2 ‘ 43 


Digitized by the Internet Archive 
in 2009 with funding from 
University of Toronto 


http://www.archive.org/details/nsquarterlyjourn6Ocomp 


Pig ‘ Nes A : arin 


Lf 7 
7h nd 
i] 


us 


a a 
iN ae a 
Gian Pane my i 
i" 


ae 
ia 
oa v 
ye 
fi 


az 


a jt 
Sy a) 
ay ws aay 
Bus 


QUARTERLY JOURNAL 


OF 


MICROSCOPICAL SCIENCE. 


EDITED BY 


pir RAY LANKESTER, K.C.B., M.A., D.Sc., LL.D., F.RB.S., 


NONOKARY FELLOW OF EXETER COLLKEGK AND HONOKAKY STUDENT OF CHRIST CHURCH, OXFORD; 
MEMBER OF THK INSTITUTK OF FRANCK (ASSOCIE ETRANGER DE 1’ACADEMIE DES SCIENCES) 3 
CORKKEKSPONDENT OF THE IMPERIAL ACADKMY OF SCIKNCKS OF 8ST. PETEKSBUKG, AND OF THR 
ACADKMY OF SCIKNCKS OF PHILADELPHIA, AND OF THE KOYAL ACADEMY OF SCIENCES 
OF TURIN FORKIGN MEMBER OF THK KOYAIL SOCIETY OF SCIENCES OF 
GOTTINGEN, AND OF THER KOUYAL BOUEMIAN 8SO0CIKTY OF SCIENCES, AND 
OF THK ACADEMY OF THE LINCEI OF KOMK, AND OF THK AMERICAN 
ACADEMY OF AKTS AND SCIENCES OF BOSTON . ASSOCIATE OF THK 
ROYAL ACADEMY OF BELGIUM. HONORAKY MEMBER OF THK 
NEW YORK ACADEMY OF SCIKNCES, AND OF THK 
CAMBKIDGE PHILOSOPILLCAL SOCIKRTY, AND OF 
YUK ROYAL PHYSICAL SOCIETY OF EDIN- 

BURGH, AND OF THE 
BIOLOGICAL SOCIKTY OF PAKIS, AND OF THR CALIFOKNIA ACADEMY OF SCIKNCES OF SAN FRANCISCO, AND 
OF THK KOYAL ZOOLOGICAL AND MALACOLOGICAL SOCIETY OF BELGIUM; 
COKKESPONDING MEMBER OF TUK SENKENBERG ACADEMY OF FRANKFURT-A-Mj3 
FORKKIGN ASSOCIATK OF TILK NATIONAL ACADEMY OF SCIENCES, U.S., AND MEMBER OF THK 
‘ AMERICAN PHILOSOPHICAL SOCIRTY 5 
HONOKARY FELLOW OF THE ROYAL SOCIETY OF EDINBURGH; 

LATE PIKKCTOK OF THK NATUKAL HISTORY DEPARTMENTS OF THE BRITISH MUSEUM, LATE PRKSIDENT oF THE 
BKITISH ASSOCIATION FOR THE ADVANCEMENT OF SCIENCR; LATK FULLRKIAN PRKOFRSSOR OF 
PHYSIOLOGY IN THK ROYAL INSTITUTION OF GREAT BRITAIN, 

LATK LINACKK PROFPKSSOR OF COMPARATIVE ANATOMY AND FELLOW OF MERTON COLLEGRK, OXFORDS 
EMEKITUS PROFESSOR OF ZOOLOGY AND COMPARATIVE ANATOMY IN UNIVERSITY COLLEGEK, UNIVERSITY OF LONDON 


WITH THE CO-OPERATION OF 


SYDNEY J» HICKSON, McA. FR 


BEYRR PROFKSSOK OF ZOOLOGY IN THK UNIVERSITY OF MANCHKRTKK, 


E.. A. MINCHIN, M.A., F.RS., 


PROFESSOK OF PROTOZOOLOGY IN THE UNIVERSITY OF LONDON; 


GILBERT C. BOURNE, M.A., D.Sc., F.R.S., 


LINACKEK PROPRKSSOK OF COMPAKATIVE ANATOMY, AND FKLLOW OF MERTON COLLEGE, OXFORD 5 


D 


J. GRAHAM KERR, M.A., F.R.S., 


KEGIUS PROFESSOR OF ZOOLOGY IN THK UNIVERSITY OF GLASGOW, 


VOLUME 60.—New SeErRIEs. 


| a 
LONDON: 
J. & A. CHURCHILL, 7, GREAT MARLBOROUGH S 
1915. 


