from tip of rostrum to
tip of telson.

1 The extremes of measurement vary from these mean averages 20% in some
cases, which is probably due to slight differences in the age of larvae, or to indi-
vidual differences independent of age.

WAITE : ANTENNAE GLANDS IN HOMAKUS AMERICANUS. 185

2. Histology.

The histological character of the endsac differs distinctly from that of
the ectodermic sac, and the dividing mes-ectal line is sharply marked
(Plate 6, Fig. 51, In. mes-ec). The valve-like flap (vlv.) in the orifice
between the two sacs lies on the dorsal side, and its cells are connected
directly with those of the wall of the endsac, and are of the same na-
ture. In the region where the walls of the two sacs are adjacent,
they are not applied to each other continuously, but are separated by
spaces (Figure 51, lac. sng.) which constitute blood lacunae. This is
proved by the frequent presence in them of blood corpuscles (Figure
59), the structure of which is characteristic. These blood spaces are
separated at intervals by partitions or pillars (Figures 49, 51) connect-
ing the basement membranes of the two sacs. In structure these par-
titions resemble closely the tissue of the endsac. I have not traced the
continuity of these blood spaces with the general circulatory system.
They are not lined by an endothelial layer, as far as I am able to
discovei*.

The cells of the endsac are all of the same general type, but show
two different forms, one (Fig. 52, Plate 6) being more flattened than the
other (Fig. 53, Plate 6). The more flattened cells occur in those regions
of the wall which are not adjacent to the wall of the ectodermic sac
(Fig. 51). All of the cells of the endsac have large nuclei, which lie close
against the free wall of the cell and in most cases cause a protrusion of
the cell into the lumen (Figs. 51, 53). The nuclei in the flattened cells
are more elongated in a direction parallel to the surface of the wall than
is the case in the more rotund cells of the other region (cf. Fig. 52 with
Fig. 53). Both kinds show a chromatic network. By the methods used
it was impossible to demonstrate any definite lateral cell boundaries, but
the extent of the cell territories is indicated by the position of the nuclei
and by the contour of the free face of the cells. The cytoplasm is finely
granular, and in most cases distinctly striated, the rows of granules be-
ing perpendicular to the surfacs of the wall of the sac (Fig. 53). This
granular striation would seem to indicate a secretory activity. There
is, however, no bounding striate cuticula on the inner face of the cells.
The wall of the sac presents a basement membrane, which is however
much less marked than that of the wall of the ectodermic sac, and
appears as a single line even under the amplification of a ^ oil immer-
sion (Zeiss) and ocular 4.

The cells of the ectodermic sac are very different in character from

186 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY.

those of the endsac, and the transition at the mes-ectal line is, as I have
said, abrupt, there being no cells »f intermediate character (Fig. 51, In.
mes-ec). The cells of this ectodermic part differ considerably in size and
shape, although not essentially in structure. Figures 54 to 57 (Plate 6)
are all drawn from the same series of sections, and with the same ampli-
fication, but from different regions in the wall of the ectodermic sac ;
they fairly illustrate the variation in the size and the shape of the cells.
The cells from the dorso-median wall (Fig. 56) are more columnar, and
the nuclei are relatively farther from the basal end of the cell than in
other regions. Those from the ventral face (Fig. 55) are much smaller
and show fewer striations. The cytoplasm in all these regions is granu-
lar, more densely so than in the endsac, and in all but the ventral wall
(Fig. 55) the granules are arranged in rows perpendicular to the base-
ment membrane and separated by clearer areas of periplasm. This
striation is found both in the basal and free ends of the cell. In some
cases the clear area between rows of granules is expanded to form large
vacuoles (Fig. 56, vac). The nuclei are bounded by a definite nuclear
membrane and have a coarse chromatic network, which stains deeply
in hematoxylin. Where the strands of this network cross, there are
formed chromatic masses of greater density ; these are usually triangular
in section (Figs. 54, 57, and 58). Nucleoli are found in some cells
(Fig. 55).

Mitotic phenomena, though not abundant in these cells, are now and
then met with. One such cell in pi'ocess of division is shown in Figure
57. The plane of division is in all cases perpendicular to the basement
membrane, so that the tissue remains only one layer thick. The rarity
of these indications of cell division, in spite of the considerable increase
in the size of the gland in the beginning of the second larval stage, is
perhaps an indication that the cell increase in the organ is more rapid
in the latter part of the first larval period than in the earlier part, at
which time my material was killed.

The basement membrane is seen in all cases. Under high powers it
has an appreciable thickness (Fig. 58, Plate 6, mb. ba.). I have seen
no trace of nuclei in it to indicate an endothelial nature.

All the cells of the ectodermic sac agree in having a vertically striate
cuticula on the free face (Figs. 54-58, eta.). Under high magnification
this cuticula is seen to be made up of bundles of rods or strings of gran-
ules grouped into the shape of an hour-glass, the clear spaces between
these bundles having in section a lenticular shape (Fig. 58, eta.). This
cuticula agrees in a measure with that in the walls of the labyrinth of

WAITE: ANTENNAL GLANDS IN HOMARUS AMEEIGANUS. 187

the adult gland, but iu this larval stage there are no globular promi-
nences extending out from it into the lumen.

The histological structure of the duct is shown in Figure 60 (Plate 6),
which is from a longitudinal section of the appendage. Its wall is di-
rectly continuous with the body wall. The dorsal lip of the orifice {of.
ex.) shows no abrupt transition from the wall of the duct to that of the
body. On the ventral lip the limit is somewhat more precise owing to
the sharp folding, but it is difficult to say to which wall the apical cells
of the ventral lip belong. The cells of the dorsal wall show differences
in size and shape among themselves, and a considerable difference from
the cells of the ventral wall ; these being smaller and imbricated. The
cells of the duct do not show the granular striatum characteristic of the
cells of the ectodermic sac, nor is there a striate cuticula on the face
of these cells. These two characters serve to distinguish the cells of
the duct from those of the ectodermic sac. and lead to the inference that
the former have no secretory function. However, as one follows the
wall of the ectodermic sac as it merges into the wall of the duct, these
two characters gradually disappear, and it is impossible to say that there
is any definite line where ectodermic sac ends and duct begins, — a con-
dition to be expected from the fact of their having had a common origin.

The cuticular layer of the external shell bends into the duct (Fig. 60),
but soon thins out and disappears. It extends for a greater distance
along the dorsal than along the ventral wall.

The lumen of the ectodermic sac contains a granular coagulum, which
resembles in staining properties that seen in various parts of the adult
gland, but there is no trace of the globular structures there seen (p. 166).
This coagulum is in all probability the secretion of the walls of the ecto-
dermic sac, indicating that the functional activity of the gland has
already begun ; but that function must differ in some respects from the
function in the adult, as indicated by the absence of the globules. This
coagulum is not seen in the lumen of the endsac. The identity of the
substance found here with that found in the adult gland can only be
established by micro-chemical tests, and for that my material is not
suited.

It seems probable that the function, although possibly somewhat
different from that of the adult, is of essentially the same nature, i. e.
excretory. This would be expected both upon a priori grounds and
from the structure of the cells as described. At this stage metabolism,
consequent upon the activity of the larva, is constant, and as all the
structural conditions appear favorable, it seems as if this organ, which

188 bulletin: museum of comparative zoology.

later is certainly the chief, if not the only, excretory organ of the body,
must be functional and functional in the way of its later activities. I
therefoi'e believe that the antennal gland in the first larva is a functional
excretory organ.

(b.) In Older Larvae.

The antennal gland in the second larva shows no marked difference
from that in the first, either in shape or in histology. There is, how-
ever, as compared with the first larva (see page 184), a difference in size,
which again taking averages, is now :

Maximum proximo-distal axis, 0.3 mm.
" dorso-ventral axis, 0.3 "
" latero-median axis, 0.25 "

If we assume, in view of the simplicity of the organ at these stages,
that its secreting surface is about equal to the surface of a parallelopipedon
having the dimensions of the whole organ, we find that the area of the
secreting surface in the second larva is 40-50 per cent greater than in the
first. The average length of the second larva (see Herrick '95, Table 34)
is only about 13 per cent more than that of the first. If we compare the
cubes of these dimensions we get an increase of something over 40 per
cent in the bulk of the larva ; consequently the increase of secreting
surface in the antennal gland simply about keeps pace with the increase
in bulk of the larva.

In the third larva we get the beginning of that process of complication
by which the antennal gland of the adult animal is developed from the
relatively simple organ of the time of hatching. Between the conditions
in the first and second larvas there is only the difference of size. Be-
tween the conditions of the second and third larvae there is a difference
due to increase in size, by reason of which the gland comes to lie largely
in the body instead of in the appendage only, and in addition a differ-
ence resulting from the development of long slender evaginations on
the dorsal and median faces of the ectodermic sac. These evaginations
have only a very small lumen, but the lumen is always in connection with
the main cavity of the ectodermic sac, as is shown in Figure 50 (Plate
6, evacf.). By such evaginations the secreting surface is greatly increased
without a correspondingly large demand for space in the body cavity.

In the third larva there also begin to appear at the free ends of the
cells composing the wall of the ectodermic sac the peculiar globular
vesicles which form so marked a feature of the adult gland. They are
less well marked and less abundant here thau in the older larvse.

waite: antennal glands in homaeus amekicanus. 189

Iii this stage, the vessels and blood spaces {lac. sng.) which lie between
the endsac and the ectodermic sac are for the first time seen to be
lined with endothelial cells. This vascular tissue has arisen from the
cells which are between the endsac and the ectodermic invagination, and
which are seen at later embryonic stages (Plate 3, Fig. 30) as attenu-
ated cells forming a sort of sheath around the endsac.

In the fourth larva, the evaginations from the ectodermic sac have
further increased in number; those which occupy the positions of the
ones found in the third larva have elongated, and in some cases the free
end is folded back upon itself. On the dorsal face of the ectodermic sac,
well forward, there is an evagination which is larger and more rounded
than the others. It contains a large cavity which is comiected with the
main cavity of the ectodermic sac by a narrow duct. This evagination
is directed backward and dorsad ; it is the first differentiation of the
vesicle of the adult gland.

The globular vesicles on the free ends of the cells in the wall of the
ectodermic sac are more marked, and appear very much like the con-
dition in the adult.

My material of the older larval stages is meagre and not in good his-
tological condition. It is difficult, in the absence of data as to the num-
ber of moults passed, to determine to what larval stage a given specimen
belongs. The sole criterion available is therefore that of length.

In larvce 1^. millimetres long, the dorsal rounded evagination represent-
ing- the vesicle has increased much in size, extending caudad and dor-
sad, until its walls are in contact with those of the masticatory stomach.
The narrow slender evaginations have increased in number, especially in
the ventral part of the organ, and are so entangled that it is futile to at-
tempt to follow details.

Larvce 18 millimetres long are the oldest that I have. According to
Herrick ('95, Table 34), this length indicates the seventh larval stage,
and life at the bottom of the sea. In these larvae the vesicle has enlarged
still more, its walls becoming thinner, until it covers dorsally nearly all
the organ. The complication of the slender evaginations is now still
greater, and anastomoses between adjacent evaginations are established
at many poiuts by a breaking through of their walls. This condition is
found sparingly in the fourth larva, but in this (seventh T) stage it is
quite common.

I have no material of adolescent stages in which to follow this process,
but there seems every reason to believe that it is by these evaginations,
with subsequent anastomoses, that the complication of tubules in the

190 BULLETIN: MUSEUM OF COMPAEATIVE ZOOLOGY.

labyrinth of the adult gland is attained, and not by any system of coilin"
of one or a few tub ides.

I have already (p. 180) called attention to the fact that in the embryo
the endsac is ventral to the ectodermic sac, whereas in the adult organ
the endsac is dorsal to the mass of ectodermic tubules (labyrinth) which
are developed from the ectodermic sac. How is this change of position
brought about 1 In the first larva — as is seen in Figures 33-35 (Plate 4)
and Figures 40, 44-48 (Plate 5) — the endsac lies chiefly ventral to the
ectodermic sac, but also extends in part dorsad up along the lateral face
of the latter. Sections of later larvae show that the chief region of
growth in the endsac is in this dorso-lateral projection, while in the
ectodermic sac the chief area of growth and complication by evagination
(see p. 189) is in the ventral i-egion, by means of which it extends laterad
ventral to the endsac and between it and the ventral wall of the append-
age. There is, then, by this differential growth a rotation of the centre
of mass of each of these parts of the organ. This rotation is around the
antero-posterior axis of the gland, and, if one imagine himself viewing
this from an anterior point, this rotation is clockwise in the right gland,
and the reverse in the left. There is no actual twisting, but only a
change in position of the centre of mass of endsac and of ectodermic
sac. In the larva 18 millimetres long this process has progressed so far
that the plane which separates endsac from ectodermic sac is approxi-
mately parasagittal in position. Since the antero-posterior extent of the
endsac is much less than that of the ectodermic sac, the former can grow
dorsad and then mediad without disturbing the stalk which connects the
evaginated vesicle with the ectodermic sac, because the endsac lies en-
tirely posterior to this stalk. This process also makes it easy to under-
stand how it comes about that the endsac in the adult gland empties
into the lateral anterior lobe of the labyrinth (see p. 164), but has no
direct connection with the median anterior lobe.

I have not at my command the material to enable me to follow this
process of differential growth in the adolescent stages; but as far as I
have been able to examine the larvae, the conditions found agree with
the process which I have mentioned, and I believe that it is in this way
that the relative position of ectodermic sac and endsac as found in the
embryo comes to be reversed in the adult, where the endsac is dorsal to
the ectodermic sac or labyrinth.

waite: antennal glands in homarus americanus. 191

III. Theoretical Considerations.

A. Homology of the Antennal Glands with the Nephridia of

Annelids.

Leydig ('60, p. 28) was the first to suggest that the antennal glands
of Crustacea and the nephridia of Annelids performed similar duties.
He, however, ascribed to both a respiratory instead of an excretory
function. Since that time there has been much discussion in regard
to the functional likeness of these organs in the two groups of animals.
Naturally, this consideration has led to the discussion of their possibly
being homologous. This discussion was inaugurated by Kowalevsky
('71). Upon grounds of analogy, and from a comparison of the struc-
ture of the adult organs, it came to be pretty generally accepted that
this homology was valid. Confidence in this view was, however, shaken
by the publication of Reichenbach's ('77) description of the ectodermic
derivation of the antennal gland in Astacus. Kowalevsky (71) had
given proof of the mesodermic origin of the nephridia in Euaxes (p. 19)
and in Lumbricus (p. 25) and had formulated the germ-layer theory,
holding "that the homologies of the germ layers in different types
afford a scientific basis for comparative anatomy and embryology, and
must be recognized as the starting point for the proper understanding
of the relationships of the types " (p. 60). So great was the influence
of the germ-layer theory that the results of Reichenbach in deriving the
antennal gland of Astacus entirely from the ectoderm served, for the
time being, completely to check its comparison with the mesodermic
nephridia of Annelids.

Grobben ('79) showed, however, that the shell gland of Moina was
in part mesodermic, and thus partially restored the grounds for homolo-
gizing it with the nephridia of Annelids ; but Ishikawa's ('85) evidence
that the antennal gland in Atyephira is derived solely from the ecto-
derm, and the conclusions reached soon after in Reichenbach's ('86)
memoir, — in which the ectodermic origin of the antennal gland in
Astacus was described and figured in detail, — not only reaffirmed the
obstacles to establishing a homology between the antennal glands and
the nephridia of Annelids, but also placed in different categories the shell
glands of Entomostraca and the antennal glands of Malacostraca.

The contrary conclusions reached by Kingsley ('89) in his descrip-
tion of the development of the antennal gland in Crangon threw some
doubt upon the accuracy of the observations and interpretations of

192 bulletin: museum of comparative zoology.

Reichenbach, and served to restore tentatively confidence in the ho-
mology of the antennal glands of Malacostraca with the nephridia of
Annelids. Since that time the idea that the antennal gland is in part
mesodermic has gradually gained ground, but the question is still
treated in current writings as in a degree an open one, though nearly
all the text-books, following Reichenbach, state that it is entirely ecto-
dermic in origin.

No one has repeated, as far as I know, Reichenbach's work on the
development of the gland in Astacus. With the exception of the work
of Boutchinsky ('95), 1 — which unfortunately is little known because
of its being in Russian, — no satisfactory account of the embryology of
the antennal gland in any Malacostraca has appeared since Kingsley
('89) wrote. It seems probable, in view of the agreement of Kingsley,
Boutchinsky, and myself, that Reichenbach's figures and descriptions
are misleading, and that he entirely failed to see the endsac. Lebe-
dinski ('92, p. 231), however, endeavors to reconcile Reichenbach's
results with those of Kingsley, and with his own on the shell gland of
Eriphia, by holding that the mesodermic sheath described by Rei-
chenbach ('8G, p. 98) is homologous with the mesodermic constituent
forming the endsac in Crangon and in Eriphia. I do not think that
Lebedinski's point is well taken, because in Crangon, Eriphia, Gebia,
and Homarus the endsac appears before the ectodermic invagination,
whereas in Astacus — accepting for the moment Lebedinski's interpreta-
tion — it arises much later. Further, the mesodermic sheath described
by Reichenbach is, according to his own words, only an enveloping
sheath of connective-tissue elements and not at all glandular. This
sheath is recognizable in Homarus, where it becomes in part the in-
vesting sheath of the gland proper and of the vesicle, and in part the
tissue of the blood vessels, but has no part whatever in the formation
of the secreting epithelium.

We now know that the antennal gland is of double origin — meso-
dermal and ectodermal — in Crangon (Kingsley), Gebia (Boutchinsky),
and Homarus. These, to be sure, are all in a rather limited group, the
Macrura ; but it is my belief that similar conditions will be found in
other Malacostraca.

The ontogenetic development of the nephridia of Annelids has re-
ceived attention from many investigators since the pioneer work of Ko-
walevsky ('71). The stimulation to these researches in the earlier part
1 The preliminary paper was published a year earlier, Boutchinsky ('94).

waite: antennal glands in homarus americanus. 193

of the period came from the problematical nature of the organ, and to
this was later added the interest occasioned by the prominent place
given this system of organs in the discussion of the Annelid ancestry
of Vertebrates.

Since no one has held that the endoderm takes part in the formation
of the nephridia in Annelids, the organ presents three possibilities as to
origin: (1) that it arises entirely from the ectoderm; (2) that it is
wholly mesodermic ; or (3) that both ectoderm and mesoderm contri-
bute to its formation. The first of these possibilities has not been ad-
vocated by any one, although Wilson ('89) * comes very near to such an
interpretation. He says (p. 423) that the nephridia "arise in connec-
tion with a continuous cell-cord of ectoblastic origin. . . . Each neph-
ric cord terminates behind in a pair of teloblasts derived from the
ectoblast. The entire nephric cord is formed by the continued divis-
ions of these ' nephroblasts,' which agree precisely with the ' neuro-
blasts' in structure, action, and mode of origin." The nephridium
(p. 424) is later invested by a sheath of mesoblastic cells which become
the peritoneal investment. Wilson believes that the funnel and the
investing cells alone arise from the mesoblast, i. e. separately from the
rest of the nephridium, and his conclusions are in the main in agree-
ment with those of Whitman ('87, p. 161) in the case of Clepsine.
Wilson (p. 426) admits that, in view of the observations of Vejdovsky,
he cannot " deny the possibility that the glandular part may be differ-
entiated from the somatic mesoblast at a very early period, fusing im-
mediately with the cells of the nephric cord, which may give rise only
to the end vesicle." If this be admitted, his results agree in general
with those of Vejdovsky (see p. 194).

The other extreme, represented by the second possibility, is reached
by Bergh ('88). He finds (pp. 226, 227) in Criodrilus a single funnel-cell
differentiated at the point of origin of the septum, from which is prolif-
erated a cord of cells extending posteriorly. This is the beginning of
the segmental organ, which soon becomes covered with a sheath. The
external terminal part (Endstiick) is in no way connected with the epi-
dermis (p. 229), the segmental organ arising entirely from the " Haut-
muskelplatte." The same condition is found by Bergh ('90, p. 501) in
Lnmbricus, in the development of the nephridia in which " Trichter-,
Schlingen- und Endabschnitte differenziren sich aus einer einheitlichen
Anlage heraus, die in den inneren Muskelplatten ohne Betheiligung der
Epidermis entsteht." Bergh finds support in Lehmann's ('87, p. 348)

1 I refer to his later paper only, as it includes all the essentials of the earlier
(Wilson, '87) one.

194 bulletin: museum of compaeative zoology.

statement that the epidermis takes no part in the formation of the gland,
but concedes ('88, p. 229) that in certain Hirudinea the Endstiick arises
by an invagination of the epidermis.