‘REET. 


CO NEE NTs: 


CONTENTS OF No. 237, N.S., APRIL, 1914. 


MEMOIRS: 
On the Anatomy of Conus tulipa, Linn., and Conus textile, 
Linn. By H. O. N. SHaw, B.Se., F.Z.S. (With Plates 1-6 and 
12 Text-figures) 1 
On the Relation between the smruchure and bie Denelontent of 
the Centrifuged Egg of the Frog. By J. W. Jenxinson, M.A., 


oD 
D.Se., University Lecturer in Embryology, Oxford ; Fellow of 
Exeter College. (With Plates 7-12 and Text- Gears 1-18) GE 


The Sporogony and Systematic Position of the Aggregatide. By 
Hewen L. M. Prxetu-Goopricn, B.Sc., Beit Memorial Research 
Fellow. (With Plate 13) . : : : 5 BY: 


CONTENTS OF No. 238, N.S., JUNE, 1914. 
MEMOIRS : 


The Blastocyst and Placenta of the Beaver. By ArrHur WILLEy, 
F.R.S., Strathcona Professor of Zoology in McGill University, 
Montreal. (With Plates 14-21 and 6 Text-figures) . 5 ies 

On Acrossota liposclera, a New Genus and Species of Alcyo- 
narian with Simple Tentacles. By GiLtBerr C. Bourne, M.A., 
D.Sc., F.R.S., Linacre Professor of Comparative Anatomy and 


Fellow of Merton College, Oxford. (With Plate 22) . . 261 
The Proboscidian System in Nemertines. By Dr. Grrarpa 
Wynuorr, Utrecht. (With 36 Text-figures) ; PAB? 


CONTENTS OF No. 239, N.S., SEPTEMBER, 1914. 
MEMOIRS : 

Observations on the Gametogenesis of Grantia compressiz. 
By Arruur Drnpy, D.Sc., F.R.S., Professor of Zoology in the 
University of London (King’s College). (With Plates 23-26). 313 

The Chromosome Complex of Culex pipiens. By Monica 
Taytor, S.N.D., B.Sc. (With Plates 27 and 28 and 3 Text- 
figures) ; : E ; > oat 


iV CONTENTS. 


PAGE 

Studies on Avian Hemoprotozoa: No. III.—Observations on the 
Development of Trypanosoma noctue (of the Little Owl) 
in Culex pipiens; with Remarks on the Other Parasites 
occurring. By H. M. Woopncock, D.Se.Lond., Assistant to the 
University Professor of Protozoology. (With Plates 29-31 and 

1 Text-figure) . ; : ; : ; 
Studies in the Experimental Analysis of Sex. Part 11.—On 
Stylops and Stylopisation. By Grorrrey Smiru, M.A., Fellow 
of New College, Oxford, and A. H. Hamm, Assistant in the 
Hope Department, Oxford. (With Plates 32-35) : . 435 


399 


CONTENTS OF No. 240, N.S., JANUARY, 1915. 


MEMOIR: 
The Rat-Trypanosome, Trypanosoma lewisi, in its Relation 
to the Rat-Flea, Ceratophyllus fasciatus. By E, A. 
Mrncuin and J. D. Tuomson. (With Plates 36-45 and 24 Text- 
figures) s : : . 463 


Tire, INDEX, AND CONTENTS. 


CONUS TULIPA, LINN., AND CONUS TEXTILE, LINN. ] 


On the Anatomy of Conus tulipa, Linn., and 
Conus textile, Linn. 


By 
H. oO. N. Shaw, B.Sc., F.Z.8. 


With Plates 1 to 6, and 12 Text-figures. 


Since 1895, few workers on the anatomy of mollusca have 
devoted their attention to the genus Conus. Inthat year Dr. 
Bergh (8) published an extensive memoir on a large number 
of species in this genus, and his work may be considered as 
the most complete, and embracing the greatest number of 
species examined, though his description of each species was 
not exhaustive. ‘T'roschel (20) devoted most of his attention 
to the radulz of the different genera and species of which his 
excellent work is composed, and although he gives a certain 
number of figures with descriptions of various anatomical 
points, these latter are for the most part of rather a crude and 
diagrammatic kind. 

While malacologists have done a certain amount towards 
working out and elucidating the anatomy of various members 
of this genus, the conchologists, as is generally the case, have 
produced many excellent monographs, and such names as 
Reeve, Sowerby, Tryon, Weinkauff and others will always be 
remembered for the general excellence of their figures and 
descriptions of the numerous species which are contained in 
this genus. Various writers have essayed different forms of 
classification, but for the most part on purely conchological 
grounds, and when more is known about the inhabitants of 
these shells, and their different points of resemblance to one 

vot. 60, PART 1.—NEW SERIES. 1 


2 H. 0. N. SHAW. 


another, whether they vary as much or more than their shells, 
and if the conchological grouping also holds good from an 
anatomical point of view, we shall be on the high road to 
founding a solid and logical form of classification. 

The two species with which I shall deal in this paper are 
Conus tulipa, Linn.,and Conus textile, Linn. The speci- 
mens of both species were females. The Conidz belong to 
the order Prosobranchiata of the class Gastropoda, are 
dicecious, and according to the most widely accepted form of 
classification, Conus tulipais included in that sub-genus, or, 
as some would have it, section, of the genus Conus, called 
Rollus by Montfort, while Conus textile forms the type 
of the sub-genus or section Cylinder, of the same author. 

Rollus was first described by Montfort in 1810 (15, p. 395), 
with Conus geographusas the type, and the sub-genus was 
characterised as follows: “ Shell light, sub-cylindrical, spire 
short but pointed at the summit, whorls slightly coronated, 
aperture effuse, emarginate in front, columella smooth, outer 
lip with a wide but not deep notch at the suture.” 

This group corresponds to Nubecula of Klein, 1753 ; but 
owing to his being pre-Linnean and non-binomial, this designa- 
tion cannot be accepted. Utriculus, of Schumacher, 1817, 
and Tuliparia, Swainson, 1840, are synonymous. 

Cylinder was described on p. 391 of the same work. 

“Shell sub-conic, smooth, spire elevated, pointed, whorls 
numerous, body whorl ventricose, notched at the suture, 
aperture effuse at the fore part.”’ 

Textilia, Swainson, 1840,is a synonym. 

One of the difficulties to be contended with in working on 
the anatomy of tropical molluscs is the trouble in getting 
sufficient material, and in a good state of preservation. The 
two specimens here described are from the British Museum 
(Natural History), and had been there in spirit fora good many 
years, and I had started working on these before some recently 
collected specimens came to hand. 

My thanks are due to J. Hornell, Esq., Pearl and Chank 
Fisheries, Tuticorin, for sending me specimens which I have 


CONUS TULIPA, LINN., AND CONUS TEXTILE, LINN. 3 


not yet worked out, also to E. A. Smith, Hsq., I.8.0., of the 
Natural History Museum, for kindly allowing me to use some 
of the Museum specimens. My especial thanks are due to 
Prof. G. C. Bourne, Merton College, Oxford, for his kindly 
help and advice on many points connected with this paper, for 
which I am much indebted. 

With the exception of Conus mediterraneus, which is 
found all round the Mediterranean and west coast of Spain, 
no other species inhabit European waters, though this large 
genus is plentifully represented in the tropical seas, and round 
the coasts of Australia, Japan and America. It generally 
lives in fairly shallow water, and is found on reefs and in pools 
under stones, corals, etc., and is supposed to be able to inflict 
a poisonous bite. I have made inquires from those who have 
collected and handled them alive. They tell me that they 
have never had this experience.! The animal is extremely 
timid ; on the slightest touch it withdraws itself into its shell, 
and will remain in this retracted condition for a considerable 
time. 

The operculum is generally elongate or unguiform, and so 
small that it is useless for closing the mouth of the shell when 
the animal has withdrawn itself inside. 