The middle ground between these extremes is the position taken by
Vejdovsky ('84 and '92). In early stages of Rhynchelmis ('84, p. 123),
on the posterior face of the dissepiment there is a large cell, from which
there is proliferated posteriorly a solid cord of cells. These cells later
form a double row and acquire a lumen. This becomes the glandular
portion of the organ, which therefore arises from the mesoblast. The
nephrostome appears independently on the anterior face of the dissepi-
ment, and later joins the glandular portion by piercing the dissepiment.
The cord of cells proliferated posteriorly from the dissepiment is met, at
a later stage, by an ectoblastic invagination in the form of a solid cord.
From this invagination is derived the lining of the end vesicle and the
efferent duct. This ectoblastic lining is, at a subsequent period of devel-
opment, surrounded by mesoblastic tissue in the form of a sheath. In
his later paper ('92, pp. 339-342) he reaffirms the results of his earlier
publication. This general plan of development, with slight modification
of details, holds, in his opinion, for the nephridia of all Oligochseta ('84,
p. 123), as also for Polycha3ta and Hirudinea ('92, p. 357).

Vejdovsky's results are in agreement with those of Kowalevsky ('71)
and Boutchinsky ('81), and in part are confirmed by the work of several
other writers. It will be noticed that they do not depart widely from
Bergh ('88), except in the conditions at the peripheral end.

After comparing these various papers it seems to me that Vejdovsky's
descriptions and figures are more convincing of accuracy of observation
and reliability of interpretation than are those of the writers who adopt
one or the other of the extreme views ; and yet it is not permissible to
disregard the work of such able investigators as Bergh and Wilson. If
we except the positive statements of Bergh, which lack the corrobora-
tion of later investigators, we may conclude that the evidence goes far
to show that both mesoderm and ectoderm share, though unequallv, in
the development of the nephridium in Annelids, and therefore that, as
far as origin from germ layers is concerned, there is no insurmountable
obstacle to homologizing the nephridia of Chsetopod Worms with the
antennal glands of Macruran Crustacea.1

1 Postscript. — Bergh ('99) has repeated Vejdovsky's work on Rhynchelmis,
but disagrees with him in regard to the conditions and the interpretation to be put
on them: "Nach alledem meine ich meine urspriingliche These, dass Trichter-,
Sehlingen- und Endabschnitt bei den Oligochaten aus einer einheitlichen Anlage
hervorgehen, audi fiir Rhynchelmis festhalten zu miissen " (p. 446).

waite: antennal glands in homarus americanus. 195

Granting, then, that in the antennal glands of Macrura and in the
nephridia of Chaetopoda both ectoderm and mesoderm are involved, can
these organs be compared part for part '?

The chief point is to determine in each case the boundary between
the ectodermic and mesodermic constituents in the two sets of organs.
Where is this in the antennal gland of Macrura 1 In Homarus, as I
believe I have shown, it lies where the luruina of the endsac and of the
ectodermal sac (labyrinth) become confluent. Kingsley's results ('89,
pp. 29, 30, Figs. 61, 74) in Crangon seem to me to be capable of the same
interpretation, the mes-ectal line being at the junction of endsac and
ectodermal sac (canal). Boutchinsky's ('95) results on Gebia, since
they do not go beyond the earlier embryological stages, are not so
precise, but I find nothing in his descriptions or figures irreconcilable
with my conclusion, that the epithelium lining the endsac is also of
mesodermic origin, while that lining the labyrinth is derived from the
ectoderm.

There is difference of opinion as to the corresponding boundary in the
nephridia of Annelids. In the permanent nephridia of Rhynchelmis, as
well as all other Annelids, according to Vejdovsky ('84 and '92), the
lining of the efferent duct and of the contractile endsac is, as has been
said, ectoblastic, whereas all the glandular part and the nephrostome are
mesoblastic in origin. Wilson ('89, p. 425), as we have seen, concludes
that the nephrostome and the investing peritoneal sheath of the glandu-
lar portion are alone mesoblastic, whereas the epithelium lining the glan-
dular portion, the duct, and the end vesicle are ectoblastic, arising from
the " nephric cord." Finally, Bergh ('88, p. 230, '90, p. 501) concludes
that in Criodrilus and Lumbricus the entire organ is mesoblastic, there
being no ectoblastic constituent whatever. Thus according to Vejdovsky
the mes-ectal line is at the junction of the glandular region and the
efferent duct ; according to Wilson it is at the base of the nephrostome,
and according to Bergh the gland presents no such line.

It seems impossible to draw any satisfactory conclusion from this con-
flicting evidence. These differences of opinion are not solely explicable
upon the assumption of differences of methods in accomplishing the re-
sult in different worms, for all of these observers have worked and based
their conclusions in part iipon Lumbricus. The only point of complete
agreement is that the nephrostome is mesoblastic.

If it is impossible to determine so fundamental a question as where
mesoblast ends and ectoblast begins, it is idle to attempt more detailed
comparisons. In the present unsettled state of knowledge as regards the

196 bulletin: museum of compaeative zoology.

origin of the nephridia of Annelids, the most that can be said is, that
the endsac of the antennal gland in Crustacea may be homologous with
the nephrostome of the nephridium in Annelids, together with perhaps a
part or all of the nephridium peripheral to the base of the nephrostome.

In some points there is considerable difference between the nephro-
stomic end of a nephridium in an Annelid and the endsac of an antennal
gland in Macrura. The lumen of the former is usually in direct contin-
uity with the coelom, and is ciliated ; the latter is closed and non-ciliated.
After the work of Grobben ('79) on the shell gland, this question occa-
sioned considerable discussion, and it was claimed that these differences
must invalidate any proposed homology between the two organs ; but
later investigation has shown that the blood sinuses surrounding the
endsac of the antennal gland are not homologous with the nephrostomic
spaces in Annelids, and therefore the premise upon which this discussion
was based is destroyed.

Eisig ('87) holds that the nephridium of Annelids has two functions :
(1) the elimination of solid particles from the coelomic spaces, a duty
performed by the nephrostome and in which the cilia take an active
part; and (2) the excretion of soluble products from the blood and
coelomic fluid, a function exercised by the excretory cells. If this view
be correct, — and there are many things which support it, — the neph-
rostomic function is not directly represented in the autennal gland of
Crustacea, and we therefore cannot, expect close morphological resem-
blance. Hence it seems probable that the closed endsac, in place of an
open nephrostome, does not necessarily invalidate a general homology
between the two sets of organs.

B. The Number of Metameric Organs of this Nature in Crustacea.

Thus far I have considered the glands in Crustacea from one segment
only, — that of the second antenna. In Annelids the nephridia are re-
peated through a relatively large, although varying, number of segments.
Is such metameric repetition realized in Crustacea ?

Boutchinsky ('95, p. 170, Tab. VI. Fig. 145, Tab. VII. Figs. 157, 162)
has described in the first maxilla of Gebia the development of a meso-
dermic structure which takes on the form of a tubule closely resembling
the antennal gland in histological detail, inclusive of staining qualities.
Only the earlier stages of this organ are described, but from this evi-
dence the author thinks the organ is probably excretory, and that it
belongs to the same series as the antennal gland.

waite: antennal glands in homakus americanus. 197

Tlie shell gland of the second maxillary segment has long been con-
sidered as representing an Annelid nephridium, both from its resemblance
in function to that organ and from the fact that it is derived from both
ectoderm and mesoderm. This gland is widely distributed in Crustacea.
It is known as an adult organ, the shell gland, in many Eutomostraca
and some Malacostraca ; moreover, an organ similar in development aud
appearance is found during embryonic and larval life in the basal seg-
ment of the second maxilla of certain Malacostraca, although it disap-
pears before the adult stage is reached. As a larval organ the shell gland
has been described in the fifth segment of Palsemonetes (Allen, '93a) and
in some Schizopods and Decapods (Claus).

Thus in many Eutomostraca and Malacostraca glands closely re-
sembling one another are developed in both the second (antennal) aud
fifth (second maxillary) segments ; but in some cases — chiefly Euto-
mostraca — that of the fifth segment remains as an adult organ (shell
gland), while that of the second segment atrophies. On the other hand,
in most of the higher Malacostraca, it is the organ of the second seg-
ment (antennal gland) which persists in the adult, while the organ of
the fifth segment, if it appears at all, degenerates before adult charac-
ters are assumed.

According to the comparative anatomical work of A. Dohrn, Claus, and
Grobben, the shell gland is composed of the same parts as is the anten-
nal gland.

The development of the shell gland in the different families of Crus-
tacea has been treated of by several authors. It appears that the gland
arises from two sources ; the mesodermic part forms the' endsac, the
ectodermic portion the " canal." Hence it agrees in the main with the
development of the antennal glands. We may therefore conclude that
the antennal and shell glands in Crustacea are serially homologous
organs, being similar in structure and development.

Lebedinski ('89, p. 197, Tab. III. Figs. 79-81, '90, p. 184) describes in
Eriphia spinofrons the development of a " Segmentalorgan " as an evagi-
nation of the somatopleure, which becomes tubular and grows forward
until it ends blindly within the base of the first maxilliped. There arises
concurrently an ectodermic invagination in the wall of the coxopodite of
this appendage, which meets the evagination from the somatopleure, and
their lumina become continuous. We have, then, in this Crustacean a
gland in the sixth segment, which is apparently of the same series as the
antennal and shell glands.

The branchial glands of Crustacea were first noted in Crayfish, Crabs,
vol. xxxv. — no. 7. 4

198 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY.

and Paguridse by Cuenot ('87, p. xlv), but were considered as of a
lymphatic nature. In his later and more complete work Cuenot ('91,
pp. 76-80) describes these glands in several Decapods as found between
the two branchial blood channels (Crabs and Pagurus), or as closely
investing the walls of either the efferent vessels (Palsemon), or of
the afferent vessels. He still ascribes to them lymphatic functions, and
makes no mention of Kowalevsky's paper, which had appeared two years
before.

Kowalevsky ('89) studied these organs from a physiological point of
view. He found (pp. 35-42) that by intra vitam injection of a one per
cent solution of ammonium carminate into the body spaces of Astacus,
the endsac of the antennal gland soon became red. If indigo-carmine be
injected, the labyrinth becomes colored blue, but the endsac remains un-
colored ; and if a thorough mixture of these two pigments be used, the
endsac and labyrinth rigidly select the colors. Injection of tincture of
litmus gave reactions showing the endsac to be acid, the labyrinth in
part alkaline and in part neutral. Palsemon yielded almost identical
results, as far as regards the antennal glands, but in addition there were
produced by the injection of tinctui'e of litmus two red streaks in each
gill, one on each side of the shaft. There were eight double streaks on
each side of the body, viz. on the five gills and the appendages of the
three maxillipeds. This coloration was caused by rows of cells in which
were the same color appearances as in the cells of the endsac of the
antennal glands. Further, these cells acted toward the mixture of am-
monium carminate and indigo-carmine precisely as did the cells of the
endsac, from which it is to be inferred that there is in each gill tissue of
the same function as that of the endsac. These glandular regions are the
so called branchial glands.

The purely physiological work of Kowalevsky was supplemented by
Allen ('92, p. 79), who describes the structures of these glands in the
gills of Palsemonetes. They are spherical and composed of conical cells,
the apices of which border a central area. There are two varieties of
the gland cells, somewhat different in shape and in position. Allen ten-
tatively considers these spherical branchial glands as belonging to the
series of ectodermal glands. It should, however, be noticed that, ac-
cording to Cuenot and Kowalevsky these glands agree in function with
the mesodermic portion of the antennal glands.

The description and figures given by Allen strongly remind one of the
tegumental glands described by Herrick ('95, p. 125, Plates A and 49)
from the pleopods of the female Homarus, the function of which is the

waite: antennal glands in homarus amekicanus. 199

secretion of the glutinous material by which the eggs are fastened to
the hairs on the pleopods. Glands of the same general structure are of
wide distribution in the integument in Homarus and other Crustacea, as
well as in the alimentary tract. Laug ('89-94, p. 338) says that the
dermal glands of Crustacea take part in excretion.

It is still questionable whether the branchial glands are modified
tegumental glands, or are segmental and therefore belong to another
category. The physiological evidence points strongly to their being
nephridial organs, and this is strengthened by the fact that Cuenot
('94, p. 249) has found in the branchial glands chemical products
closely resembling those which, according to Marcbal ('92), are found in
the antennal glands. In histological structure, too, they resemble closely
the endsac of the antennal gland. They possess, however, no duct to the
exterior, and Cuenot suggests that the products elaborated in these
glands may, without chemical change, be carried by the blood to
the antennal glands there to be eliminated.

Boutchinsky ('95, p. 1G8, Tab. VII. Figs. 163, 164) has described
the development by invagination from the ectoderm of certain glands
whose ducts lead to the gill chamber. These lie in the gill cavity,
commonly in the dorsal wall, and he considers them as belonging in
the same series as the branchial glands. But even if these are to be
classed with branchial glands, they certainly are not, like the branchial
glands, homologous with the endsac of the antennal glands, but rather
with its labyrinth. I believe, however, that Boutchinsky has erred in
classing these with the branchial glands of Cuenot, Kowalevsky, and
Allen, for they seem to me to belong rather to the category of tegumen-
tal glands.

The position of the branchial glands in the axis of the gill, and in
close relation to the blood vessels, as described by Cuenot and Allen,
leads one to believe them to be mesodermic rather than ectodermic
structures, and with this their physiological activities are in harmony.
They have no duct as far as known, and there are several of them in a
single gill. If in each gill they represent a single segmental organ,
this organ must be in a diffuse condition.

So far as the meagre evidence goes, it shows, in my opinion, that the
branchial glands belong to the same category as the antennal and shell
glands, but that they represent only the mesodermic parts (endsac)
of these. It must, however, be confessed that the evidence for this
conclusion is not very satisfactory. A careful examination of the de-
velopment of the branchial glands will go far towai'd establishing

200 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY.

or invalidating this homology. I hope soon to undertake such a
study.

Kingsley ('89, p. 32, foot-note) has suggested that the genital outlets
represent part of a metamerically repeated series of ducts. The female
and male genital openings in Decapods, for example, occupy the bases
of the eleventh and thirteenth pairs of appendages respectively ; their
similar position is, in Kingsley's opinion, "inexplicable upon any other
ground than that the oviducts and vasa deferentia are themselves modi-
fications of pre-existing metameric organs, and the only organs in the
Annelids which would answer the requirements of the case are the
nephridia."

The first part of his proposition is strengthened by evidence given by
Eateson ('94, pp. 152-155), who found that in over three per cent of
the females of Astacus fluviatilis examined, there appeared supernumer-
ary oviducal openings. These were either bilaterally symmetrical, or
on one side only, and were not necessarily situated in the segment
adjacent to the normal opening, but were sometimes removed to the
thirteenth segment. Dissection showed that, in cases where such su-
pernumerary openings appeared, the oviduct was branched and com-
municated with all the openings, which were pi'esumably functional.
Oviducal openings wei*e found on the eleventh (normal), twelfth, and
thirteenth segments, but no such repetition of genital openings was
noted in males. It is probable, then, that Kingsley's proposition that
the genital ducts are members of a metameric series is valid ; but
whether or not they are homologous with the nephridia of Annelids is
quite another question. There are in Annelids other metameric organs
of somewhat similar position, of which these ducts in Crustacea might
be representatives, e. g. the setigerous glands or the dorsal pores. The
fact that in Annelids the genital products are in many cases carried to
the exterior by the nephridia, or by modified nephi'idial ducts, lends
support to Kingsley's contention ; but we cannot rely solely upon the
evidence of analogy from which to draw conclusions as to homology.

The evidence on the ontogeny of the genital ducts and their open-
ings in Crustacea is not very complete, but in the main it points to
their peripheral portions at least as being ectodermic invaginations
entirely distinct from the primary reproductive organs. If this be
confirmed, it will add support to the view that they are homologous
with the ectodermic portion of the nephridium of Annelids ; but it
must be remembered that the setigerous glands of Annelids also have
a similar origin.

waite: antennal glands in homarus americanus. 201

It seems possible that these ducts may hereafter be shown to repre-
sent the ectodermic part of nephridia, while the branchial glands repre-
sent the mesodermic portion, in which event the homologs of the entire
nephridium would exist in the eleventh and thirteenth segments in
Decapods, though the ectodermic and mesodermic portions do not come
into conjunction as in Anuelids ; and though, further, the mesodermic
constituent is separated into several masses. Such a scheme must
remain purely hypothetical until the conditions can be thoroughly
studied from the standpoints of both development and comparative
anatomy.

To sum up the foregoing discussions, I believe that there is suffi-
ciently good evidence that representatives of the nephridium of Anne-
lids are found in Crustacea in the somites bearing the second antenna
(antennal gland), the second maxilla (shell gland of Entomostraca and
some Malacostraca), and the first maxillipeds (" Segmentalorgan " of
Lebedinski) ; and, further, that there is partial evidence of such repre-
sentatives in the somites of the first maxilla (Boutchinsky), and in the
eight thoracic somites of the maxillipeds and pereiopods of Malacostraca,
namely, the branchial glands which exist in all these somites, and the
genital ducts which are found in part of them.

Summary.

1. The antennal gland consists of three parts : gland proper, vesicle,
and duct.

2. The gland proper consists of endsac (dorsal) and labyrinth (ven-
tral), whose lumina are connected by a single small orifice, which is in
the lateral anterior part of the organ.

3. The lumen of the endsac is connected with the exterior only in-
directly, by way of the labyrinth.

4. The labyrinth leads into the duct by a series of converging cana-
liculi, which open by separate pores through the dorso-median wall of the
duct. These canaliculi form the "white lobe," which projects ventrad
from the median anterior region of the labyrinth.

5. The duct is short and opens through a tubercle on the ventral side
of the coxopodite of the second antenna. The opening is guarded by an
operculum.

6. The vesicle is a dorsal diverticulum from the duct, and has no
direct connection with the gland proper.

202 BULLETIN : MUSEUM OF COMPAEATIVE ZOOLOGY.

7. The blood supply of the organ is from the antennal and sternal
arteries.

8. The nerve supply is from the antennal nerve.

9. The endsac encloses a single chamber with numerous radiating out-
pocketings ; the labyrinth is a complicated mass of anastomosing tubules,
which are short and of varying calibre.

10. The epithelial cells of endsac and labyrinth are of distinctly differ-
ent types, and there are no cells of an intermediate nature.

11. The cells of both endsac and labyrinth take pai't in the secretion.
This is in large measure given off in the form of globular vesicles con-
taining an irregular granular mass or clot ; these are constricted off from
the free ends of the cells.

12. The wall of the vesicle is made up of a single layer of epithelium
without folds, surrounded, except on the ventral side, by a muscular
sheath, which is variable in thickness. It is presumably by the con-
traction of this sheath that the secretion is forced out of the external
orifice of the duct. Between the epithelial and muscular layers there is
a vascular layer.

13. The endsac arises from the mesoderm occupying the axis of the
second antenna, when the embryo is 15 to 17 days old. At first there
are but one or two cells differentiated.

14. By nuclear division without corresponding formation of cell walls
there is formed a solid multinuclear body, — a syncytium.

15. Vacuoles appear in the syncytium and probably by confluence form
a single large vacuole.

16. The belated cell walls form around this vacuole, so that the vacuole
becomes an intercellular space, — the lumen of the endsac.

17. The ectodermic ingrowth — which in the adult becomes the laby-
rinth — occurs on the median ventral face of the antenna, and appears
when the embryo is 28 to 30 days old.

18. This ingrowth passes, first, proximad and then around the anterior
side of the endsac ; by separation of its cells a lumen is formed near its
deep end.

19. The separation of cells progresses rapidly toward the exterior,
which is reached by the end of the sixth week. The lumen thus formed
becomes the permanent duct and the lumen of the labyrinth.

20. The lumina of the endsac (mesodermic) and of the ectodermic
sac (labyrinth) do not become continuous until the embryo is 273 to
303 days old.

waite: antennal glands in homarus americantjs. 203

21. The adjacent walls of the two sacs break through at a single
point. There is consequently an abrupt transition — marked by the
mes-ectal line — from the mesodermal to the ectodermal part of the
organ, the epithelial cells of the two parts being from the beginning of
different types. This boundary persists in the adult.

22. The organ probably does not secrete until after hatching.

23. At hatching the lumen of the ectodermic sac increases greatly in
size.

24. In the first larva the endsac is for the most part ventral to the
ectodermic sac. The duct leads from the ventral region of the ecto-
dermic sac to the exterior.

25. At this stage the cells of both endsac and ectodermic sac (laby-
rinth) show evidence of secretory activity. The lumen of the ectodermic
sac alone contains a granular coagulum.

26. The epithelial cells of the ectodermic sac in the first larva have a
vertically striate cuticuloid layer on the free end. Such a structure is
not present in any cells of the endsac.