The shells are covered with yellowish periostracum, which 
in some species is only a thin, smooth, transparent, but tough 
coating. In others, as in C. tulipa, the periostracum is 
exceedingly thick and of a dark-brown colour. It is rough, 
furrowed longitudinally, and of a leather-like texture, and has 
tufts or outgrowths disposed in even rows along its surface. 
When dry, this thick periostracum becomes very brittle and 
peels off the shell. 

No doubt this shell covering, which, as I have said, in some 
species is tufted and of a leathery formation, is both a 
protection to the shell and also a form of protective colora- 
tion for the animal. C.tulipa has one of the lightest and 

' While this paper was in the press, a note appeared in ‘ The Nautilus,’ 
vol, xxvii, pt. 10, pp. 117-120, 1914, ‘‘ Poisoning by the Bite of Conus 
geographus.” 


4. H. O. N. SHAW. 


thinnest of shells, and at the same time one of the thickest 
and most tufted periostraca. As the animal grows, and 
outer whorls are formed surrounding the inner ones, the walls 
of the internal convolutions of the shell are so reduced in 
thickness as to be hardly as thick as a sheet of paper, and 
semi-transparent. 


Description oF Conus TEXTILE. 


Maximum length of shell 24} in., maximum breath 1¢ in. 
Operculum long and narrow, and slightly reflected outwards 


Trxt-Fic. 1. 


inv: 


Conus textile. The siphon, with one side reflected to show 
ridges and furrows, and invagination (cnv.). 


at its anterior end. The foot, which is large and muscular, 
when withdrawn completely fills the opening of the shell. 
Along the internal edge of the foot (next the columella) is a 
row or ridge of large tubercles or nodules of a reddish-brown 
colour. he opening of the branchial cavity extends along 
the right-hand side of the animal, and the cavity is enclosed 
on its dorsal surface by a thin covering of membrane which 
along its outer or free edge is considerably thickened and 
muscular for a space of about Jin. This covering is attached 
at its posterior end to the body-wall, and has no attachment 
forward till it reaches the siphon, round which itisfused. It 
extends forward as a kind of flap beyond its point of attach- 
ment round the siphon, and overhangs the latter about + in. 

The siphon (Text-fig. 1)"is large, and has thick muscular 


CONUS TULIPA, LINN., AND CONUS TEXTILE, LINN. 3) 


walls which are folded over and form a funnel, open in front 
and beneath. The funnel isin the shape of a triangle, with its 
base above, and apex ventral, the apex being formed by the 
two free edges of the siphon, which run parallel and slightly 
to the right of the foot. The interior and dorsal surface of 
the siphon has a number of deep grooves and ridges running 
across and at right angles to its longitudinal axis, and also 
extending down its sides. They start near the anterior and 
free end, and continue about half its length backwards. At 
its junction with the body, the ventral edge of the left-hand 
fold of the siphon is sharply reflected upwards and at right 
angles into the funnel, forming a V-shaped invagination or 
elbow (wv.), which closes up more than half the passage, and 
fits into a depression in the foot. 

The eyes are situated on the tentacles, about the middle, 
and on the outer side, this position being due to the tubercles 
on which they were borne having fused with the tentacles. 
The anterior end of the foot has a glandular groove running 
at right angles and across it, with a reflex fold above the 
groove. About a quarter of an inch behind this groove, and 
on the ventral surface, and equi-distant from both sides of 
the foot, is the orifice of the pedal gland, or pedal sinus. 

When, as in PI. 1, fig. 1, the branchial cavity (br.c.) is opened 
from above, with the siphon and buccal openings turned away 
from the observer, by cutting through the mantle (ml.) 
longitudinally and about three quarters of an inch to the left 
of the branchial opening, the ctenidium (ct.) is seen starting 
just behind the opening of the siphon (sz.) into the branchial 
cavity, running backwards and across the latter in a curve to 
its dorsal attachment to the mantle, with the “ fausse- 
branchie” on the right, and inside the ctenidium. ‘The 
“ fausse-branchie,” or osphradium, is trifid, the result of 
specialisation from the more archaic form, where the osphra- 
dium is merely a filiform epithelial ridge. This specialisation 
is common to most of the Gastropod Toxoglossa. 