27. At this stage the epithelium of the duct differs from that of the
ectodermic sac in having neither a striated cuticuloid layer nor striated
cytoplasm, but there is a gradual transition from one kind of epithelium
to the other.

28. The gland of the second larva differs from that of the first chiefly
in size, which increases proportionally to the increase in the size of
the larva.

29. The development of evaginations from the ectodermic sac (laby-
rinth) begins in the third larva. In later larvae these evaginations in-
crease in number and size, and form anastomoses. Thus the complications
of the adult labyrinth result from evagination and anastomosis, not from
the coiling of a single tubule.

30. Secretion by globules (as in the adult) begins in the third larva
and increases in amount in older larvae.

31. The vesicle first appears in the third larva, as a large, open dorsal
evagination from the anterior portion of the ectodermic sac. In older
larvae it increases rapidly in volume and grows caudad.

32. During larval life the chief growth of endsac and ectodermic sac
are in different directions. The result is that the two parts of the ^land
revolve around an approximately antero-posterior axis, so that the endsac
comes to lie, as in the adult, dorsal to the ectodermic sac.

b

204 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY.

33. The conflicting evidence upon the development of the nephridium
in Annelids makes it impossible closely to homologize this structure and
the antennal glands of Macrura. The endsac may be homologous with
the nephrostome of the nephridium of Annelids together with (or) none,
(b) a part, or (c) all of the remainder of the organ, depending upon
which is the correct view in regard to the development of the nephridium
in Annelids.

34. The nephridium of Annelids is probably represented in Crustacea
in the second (antennal) segment by the antennal gland of Malacostraca ;
in the fifth (second maxillary) segment by the shell gland of Entomos-
traca and some Malacostraca ; in the sixth (first maxillipedal) segment
of some Malacostraca by the " Segmentalorgan " of Lebedinski ; it is
possibly represented in the fourth (first maxillary) segment by the ex-
cretory organ described by Boutchinsky, and in the sixth to thirteenth
(maxillipedal and pereiopodal) segments in part by the branchial glands,
and in part (in the eleventh and thirteenth segments^ by the genital
ducts.

Mat 1, 1898.

WAITE : ANTENNAL GLANDS IN HOMARUS AMERICANUS. 205

BIBLIOGRAPHY.

Allen, E. J.
'92. On the Minute Structure of the Gills of Palsemonetes varians. Quart.
Jour. Micr. Sci., Vol. 34, pp. 75-84, PI. 10.

Allen, E. J.

'93. Preliminary Account of the Nephridia and Body Cavity of the Larva of
Palsemonetes varians. Proc. Roy. Soc, London, Vol. 52, pp. 338-342.

Allen, E. J.

'93\ Nephridia and Body-Cavity of some Decapod Crustacea. Quart. Jour.
Micr. Sci., Vol. 34, pp. 403-426, Pis. 36-38.

Bateson, W.
'94. Materials for the Study of Variation, pp. xvi + 598. London.

Bergh, R. S.
'88. Zur Bildungsgeschichte der Exkretionsorgane hei Criodrilus. Arbeit.
Zool.-Zoot. Inst. Wiirzburg, Bd. 8, pp. 223-248, Taf. 13, 14.

Bergh, R. S.
'90. Neue Beitr'age zur Embryologie der Anneliden. I. Zur Entwicklung
und Differenzirung des Keimstreifens von Lumbricus. Zeitschr. f. wiss.
Zool., Bd. 50, pp. 469-526, Taf. 19-21.

Bergh, R. S.
'99. Nochmals iiber die Entwicklung der Segmentalorgane. Zeitschr. f. wiss.
Zool., Bd. 66, pp. 435-449, Taf. 30.

Boutchinsky, P.
'81. Kt> Bonpocy o pa3Bimft jjoawBaro lepBAKa (Lumbricus terrestris). [On
the Question of the Development of the Earthworm (Lumbricus terrestris).
3anucKH HoBopocc. 06mecTBa EcTecTBoncnHTaiejefi. [Mem. Soc. Nat.
New-Russia.] Tome 7, Pt. 2, pp. 11-71, Tab. 1-3.

Boutchinsky, P.
'94. Zur Entwicklungsgeschichte von Gebia littoralis. Zool. Anz., Jahrg. 17,
pp. 253-256.

Boutchinsky, P.
'95. Hafijuojuema Hajri 3M6pioHajiBHHMT> pa3BirrieMi> Malacostraca. [Ob-
servations on the Embryonic Development of Malacostraca.] 3airacRH
HoBopocc. 06m,ecTEa EcTecTBoncnHTaTeaeft. Mem. Soc. Nat. New-Rus-
sia.] Tome 19, Pt. 2, pp. 1-216, Tab. 1-7.

206 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY.

Brandt, J. F., und Ratzeburg, J. T. C.

'33. Mediziniscbe Zoologie. 2 Bde, 4to, Berlin. Der Flusskrebs. Bd. 2,
pp. 58-70, Taf. 10, 11.

Bumpus, H. C.

'91. The Embryology of the American Lobster. Jour. Morph., Vol. 5, pp.
215-262, Pis. 14-19.

Butschinsky, P. See Boutchinsky, P.

Claus, C.

'63. Ueber einige Schizopoden und niedere Malacostraken Messina's.
Zeitschr. f. wiss. Zool., Bd. 13, pp. 422-454, Taf. 25-29.
Cu6not, L.

'87. Etudes sur le sang, son role et sa formation dans la serie animale.
Partie 2, Invertebres. Note prelimiuaire. Arch, de Zool. exper. et gen.,
ser. 2, Tome 5, Notes et Revue, pp. xliii-xlvii.

Cuenot, L.

'91. Etudes sur le sang et les glandes lymphatiques dans la serie animale.
Partie 2, Invertebres. Arch, de Zool. exper. et gen., ser. 2, Tome 9, pp.
13-90, 365-475, 593-670, Pis. 1-4, 15-18, 23.

Cue"not, L.
'94. Etudes physiologiques sur les Crustaces Decapodes. Arch, de Biol.,
Tome 13, pp. 245-303, Pis. 11-13.

Dohrn, A.
'70. Untersuchungen iiber Bau und Entwicklung der Arthropoden. 6. Zur
Eutwicklungsgeschichte der Panzerkrebse (DecapodaLoricata). Zeitschr.
f. wiss. Zool., Bd. 20, pp. 248-271, Taf. 16.
Dohrn, H.
'61. Analecta ad historiam naturalem Astaci fluviatibs. Inaug.-Dissert., 8vo,
28 pp., Berlioni.

Eisig, H.
'87. Monographic der Capitelliden des Golfes von Neapel. Fauna und Flora
des Golfes von Neapel, Mem. 16, pp. xxvii + 906, Taf. 1-37. 4to,
Berlin.

Gerstaecker, A.

'95. Arthropoda : Besondere Driisen. Bronn's Klassen und Ordnungen des
Thierreiches, Bd. 5, Abth. 2, Lief. 41-43, pp. 982-1010, Taf. 97, 98.

Griffiths, A. B.
'85. On the Extraction of Uric Acid Crystals from the Green Gland of
Astacus fiuviatilis. Proc. Roy. Soc. London, Vol. 38, pp. 187, 188.

Grobben, C.

'79. Die Entwicklungsgeschichte der Moina reetirostris. Arbeit. Zool. Inst.
Wien, Bd. 2, pp. 203-268, Taf. 11-17.
Also separate. 65 pp., 7 Taf., Wien, 1879.

waite: antennal glands in homarus ameeicanus. 207

Grobben, C.
"80. Die Antermendriise der Crustaceen. Arbeit. Zool. Inst. Wien, Bd. 3,
pp. 93-110, Taf. 9.

Herrick, F. H.
'95. The American Lobster : A Study of its Habits and Development. Bull.
U. S. Fish Coram, for 1895, pp. 1-252, Pis. A-J and 1-54.

Ishikawa, C.

'85. On the Development of a Freshwater Macrurous Crustacean, Atyephira
compressa, De Haan. Quart. Jour. Micr. Sci., Vol. 25, pp. 391-42S, Pis.
25-28.

Kingsley, J. S.
'89. The Development of Crangon vulgaris. Third Paper. Bull. Essex
Inst., Vol. 21, pp. 1-42, Pis. 1-3.

Kirch, J. B.
'86. Das Glykogen in den Geweben des Flusskrebses. Inaug. -Dissert.
8vo, Bonn.

Kowalevski, A.
'71. Embryologische Studien an Wurmern und Arthropoden. Mem. Acad.
Imp. Sci. St. Petersbourg, ser. 7, Tome 16, No. 12, 70 pp., 12 Taf.

Kowalevsky, A.

'89. Ein Beitrag zur Kenntnis der Exkretionsorgane. Biol. Centralbl., Bd.
9, pp. 33-47, 65-76, 127, 128.

Lang, A.

'89-94. Lehrbuch der vergleichenden Anatomie der wirbellosen Thiere.
xvi + iv + 1197 pp., Jena, 1888-94.

Lebedinsky, J.
'89. Ha6jno,neHia Hajrt pesBHTiean. KaiaeHHaro Kpa6a (Eryphia spinifrons).
[Observations on the Development of the Crab (Eryphia spinifrons).]
3anHCKH HoBopocc. OoniecTBa EcTecTBoncnHTaiejieR. [Mem. Soc. Nat.
New-Russia.] Tome 14, Pt. 2, pp. 131-200, Tab. 1-3.

Lebedinski, J.

'90. Eiuige Untersuchungen iiber die Entwicklungsgeschichte der Seekrab-
ben. Biol. Centralbl., Bd. 10, pp. 17S-185.

Lebedinsky, J.
'91. Die Eutwicklung der Daphnia aus dem Sommereie. Zool. Anz., Jahrg.
14, pp. 149-152.

Lebedinski, J.
'92. Hafijirojemfl najrt pa3BHTieMt KaMeHHaro Kpa6a (CpaBHHTejBHaa n
TeopeTn^ecKaa ladt). [Observations on the Development of the Crab
(Eryphia spinifrons). Comparative and Theoretical Part.] 3airacKH
HoBopocc. 06mecTBa EcTecTBoncnuTaTeJiefi. [Mem. Soc. Nat. New-Rus-
sia.] Tome 17, Pt. 1, pp. 137-256.

20S bulletin: museum of comparative zoology.

Lehmann, O.

'87. Beitrage zur Frage von der Homologie der Segmentalorgane und Aus-
fiihrgange der Gesclilectsprodukte bei den Oligochaeten. Jena. Zeitschr.,
Bd. 21, pp. 322-360, Taf. 20.

Lereboullet, A.
'62. Recherches d'embryologie comparee sur le developpement du brochet,
de la perche, et de l'ecrevisse. Mem. Acad. Sci. Inst. France (Savants
etrangers), Tome 17, pp. 447-802, Pis. 1-6.

Leydig, F.

'60. Naturgeschichte der Daphniden (Crustacea Cladocera). pp. iv -+- 252,
10 Taf., Tubingen.

Marchal, P.
'92. Recherches anatomiques et physiologiques sur l'appareil excreteur des
crustaces decapodes. Arch, de Zool. exper. et gen., ser. 2, Tome 10, pp.
57-275, Pis. 1-9.

Rathke, H.

'29. Uutersuchungen iiber die Bildung und Eutwicklung des Flusskrebses.
viii — {— 98 pp., 5 Tab., 4to, Leipzig.

Reichenbach, H.

'77. Die Embryonalanlage und erste Eutwicklung des Flusskrebses. Zeitschr.
f. wiss. Zool., Bd. 29, pp. 123-196, Taf. 10-12.

Reichenbach, H.
'86. Studien zur Entwicklungsgeschichte des Flusskrebses. Abhandl. Senck-
enberg. Naturf. Gesellschaft, Bd. 14, Heft 1, pp. 1-137, Taf. 1-14 [19
Taf.].

Rosel von Rosenhof, A. J.
1755. Der Fluskrebs hiesiges Landes, mit seinen merkwurdigen Eigen-
schafften. In Der monathlich-herausgegebenen Insecten-Belustigung,
Theil 3, pp. 305-350, Suppl. Tab. 54-61, Niirnberg.

Semper, C.
'61. Reisebericht (Fortsetzungj [aus Manila]. Zeitschr. f. wiss. Zool., Bd.
11, pp. 100-108.

Szigethy, K.

'85. Anatomie, Histologie und Physiologie der griinen Driise des Astacus
fluviatilis. Iilrtekezesek a Termeszettudomanyi osztaly korebol [Trans.
Dept. Nat. Sci.'], Bd. 14, 2 Taf. 1884.

Abstract in Math. u. Naturwiss. Berichte aus Ungarn, Bd. 3, pp.
108, 109.

Vejdovsky, F.
'84. System und Morphologie der Oligochaeten. 172 pp., 16 Taf., Folio,
Prag.

WAITE: ANTENNAL GLANDS IN HOMAKUS AMERICANUS. 209

Vejdovesky, F.

'92. Eutwicklungsgeschichtliche Untersuchungen. Hefte 3 u. 4, vi -4- 491 pp.,
32 Taf., 8vo, Prag.

Waite, F. C.
'98. Structure aud Development of the Autennal Glands in Homarus ameri-
canus (Abstract). Science, Vol. 8, pp. 178,179.

Weldon, W. F. R.
'89. The Ccelom and Nephridia of Palaemon serratus. Jour. Mar. Biol.
Assoc, N. S., Vol. 1, pp. 162-168, Pis. 13-15.

Weldon, W. F. R.
'91. The Renal Organs of Certain Decapod Crustacea. Quart. Jour. Micr.
Sci., Vol. 32, pp. 279-291, Pis. 21, 22.

Whitman, C. O.
'87. A Contribution to the History of the Germ-Layers in Clepsine. Jour.
Morpb., Vol. 1, pp. 105-182, Pis. 4-6.

Will, F., und v. Gorup-Besanez, E.
'48. Guaniu ein wesentlicber Bestandtheil gewissen Secrete wbelloser
Thiere. Gelehrte Anz. d. kou. bayer. Akad. Wiss., Bd. 27, No. 233, col.
825-S28.

Wilson, E. B.
'87. The Germ-Bands of Lumbricus. Jour. Morph., Vol. 1, pp. 183-192,
PI. 7.

Wilson, E. B.
'89. The Embryology of the Earthworm. Jour. Morph., Vol. 3, pp. 387-
462, Pis. 16-22.

vol. xxxv. — No. 7.

Ip

210

BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY.

EXPLANATION OF PLATES.

All figures were outlined with the camera lucida.

The orientation of figures is given for each plate separately.

ABBREVIATIONS.

art. sac.

Sacculary artery.

lob. m-a.

Median anterior lobe.

at. 1.

First antenna.

lob. p.

Posterior lobe.

at. 2.

Second antenna.

lu.

Lumen.

cl. Iby.

Cells of the labyrinth.

tnarg.

Border.

cl. sac. trm.

Cells of the endsac.

mb. ba.

Basement membrane.

cad.

Body cavity.

md.

Mandible.

cry.

Crystals.

ms'drm.

Mesoderm.

eta.

Cuticula.

mu.

Muscle.

drm.

Dermis.

mx. 1.

First maxilla.

ee'drm.

Ectoderm.

oc.

Eye.

evag.

Evagination.

of. ex.

External orifice.

fas. ax.

Axillar bundle.

of. i.

Internal orifice.

fos.

Depression.

op-

Operculum.

gib.

Globule.

par, cl.

Cell wall.

gl. e'drm.

Tegumental gland.

par. vs.

Wall of vesicle.

gn. opt.

Optic ganglion.

sac. ee'drm.

Ectodermic sac.

gn. su'ee.

Supercesophageal ganglion.

sac. trm.

Endsac.

hi.

Hilum.

set. sns.

Sensory hairs.

i'vag. ee'drm.

Ectodermic invagination.

tis. con't.

Connective tissue.

lac. sng.

Blood lacuna?.

vac.

Vacuole.

la. ex.

Non-pigmented calcified

vas. sng.

Blood vessel.

layer of the shell.

vlv.

Valve-like fold.

la. ex'.

Pigmented calcified layer

a.

In Fig. 2, marks junction

of the shell.

of walls of endsac and

la. e'th.

Epithelial layer.

labyrinth ; in Figs. 47,49,

la. mu.

Muscular layer.

marks region of commu-

Iby.

Labyrinth.

nication between lumina

lig.

Ligament.

of endsac and ectoder-

In. mes-ec.

Mes-ectal line.

mic sac.

lob. l-a.

Lateral anterior lobe.

Waits. — Antennal Glands.

PLATE 1.

(In Fig. 1 anterior is up on the plate ; in Fig. 2 the dorsal edge of section is up ;
in Fig. 14 the ventral edge of section is down, the posterior edge to the right.)

Fig. 1. Dorsal view of right gland of adult with greater part of vesicle removed.

X 2. (Drawn by Mr. K. Hayashi.)
Fig. 2. Anterior face of a transverse section through left gland of adult about

| mm. posterior to hilus. X 14.
Fig. 3. Section of a fold in the floor of the endsac, together with adjacent walls

of labyrinth tubule. X 350.
Fig. 4 Cells, showing globules, from wall of endsac. X 410.
Fig. 5. Cells, containing crystals (see Fig. 6), from wall of endsac. X 410.
Fig. 6. Crystals (probably artifacts, see p. 164 of text), from cells of endsac wall.

X410.
Fig. 7. Section of fold in wall of a labyrinth tubule showing formation of globules.

X410.
Fig. 8. Section of wall of labyrinth tubule showing striate cuticula of epithelial

cells. X 410.
Fig. 0. Surface view of epithelium from posterior region of vesicle. X 410.
Fig. 10. Transverse section of wall of vesicle through a region where muscular

layer is thick. The muscle cells are cut obliquely. X 410.
Fig. 11. Transverse section of wall of vesicle through a region where muscular

layer is thin. X 410.
Fig. 12. A single muscle cell teased from wall of vesicle. X 410.
Fig. 13. Region of nucleus from a cell similar to that of Figure 12. X 900.
Fig. 14. Lateral face of a proximo-distal section through left second antenna in

the region of the tubercle on which the duct opens. X 27.

Waite.-Antennal Gland ;

1_

i as.snq

par. i -s

or/ .

Iby. ..
I

■

lob. ma.
art sac.

nana .

santrm ..

../a.

lob.hn.

s

<t

mi fro

lob.p.

O

par v.v

■ 1

S i

i I ! :. 1 |

/2

<^ i«

$ •>■ •:-

'} Jama.

:

> .9

—fas.t

- %

<?1

las -OX.

/r/ / •//?

'/>?/,

v *c±

■ la ex

ht'.v

■

i

i' t 'Urm
SItS

' <iJ_

Waite. — Antennal Glauds.

PLATE 2

(In Figs. 15-17 anterior is up on the plate ; in Figs. 18-24 (except 23) dorsal is
up ; in Fig. 23 lateral is up; in Figs. 18, 20, 22,24, anterior is to the right ; in Figs.
19, 21, 23, anterior is to the left ; in Fig. 25 mediad is to the right. In Figs. 18-25
the mesoderm is distinguished by the darker shade.)

Fig. 15. Ventral view of an embryo 15-17 days after extrusion of egg. X 72.
Fig. 16. Similar view of an embryo 16-18 days after extrusion of egg (stage at

which gland first appears). X 72.
Fig. 17. Similar view of an embryo 23-25 days after extrusion of egg. X 72.
Fig. 18. Part of parasagittal section through second antenna of embryo 15-17 days

old, showing undifferentiated mesoderm in axis of antenna. X 410.
Fig. 19. Similar section of embryo 16 to 18 days old (Fig. 16), showing (in the

mesoderm) first differentiation of endsac with two nuclei. X 410.
Figs. 20-22. Parasagittal sections through antenna at slightly more advanced

stages. (Fig. 20 is youngest, Fig. 22 oldest.) X 410.
Fig. 23. Section inclined slightly to the frontal plane at a stage when the endsac

is ovate in form and contains in all 12 nuclei. X 410.
Fig. 24. Parasagittal section through antenna in which a vacuole appears in the

endsac. X 410.
Fig. 25. Anterior face of a transverse section through right antenna at stage when

invagination of ectoderm first begins. (Embryo 28-30 days old.)

X 410.

Waite.-A Glands.

'd

/6.

17

at J.

.at J

a ;.

-oc.

Waite. — Antennal Glands.

PLATE 3.

(Dorsal is up on the plate in all figures. In Figs. 26, 27, 28, 30, anterior is to the
right ; in Figs. 29, 31, median is to the right. Mesoderm is distinguished by the
darker shade.)

Fig. 26. Lateral face of a parasagittal section through left antenna of embryo,
showing vacuolation in the endsac and the path of the ectodermic
invagination. The invaginated cells in Figs. 26, 27, are more deeply
shaded than the rest of the ectoderm. X 410.

Fig. 27. Similar section in slightly older embryo, showing expansion at deep
end of ectodermic invagination, and a lumen formed in the endsac.
X410.

Fig. 28. Similar section in embryo about 35 days old, showing position of future
lumen of ectodermic sac. X 410.

Fig. 29. Anterior face of transverse section through right antenna of an embryo
about 40 days old, showing lumina of endsac and ectodermic sac and
duct to exterior. X 410.

Fig. 30. Lateral face of a parasagittal section through left antenna of embryo 60-70
days old, showing growth of ectodermic invagination around the end-
sac. X 410.

Fig. 31. Anterior face of transverse section through right antenna of embryo 3 to 4
months old, showing extension of ectodermic invagination. X 410.

^TE 3.

(/nil

i ,,ry,,r/rni.

sar./rm

■

•

tus'drm

lig

drm

28.

Waits. — Autcuual Glauda.

PLATE 4.

(Dorsal is up on plate in all figures. In Tigs. 32-38 anterior is to the left ; in
Fig. 39 anterior is to the right.)

Fig. 32-38. Lateral face of the second, third, fifth, sixth, eighth, eleventh, and
twelfth sections respectively from a series of thirteen parasagittal
sections through the left antennal gland of* the first larva. X 128.

Fig. 39. Median face of a nearly parasagittal section through the left gland and
duct of the first larva. X 128.

Watte.- Anteh

gn.suoc

sin tnn .

i"t -...

surjnii

36

sac.a

37.

qn.siVoe.

"'( ilrtn.

38

qn.su or.

19

,11, SUOt

vc'drm

Uq.

sac <c' drill .

Waits. — Antewial Glands.

PLATE 5.

(Dorsal is up on all figures. Median is to the right in Figs. 41-48.)
Fig. 40. Anterior face of a transverse section through the region of the base of thc
second antennae in the first larva. (The dorso-ventral axis has been
slightly shortened in cutting.) x 42.
Fig. 41-48. Anterior face of the third, fifth, seventh, ninth, tenth, twelfth four-
teenth, and seventeenth sections respectively from a series through the
right antennal gland of the first larva. X 128.

e

Wait:

grt-opt

Waite. — Antennal Glands.

PLATE 6.

(In Fig. 49 anterior is up, lateral to the right; in Figs. 50, 51, 60, dorsal is up;
in Figs. 50, 60, anterior is to the right ; in Fig. 51 anterior is to the left.)

Fig. 49. Dorsal face of a frontal section (the tenth of a series of eighteen) through
the right antennal gland. X 128.

Fig. 50. Lateral face of a parasagittal section through right gland and duct in the
third larva. X 128.

Fig. 51. Parasagittal section through the orifice connecting the lumina of endsac
and ectodermic sac in the first larva. X 350.

Fig. 52. Cell from section of wall of endsac of the first larva in a region not adja-
cent to the ectodermic sac. X 900.

Fig. 53. Cells from section of wall of endsac in the first larva in a region adjacent
to the ectodermic sac. X 900.

Fig. 54. Cells from section of dorsolateral wall of ectodermic sac in first larva.
X 900.

Fig. 55. Cells from section of ventral wall of ectodermic sac in first larva. X 900.

Fig. 56. Cells from section of dorso-median wall of ectodermic sac in first larva.
X 900.

Fig. 57. Cells from section of dorsal wall of ectodermic sac in first larva, one show-
ing mitotic figures. X 900.

Fig. 58. Cell from section of dorsal wall of ectodermic sac in first larva. X 1600.

Fig. 59. Blood corpuscle from blood lacunre of antennal gland in first larva.
X 900.

Fig. 60. Longitudinal section of duct of the antennal gland in first larva. X 350.

Wait1

6,

at
eta.

■

•JSi

nib bit

i

uilt&S&ii -

lac. snff.
sac. a drm

•to.-

-

■

• &"•

'Hi

m **&.

coel.

5

•

■

%

- •

cn'ei

-

622

"*3 •

^

s

-

The following Reports have been published oh are in prepa-
ration on the Dredging Operations off the West Coast of
Central America to the Galapagos, to the West Coast of
Mkxico, and in the Gulf of California, in Charge of Alex-
ander Agassiz, carried on by the U. S. Fish Commission
Steamer " Albatross," during 1891, Lieut. Commander Z. L.
Tanner, U. S. N., Commanding.

A. AGASSIZ. II.1 General Sketch of the
Expedition of the "Albatross," from
February to May, 1891.

A. AGASSIZ. The Pelagic Fauna.

A. AGASSIZ. The Deep-Sea Pananiic Fauna.

A. AGASSIZ. I.2 On Calainocrinus, a new
Stalked Crinoid from the Galapagos.

A. AGASSIZ. XXIII.23 The Echini.

JAS. E. BENEDICT. The Annelids.

R. BERGH. XIII.13 The Nudibranchs.

K. BRANDT. The Sagittse.

K. BRANDT. The Thalassicolae.

C. CHUN. The Siphonophores.

C. CHUN. The Eyes of Deep-Sea Crustacea.

S. F. CLARKE. XI.11 The Hydroids.

W. H. DALL. The Mollusks.

W. FAXON. VI.3 XV.'S The Stalk-eyed
Crustacea.

S. GARMAN. The Fishes.

W. GIESBRECHT. XVI.'S The Copepods.

A. GOES. III.4 XX. =o The Foraminifera.

H. J. HANSEN. XXII.22 The Cirripeds
and Isopods.

C. HARTLAUB. XVIII.'8 The Comatulse.

W. A. HERDMAN. The Ascidians.

S. J. HICKSON. The Antipathids.

W. E. HOYLE. The Cephalopods.

G. von KOCH. The Deep-Sea Corals.

C. A. KOFOID. Solenogaster.

R. von LENDENFELD. The Phosphores-
cent Organs of Fishes.

IV.6 XII.14 The Holothu-

H. LUDWIG.

rians.
C. F. LUTKEN and TH. M0RTENSEN.26

The Ophiuridse.
OTTO MAAS. XXI.21 The Acalephs.
J. P. McMURRICH. The Actinarians.
E. L. MARK. XXIV.24 The Cerianthidas.
GEO. P. MERRILL. V.a The Rocks of the

Galapagos.
G. W. MULLER. XIX.19 The Ostracods.
JOHN MURRAY. The Bottom Specimens.
A. ORTMANN. XIV. « The Pelagic Schi-

zopods.
ROBERT RIDGWAY. The Alcoholic Birds.
P. SCHIEMENZ. The Pteropods and Het-

eropods.
W. SCHIMKEWITSCH. VIII.8 The Pyc-

nogonidae.
S. H. SCUDDER. VII.7 The Orthopteraof

the Galapagos.
W. PERCY SLADEN. The Starfishes.
L. STEJNEGER. The Reptiles.
TH. STUDER. X.10 The Alcyonarians.
C. H. TOWNSEND. XVII.17 The Birds of

Cocos Island.
M. P. A. TRAUTSTEDT. The Salpidse and

Doliolidse.
E. P. VAN DUZEE. The Halobatidse.
H. B. WARD. The Sipunculoids.
H. V. WILSON. The Sponges.
W. McM. WOODWORTH. IX.9 The Pla-

narians and Nemerteans.
W. McM. WOODWORTH. XXVII.27

Planktonemertes.

89 pp

Bull M.C

, 22 Plates.

Mem. M. C.

Bull. M. C.

Bull. M. C.

Bull. M. C.

Bull M. C.

Bull. M. C.

Bull. M. 0.

Bull M. 0.

Bull. M 0.

Bull. M. C

Bull. M. C.

Bull. M. C

Mem. M. C.

Bull M. C

Mem. M. C.

Bull. M. C.

Bull. M. C.

Bull M C.

Bull. M. C.

Mem. M C

Bull. M. ti

Bull. M. C

Bull. M C

Mem M C

Bull. M. C

Z., Vol. XXT., No. 4, June, 1891, 16 pp. ; and Vol. XXIII., No 1, February, 1892,

Z., Vol.
Z., Vol.
Z., Vol.
Z., Vol.
Z., Vol.
Z.,Vol.
Z . Vol.
Z., Vol.
Z., Vol.
Z., Vol.
Z., Vol.
Z., Vol

Z., Vol.

Z , Vol.

Z ., Vol.

Z. , Vol.

Z., Vol

Z , Vol.

Z., Vol.
. Z , Vol.

Z , Vol
. Z., Vol
. 7, , Vol.

Z , Vol.
Z., Vol.

XVII., No. 2, January, 1892, 95 pp., 32 Plates.

XXIV., No. 7, August, 1893, 72 pp.

XXIII., No. 5, December, 1892, 4 pp., 1 Plate.

XXIV.. No 4, June, 1893, 10 pp [Zool. Anzeig., No. 420, 1893.]

XVI., Mo. 13, July, 1893, 3 pp

XXV , No. 1 September, 1893, 25 pp.

XXV., No. 2, December, 1893, 17 pp., 2 Plates.

XXV., No. 4, January, 1894, 4 pp., 1 Plate.

XXV., No. 5, February, 1894, 17 pp.

XXV., No. 6, February, 1894, 7 pp., 5 Plates.

XXV., No. 8, September, 1894, 13 pp., 1 Plate.

XXV., No. 10, October, 1894, 109 pp., 12 Plates.

XVII., No. 3, October, 1894, 183 pp., 19 Plates.

XXV, No. 12, April, 1895, 20 pp., 4 Plates.

XVIII., April, 1S95. 292 pp., 67 Plates, 1 Chart.

XXVII., No. 3, July, 1895, 8 pp., 2 Plates.

XXVII., No 4, August, 1895, 26 pp., 3 Plates.

XXVII., No. 5, October, 1895, 14 pp., 3 Plates.

XXIX. , No 1, March, 1896, 103 pp., 9 Plates, 1 Chart.

XXIII., No. 1, September, 1897, 92 pp., 15 Plates.

XXXI , No 5, December, 1897. 37 pp., 6 Plates, 1 Chart.

XXXTT., No. 5, May, 1898, 18 pp., 13 Plates, 1 Chart.

XXXII.. No. 8. August, 1898, 8 pp., 3 Plates.

XX'I , No. 2, November, 1899, 116 pp., 22 Plates, 1 Chart.
XXXV., No. 1, June, 1899, 4 pp., 1 Plate.

PUBLICATIONS

OF THE

MUSEUM OF COMPARATIVE ZOOLOGY

AT HARVARD COLLEGE.

There have been published of the Bulletins Vols. I. to XXXIV. ;
of the Memoirs, Vols. I. to XXII.

Vol. XXXV. of the Bulletin, and Vols. XXIII. and XXIV. of
the Memoirs, are now in course of publication.

The Bulletin and Memoirs are devoted to the publication of
original work by the Professors and Assistants of the Museum, of
investigations carried on by students and others in the different
Laboratories of Natural History, and of work l>3r specialists based
upon the Museum Collections.

The following publications are in preparation : —

Reports on the Results of Dredging Operations from 1877 to 1880, in charge of
Alexander Agassiz, by tlie U. S. Coast Survey Steamer " Blake," Lieut.
Commander C. I). Sigsbee, U. S. N., and Commander J. R. Bartlett, U. S. N.,
Commanding.

Reports on the Results of the Expedition of 1891 of the U. S. Fish Commission
Steamer " Albatross," Lieut. Commander Z. L. Tanner, U. S. N., Com-
manding, in charge of Alexander Agassiz.

Contributions from the Zoological Laboratory, in charge of Professor E. L.
Mark.

Contributions from the Geological Laboratory, in charge of Professor N. S.
Shaler.

Studies from the Newport Marine Laboratory, communicated by Alexander
Agassiz.

Subscriptions for the publications of the Museum will be received
on the following terms : —

For the Bulletin, $4.00 per volume, payable in advance.
For the Memoirs, $8.00 " '>

These publications are issued in numbers at irregular inter-
vals ; one volume of the Bulletin (8vo) and half a volume of the
Memoirs (4to) usually appear annually. Each number of the Bul-
letin and. of the Memoirs is also sold separately. A price list
of the publications of the Museum will be sent on application
to the Librarian of the Museum of Comparative Zoology, Cam-
bridge, Mass.

MAY 8 1900

Bulletin of the Museum of Comparative Zoology
AT HARVARD COLLEGE.

Vol. XXXV. No. 8.

MATURATION AND FERTILIZATION IN PULMONATE

GASTEROPODS.

By Henry R. Linvillk.

With Four Plates.

CAMBRIDGE, MASS., U. S. A. :

PRINTED FOR THE MUSEUM.
May, 1900.

Bulletin of the Museum of Comparative Zoology
AT HARVARD COLLEGE.

Vol. XXXV. No. 8.

MATURATION AND FERTILIZATION IN PULMONATE

GASTEROPODS.

By Henry R. Linville.

With Four Plates.

CAMBRIDGE, MASS., U. S. A. :

PRINTED FOR THE MUSEUM.
May, 1900.

MAY 8 1900

No. 8. — Maturation and Fertilization in Pulmonate Gasteropods.1

By Henry E. Linville.

CONTENTS.

Introduction 213

Collection of Material .... 214
Technique 215

A. Maturation 217

I. General Account 217

1. Limax Eggs in the Oviduct 217

2. Limax Eggs in the Albumen

Gland 218

3. Limax Eggs in the Uterus . 218

4. Eggs of Limnaea .... 219
II. Centrosome and Centrosphere 219

1. Limax 220

(a) First Maturation Spindle 220

(b) Second Maturation Spin-

dle 222

2. Limnasa elodes 222

(a) First Maturation Spindle 222

(b) Second Maturation Spin-

dle 226

III. The Nucleus 228

1. Division of Chromosomes . 228

2. Nucleoli, or Karyosomes . 232
B. Fertilization 233

I. The Early History of the Sper-
matozoon in the Egg . . . 233

1. General Description . . . 233

2. Changes in the Sperm-

Head 235

3. The Sperm-Head and the

Beginning of the Sperm-

Centrosome 236

II. The Sperm-Centrosome . . 239
III. The Origin of the First Cleav-
age Spindle 241

Summary 244

Bibliography 246

Explanation of Plates 248

INTRODUCTION.

In the Autumn of 1896 I undertook this investigation at the sugges-
tion of my teacher, Professor Mark, to whom I am under mauy obliga-
tions for helpful suggestions and criticisms.

In the beginning my plan was to study the maturation and fertilization
of Limax maximus L. and Limax agrestis Muller, but owing to ill fortune
and inexperience I was unable to preserve a sufficient amount of
properly prepared material of the desired stages, and consequently I
have been obliged to supplement my work with the results obtained

1 Contributions from the Zoological Laboratory of the Museum of Comparative
Zoology at Harvard College, under the direction of E. L. Mark, No. 109.
vol. xxxv. — no. 8.

214 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY.

from a study of two species of Lirrmsea. The results on Limnsea to
some extent repeat the most excellent work by Kostanecki und Wier-
zejski ('96) on Physa, a near relative of Limnsea. However, additional
points of interest in regard to chemical phases in the centrosome, to-
gether with some facts concerning reduction division in the second
maturation spindle, make this something more than a mere repetition
of their work.

Since the publication of Professor Mark's monograph on Limax
campestris, in 1881, the technique of embryology has been so much
developed that a re-investigation of the maturation and fertilization of
this genus of pulmouates with special attention to the history of the
centrosome seemed likely to yield interesting and important results.
The incompleteness of my series of stages in Limax material will, how-
ever, necessitate further work on this genus, which I hope to complete
at some time in the near future.

By each addition to the already great mass of cytological litera-
ture, it becomes more apparent that if the centrosome is a perma-
nent organ of the cell, it is an organ which goes through various and
complicated phases. Certain facts impel us to believe that there
is a variation in the chemical condition of the centrosome. For
example, in some preparations, at a certain stage of maturation,
there is no centrosome (nor centriole) to be distinguished as a
deeply staining body at the poles of a perfectly formed spindle, whereas
in other preparations at apparently the same stage and treated in the
same manner, one may observe centrosomes which are deeply stained and
of enormous size in comparison with the volume of the cell. This great
variation of the centrosome is, however, not inconsistent with the idea
of its permanence in the first maturation spindle, where the astral rays
always indicate its presence. When, however, no astral rays are visible,
the difficulties in the way of identifying the centrosome are exceedingly
great, so great in fact as to render it well-nigh impossible to distinguish
it from the yolk granules.

Collection of Material.

I have collected and preserved eggs of Limax agrestis, Limax maximus,
and of two species of Limnsea. Limax maximus lays in the vicinity of
Cambridge, Mass., during October, November, and December. Limax
agrestis begins laying in early summer, and may, under especially favor-
able circumstances, continue to lay late into the winter. The eggs of

linyille: pulmonate gasteropods. 215

Limax maximus may be found in the vicinity of Cambridge in damp,
protected places, under rotting wood, or under waste lumber and straw,
and especially about greenhouses, where these slugs, and also Limax
agrestis, are found most abundantly. The eggs of Limax maximus are
not covered with earth, as are those of Limax agrestis. It is seldom
that one finds the eggs of Limax agrestis lying exposed on the surface of
the ground. Limax maximus in captivity apparently lays its eggs wher-
ever it happens to be ; sometimes on the bare side of the box or can in
which it is confined ; at other times under a piece of rotting wood or
other protection. Limax agrestis, on the other hand, almost invari-
ably bores into the loose soil in the box, sometimes nearly an inch, and
lays its eggs in a single heap. Its eggs do not cohere as do those of
Limax maximus.

In general, early morning is the time when eggs are laid by both
species of Limax, although the laying may take place at auy time of
the day or night. The duration of the time of laying is not great. Xo
continuous observations of the time consumed in laying were made,
because of the desirability of getting eggs in as early stages as possible,
the animal being killed as soon after the beginning of laying as possible.
Considering the time required to extrude several eggs one after
another, however, it does not seem likely that more than thirty minutes
would be required to complete the extrusion of the largest number
deposited at a single laying.

About the 15th of March, 1897, I collected a large number of Limnaea
elodes Say, one of the common pond snails. Many pairs were found in
the act of copulation. In the course of two days eggs were laid in the
aquarium in great abundance. These snails can be stimulated to lay
simply by supplying plenty of fresh water and keeping the vessels free
from any decaying matter. Laying usually takes place early in the
morning, but a sudden change from impure water to pure water will
cause them to lay at any time of day.

Technique.

The eggs of Limax were taken either just after laying or from the
sexual organs before being laid. The latter were obtained from the
uterus, from the albumen gland, or from the oviduct. The eggs just
deposited and those abstracted from the uterus, where they lie one after
another ready to pass to the exterior, were " shelled," freed from the
albumen and fixed. The albumen gland and the oviduct of individuals

216 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY.

which had begun to lay were sectioned ; eggs were found in both of
these regions.

In freeing the egg cell from the shell and the albumen, I have found
the best method to be the one recommended by Kofoid ('95). This
consists in placing the eggs, a few at a time, in a watch glass containing
normal salt solution, care being taken not to let any eggs remain in the
solution longer than ten or fifteen minutes. Taking an egg carefully in
a pair of fine forceps, one can either snip the membrane with a pair of
sharp-pointed scissors, or, with the aid of a sharp needle, rend it by
catching it between one arm of the forceps and the needle. After the
egg has escaped through the opening in the egg-shell, the albumen can
be washed away from it by a gentle current produced by a pipette. As
soon as the eggs are free from albumen they are transferred to the fixing
solution. Sometimes I have thrown the entire egg, without shelling it,
into the fixing solution, and have subsequently removed the membrane
and as much of the albumen as could be taken away with safety to the
yolk. The egg membrane must be removed within a short time after
the egg has been thrown into the fixing fluid, because otherwise it be-
comes too hard to be cut successfully. Eggs that have been killed (fixed)
without shelling, then washed to remove the killing agent, and dehy-
drated in alcohol, may be returned through weaker grades of alcohol
to water for the purpose of shelling and removiug all but a small por-
tion of the surrounding albumen ; but far better conditions of yolk and
cytoplasm were obtained by removing the albumen frcm the egg as soon
after killing as possible, and usually the result was better still, if the
egg was freed from albumen before it was killed. The latter, indeed, is
the best of all methods for Limax, but iu the case of the eggs of Limnsea,
which are quite small, much time can be saved, and apparently quite
as good results obtained, by thoroughly fixing the eggs before shelling,
providing the egg-shell is removed before the albumen has had time to
harden.