Below the large branchial cavity, and separating it from 
the body-cavity, is a thin membrane or covering, which 


6 H. O. N. SHAW. 


gradually becomes thicker on each side, where it is attached 
to the body-wall. Immediately beneath this membrane, and 
at the posterior end of the body-cavity, lies a large yellowish 
mass, the poison gland (p.g.). Anterior to this latter are the 
numerous coils of the duct (p.g.d.) leading from the gland 
into the cesophagus (w.); and anterior again to these and 
over-lapping them is the radula-sac (7.s.). The cesophagus 
is continued past the radula-sac and so to the mouth (m. h.). 

The duct of the poison-gland enters the cesophagus close 
behind the opening of the radula-sac, and passes backwards 
and to the right of the body-cavity, where it is twisted into a 
large coil. After leaving this coil it runs forwards and 
downwards parallel to the cesophagus, both it and the latter 
organ being surrounded by the nerve collars. Having passed 
backwards through the nerve collar, the duct is composed of 
very numerous and tightly twisted coils and knots, which are 
situated in front of and under the poison gland. The duct 
then straightens, passes to the right across the body-cavity, 
and enters the gland at its right-hand extremity. For about 
half an inch before the opening into the gland the duct is 
much constricted (PI. 1, fig. 2). The whole of the poison duct 
is firmly bound together by connective tissue, which also sur- 
rounds the nerve collars and nerves given off from it, and 
binds all these organs tightly to the cesophagus. ‘The poison 
gland (PI. 1, fig. 2, p. g.) isa long narrow mass, nearly circular 
in section, pointed at each end, and slightly curved. It lies 
directly across the body-cavity with its two curved ends 
pointing downwards. 

It is impossible to completely straighten out the numerous 
kinks and twists of the poison duct (PI. 1, fig. 2) and so measure 
its length accurately, 270 mm. being as near as possible correct. 
The length of gland is 17 mm. and maximum width 5°5 mm. 
The length of the duct is about five times the total length 
of the animal, and the duct and gland together occupy the 
greater part of the body-cavity. 

The salivary gland (PI. 1, figs. 1, 2, s.g.) isa small yellowish, 
rather oval-shaped body, situated to the left of the body- 


CONUS TULIPA, LINN., AND CONUS TEXTILE, LINN. a 


cavity, and in front of the poison ducts. It is provided with a 
pair of extremely fine thread-like ducts (s.d.), which open 
into the right side of the gland, one above and the other 
below. The lower one passes under the cesophagus, and the 
other above, and both enter the base of the V of the radula- 


TEXT-FIGS. 
33. 4. 


bo 


The tops of teeth of Conus textile. bs. Barbs. 7.f. Row of 
denticulations. 


sac (7. s.), one above the other, as in the case of their gland- 
openings. ‘These two ducts lie at right angles to the body axis. 

The radula-sac is in the shape of a V; the right arm is 
elongated and about twice the length of the left, and ends in 
a cul-de-sac; the base is produced downwards and forms a 
knob, the left arm being the opening into the cesophagus. 
The long or right arm lies across the body-cavity and over 
the cesophagus, and behind the nerve collars. It is joined on 
the right by the left arm, which runs forward and downwards 


8 H. O. N. SHAW. 


where it joins the esophagus. The long arm is slightly 
curved, and its length is 20 mm. The radula-sac is thick- 
walled and muscular. 

The radular teeth (Pl. 1, fig. 4) are long, thin-walled tubes 
composed of chitin, and are very brittle, transparent, generally 


Trext-FIGs. 5 and 6. 


A view of the base of two teeth of Conus textile. b.a. Showing 
basal attachments. 


light yellow, but sometimes quite dark. When placed in 
xylol to clear them for mounting purposes, they immediately 
underwent various contortions, tying themselves in knots 
and twisting into circles, etc. They are provided at their 
anterior ends with a flat lance-like point (Text-figs. 2, 3, 4), 
and on each side have a large and hooked barb (bs.), pointing 
backwards. The teeth are slightly curved and the barbs are 
placed one on each side of the curve, the one on the outside 
of the curve being about equi-distant between the one on the 


CONUS TULIPA, LINN., AND CONUS TEXTILE, LINN. 9 


inside and the lance-point. The average total length is 
8-10 mm., and the diameter at the base, above the attach- 
ment,‘2mm. The attachment variesa good deal in shape, but 


TEXT-FIG. 7. 


Conus textile. The base of rostrum, proboscis, mouth and 
esophagus. mh. Mouth, rm. rostrum, ps. proboscis (with 
thick reflected lips), 0. .s. opening of radula-sac into esophagus, 
e., 0.p.d. opening of poison duct (p.g.d.) into esophagus, 
n.c. nerve collars (only one represented). The shortness of the 
proboscis, and thick wall through which the radula-sac opens, 
will be noticed. (Diagrammatic.) 


is generally a swelling of the base (Text-figs. 5, 6, b.a.), and 
thickened in parts. The points and barbs hardly vary at all. 
The lance-points in the radulz I have seen, are not serrated 
as shown by Bergh (8), tab. vi, figs. 143, 144,145. The 
radular teeth are placed in two groups in the radula-sac with 
their bases in the angle formed by the union of the two arms. 


10 H. O. N.. SHAW. 


They are so arranged that one group have their barbs in the 
cul-de-sac of the long arm, while the points and barbs of the 
other protrude beyond the radula-sac opening, and into the 
cesophagus. 

The teeth are fixed or anchored by means of an attachment 
or ligament which is firmly connected to their bases, and also 
to the wall of the radula-sac, but is of sufficient length to allow 
each tooth to move backwards and forwards. The teeth are 
placed in rows one behind the other for a short way up each 
side of the radula-sac but are all of nearly equal length, and 
are surrounded and connected with one another by a stout 
layer of connective tissue. 

The rostrum (Text-fig. 7, and Pl. 1, fig. 1, rm.) is non- 
retractile, and forms what Gray (10) calls a veil ; it is thick- 
walled, muscular, and longitudinally ridged. The open or 
free end of the veil is provided with a single row of tentacles 
(ts.), by means of which the animal can attach itself to its 
prey. The posterior and internal wall of the veil is reflected 
forward, and, with the mouth (mh.), forms a retractile 
proboscis (Text-fig. 7, and Pl. 1, fig. 1, ps.). The edges of the 
mouth or lips are thick, ridged and corrugated and reflected 
internally (Text-figs. 7, 8). The retractile proboscis is in this 
species short, and after passing through the posterior wall of 
the veil, opens out into the cesophagus. The opening of the 
mouth is in the shape of a funnel (‘T'ext-fig. 7), the outer edge 
being formed by the thick lips, and as the cavity of the funnel 
narrows down internally, so the walls get thicker. 

The cesophagus commences behind the posterior wall of 
the veil, where it immediately swells out, and as it increases in 
size, the funnel opening inside becomes smaller and the walls 
exceedingly thick and muscular, until at the base of the funnel 
only a very small passage to the mouth is left. These thick 
cesophageal walls end suddenly as though cut across at right 
angles just in front of where the cesophagus is slightly con- 
stricted owing to its being surrounded by the nerve-collars. 
The walls now become quite thin, with the result that the 
much constricted opening at the end of the funnel suddenly 


CONUS TULIPA, LINN., AND CONUS TEXTILE, LINN. 11 


opens out into a large cavity. The edge of this thick wall 
forms a ring or lip round the opening which hangs down into 
the cavity. The outside edge of the lip is reflected back- 
wards, so forming a minute funnel. The radula-sac opens on 
the right (0. 7. s.) in the thick-walled part, and the poison 
duct (0. p. d.) opens immediately behind it on the same side 
into the thin-walled cavity. Behind this last the cesophagus 


TEXT-FIG. 8. 


Conus textile. The proboscis and mouth showing thick 
reflected lips. The rostrum has been removed. p.s. Pro- 
boscis, mh. mouth, rv. s. radula-sac, p.g.d. poison gland duct, 
e@. esophagus (see diagrammatic, fig. 7). 


is shghtly constricted and the poison duct runs parallel and 
close to it, and both are surrounded by the nerve collars 
(n.c.). The cesophagus now expands into a large and thick- 
walled dilatation which corresponds to the stomach; it is 
nearly circular and passes backwards and dorsad of the liver 
(Pl. 1, fig. 1, /. 1). Here it narrows down, descends over the 
edge of the liver and turns with a sharp bend to the left, and so 
to the anus, situated at the posterior and right end of the 
recto-genital mass (7.g. m.). Where the stomach passes over 
the liver there is a large depression on the dorsal surface of 
the former, in which lies the left side of the poison gland. 