For killing and fixing, the following solutions were used : Saturated
aqueous solution of corrosive sublimate with 3 per cent to 5 per cent
acetic acid, Flemming's fluid, Perenyi's fluid, and three of the mixtures
employed by Kostanecki und Siedlecki, viz. (I) a mixture consisting of
saturated aqueous solution of corrosive sublimate one part, 3 per cent
nitric acid one part, absolute alcohol one part ; (II) a solution similar
to the last in which acetic acid is substituted for the nitric acid; and
(III) a simple 3 per cent solution of nitric acid. Sublimate-acetic is
well known for its good preservation of cytoplasmic structures. Flem-

LINVILLE: PULMONATE GASTEKOPODS. 217

ming's fluid has no advantage over other methods, and has the
decided disadvantage of causing very great brittleness. Series of sec-
tions, even after being mounted and subsequent to the hardening of the
balsam, frequently break into small bits under very slight pressure.
Perenyi's fluid was used much more successfully. The excessive brittle-
ness noticed in the Flemming preparations was not present in those
made with this mixture ; moreover, with Perenyi's fluid nuclear struc-
tures were well preserved, and when stained the elements came out
clearly. Cytoplasmic structures, including the astral rays, were in most
cases sharply marked. The methods proposed by Kostanecki und Sied-
lecki were not tried with Limax material. In the eggs of Limnaea the
nuclear and the cytoplasmic structures were preserved very well by the
three methods of Kostanecki und Siedlecki ; of the three, the 3 per cent
nitric acid solution gave the best results, the preservation by this method
being exceptionally good.

Heidenhain's iron-hsematoxylin was used exclusively in staining.
Slides with sections affixed were immersed in the 2 per cent iron-alum
mordant for a period varying from three to twelve hours, and after
washing in a gentle current of tap water for several minutes, were
placed in the \ per cent aqueous solution of hematoxylin and left for
periods varying from eighteen to forty-eight hours. A 2 per cent iron-
alum solution was used for decolorizing, the process being carefully
watched by frequent examination under a low power of the microscope.
In most cases the yolk could be decolorized sufficiently to disclose the
ceutrosomes, for example, without decolorizing the centrosomes them-
selves.

A. MATURATION.
I. General Account.
1. Limax Eggs in the Oviduct.

My study of the eggs of Limax as they occur in the hermaphrodite
gland has been very limited. As far as the position of the nucleus in
such eggs is evidence, I cannot discover that the egg has any pre-estab-
lished axis, the nucleus being always central.

The earliest observed stage in the first maturation spindle was seen
in eggs found in the oviduct and apparently not long freed from the
hermaphrodite gland. Indeed, judging from the proximity of these
eggs to the hermaphrodite gland, it seems highly probable that changes

218 bulletin: museum of comparative zoology.

leading to the formation of the first maturation spindle begin before the
egg is set free from that organ. The least developed of the eggs found
in the oviduct showed the first maturation spindle already established,
and nearly the whole of the germinative vesicle involved in the spindle.

2. Limax Eggs in the Albumen Gland.

I found a few eggs in the albumen gland imbedded in a small mass of
albumen ; there was no trace of an egg membrane. In these eggs the
first maturation spindle was in every case completely formed and lying
near the middle of the egg. There was no indication of the presence of
a spermatozoon in the eggs found either in the oviduct or in the albumen
gland.

3. Limax Eggs in the Uterus.

In the uterus of Limax agrestis I found a few eggs in which the first
maturation spindle had not yet begun to move toward the periphery.
The earliest stage of the eggs of Limax maximus secured was found in
the uterus, no eggs of this species having been obtained either in the
oviduct or in the albumen gland. In these (L. maximus) eggs the first
maturation spindle was eccentric in position, the centre of one centro-
sphere being near the periphery of the egg. Uterine eggs were kept
separate from eggs already laid, and likewise from eggs obtained from
the oviduct or the albumen gland, but no note was taken of the exact
location of eggs in the uterus, whether they were nearer the albumen
gland or the external opening of the uterus.

The eggs of Limax maximus found in the uterus ranged from a stage
in which one centrosphere of the first maturation spindle was nearly
peripheral in position to a stage in which the first polar cell was com-
pletely formed. Since these eggs were found in the uterus of an animal
killed while engaged in laying, one is safe in assuming that the earliest
stage likely to be found in an egg of Limax maximus already laid is one
in which the first polar cell has been formed. Unfortunately, not a
sufficient number of eggs of Limax agrestis were preserved to furnish any
definite notion of the earliest and latest stages to be found in the uterus
of this species. In the few eggs from the uterus that were sectioned and
examined, the first maturation spindle was nearly central.

The earliest indication of a spermatozoon within the egg was noted in
uterine eggs. The fact that in some cases (L. agrestis) the head of the
spermatozoon was still attached to the filament indicates that penetra-
tion had taken place only a very short time before fixation by the killing

linville: pulmonate gasteropods. 219

fluid. In sections of the oldest uterine eggs from Limax maximus (first
polar cell completely formed) the head of the spermatozoon — still a
single, oval, homogeneous, and slightly swollen body — lies near the
centre of the egg, its long axis being directed more or less definitely
toward the egg-aster.

4. Eggs of Limn m a.

All the material of Limnsea elodes was obtained from eggs already
laid. In the earliest stages thus secured, the first maturation spindle
had begun to move from the centre of the egg. In only one instance
(Plate 1, Figure 2) have I seen distinct remnants of the germinative
vesicle.

II. Centrosome and Centrosphere.

Recent investigation of the nature of the centrosome and the centro-
sphere has thrown considerable light on the variable nature of these
structures. The centrosphere cannot at present be considered a perma-
nent structure, but merely a temporary manifestation of an unknown
force. As long as the centrosphere appeared to be the region of the be-
ginning of the astral rays, there was good reason for assigning to it a
considerable degree of importance. In the light of the recent work on
Physa by Kostanecki und Wierzejski ( '9S), and on Ascaris by Kos-
tanecki und Siedlecki ( '93), however, the centrosphere becomes of less
significance. They find that the astral rays extend into the centrosphere
and even to the centrosome itself. The centrosphere, according to these
authors, with whom I agree, is formed merely by a thickening of the
rays. Wilson ('96, p. 234) calls attention to the concentric rings of
microsomes on the astral rays in the spermatogonium of Salamandra as
seen by Drtlner, and says that the innermost two rings, being especially
prominent, mark off a centrosphere composed of a medullary and a cor-
tical zone. Another case in point which Wilson discusses is that of the
rings of microsomes found by Heidenhain in leucocyte asters. The con-
dition shown in leucocytes, with the astral rays beginning at the centro-
some (the ultimate structure at the centre), tallies well with what
Kostanecki und Wierzejski found in the egg of Physa. If the centro-
sphere is to be taken out of the category of the permanent organs of the
cell, it nevertheless represents a condition in the chemical and physical
phases of the cell which is worthy of further investigation.

In the introduction to this paper, I have referred to the variable size
and condition of the centrosome. This is so intimately connected with

220 BULLETIN: MUSEUM OF COMPAEATIVE ZOOLOGY.

the variation in the centrosphere as to render it advisable, and even
necessary, to discuss the two structures together. I shall first take up
the results obtained from the study of the maturation of Limax maximus,
and then follow with a more complete account of the history of the cen-
trosome and centrosphere in Limnaea elodes.

1. Limax.
(a) First Maturation Spindle.

The earliest stage of the egg of Limax maximus that I have found is
one showing the first maturation spindle fully formed (not figured ;
compai'e with next older stage, Plate 3, Figure 16). The spindle, with
all the chromosomes in the telophase, has moved half the length of the
egg radius toward the animal pole. At either end of the spindle is to
be seen a large distinct centrosphere, composed of a central, pale, reticu-
lated area, and a very thick wall. The wall is nearly one third as thick
as the diameter of the whole sphere. Figui'e ]Q represents a later stage,
but does not differ essentially from the earlier one as far as the condition
of the centrosphere is concerned. Careful examination of the centro-
sphere in both shows that the very fine reticulum of the clear region at
the centre is, to all appearances, continuous with the more compact re-
ticulum composing the wall. The astral rays emerge from the outer
portion of this wall, but, because of the density of the substance com-
posing the wall, it is impossible to say whether or not the central retic-
ulum is continuous through the wall of the centrosphere with the astral
fibres. The thickened wall of the centrosphere and the fine central <
reticulum are not occasional phenomena, but are present in every first
maturation spindle in Limax maximus up to the stage represented in
Plate 3, Figure 16. The most thorough search through specimens fixed by
different methods and stained for varying lengths of time has failed to re-
veal any trace of a centrosome in the hind of centrosphere just described.
Between the stage represented in Plate 3, Figure 16, and the one repre-
sented in Plate 3, Figures 17 and 21, is a gap in the condition of the
centrosome which I am not able to bridge over by intermediate stages.
However, the stages are not so far apart that the changes through which
the centrosphere would pass may not be fairly inferred.

Whatever the influence that causes the condition of the centrosphere,
it is clear that the process is one of increase in volume of the central
finely reticulated area. Figures 17 and 21 are particularly instructive
in this regard. In each two centrosomes have made their appearance.

LINVILLE : PULMONATE GASTEROPODS. 221

In addition to the fact that in Figure 21 the centrosomes have moved
further apart than in Figure 17, the more completely fused condition of
the elements composing the " Zwischenkorper " in Figure 21 gives con-
clusive evidence that it has progressed in development further than
Figure 17. I have not been able to follow closely the fate of this cen-
tral reticulated area after the stage represented in Figure 21.

It will be noticed that in Figure 1 7 the long diameter of the centro-
sphere is nearly coincident with the chief axis of the egg, and likewise
with a line connecting the two minute centrosomes. In Figure 21, on
the contrary, the long diameter of the centrosphere is nearly perpen-
dicular to the chief axis of the egg, and forms a very sharp angle
with the line connecting the two small centrosomes. This relation of
centrosphere-axis with centrosomes suggests a causal connection, but
when compared with the position of the centrosomes in Plate 3, Figure
22, it seems probable that the reason for the variation in the direction
of the long axis of the centrosphere, as compared with the chief axis of
the egg, in these three cases must be sought in something else than the
position of the centrosomes.

Before turning to the consideration of the centrosome, it will be inter-
esting to call attention to the astral rays of Figure 17. Here one set of
the astral rays begins at the centrosphere and extends two thirds of the
way to the vegetative pole, while another set is composed of very short
rays scattered among the long rays. Beginning near the distal end of
the long rays and continuing in the same general direction to the
periphery of the egg are rather broad, indistinct band-like radiations
(unfortunately not well repi'oduced from the drawing). On examination
with a one-twelfth inch immersion lens, the bands were found to be
composed of exceedingly fine granules, different from the ordinary yolk
granules, and distinct also from the " microsomes " which are visible
among the long astral rays. The same faint bands appeared in other
eggs of the same stages as those shown in Figures 17 and 2.1.

There is a point concerning the astral rays in Figure 21, to which I
wish to call particular attention, since it may serve to throw new light
on the question of the centrosphere. Beginning at the periphery of the
central finely reticulated area, the rays extend outward a short dis-
tance as a set of extremelv fine fibres, then suddenlv become thicker,
and retain this condition to their peripheral ends. I cannot demonstrate
any ring of microsomes at the region of transition from the finer to the
coarser fibres, but nevertheless the zone embracing the finer rays is a
distinct modification of the aster, and is of interest because of its bear-

222 bulletin: museum of comparative zoology.

ing on the question of the limits of the centrosphere. Other instances
of this phenomenon will be referred to further on.

It will be remembered that no centrosome, as that structure is gen-
erally understood, is to be made out in connection with the first matu-
ration spindle of Limax maximus. After the formation of the first polar
cell, however, we have (in Figures 17 and 21) within the centrosphere
the centrosome already divided, presumably in preparation for the
formation of the second maturation spindle. In Figure 17 the centro-
somes are extremely small, and only one of them gives indication, even
with the best immersion lens, of having rays in connection with it, but
in Figure 21 the small dense bodies have all the characteristics of
centrosomes. Although the centrosomes in Figure 21 are very minute,
they may still be made out with certainty at points from which several
very delicate rays diverge. These rays may have some connection with
rays outside the reticulated area, but if they do, it is only a secondary
connection. The rays extending from either centrosome toward the
other have united into a very small but fairly distinct spindle (not well
shown in the figure).

(b) Second Maturation Sjnndle.

My study of Limax maximus points to the conclusion that after the
formation of the first polar cell, and possibly during that process, the
centrosphere remaining in the egg increases in size to many times
the volume it had in the first maturation spindle. Neither after the
formation of the first polar cell, nor after the formation of the second,
have I been able to trace continuously the modification of the centro-
sphere. In Figure 20, however, the periphery of the centrosphere is
very faint ; and it seems as if the whole structure were on the verge of
disappearance. In a subsequent part of this paper I shall take up the
discussion of the fate of the inner centrosphere and also the centrosome
of the second maturation spindle.

2. Limn^ea Elodes.

(a) First Maturation Sjnndle.

My observations on the first maturation spindle of Limnaaa elodes
have resulted in a number of interesting facts bearing on the relation of
centrosphere to centrosome. As stated in the introduction, the centro-
some sometimes appears stained faintly, sometimes very deeply. "When
deeply stained, it varies in size from an extremely small body to one of

lixville: pulmonate gasteropods. 223

the size of the centrosphere itself. The existence of the latter extreme
affords strong reason for believing that the entire centrosphere may be-
come stained, for, by a proper serial arrangement of several prepara-
tions, one can see that the clear region of the centrosphere surrounding
the centrosome is gradually encroached upon by the centrifugal advance
of the stainable region, until the entire centrosphere is deeply and
homogeneously stained. It is quite possible that the cases of faintly
stained centrosomes may he due either to understanding, or, what is
more probable, to protracted decolorizing.

In taking up the detailed description of the maturation spindles, I
shall follow the order indicated in the discussion of the same structures
in Limax maximus ; that is, I shall begin with the earliest stage in the
first maturation spindle. It will he remembered that in Limax maxi-
mus the earliest stage obtained was one in which the condition of the
chromosomes of the first maturation spindle already indicated the
telophase. That was a uterine egg.

All the eggs of Limnaea elodes were taken after being laid. The
earliest stage obtained was that shown in Plate 1, Figure 2.1 The spin-
dle is fully formed ; it surrounds the disintegrating germinative vesicle,
and already has a radial, though deep, position. At either pole of the
spindle is a well-developed aster, at the centre of which appears a mass
of minute granules, stained yellowish-brown. In the midst of this mass
it is possible with an immersion lens to distinguish a very small and
faint centrosphere, containing an extremely small centrosome (not to be
seen in the drawing). The inner aster is curiously modified by the
sperm aster, a phenomenon to which I shall refer later on.

A stage in the egg of Limax agrestis, similar to this, is shown in
Plate 3, Figure 14. In this case the germinative vesicle, judging from
the rather uneven arrangement of the chromosomes, has just disappeared.
The spindle here is central, and in this respect differs from the condition
shown for Limnaea in Plate 1, Figure 2, where one pole of the spindle is
near the centre of the egg. In Figure 14 both the position of the spin-
dle and the condition of the spermatozoon indicate a stage younger
than that of Limnaea (Figure 2), but the condition of the germinative
vesicles indicates that it is older than Limnaaa.

In the eggs of Limnaea I have seen frequently the typical flattening
of the animal pole preceding the formation of the first polar cell (Plate 1,

1 (Plate 4, Figure 25, represents the egg of Limax agrestis before it leaves the
hermaphrodite gland, and shows the condition of the germinative vesicle at this
stage.)

224 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY.

Figure 1). The " polar depression " described by Kostanecki und
Wierzejski ('96) for Physa is also not a rare phenomenon in Limnaea.
Conkliu ('94) suggests that the flattening of the animal pole preceding
the formation of the first polar cell is caused by the contraction of the
spindle fibres. Perhaps, in a similar way the " polar depression " is
produced by a more active contraction of the spindle fibres which lie in
the prolongation of the axis of the spindle. But even before the outer
centrosphere has reached the cell membrane (Plate 1, Figure 1), the
polar depression has disappeared, and a marked flattening of the ceiitro-
sphere accompanies that of the animal pole of the egg. On either side
of the outer centrosphere, as seen in optical longitudinal section of the
spindle (Figure 1), there is a projecting "wing" of deeply staining
substance. Careful examination shows these wings to be composed of
closely crowded astral rays, which have stained near their proximal ends.
Since the "wings" are to be seen in all longitudinal sections of this
spindle, it follows that the appearance is due to the presence of a
continuous disk of staining substance.

Studying the different figures of the first maturation spindle of
Limnaea with special regard to the deeply staining portion at the poles,
one must, I think, conclude that both centrospheres shown in Figure 1
have taken the stain throughout their whole extent. The roughness of
the outline of the stained centrospheres, as compared with those of
smaller size, is due to the increasing distance between the astral rays,
as one passes outward from the centre. This condition suggests the
idea that even parts beyond the limit of the centrosphere may have
been stained ; for if only the centrosphere were stained, the outline
should be more regular. Those investigators who have discovered
enormous centrosomes should examine Plate 4, Figure 23. The dense
mass shown there at each pole of the spindle looks more like a precipi-
tate lodged about a central point than like an organ of the cell. In
Figure 24 the outlines of the centrospheres may be seen. The great
irregularity in the form of the stained portions proves beyond question,
it seems to me, that the dense masses at the poles of the spindles repre-
sented in Plate 1, Figures 1 and 3, and in Plate 4, Figure 23, are
simply portions of the cell protoplasm which have not been decolorized.

Many investigators, notably Mark, Garnault, and Kostanecki und
Wierzejski, have given in detail for pulmonates the process of the actual
formation of the polar cells. It is my purpose in this division of the
paper to note only the modifications of the centrosome and centrosphere.
In the part devoted to the discussion of the nucleus, I shall make special

LIXVILLE: PULMONATE GASTEEOPODS. 225

mention of the differences in the character of the division of the chro-
mosomes in the first as compared with the second maturation spindle.

Turning now to the question of the centrosome, I find that after the
first polar cell has been nearly or quite formed the centrosphere belong-
ing to it, when it can be made out at all, gives evidence of disintegra-
tion. This is shown in Plate 2, Figure 8. Nothing which can be
maintained to be a centrosome appears in the outer centrosphere of the
polar cell. There is, however, in Plate 2, Figure 12, in the only polar
cell represented in that figure, a distinct centrosome, with fibres radiat-
ing to the chromosomes. The region of the supposed connection of the
polar cell with the egg in this section was covered by a mass of foreign
substance, so that it is not certain, though highly probable, that the
polar cell was joined to the egg. The aster remaining in the egg shown
in Figure 8 resembles the condition of the centrosphere shown in Figure
21. In the present case (Figure 8) the centrosphere may be said to be
composed of two parts, a very small, central clear area, of spherical
form, and a very much larger non-spherical enveloping structure. The
two are not concentric, the inner sphere being much nearer to the
peripheral flattened wall of the enveloping structure than to its deep
rounded extremity. It is to be observed that in Figure 8, at the exact
centre of the inner, small centrosphere, there is a minute centrosome,
but with no indication of division in preparation for the next matura-
tion spindle. The inner centrosphere, though small, is very distinct.
Within this centrosphere no rays could be distinguished ; in fact, the
contents, under the highest magnification, appeared to be entirely
homogeneous. Eather prominent rays, few in number, can be seen
passing from the periphery of the inner centrosphere out through the
outer or enveloping structure. This outer structure stretches from the
plane of the deep ends of the series of chromosomes belonging to the egg,
toward the centre of the egg, a distance equal to nearly twice its width.
The walls, beginning at the chromosomes, run for a short distance per-
pendicular to the plane of the chromosomes and then gradually converge
and nearly come together, but the lines bounding this outer centro-
sphere decrease in distinctness in passing from the plane of the
chromosomes. The outer structure is to be conceived of as a cylinder,
truncate at the peripheral end, dome-shaped at the deep end, and con-
taining a spherical inner centrosphere located much nearer the truncate
than the opposite end.

The enlargement of the centrosphere during the completion of the
polar cell and after its formation — an occurrence so striking in Limax

226 bulletin: museum of comparative zoology.

maximus — has not yet begun in the specimen from which Figure 8
was drawn. In many other cases, however, at a slightly later stage the
centrosphere shows a decided increase in size, as is seen, for example,
in Plate 2, Figure 7. The peculiar arrangement of the chromosomes,
the disappearing spindle fibres, and the enlarged centrosphere shown in
Figure 7 represent a stage I have seen many times. In earlier stages

I have never seen the centrosphere as large as it is in Figure 7. This
phenomenon of an increase in the size of the centrosphere which has
performed its function, was described by Miss Esther F. Byrnes for
Limax agrestis in a paper read at a meeting of the American Morpho-
logical Society in December, 1896.

b. Segond Maturation Spindle.

I have searched with considerable care for evidence of two centro-
somes within the enlarged centrosphere of the first maturation spindle of
Limnsea elodes, but I have found them in only a single case (Plate 1,
Figure 4), a very early stage in the formation of the first maturation
spindle. There is no evidence of a spindle forming between the centro-
somes, as, indeed, nothing of the kind could be expected at so early a
stage. Hence I am unable to say, as far as Limnsea is concerned, what
relation the enlarging centrosphere bears to the formation of the second
maturation spindle ; whether the new spindle is formed de novo within
the still persisting centrosphere, or whether the centrosphere disappears
before the second maturation spindle comes into existence.