12 H. O. N. SHAW. 


The whole of the stomach, cesophagus and duct are pleated 
or corrugated, with the plications running parallel to their 
length, and the thick and deep ridges of the mouth are 
simply a continuation of these ridges in the cesophagus 
(Pls2, fig. 15). 


HistToLoay. 


On account of the age and state of its preservation this 
specimen was not satisfactory for histological purposes. The 
sections were very hard to cut on account of the preparation 
being macerated, and were very difficult to stain. Various 
stains were tried, and those which gave the best results were 
Ehrlich’s hematoxylin and eosin, but even with this method 
the nuclei were very badly defined, and in some cases were 
unstainable. 


Porson Guanp, oR GLANDE DE LerpBLEIn (PI. 2, fig. 9). 


Down the centre of this gland runs a nearly circular canal 
(cw.). The poison duct opens into this canal at the right- 
hand extremity of the gland. The duct joins the gland in 
front and at right angles, so that the opening into the canal 
(Pl. 2, fig. 14, 0. p. d.) is in reality to the side of one of its 
extremities and not at the end. A considerable portion of 
the coils of duct lies under and so ventral to the gland. 

The canal in the gland is lined internally with non-ciliated 
epithelium (PI. 2, fig. 9, /.ep.}; this is surrounded by a thin ring 
ot circular muscular-fibres (¢. m. s.), external to which is a 
layer of longitudinal muscle-fibres and connective tissue 
(J. m. s.). This is again surrounded by another layer of 
circular muscular tissue (c.m.s.),of about a third the thickness 
of the preceding one. The last or external layer (Ul. m. s.) 
is about three times as thick as the other four layers taken 
together, and is composed of longitudinal muscular fibres and 
connective tissue. In the two layers of longitudinal muscles, 
1. e. the external and the middle layers, the muscle-fibres, 
though all running longitudinally and parallel to the length of 


CONUS TULIPA, LINN., AND CONUS TEXTILE, LINN. 13 


the gland, are disposed in bands of different depth and width 
so that they converge and overlap one another (PI. 2, fig. 14). 
These thick muscular layers forming the walls of the poison 
gland serve to eject the secretion along the duct to its 
opening in the cesophagus whenever its use is required. 


Poison Duct (PI. 2, fig. 10). 


The wall of this duct is built up of two layers: an external 
layer (J. 1. m.) of longitudinal muscular fibres and connective 
tissue, and an internal layer (l.c.m.) of connective tissue and 
circular muscular fibres. The internal layer is about half the 
thickness of the external one. The edges of these layers are 
well defined and do not merge intooneanother. This internal 
layer is again lined with a thick epithelium (ep. cs.), composed 
of very elongated club-shaped cells having basal nuclei which 
vary slightly in their position. The cells are filled with a 
fine granular substance. Interspersed among the cells of the 
epithelium are anumber of granular vesicles. The club cells, 
which are extremely attenuated, are as much as ‘2 m. in length. 
The average external diameter of the duct is “65 m. An 
invagination of epithelium (nv. ca.) hangs down into the free 
passage of the duct. ‘This passage or channel (ca.) along the 
centre of the duct is of a very irregular shape, having deep 
grooves or arms running into the epithelium. 

The numerous coils of the poison duct are firmly bound 
together by connective tissue. 


SromacH anp (Hsopuacus (Pl. 2, fig. 15). 


The stomach is nearly circular in section and the wall is of 
considerable thickness. The exterior layer (J. 1. m.) is 
composed of longitudinal intermingled with a few transverse 
muscle-fibres and connective tissue. Beneath this layer and 
running into it is a complex network of transverse and 
circular muscular fibres (/.c.l.m.),intermingled with connective 
tissue and a few longitudinal fibres. Between this last layer 
and the thin coat of epithelium (J. ep.) lining the internal 


14 H. O. N. SHAW. 


surface of the stomach and cesophagus is a thick layer (J. 1. m) 
of, for the most part, connective tissue and longitudinal muscle- 
fibres. In this there are a certain number of circular and 
transverse fibres disposed about the whole wall of the 
cesophagus and stomach, forming a complex mass of muscular 
fibres and connective tissue. These run to a certain extent 
in layers, one outside the other, and in some places are more 
distinct than in others, but have no definite margins, and 
merge one into the other. 