A good example of a second maturation spindle near the height of its
development is shown in Plate 2, Figure 11. In this case the two astral
figures are conspicuously unlike. Frequently I have found in first matu-
ration spindles one centrosome differing from the other in size, and also
the centrospheres differing in condition, but the variation shown in Figure

II appears to be of quite another nature. In this case one centrosome,
the peripheral one, has no centrosphere surrounding it, and the centro-
sphere at the deep pole of the spindle is flattened in a plane per-
pendicular to the axis of the spindle. I have not enough material of
the proper stages to allow me to make a careful study of the second
maturation spindle with reference to the centrosomes and centrospheres,
and hence do not pretend to say whether Figure 11 represents a typical
condition.

The phenomenon of concentric centrospheres is not confined to the
first maturation spindle. The aster of the second maturation spindle
shown in Plate 2, Figure 13, is not altogether like that shown in Figure

linyille: pulmonate gasteropods. 227

8, for Figure 13 shows that the small centrosphere is very faintly out-
lined and that the centrosome is very small. There can be no doubt
that rays begin at the small centrosphere and continue through the
outer centrosphere and beyond ; in fact the outer centrosphere is limited
by a very faint outline, which does not interrupt the course of the fibres.
There appear to be vacuolations within the space enclosed by the irreg-
ular outline of the outer centrosphere, a condition which is not well
shown in the figure. The outline itself is not a distinct membrane, but
on the contrary marks the extreme limit of what seems to be a progres-
sive vacuolation, which advances outward in all directions except toward
the animal pole. Neither in the first nor in the second polar cell of the
specimen from which Figure 13 was drawn could any trace of a centro-
some or centrosphere be found.

The question of the fate of the deep centrosome of the second matu-
ration spindle involves the question of the origin of the first cleavage
spindle. The principal part of the discussion of the latter question I
shall leave for another division of this paper. There are, however, cer-
tain points which may be considered here.

Unlike the egg-nucleus in sea-urchins and tunicates,1 the egg-nucleus
in gasteropods moves but slightly from the region where the polar cells
ai'e formed. Hill ('95) has shown for Sphaerechinus that while the egg-
nucleus is in the resting stage at the centre of the egg, it has no rays
indicating the presence of a centrosome. The sperm-nucleus, with its
centrosome, may be seen at this time a sufficient distance away from
the egg-nucleus to enable one to determine readily the relationship of
the single aster present in the egg. The case with gasteropods is quite
different. Not only is the egg-nucleus eccentric in position, but the
astral rays belonging to it persist till a very late stage in the develop-
ment of the two nuclei. Kostanecki und Wierzejski find for Physa that
as the sperm-nucleus, with the aster in advance, moves toward the egg-
nucleus, the astral rays of the egg-nucleus begin to disappear, as if they
were being assimilated by the sperm-aster, while the two are still a
considerable distance apart. In both Limax and Limnpea I have seen
quite the opposite conditions and have never seen the phenomena these
authors describe. Instead of a diminution of the area affected by the
deep aster of the second maturation spindle, I find that, as the egg-
nucleus develops, the centrosphere enlarges and the extent of the rays

1 The statement with regard to the movement of the egg-nucleus in tunicates
is based on unpublished evidence, which Dr. H. E. Crampton of Columbia Univer-
sity has kindly permitted me to refer to.
vol. xxxt. — no. 8. 2

228 bulletin: museum of comparative zoology.

becomes greater (Plate 3, Figure 20). I have never seen for Limnaea
the stage shown for Limax in Figure 20. Although every indication of
an egg-aster finally disappears, the time of its disappearance is very
much delayed as compared with many other animals.

There are many interesting questions which arise in connection with
the disappearance of the egg-centrosome. There are difficulties in proving
either that the disappearance is permanent or that it is only temporary.
If one maintains that the centrosome after the formation of the sec-
ond polar cell simply goes into a resting stage, and thus becomes in-
visible, but finally reappears, the only answer that can be made is,
that such a statement is an assumption that can neither be proved
not disproved. It is an easy matter to find small dense bodies in the
region of the egg-nucleus, and even bodies surrounded by radiations.
Often many of these small dense bodies may be found in such positions
with reference to the two nuclei in the egg as to seem to be significant,
but the difficulty comes in deciding which, if any, of these many
centrosome-like structures are really centrosomes. For, not only may
these structures be found in close relation to the nuclei, but similar
appearances are frequent throughout the egg. With such difficulties as
these to contend against, it is of the greatest importance that the
phases in the regressive metamorphosis of the egg-centrosome be fol-
lowed with the closest scrutiny.

III. The Nucleus.

In my original plan of work I was little concerned with the nucleus,
but as the investigation progressed two problems of great interest and
importance claimed my attention. These are : first, the relation of the
nucleoli or karyosomes to the chromatin in the resting nuclei, and,
secondly, the question of the reduction division in the Roux-Weismann
sense ; that is to say, a reduction of qualities by a transverse division of
the single chromosomes in the second maturation spindle. I shall take
up the second of these questions first, because it comes first in the
stages of development which I have studied.

1. Division of Chromosomes.

The apparent variation in the number, size, and form of the chromo-
somes in the maturation spindles of Limax maximus made the study of
these elements a particularly difficult and perplexing one. As long as I
worked exclusively with this material, the possibility of ever being able

linville: pulmonate gasteropods. 229

to find evidence concerning a reduction division in the second matura-
tion spindle seemed to me very remote. The number of chromosomes
in the first maturation spindle of Limax maximus varies from sixteen to
twenty, and sometimes twenty-one or twenty-two of them may be seen.
Platner ('86) and Garnault ('88) found for Arion sixteen to twenty
chromosomes; but it is quite possible, as Boveri ('90) suggests, that the
normal number is really sixteen, — a number which seems to be typical
for gasteropods. The fact that in Limax maximus the chromosomes
exhibit such variation in size and form, leads me to believe that in the
telophase of the first maturation spindle of this species, and possibly cf
Arion also, not all the bodies seen are simple chromosomes, but rather
that some of them are the result of an appreciable separation of the
elements composing the " dyads," so that, while some of the supposed
chromosomes are unseparated dyads, others are simply one of the com-
ponents of a dyad. I believe the tetrad formation of the chromosomes
to be characteristic of the prophase of the first maturation spindle.
When the elements of the dyads are very close together, the resulting
appearance is that of a very large chromosome, much larger than the
smaller ones. The smaller chromosomes occur most frequently near
together, either in pairs or suggesting a paired arrangement. That they
are actually joined together, I could not demonstrate satisfactorily.
There are cases, however, in which it is a matter of considerable doubt
whether a mass of chromatin is a single body with a constriction at the
middle, dumb-bell fashion, or whether there are really two chromosomes
very near together. In one case I was able to make out the sixteen
dyads very distinctly in the polar cell (Plate 3, Figure 17), but the
chromosomes remaining in the egg were so closely massed that it was
impossible to count them. A curious condition is to be noticed in
Figure 21. The number of dyads in the polar cell is fourteen, and
the number of chromosomes remaining in the egg is also fourteen.
This is quite an unusual variation, and it is probable that two
chromosomes at either end of the spindle were obscured in some
way. The evidence afforded by a single instance, even though as satis-
factory as that shown in Figure 17, is alone not sufficient to carry con-
viction, but taken with the results I have obtained from my study of
Limnaea, affords reasonable ground for belie'ving that the explanation
which I have offered of the appearance of chromosomes in excess of
the number sixteen is the correct one.

The eggs of Limnsea are more favorable for following the phases of
the maturation divisions. Material fixed in 3 per cent nitric acid showed

230 bulletin: museum of comparative zoology.

better preservation than that fixed by any other method. The earliest
stage of the chromosomes derived from the germinative vesicle that has
come under my observation is shown in Plate 1, Figure 2. The outline
of the germinative vesicle has only partially disappeared, and the chro-
mosomes are still indefinite in number and unlike in form. Their ar-
rangement on the spindle can hardly be said to have more than begun.
I have no stages between this condition and that represented by Figure
4, but I have in other specimens abundant corroboration of the appear-
ance which Figure 4 shows in regard to the method of separation of the
chromosomes in the first maturation division (compare Figure 1). In
the first maturation spindle I have always found sixteen chromosomes,
some of them more or less completely divided, others not only divided
but separated a short distance. Examination of Figure 4 will show that
some chromosomes are considerably longer than others. The shorter
ones lie end to end on the same spindle fibre. These I take to be
chromosomes which have completed their separation, and have now
begun to move toward their respective poles. At the middle point of
the longer chromosomes, which corresponds with the equator of the
spindle, there is an " elbow." Chromosomes in which this elbow is less
prominent, are longer than those in which it is large. The meaning is
quite clear. The appearances seen in the chromosomes of the first
maturation division of Limnsea are due to the more or less complete
splitting and separation of elongated chromosomes.

Leaving now for a time the discussion of the phases immediately fol-
lowing metakinesis in the first maturation spindle, I shall take up the
consideration of the chromosomes in the prophase of the second matura-
tion division. Figure 11 (Plate 2) illustrates a condition which repre-
sents partly the prophase and partly the metaphase of division : that is to
sav, some of the chromosomes are undivided and some have just divided.
The chromosomes of this spindle are distinctly dumb-bell shaped, and
lie with their long axes parallel with the axis of the spindle. Here, as
in the first maturation spindle, I find sixteen chromosomes, all of which
are arranged on the outer, thicker fibres of the spindle, never, as is
usual in Limax, through the axis of the spindle. Careful examination
of the successive sections from which Figure 1 1 was constructed showed
that some of the dumb-bell shaped chromosomes were completely divided
across the "handle," and that migration from the equator had barely
begun. ISTot only is the form of the chromosomes in the prophase of
the second maturation division that of a dumb-bell, but even as early as
the telophase of the first maturation division all the chromosomes, both

LINVILLE : PULMONATE GASTEROPODS. 231

those in the group that is to go into the first polar cell and those that
are to remain in the egg, have this dumb-bell shape. Even in the meta-
kinesis of the first maturation division I have occasionally seen evideuce
of a transverse constriction of the chromatin rods. Thus we seem to
have in Limnsea a partially concealed " tetrad formation." The presence
of the transverse constriction so soon after the longitudinal splitting
leaves little doubt that the division as it finally takes place in the second
maturation spindle agrees in position with the early constriction.

This observation does not accord with the results Boveri ('90) ob-
tained for Carinaria, one of the heteropods. He found the chromosomes
of the first maturation spindle to be quadruple. This quadruple group
splits longitudinally in the first maturation divisiou, and then, after the
rearrangement of the chromosomes on the second maturation spindle,
the division takes place by a longitudinal splitting, exactly as in the
first maturation spindle. ^Naturally one is disposed to ask, What, then,
is the meaning of the quadruple groups'?

The thing of importance now to be decided for Limnsea is the manner
of formation of these " Vierergruppen " or tetrads. To make sure of
an exact answer to that question a study of the processes going on in
the rearrangement of the chromatin in the germinative vesicle in pre-
paration for the maturation divisions would be necessary. If the tetrads
are formed by two longitudinal splittings of segments of the original
spireme thread, as in Ascaris (Boveri, '87), and if the two pairs of ele-
ments composing the tetrad are elongated and pressed closely together,
as at least they seem to have been in the case of Cafinafia, then the
two maturation divisions would be " equation divisions," and not re-
ducing, except in the sense of a quantitive reduction. If, on the other
hand, the length of the masked tetrads represents the length of the
original spireme which, in breaking up, first divides longitudinally and
next transversely, then the second maturation division is a reduction
division in the Roux-Weismann sense. Apart from the evidence which
a study of the early stages in the formation and fission of the spireme
thread would give, I have little hesitation in holding that we have in
Limnoea a reduction division of the Roux-Weismann type, because soon
after the evident longitudinal splitting of the chromosomes there occurs
a transverse constriction in each of the resulting halves, which continues
until there is complete separation of each half into two parts, which
then move toward their respective poles.

232 BULLETIN: MUSEUM OF COMPAEATIVE ZOOLOGY.

2. Nucleoli, or Karyosomes.

After the formation of the second polar cell the egg-nucleus becomes
vacuolated, and in most of the cases that I "have observed the sperm-
nucleus also becomes vacuolated at the same time. Both contain many
nucleoli or karyosomes varying in size and situated either at the cross-
ing of linin fibres or arranged along such fibres in bead-like fashion.
The nucleoli frequently stain with varying intensity in the same nucleus,
but generally the staining of all nucleoli is very faint. The larger ones
appear to be vacuolated. I have no doubt of the integrity of these
karyosomes as distinct elements. Platner ('86) believed them to be
distinct elements; but the opposite view was maintained by Boveri
('90, p. 357). who is inclined to think that Platner's " Karyosomen "
are artefacts produced by a crowding together of net-knots and numerous
large achromatic nucleoli by an unfavorable method of preservation, and
that the mistaken interpretation is in part due to the appearance given
in very thin cross-sections. I believe, however, that the structures in
question can be shown to actually exist as distinct normal bodies, not
artefacts. Moreover, I can bring observations to support Platner's
claim that the chromosomes which are taking shape for the first cleav-
age spindle are to be seen surrounding karyosomes as rings. Platner
('86, p. 53) says, " Die Karyosomen nur erscheinen auf den ersten Blick
vollig farblos, eine aufmerksame Beobachtung lehrt aber dass die Chroma-
tinsubstanz welche anfangs diffus in ihnen vertheilt war, sich in der
Form kleiner Kornchen an der Peripherie concentrirt hat."

" Diese Chromatinelemente sind anfangs noch sehr klein und in
grosserer Anzahl vorhanden und entzishen sich dann leicht der Beobach-
tung. Spater sammeln sie sich aber zu einigen wenigen grossern Kor-
pern an und treten dann deutlich hervor. Von diesen Chromatinkornern
enthalt die Mehrzahl der Karyosomen zwei Stuck, welche meist an zwei
diametral gegeniiberliegenden Punkten der Peripherie gelegen sind. In
den kleinern erkennt man zuweilen nur ein solches Element."

I have not studied the karyosomes sufficiently either to confirm or
deny Platner's quoted statement, that the chromatin is diffused through
the karyosomes preceding the time of its accumulation into a ring on
the periphery. I should say, however, that the fact that the karyo-
somes show staining reaction before the chromatin is collected on the
periphery, and that afterwards they do not (a statement which I have
not quoted), is not absolute proof that the chromatin is contained in the
substance of the karyosomes. The variations in the stainability of these

LINVILLE : PULMONATE GASTEROPODS. 233

bodies may, it seems to me, be due to causes which have nothing to do
with the movement of the chromatin. It is a well known fact that the
ceutrosome goes through certain chemical phases, in which it does not
stain at all by a method which at other times stains it deeply. It seems
quite possible that the karyosomes may also have chemical phases inde-
pendent of any physical accumulation of chromatin.

AVhether the chromatin in the resting stage of the nucleus is distrib-
uted through the substance of the karyosomes, or whether instead it is
distributed through the nuclear sap, as Klinckowstrdm ('97) thinks is
the case with Prosthecerseus, I believe Platner is right in saying that
chromatic substance is to be found at one period in the resting stage of
the nucleus collected in rings about the karyosomes. I have frequently
noticed densely staining rings, irregular in outline, surrounding faintly
stained karyosomes. My best evidence concerning the points in ques-
tion is shown in Plate 1, Figure 6. This figure represents the formation
of the first cleavage spiudle, and, quite exceptionally, the two so-called
pro-nuclei have fused. Rays from each aster have extended into the
substance of the fused nuclei, one bundle of rays running nearly through
the nuclei. Scattered along this bundle of fibres are numerous deeply
stained, bent rods. Similar rods may be seen here and there in the
nucleus, apparently unattached. Still others are found lying very close
to the faintly stained karyosomes and in more or less intimate coutact
with them. The chromatin particles surrounding the karyosomes, those
lying loose in the nuclear sap, and those being drawn along the pene-
trating astral rays, are so evidently of the same origin that investigation
of the history of the chromatin within the resting nucleus would involve
an examination of the changes in the so-called nucleoli or karyosomes.

B. FERTILIZATION.

I. The Early History of the Spermatozoon in the Egg.

1. General Description.

As I stated at the beginning of this paper, the first evidence of the
presence of the spermatozoon in the egg of Limax that I have seen, was
in specimens taken from the uterus. Xot only was the first evidence of
penetration of the spermatozoon seen in uterine eggs, but in all. uterine
eggs that I have examined penetration had taken place. Whenever I
have found eggs in the oviduct (L. agrestis), great numbers of sperma-

234 bulletin: museum of compakative zoology.

tozoa have always been present surrounding the egg and completely
filling the oviduct. This condition suggests simultaneous movement of
the eggs and spermatozoa from the hermaphrodite gland. I have never
found either eggs or spermatozoa alone in the oviduct.

Under these conditions one might expect self-fertilization, and I have
looked carefully for evidence of it. That self-fertilization has not taken
place, at least in the oviducal eggs that I have examined, may he
explained by what seems to be a fact, viz. that the spermatozoa occu-
pying the oviduct are in an immature condition. In the developing
sperm element, as the head of the spermatid begins to take on the
form characteristic of the mature condition, the tail is continually grow-
ing in length. The end of the tail of the spermatid exhibits a large
knob. In the fully developed free spermatozoon there is no indication
of this knob. Now, in the oviducal spermatozoon of Limax agrestis I
have seen this knob-like structure and I am inclined to think all the
spermatozoa in the oviduct are in this immature condition.

So far as I know, the spermatozoon has never been observed in the
act of penetrating the egg in gasteropods. It is not very important,
however, in gasteropods that this phenomenon should be observed, since
the tail in following the head into the egg affords the observer the
means of determining the topographical relations necessary for noting
certain preliminary processes of fertilization.1

In the case of Limax maximus I have found, in one instance, a sper-
matozoon very near the periphery of the egg, which it apparently had
but recently penetrated, while the first maturation spindle was migrat-
ing to the periphery to form the first polar cell (Plate 3, Figure 1G).
When, in other instances, the spermatozoon was in practically the same
stage as represented in the figure just mentioned, I have found the first
polar cell completely formed (Plate 3, Figure 21). In one instance
(L. agrestis, Plate 3, Figure 14) I found the spermatozoon with head
and tail still connected, and the germinative vesicle of the egg just dis-
appearing. Kostauecki und Wierzejski ('96) found in Physa that a
spermatozoon may penetrate the egg and a large sperm-aster may be
developed by the time the first maturation spindle is formed. On the
other hand, they found that, in the same species, the spermatozoon may
not penetrate the egg until both polar cells have been formed.

1 I may anticipate criticism of this statement by calling attention to what is
probably a fact, that the viscidity of the egg-cytoplasm prevents great whiplash
movements of the spermatozoon; it is not likely therefore that the tail is "fixed"
in a position far from the path along which the spermatozoon has progressed.

LIXVILLE : PULMONATE GASTEROPODS. 235

A very suggestive phenomenon in the early history of the sperma-
tozoon in the egg is shown in Plate 3, Figure 15. It seems as if the
attraction which caused the spermatozoon to penetrate the egg were
only a general attraction, aud not located in a definite region of the egg ;
for after penetration the head made almost a complete circle before it
came within the more definite influence of the egg-centrosome, or before
the egg-centrosome and the sperm-centrosome had entered upon the
proper phases for attracting each other. At other times, as in Figure
14, the spermatozoon on entering moved straight ahead, not stopping
until it had come in contact with the membrane on the other side of
the esrs:. Here it would have remained until the head aud tail had
separated.

In a preceding paragraph I have referred to the preliminary processes
through which a spermatozoon in the egg may be said to go. These
processes are : first, the change in form of the sperm-head ; and,
secondly, the separation of the head from the tail. I shall describe
these processes in turn.

2. Changes in the Sperm-Head.

The sperm-head in its fully developed condition, and before it has
entered the egg, has the form of a skewer with one, more or less com-
plete, spiral turn. The sperm-head in Figure 15, although broken away
from the tail, still retains the general form it had before it entered the
egg. As a rule, however, immediately after penetration it undergoes a
modification in form. The nature of this modification is well shown in
Figure 14. The fact that the head is still attached to the tail, enables
one to see that, in this case at least, the long axis of the head is now at
right angles to its original long axis. Kostauecki und Wierzejski ('96,
Figure 1) show the same condition for Physa. Apparently the change
begins very soon after the spermatozoon enters the egg, for in only two
instances (one of them shown in Figure 15) have I seen the normal
form preserved. It also seems probable that the sperm-head, after be-
coming elliptical, does not increase in size for a considerable time, for
many sperm-heads are found having the same elliptical form, and
apparently the same size, both when they are at the periphery of the
egg (Plate 1, Figure 1, Plate 2, Figure 8, and Plate 3, Figure 17) and
when advanced in their course toward the egg-aster (Plate 3, Figure 21).
This fact is brought out better by the figures of Kostanecki und Wier-
zejski than by mine. I do not believe any especial significance can be
attached to this change in the form of the sperm-head. At first thought,

236 bulletin: museum of comparative zoology.

the comparison of the sperm-heads in Figures 14 and 21 would seem to
lead to the conclusion that the centrosome in the egg represented in
Figure 21 had moved from the side to the end of the head (if the
change in form shown in Figure 14 is typical) ; but such a change in po-
sition it seems to me would be meaningless. The difference between the
two may be explained by the fact that the sperm-head would naturally
present the smallest surface while being drawn through the yolk gran-
ules and the protoplasm.

Besides the modification of the sperm-head, there is another phe-
nomenon which, like the one just described, is incidental rather than
essential. I refer to the early appearance of a clear area about the sperm-
head. The conditions are represented in Figures 14, 15, 21, etc. The
outer limit of this area is usually marked by a fine sharp line, so that the
whole has the appearance of a clearly defined vacuole, with the sperm-
head at the centre. One can make out fine threads radiating from the
sperm-head to the margin of the vacuole. In the eggs that I have ex-
amined there is no visible modification of the form of the sperm-head
by these radiating threads in the vacuole, except in two cases. One
of these is represented in Figure 15, where I find that the radiating
threads have caused a modification of the outline of the sperm-head.
Wherever a thread comes in contact with the sperm-head, a projection,
apparently from the head, is formed. The origin and meaning of these
radiating threads seem to me of considerable interest in connection with
the subsequent changes in the sperm-nucleus, and hence I shall discuss
that subject in the next division of this paper.

3. The Sperm-Head and the Beginning of the Sperm-Centrosome.

The change of most importance in the early history of the spermato-
zoon in the egg is the breaking away of the sperm-head from the tail.
The interval of time between the penetration of the sperm-head and its
separation from the tail may be short or long according as we interpret
the conditions shown in the figures I have made. In some cases it ap-
pears probable that the spermatozoon has travelled over a considerable
part of the substance of the egg before the head is separated ; in others
(Figures 1, 7, 8, 11, 16, 17, etc.) either the sperm-head on entering has
left the tail outside, or the tail, though entering, has been resorbed.
However, the time of separation of the head from the tail is not so im-
portant as the manner of separation. It is to the manner, and still
more to the cause, that I desire now to direct attention.

LIXVILLE: PULMONATE GASTEROPODS. 237

Kostanecki und Wierzejski in their Figure 1 give a stage of an egg
of Physa which is slightly more advanced than my Figure 14. The
head in their figure, as in mine, is represented as being still attached to
the tail. Between the head and the stainable tail there is the so-called
middle piece, containing the densely stained body supposed to be the
centrosome. I have tried to demonstrate the existence of a middle
piece and a centrosome in the specimen from which my figure was
drawn. Although the sperm-head and a portion of the tail lie in a sin-
gle section, I am unable to make out any differentiation in the tail near
the point of union with the head, the only modification of the tail at that
point being a slight swelling. Unfortunately I have but this single
specimen showing the head still counected with the tail.

In the work of Wilson and others on the changes of the spermatozoon
in the egg, it has been established that the centrosome, in many animals
at least, lies in the middle piece of the spermatozoon, and the discov-
ery of any granule within the middle piece has come to be considered
as the discovery of a centrosome. Waiving the question of the existence
of a differentiated middle piece in the spermatozoon of gasteropods,
there are proofs, beside those brought forward by Kostanecki und
Wierzejski, that a centrosome exists near the base of the sperm-head.
Wilson and ^Mathews ('95) among others have shown that the sperm-
head soon after penetration turns so that its basal end is nearest the
egg-aster. At the time of turning a small aster is visible at the base of
the sperm-head. An important point in the turning of the head is,
that the centrosome is located at the base of the sperm-head. Now, if
the tail is present in the egg, it serves as a landmark to show when the
turning of the sperm-head takes place and how great is the angle
through which the axis of the sperm-head passes. In sea-urchins the
tail does not enter the egg with the sperm-head and centrosome ; hence
the degree of turning cannot be noted as accurately as is possible in
the eggs of gasteropods. The difficulty with gasteropod spermatozoa,
however, is that because of the early change in the form of the sperm-
head and its separation from the tail, the observer is unable to distin-
guish apex from base. Fortunately this difficulty does not exist in two
of the series of preparations that I have. Figure 14 illustrates one of
them. In this case the head and tail are still attached. In the other
case (Figure 15) the form of the detached sperm-head is very nearly
the same as that of a mature sperm-head before penetration. Whatever
movements the tail may have gone through while still attached to the
head, it is fair to assume that the position of the tail shown in the

238 bulletin: museum of compakative zoology.

drawing (Figure 15) is the position it had when the sperm-head sep-
arated from it. In this series of sections, and in another of mine simi-
lar to it, the base of the sperm-head is turned toward the egg-nucleus ;
in the section represented in the drawing (Figure 15) the turning has
been through an arc of 90 degrees. The fact that in both series the
base is turned toward the egg-aster shows that the force which caused
the turning must have acted at the base. The difficulty comes now in
finding the structure in which that force resides. I have said that I
have been unable to find a centrosome in the series represented in
Figure 14, and I am not certain of having seen it even in the prepara-
tion from which Figure 15 was drawn. However, from what we know
of the turning of the sperm-head in other forms, we can assume that
the centrosome exists in this case, even if it is not visible. I have
shown in the figure two structures which may be centrosomes. Extend-
ing from the base of the sperm-head in advance and to the left of it I
have shown two fibres, thicker than the others that surround the sperm-
head, each containing at its middle point a minute granule. On exami-
nation of the specimen with a T'g inch homogeneous Zeiss lens these two
fibres are seen to be composed of numerous very fine fibrils. This recalls
the condition that Kostanecki ('96, Figurel) shows for the free sperm-head
of the sea-urchin. In the sea-urchin, however, there is only one bundle
of rays, and that extends from the base directly toward the centre of
the sperm-aster. In the other series (not figured) of the same stage as
the one represented in Figure 15, there is evidence of a granule in a
bundle of fine fibres which extends from the base of the sperm-head.
Fibres can also be seen radiating from this granule into the surround-
ing vacuole, but these radiations have no connection, as far as I can
determine, with the more prominent threads extending from the sperm-
head through the vacuole. Since in Figure 15 — a stage in which
the centrosome is known to be active — there is no evidence of connec-
tion between a centrosome and the fibres radiating from the sperm-head,
the radiating threads about the sperm-head shown in Figure 14 can
hardly be said to be due to the influence of a centrosome.

The evidence I have of the existence in the egg of a structure which
I can say positively is a sperm-centrosome in its earliest stages is very
meagre. I have seen in the egg many appearances in close proximity
to the free sperm-head which suggested an aster, but examination in
other parts of the cell afforded many examples of similar ill-defined
aster-like structures. In Figure 211 have drawn a structure that may
be the sperm-aster in an early condition; it lies within the vacuole of

LINVILLE : PULMONATE GASTEROPODS. 239

the sperm-head at the end nearest the egg-aster. The fact that it lies
on the side of the sperm-head which is directed toward the egg-aster is,
I believe, a strong argument that this particular structure is the sperm-
aster. The discussion of examples of undoubted sperm-asters I shall
leave for the next sub-division of this paper.

II. The Sperm-Centrosome.

Kostanecki und "Wierzejski ('96) have shown that after the spermato-
zoon has entered the egg and the sperm-head has taken on a spherical
form, the centre of activity is within the sperm-aster. These authors
have also shown that at the earliest appearance of the sperm-aster in the
egg it is very small and near the sperm-head. As the aster increases
in size, it is found sometimes at a considerable distance from the sperm-
head. Later in the progress of development the sperm-aster and sperm-
head (the latter now considerably vacuolated) are found near together
moving toward the egg-aster, the sperm-aster leading the way. Fre-
quently, even at a very early stage, the sperm-centrosome divides into
two parts, and a gradually developing spindle forms between them.

It has been supposed generally that the sperm-aster in its course
toward the egg-aster takes a position somewhere between the egg-aster
and the sperm-head. Some of the figures by Kostanecki und Wierzejski
show exceptions to this. For example, in their Figure 3, showing the
first maturation spindle at the top of the figure, the sperm-head lies to
the left near the periphery of the egg, but the sperm-aster is below the
centre. Likewise in their Figure 14, the sperm-aster, instead of lying
between the sperm-nucleus and the deep end of the second maturation
spindle, is situated so that the line joining it to the sperm-nucleus
makes a right angle with the line uniting sperm-nucleus and the inner
end of the second maturation spindle. At first I was inclined to think
that a second sperm-head in the egg had escaped their observation ; but
since that time I have found that in my own material it is far easier
to make out sperm-nuclei than sperm-asters. Furthermore, I am able
to corroborate for Limnasa (Figure 7) some of the conditions which
they have shown for Physa. I have no important additions to make to
the accounts of these authors on the early changes of the sperm-centro-
some in the egg. The facts I have collected are all corroborative of the
results obtained by Kostanecki und Wierzejski. I shall therefore de-
scribe my observations only briefly.

The sperm-nuclei shown in Figures 5 and 7 are pointed on the side

240 BULLETIN : MUSEUM OF COMPAKATIVE ZOOLOGY.

nearest the sperm-aster. This condition was first noticed by Mark for
Limax campestris in 1881. In Figures 13, 20, 29, of Kostanecki nnd
Wierzejski are shown striking modifications in the form of the egg-aster
and sperm-aster. Apparently each aster is repelled or possibly, as Kos-
tanecki und Wierzejski suggest, is being assimilated by the other. The
interesting facts shown in these figures and in my own Figures 2 and
7 are: first, in each of the five cases the process of maturation is not
completed, and, secondly, there has been an attraction existing, else the
sperm-aster and the egg-aster never would have come as near together
as they have. The fact that the functions of the egg-aster are not yet
completed, may be sufficient to account for the temporary repulsion.

Considerable attention was given in Part A of this paper to the vari-
able form of the centrosome. In the study of the egg-centrosome, the
centrosphere afforded a sort of standard for comparison — a very unsat-
isfactory one, however. The sperm-centrosome has no centrosphere.
The condition of the sperm-centrosome in the specimens I have observed
which is most nearly typical of the conditions of centrosome structures
in general is that shown in Figure 5. The centrosome in that figure is
a small dense granule, which is at the point of origin of the astral rays.

The sperm-centrosome shown in Plate 2, Figure 10 is the largest I
have seen. In fact, I imagined when I first saw it that it was a sperm-
head with possibly a centrosome beneath it, but I now believe that the
dark body seen in close proximity to the radiations from the large cen-
trosome is the sperm-nucleus belonging to it. Another sperm-head is
visible in the egg, but it is near the periphery and is still homogeneous,
whereas the more central sperm-head gives evidence of being vacuolated
(a condition not shown in the figure).

Another condition of the sperm-aster is shown in Figure 2. I have not
been able to make out anything in the central region of the aster that
has the appearance of a centrosome, there being instead a large, faintly
stained, thickly reticulated area, from which astral rays extend. This
central region is very much flattened, apparently owing to a repulsion
between this aster and the deep aster of the maturation spindle. The
repulsion apparent in the astral rays of both these asters is manifest
also in the flattened condition of their central reticulated areas. The
peculiar distortion and partial displacement of the reticulated area of
the deep maturation spindle affords perhaps the strongest evidence
I have presented of a repulsion of one aster by another. The centro-
somes could be found only in the peripheral aster of the maturation
spindle.

LINVILLE: PULMONATE GASTEROPODS. 241

III. The Origin of the First Cleavage Spindle.

The fact that a fully formed sperm-spindle exists iu the egg while the
processes of maturation are going on, does not of itself prove that the
sperm-spindle will become the first cleavage spindle ; but nevertheless a
strong presumption is created that such is the case. If a spindle
arising from the division of the egg-centrosome should also be found
after the formation of the second polar cell, the reason for the existence
of a sperm-spindle could not be easily explained. As far as I know,
however, the egg-centrosome has never been found to divide and produce
a spindle after the formation of the second polar cell; but that does not
end the matter. In the first place, it is impossible to know whether
every sperm-centrosome divides to form a spindle ; and in the second
place, the sperm-spindle, whether formed, or only potentially existing in
the sperm-centrosome, nearly disappears from view during the resting
stage of the germ-nuclei.

Kostanecki und Wierzejski maintain that they have followed the his-
tory of the sperm-centrosome through even this almost quiescent stage
to the actual formation of a cleavage spindle. Their Figure 33a repre-
sents the centrosome as having nearly disappeared. Even on this
evidence, how is one to prove that the egg-centrosome has totally
disappeared " because of inability to carry on the division of the cell,"
while the sperm-centrosome, a new organ, becomes predominant and de-
velops into the first cleavage spindle 1 If at any time a link is missing
in the chain of the history of the sperm-spindle, it seems to me we can-
not exclude the possibility that the egg-centrosome may come again to
view, and actually take part in forming at least a part of the first
cleavage spindle.

In my discussion of the changes of the egg-centrosome, I have said
that, although the disappearance of the egg-centrosome and the astral
rays is long delayed, they nevertheless come to a condition in which it
is impossible to distinguish them from the yolk granules and the proto-
plasm. I have also said that I have not seen a stage such as Kosta-
necki und Wierzejski represent in their figures, where the rays of the
egg-aster are apparently being assimilated by the rays of the sperm-
aster ; but I have not seen exactly the conditions of germ-nuclei which
these authors find at the time assimilation is going on. However, their
figures as a whole do not bear out the " assimilation " theory completely,
as, for example, in Figure 27, where the egg-aster (even though the
germ-nuclei are nearer together than in Figure 25 or 28) is still appar-

242 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY.

ently active. When we come to the stage represented in their Figure
33a, and still further to conditions such as [ have found where no aster
can be detected in connection with either egg-nucleus or sperm-nucleus,
then we may well question the value of the " assimilation " theory.

I have referred in an earlier part of this paper to the clear evideuce
brought forward by Hill ('95) to show that in the egg of Sphserechinus
the egg-aster completely disappears. The sperm-aster, from the time it
appears at the base of the sperm-head till the first cleavage spindle is
formed, is clearly visible ; and the sperm-centrosome and its product,
the sperm-spindle, are accompanied by the only astral structures that
are visible. Fick ('93) finds the first cleavage spindle to be of spermatic
origin, but the proof to be found in the existence of an " archoplasmic "
mass is not so satisfactory as the evidence produced by Hill.

Some of the most recent work on the history of the centrosomes in
the egg tends to re-open the question of the origin of the first cleavage
spindle. Foot ('97) maintains for Allolobophora that whatever the evi-
deuce of an aster about the spermatozoon .in the egg, the cleavage aster
comes from the egg itself. According to this investigator, the sperma-
tozoon gives to the developing embryo only the sperm-head. The first
cleavage spindle comes fi'om the egg-aster, although that structure is
invisible during the resting stage of the egg-nucleus. Klinckowstrom
('97) finds in a planarian that both egg-aster and sperm-aster disappear
during the resting stage of the germ-nuclei. When two asters make
their appearance to form the first cleavage spindle, they are at a con-
siderable distance from each other and in such relation to the two germ-
nuclei that it is impossible to decide their origin.

The conditions encountered by Klinckowstrom are apparently the
same as those represented in my Figures 12, 18, and 19, referred to above ;
and since he, as I, could not follow the changes in the centrosome con-
tinuously, we can not say, as others have said of centrosomes in practi-
cally the same relation to nuclei, that both are of spermatic origin.
Hence, if anything leading toward conclusive proof is to be known, it
must be learned from facts striking enough to counterbalance the weak-
ness incident to a break in the series of phases through which the cen-
trosomes are seen to pass. I believe I have in Plate 4, Figures 26, 27,
and 28, evidence which proves, just at the stage of development when
evidence is most conclusive, that the first cleavage spindle is wholly of
spermatic origin.

The relation of egg-nucleus and sperm-nucleus in the egg of Limnaea
is so constant that the sperm-nucleus is never found between the polar

LINVILLE : PULMONATE GASTEROPODS. 243

cells and the egg-nucleus. Frequent observation of the stages in the
development of the sperm-nucleus enables one to know, even by a hasty
examination of the series of sections, which nucleus is the sperm-nucleus.
When, therefore, we find, as in Figure 27, that the sperm-nucleus,
although partly overlying the egg-nucleus, is nearer to the two cen-
trosomes, and indeed, between the egg-nucleus and the centrosomes, we
know that both centrosomes more definitely belong to the sperm-nucleus
•than to the egg-nucleus. More important than the position of the cen-
trosomes with reference to the sperm-nucleus, and still dependent on
that position, is the fact that the astral rays from either centrosome
have extended toward the sperm-nucleus ; and, having touched the
nuclear membrane, have caused indentations in it, and have even pene-
trated the nucleus itself and united to form a spindle there. The most
careful scrutiny of the sections through the egg-nucleus fails to reveal
the slightest trace of its being affected yet by the presence of the
asters.

Although many changes take place in the nucleus between the stages
represented in Figures 27 and 26, the formation of the cleavage spindle
has proceeded just far enough to show that the process of first involving
the sperm-nucleus in the spindle continues without interruption until that
nucleus is wholly within the spindle, while the egg-nucleus is still out-
side of it. The series of changes leading up to the formation of a perfect
first cleavage spindle with the chromosomes in their prophase position is
nearly completed in the stage represented in Figure 28.

In conclusion I desire to answer an objection that may possibly be
raised and then to call attention to an important additional fact. It
may be urged that at least one of the centrosomes represented in each
of the Figures 12, 18, and 19 may still be the egg-centrosome, since it is
as near to the egg-nucleus as it is to the sperm-nucleus. In reply to
this possible contention, I should say, that, while the various positions
which the two nuclei may occupy with reference to each other, would
easilv brinsr one centoosome into a situation where it would be as near
to the egg-nucleus as to the sperm-nucleus, nevertheless such a position
in the light of the subsequent history of the centrosome can be shown
to be temporary and incidental. It must be remembered also that the
sperm-centrosome, sometimes single and sometimes divided, precedes the
sperm-nucleus as the two move toward the egg-nucleus in the early
stages of fertilization. The fact that one and sometimes both centro-
somes later migrate from a point between the two nuclei and away from
the egg-nucleus toward the sperm-nucleus tends to emphasize the inter-
vol. xxxv. — no. 8 3

244 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY.

pretation I have given : — that the centrosomes of the first cleavage
spindle are related to the sperm-nucleus alone. Although one centro-
some is often found very near the egg-nucleus, and even extends astral
rays toward it, I have never found both centrosomes cooperating to involve
the egg-nucleus in a spindle before the sperm-nucleus was involved. This
seems to me a point of vital importance, precluding, as it does, the possi-
bility of even one centrosome, being of other than spermatic origin ; for
if in Limnaea elodes one of the centrosomes comes from the egg, then
the incipient first cleavage spindles would be as likely to involve the
egg-nucleus as the sperm-nucleus first.

SUMMARY.

The centrosome and the centrosphere are extremely variable struc-
tures, both in size and in their reaction to stains.

In the processes leading to the formation of the first polar cell in
Limax maximus, no centrosome is visible. The astral rays apparently
begin in the thickened wall of a pale centrosphere.

After the formation of the first polar cell, and also after the formation
of the second polar cell, the centrosphere is found to be greatly enlarged.
This condition is characteristic of both Limax maximus and Lirnnaea
elodes.

In the egg cell of Limnaea elodes great variation in the condition of
the centrosome and the centrosphere is observed. Two facts tend to
show that the centrosphere is not a permanent organ of the cell : First,
the centrosphere is sometimes invisible on account of the increased area
at the centre of the aster which reacts to the stain, and, secondly, astral
rays beginning in the centrosome continue through the centrosphere.
The centrosphere then appears to be no more than a region of thicken-
ings in the astral fibres.

The centrosome in the first maturation spindle of Limnaea, when it is
visible, varies from a very minute granule to a condition in which its
diameter is as great as the transverse dimension of the spindle itself,
or even greater. In the first maturation spindle of Limnaea the centro-
some was invisible in only a few cases, the astral rays usually giving
evidence of its existence. The centrosome of the second maturation
spindle of Limnaea never appeared as a large structure ; the number of
preparations, however, was limited.

In both Limax and Limuaea after the formation of the second polar

linville: pulmonate gasteropods. 245

cell, the disappearance of the egg-centrosome and the egg-aster, though
long delayed, is complete.

A reduction division of the chromosomes in the Roux-Weismann sense
was observed in the second maturation division of the egg of Limnaea
elodes. The process consisted of a transverse division of the dyads re-
sulting from the longitudinal splitting of the masked tetrads of the first
maturation division.

In fertilization the tail follows the sperm-head into the egg, but later
is resorbed by the egg cytoplasm. The sperm-head, after breaking away
from the tail begins, under the influence of the sperm-centrosome, to
move, with its base in advance, toward the egg-aster. At the beginning
of this movement the sperm-centrosome is not visible ; later it becomes
distinctly visible, and often very large, giving evidence of variations
comparable in a measure to the variation in the egg-centrosome of the
first maturation spindle of Liinnrea.

Occasionally the sperm-centrosome of Limnasa divides into two while
on its way toward the egg-nucleus, a process prophetic of the existence
of these centrosomes at either pole of the first cleavage spindle.

Ou account of the fact that in Limnaea both egg-centrosome and
sperm-centrosome at certain stages disappear for a time, it is impossible
to ascertain the source of the first cleavage spindle by following the
history of the centrosomes. However, reasonably satisfactory proof
that the first cleavage spindle is wholly of spermatic origin is found in
the facts that the incipient cleavage spindle involves the sperm-nucleus
first, and that the egg-nucleus, so far as my observations have extended,
is never involved in the spindle before the sperm-nucleus is entirely
drawn upon it.

2-46 bulletin: museum of compakative zoology.

BIBLIOGRAPHY.

Boveri, T.

"87. Zellen-Studien. Heft 1. Die Bildung der Richtungskorper bei Asearis
megalocephala und Asearis lumbricoides. Jena. Zeit. Bd. 21, pp. 423-
515. Taf. 25-28. Also separate. 93 pp., 4 Taf. Jena, 1887.

Boveri, T.

'90. Zellen-Studien. Heft 3. Ueber das Verhalten der chromatischen
Kernsubstauz bei der Bildung der Richtungskorper und bei der Befruch-
tung. Jena. Zeit. Bd. 24, pp. 314-401. Taf. 11-13. Also separate.
88 pp., 3 Taf. Jena, 1890.

Conklin, E. G.
'94. The Fertilization of the Ovum. Biol. Lectures, Mar. Biol. Lab. Wood's
Holl. Vol. 2, pp. 15-35. 10 Figs, in text.

Pick, R.

'93. Ueber die Reifung und Befruchtung des Axolotleies. Zeit. f. \riss. Zool.
Bd. 56, pp. 529-614. Taf. 27-30.

Foot, K.
'97. The Origin of the Cleavage Centrosomes. Jour. Morph. Vol. 12,
pp. 809-814. PI. 39.

Garnault, P.

'88. Sur les phenomenes de la fecondation chez l'Helix aspersa et l'Arion
empiricorum. 1. Zool. Anz. Jahrg. 11, pp. 731-736. 2. Zool. Anz.
Jahrg. 12, pp. 10-15. 3. Zool. Anz. Jahrg. 12, pp. 33-3S.

Hill, M. D.

'95. Notes on the Fecundation of the Egg of Sphserechinus granulans, and
on the Maturation and Fertilisation of the Egg of Phallusia mammillata.
Quart. Jour. Micr. Sci. Vol. 38, pp. 315-330. PI. 17.

Klinckowstrom, A. von.

'97. Bcitrage zur Kenntniss der Eireifung und Befruchtung bei Prostheceraeus
vittatas. Arch. f. mikr. Anat. Bd. 48, Heft 4, pp. 587-605. Taf. 28-29,
3 Textfiguren.

Kofoid, C. A.

'95. On the Early Development of Limax. Bull. Mus. Comp. Zool. Harvard
Coll. Vol. 27, No. 2, pp. 35-118. PI. 1-8.

LINVILLE: PULMONATE GASTEROPODS. 247

Kostanecki, K.
'96. Ueber die Gestalt der Centrosomen im befruchteten Seeigelei. Anat.
Hefte, Abth. 1, Heft 22 (Bd. 7, Heft 2), pp. 217-237- Taf. 13-14.

Kostanecki, K., und Siedlecki, M.
'96. Ueber das Verhaltniss der Centrosomen zura Protoplasma. Arch. f.
• mikr. Anat. Bd. 48, pp. 181-273. Taf. 10-11.

Kostanecki, K. von, und Wierzejski, A.

'96. Ueber das Verhalten der sogen. achromatischen Substanzen im befruch-
teten Ei. Nach Beobachtungen an Pbysa fontinahs. Arch. f. mikr. Anat.
Bd. 47, pp. 309-386. Taf. 18-20.

Mark, E. L.

'81. Maturation, Fecundation, and Segmentation of Limax campestris, Bin-
ney. Bull. Mus. Comp. Zool. Harvard Coll. Vol. 6, No. 12, pp. 173-625.
Pis. 1-5.

Platner, G.
'86. Ueber die Befrucbtung bei Arion empiricorum. Arch- f. mikr. Anat.
Bd. 27, pp. 32-72. Taf. 5-6.

Sobotta, J.
'95. Die Befrucbtung und Furcbung des Eies der Maus. Arcb. f. mikr. Anat.
Bd. 45, pp. 15-93. Taf. 2-6.

Wheeler, W. M.
'95. The Behavior of the Centrosomes in the Fertibzed Egg of Myzostoma
glabrum, Leuckart. Jour. Morph. Vol. 10, No. 1, pp. 305-311. 10 Figs,
in text.

Wilson, E. B., and Mathews, A. P.

'95. Maturation, Fertilization, and Polarity in the Echinoderm Egg. Jour.
Morph. Vol. 10, No. 1, pp. 319-342. 8 Figs, in text.

v

248 bulletin: museum of compaeative zoology.

EXPLANATION OF PLATES.

The drawings on Plates 1, 2, and 4 represent a magnification of 575 diameters ;
those on Plate 3 about 400 diameters. Projection of the image was made by an
Abbe' camera lucida. All figures except 3, 5, and 9 are composites of structures
extending through several, in some cases nine, sections.

Linville. — Maturation Pulmonates.

PLATE 1.

All Figures are of eggs of Limncea elodes.

Fig. 1. First maturation spindle, telophase. Centrosouies of large size. Sperm-
head near vegetative pole. Breaking of membrane at animal pole acci-
dental; the flattening of the egg normal.

Fig. 2. First maturation spindle. Portions of the membrane of germinative vesi-
cle still visible. Form of deep aster of spindle modified by sperm-aster.

Fig. 3. Deep centrosome and aster, to show the rays beginning at centrosome and
extending through and beyond the centrosphere.

Fig. 4. First maturation spindle. The sixteen chromosomes (concealed tetrads)
splitting longitudinally. Deep and peripheral centrosomes unlike.

Fig. 5. Sperm-aster and sperm-nucleus. Nucleus elongated in the direction of its
aster, and pointed. Position of egg-aster indicated by dotted outline.

Fig. 6. Germ-nuclei fused into a single (?) cleavage nucleus. First cleavage spin-
dle forming. Chromosomes being drawn to the spindle by the rays
which penetrate the interior of the fused germ-nuclei. Occasional
nucleoli show chromosomes attached to them.

LLE.— MaTURA] :aTES.

Plate I.

•X

,1 Hi »',•!*

#

y/.

■■■*'

*

-■ •■■: • •".'•..

-

- ■

ttWKlW**

3.

■

4

■

6.

L id

Heliotype

Linvillb. — Maturation Pulmouates,

PLATE 2.

All Figures are of eggs of Limncea elodes.

Fig. 7. First polar cell destroyed. Its position indicated by dotted outline. En-
larging of centrosphere begun. Ceutrosome of sperm-aster divided into
two, but surrounded by a single system of astral rays. Sperm-head
pointed on side toward the distant sperm aster.

Fig. 8. First polar cell ; the peripheral centrosphere disintegrating. Deep centro-
some surrounded by a small spherical centrosphere ; this surrounded, in
turn, by an eccentric larger centrosphere.

Fig. 9. Transverse section through the peripheral chromosomes of the first
maturation spindle. Sixteen chromosomes present.

Fig. 10. To show large sperm-eentrosome and, near by, the slightly vacuolated
sperm-head. Another sperm-head at periphery of egg. The polar cell
in outline projected on the plane of this section. The egg-nucleus is
not shown, as it lies in another section.

Fig. 11. Second maturation spindle. Sixteen dumb-bell shaped chromosomes
(dyads) about to divide by transverse division. Constructed from suc-
cessive sections.

Fig. 12. Only one polar cell (second) recognized. Its relation to egg obscured
by accidental presence of a foreign body. Formation of first cleavage
spindle.

Fig. 13. Telophase of second maturation spindle. Deep centrosphere surrounded
by an eccentric outer centrosphere. Sperm-nucleus without sperm-aster.

Linvllle.— Maturation. Pulmonati Plate 2.

m

£&*/,

9

■ ':■-

■

'',M-<

f"^

&

•*?'

JO. /J.

I I • ■ » • •

■ ■ ■

■ ■■ ; •

0 ■■-•;. :> -

MP I

Linville. — Maturation Pulmonates.

PLATE 3.

Fig. 14. Limax agrestis. First maturation spindle, germinative vesicle having just
disappeared. Head and tail of spermatozoon still connected.

Fig. 15. Limnceasp.? First polar cell remains. Second polar cell formed, but broken
away from the egg; its " Zwischenkurper " remains near that of the
first polar cell. Sperm-head just broken off from the tail, and with base
turned toward the egg-nucleus. The sperm-head lies in a vacuole. Two
centrosomes (?) may be seen each in a bundle of rays in the vacuole on
the side away from the proximal end of the tail.

Fig. 16. Limax maximus. Formation of the first polar cell. Centrosphere with thick
reticulated wall ; no centrosome is visible. Clear area of centrosphere
finely reticulated. Sperm-head in lower part of figure.

Fig. 17. Limax maximus. First polar cell just formed. Vacuolation or enlarging
of centrosphere begun. Two centrosomes (?) in the centrosphere.
Sperm-head in usual vacuole. Sixteen dyads in first polar cell.

Fig. 18. Limax maximus. Centrosomes about to form the first cleavage spindle.

Fig. 19. Limax agrestis. Polar cells lost. First cleavage spindle forming.

Fig. 20. Limax maximus. First polar cell lost ; second polar cell still united to the
egg. The germ-nuclei are nearly surrounded by numerous small bodies
resembling centrospheres.

Fig. 21. Limax maximus. First polar cell just formed. Enlargement of centro-
sphere further advanced than in Figure 17. Two minute centrosomes
with radiations within the centrosphere. Sperm-head with problematic
centrosome within the sperm vacuole Blunt end of tail is the one to
which the sperm-head was attached.

Fig. 22. Limax maximus. Telophase of first cleavage spindle. Nuclear membrane
already formed about the chromosomes. Centrosomes divided in anti-
cipation of the next cleavage.

' [AXES.

*

•#

Plate 3.

74

75

.■■■*: ■'»

7<2

7/

3L

7v

<?<2

' .- . .'*

//>.

•■ \

> • '• • ■<.*

*$*£&;

'

V

■ ,*A

■ A

.

■ - - - , \

•.

. £

■

/

:■ . ■

•V- .

V'» ..••.-. ■ ,v

,S03tO

Linvillb. — Maturation Pulmonates.

PLATE 4.

All Figures, except Figure 25, are of eggs of Limncea elodes.

Fig. 23. First maturation spindle. At either pole of the spindle is collected a mass

of precipitate, the whole resembling a so-called centrosphere.
Fig. 24. First maturation spindle. Apparently the centrospheres were in the process

of being decolorized when the process of decolorizing was checked.
Fig. 25. Limax agrestis. The germinative vesicle contains two nucleoli and

" Linin " fibres with elements of future chromosomes arranged in

" skein " stage.
Fig. 26. First cleavage spindle. Structures that seem to be the sperm-chromosomes

now occupy the middle of the spindle, and the egg-chromosomes are

about to be drawn into it.
Fig. 27. First cleavage spindle. Rays near the incipient spindle are beginning to

penetrate the sperm nucleus, while the egg-nucleus is still unaffected.

The chromosomes are not yet formed.
Fig. 28. First cleavage spindle. The sperm-chromosomes occupy the whole breadth

of the spindle, while the egg-chromosomes are at one side of the spindle.

The granular funnel beneath the polar cells probably represents the path

through which the egg-chromosomes were drawn to the spindle.

Linville. — Maturatio'n Pulmonates.

E 4.

:v

* •>

■3

-

^

i><3

24.

»&*•

■'V. *i>

I I ©J

V

-■tsvuj.V.

4<»"!

c-c

<?«£

.^:

>z

URL

28.

Heliotype Co.,B<

The following Reports have been published or are in prepa-
ration on the Dredging Operations off the West Coast of
Central America to the Galapagos, to the West Coast of
Mexico, and in the Gulf of California, in Charge of Alex-
ander Agassiz, carried on by the U. S. Fish Commission
Steamer "Albatross," during 1891, Lieut. Commander Z. L.
Tanner, U. S. N., Commanding.

A. AGASSIZ. II.1 General Sketch of the
Expedition of the "Albatross," from
February to May, 1891,

A. AGASSIZ. The Pelagic Fauna.

A. AGASSIZ. The Deep-Sea Panamic Fauna.

A. AGASSIZ. I." On Calamocrinus, a new
Stalked Crinoid from the Galapagos.

A. AGASSIZ. XXIII.23 The Echini.

JAS. E. BENEDICT. The Annelids.

R. BERGH. XIII.13 The Nudibranchs.

K. BRANDT. The Sagittee.

K. BRANDT. The Thalassicolse.

C. CHUN. The Siphonophores.

C.CHUN. The Eyes of Deep-Sea Crustacea.

S. F. CLARKE. XI.11 The Hydroids.

W. H. DALL. The Mollusks.

W. FAXON. VI.3 XV." The Stalk-eyed
Crustacea.

VRMAN. XXVI.28 The Fishes.
ESBRECHT. XVI.« The Copepods.
JES. III.* XX. 20 The Foraminifera.
H. J. HANSEN. XXII.22 The Cirripeds

and Isopods.
C. HARTLAUB. XVIII.'8 The Comatulse.
W. A. HERDMAN. The Ascidians.
S. J. HICKSON. The Antipathids.
W. E HOYLE. The Cephalopoda.
G. von KOCH. The Deep-Sea Corals.
C. A. KOFOID. Solenogaster.
R. von LENDENFELD. The Phosphores-
cent Organs of Fishes.

H. LUDWIG. IV.« XII." The Holothu-

rians.
C. F. LUTKEN and TH. MORTENSEN.

XXV.25 The Ophiuridse.
OTrO MAAS. XXI 21 The Acalephs.
J. P. McMURRICH. The Actinarians.
E.L.MARK. XXIV.24 The Cerianthidse.
GEO P. MERRILL. V.8 The Rocks of the

Galapagos.
G. W. MULLER. XIX.19 The Ostracods.
JOHN MURRAY. The Bottom Specimens.
A. ORTMANN. XIV. 12 The Pelagic Schi-

zopods.
ROBERT RIDGWAY. The Alcoholic Birds.
P. SCHIEMENZ. The Pteropods and Het-

eropods.
W. SCHIMKEWITSCH. VIII.8 The Pyc-

nogonidse.
S. H. SCUDDER. VII.' The Orthoptera of

the Galapagos.
W. PERCY SLADEN. The Starfishes.
L. STEJNEGER. The Reptiles.
TH. STUDER. X.10 The Alcyonarians.
C. H. TOWNSEND. XVII." The Birds of

Cocos Island.
M. P. A. TRAUTSTEDT. The Salpidee and

Doliolidse.
E. P. VAN DUZEE. The Halobatida.
H. B. WARD. The Sipunculoids.
H. V. WILSON. The Sponges.
W. McM. WOODWORTH. IX.9 The Pla-

narians and Nemerteans.

W. MoM. WOODWORTH. XXVII.2'
Planktonemertes.

t

Bull

M.

C.

7,

89 pp

, 22 Plates.

>2

Mem

M.

C.

z

3

Bull.

M.

c.

z

4

Bull.

M.

c.

z

5

Bull.

M

c.

z.

6

Bull.

M.

c

z.

7

Bull.

M.

c.

/,.

8

Bull.

M.

0

z

9

Bull

M

C

z

10

Bull.

M.

C.

z

Vol. XXI., No. 4, June, 1891, 16 pp. ; and Vol. XXIII., No 1, February, 1892,

, Vol. XVII., No. 2, January, 1892, 95 pp.. 32 Plates.
, Vol XXIV., N •. 7, August, 1893, 72 pp.
, Vol. XXIII., No. 5, December, 1892, 4 pp., 1 Plate.
Z., Vol. XXIV.. No 4, June, 1893, 10 pp [Zool. Anzeig., No. 420, 1893.]

Vol. XVI., No. 13, July, 1893, 3 pp.

Vol. XXV , No. 1 September, 1893, 25 pp.

Vol. XXV., No. 2. December, 1893, 17 pp., 2 Plates.

Vol. XXV., No. 4, January, 1894, 4 pp., 1 Plate.

Vol. XXV., No. 5, February, 1894, 17 pp.

11 Bull.'M. C Z., Vol. XXV., No. 6, February, 1894, 7 pp., 5 Plates.

12 Bull. M. C. Z., Vol. XXV., No. 8, September. 1894, 13 pp., 1 Plate.

13 Bull. M C Z , Vol XXV.. No. 10, October, 1894, 100 pp , 12 Plates.
Z., Vol. XVII., No. 3, October, 1894, 183 pp., 19 Plates.
Z , Vol XXV., No. 12, April, 1895, 20 pp., 4 Plates.

Vol. XVTTI., April, 1895. 202 pp., 67 Plates, 1 Chart.

Vol XXVH , No. 3, July, 1895, 8 pp., 2 Plates.

Vol XXVII., No 4, August, 1895, 26 pp., 3 Plates.

Vol. XXVII., No. 5, October, 1895, 14 pp.. 3 Plates.

Vol. XXIX., No 1, March, 1896, 103 pp., 9 Plates, 1 Chart.

Vol. XXI II., No. 1, September, 1897, 92 pp., 15 Plates.

Vol. XXXI , No 5, December, 1897. 37 pp., 6 Plates, 1 Chart.

Vol. XXXII., No. 5, May, 1898, 18 pp., 13 Plates, 1 Chart.
Arol. XXXII., No. 8, August, 1898, 8 pp., 3 Plates.

Vol. XXII , No. 2, November. 1899. 116 pp., 22 Plates, 1 Chart.

Vol. XXIV , December, 1899. 431 pp., 97 Plates, 1 Chart.

Z
Z
Z

'* Mem. M. C
'5 Bull. M. C
'5 Mem. M. i'
" Bull. M. C
" Bull. M. C

19 Bull M C. Z

20 Bull M. 0 Z

21 Mem M C. Z.

22 Bull. M. H. Z

23 Bull. M. C. Z
2* Bull. M C. Z
25 Mem. M. C. Z
28 Mem. M O. Z

2' Bull. M. C. Z., Vol. XXXV., No. 1, June, 1899, 4 pp., 1 Plate.

PUBLICATIONS

OF THE

MUSEUM OF COMPARATIVE ZOOLOGY

AT HARVARD COLLEGE.

There have been published of the Bulletins Vols. I. to XXXV. ;
of the Memoirs, Vols. I. to XXIV.

Vol. XXXVI. of the Bulletin, and Vol. XXV. of the Memoirs,

are now in course of publication.

The Bulletin and Memoirs are devoted to the publication of
original work bjT the Professors and Assistants of the Museum, of
investigations carried on by students and others in the different
Laboratories of Natural Histoiy, and of work by specialists based
upon the Museum Collections.

The following publications are in preparation : —

Reports on the Results of Dredging Operations from 1877 to 1880, in charge of
Alexander Agassiz, by the U. S. Coast Survey Steamer " Blake," Lieut.
Commander 0. 1). Sigsbee, U. S. N., and Commander J. R. Bartlett, U. S. N.,
Commanding.

Reports on the Results of the Expedition of 1891 of the U. S. Fish Commission
Steamer " Albatross," Lieut. Commander Z. L. Tanner, U. S. N., Com-
manding, in charge of Alexander Agassiz.

Contributions from the Zoological Laboratory, in charge of Professor E. L.
Mark.

Contributions from the Geological Laboratory, in charge of Professor N. S.
Shaler.

Studies from the Newport Marine Laboratory, communicated by Alexander
Agassiz.

Subscriptions for the publications of the Museum will be received
on the following terms : —

For the Bulletin, $4.00 per volume, payable in advance.

For the Memoirs, $8.00 " " "

These publications are issued in numbers at irregular inter-
vals ; one volume of the Bulletin (8vo) and half a volume of the
Memoirs (4to) usually appear annually. Each number of the Bul-
letin and of the Memoirs is also sold separately. A price list
of the publications of the Museum will be sent on application
to the Librarian of the Museum of Comparative Zoology, Cam-
bridge, Mass.

t

Harvard MCZ Libra

3 2044 066 300 450

Do not circulate