The epithelium is in a very macerated state, but is not 
detached from the wall. 

The projections or invaginations on the inner surface take 
the form of rounded lobes or ridges and have no very deep 
recesses between them. 

The average external diameter of a transverse section is 
2°75 mm., length4mm. Internal diameter of passage 1 mm., 
length 2°5 mm. 

The sections were by far the worst of all, the stain hardly 
taking, nuclei being invisible. 


SALIVARY GLanp (PI. 2, figs. 16, 17). 


This gland is a small yellowish-white looking body which 
lies to the left of the cesophagus, and is connected with the 
latter by two very fine and twisted ducts, which enter at the 
base of the radula-sac. The maximum length of the gland is 
approximately 4 mm. 

The two ducts (s.d.) opening into it are lined with cubical 
ciliated epithelium, and some little way after their entry into 
the gland branch off into two smaller ducts. These ducts are 
again split up into smaller ones, which in turn are divided up 
into still smaller ducts, so that the whole of the gland has a 
sponge-like appearance, and is composed of an extremely fine 
network of glands and ducts. The two main ducts enter some 
distance into the gland before receiving any branches. 

In the ordinary way one would expect to find these minute 
glands grouped together in little bunches which empty into 


CONUS TULIPA, LINN., AND CONUS TEXTILE, LINN. 15 


one duct, and on account of their resemblance to a bunch of 
grapes are known as “acinous glands.” 

In the case of Conus textile, however, this is not so; the 
glands are unicellular, of various irregular shapes and sizes, 
placed side by side, and each cell has a separate duct 
(dt. op.), which empties into a larger one. 

The cells are filled with a very fine granular substance 
(g. c. c.), which is secreted by their lining, and discharged 
through each individual duct (dt. op.) into larger ones, and so 
eventually finds its way through one or other of the main 
salivary ducts to the cesophagus. 

This system of grouped unicellular glands is extremely rare, 
the acinous arrangement being much more common. Owing 
to the inferior state of the preservation of this material | have 
been unable to work out this point as minutely as I could 
wish. 


Circutatory System or Conus Textite (PI. 5, fig. 23, and 
Pl. 6, fig. 24). 


Owing to my researches on the nervous system, and one 
specimen only being available, I have been unable to do much 
with regard to the blood circulation and supply. There is 
one artery (g.), the only one I have been able to follow out, 
which passes from the heart under the cesophagus in an 
oblique manner to the right. The artery is of considerable 
size, is oval in section, attached to the cesophagus by connec- 
tive tissue, and passes forward with it through the nerve 
collars. The artery now leaves the under-surface of the 
cesophagus, passes to the right and across, anterior to the 
left pedal ganglion, to divide slightly to the right of the latter 
into three large branches, forming a cross, with its arms at 
right angles. The left branch (z.) passes to the base of the 
siphon, the central one (y.) plunges down to the anterior and 
central portion of the foot, while that on the right (p. ft.) 
proceeds to the extreme posterior and dorsal surface of the 
foot, along the right side of the latter, and only a short 


16 H. O. N. SHAW. 


distance below the external surface. About one quarter of its 
length from its posterior extremity a branch emerges from 
this last artery and passes downwards into the foot. There 
are no branches given off from the other two arteries, and in 
all cases they end abruptly. Just before the main artery 
passes through the nerve collar formed by the pleuro-pedal 
connectives, a right-angled branch is given off which is 
directed upwards. After a short space this branch is again 
divided at right angles, the one (j.) passing backwards. This 
latter runs parallel to the main artery (g.), and both are con- 
nected to the esophagus, and, together with it, are surrounded 
by the pleuro-subintestinal collar. The branch artery, having 
passed backward for a short distance beyond this last nerve 
collar, is sharply reflected to the left and then turns forward 
again, forming aU, which in section is flat, and closely 
attached to the cesophagus. From its base and extremity 
three small branches pass over the surface of the cesophagus. 
The artery (w.) forming the ri, it arm of the T, bifurcates, 
and its two branches are attached to the radula-sae on its 
ventral surface. Anterior to the nerve collars a large artery 
(c. ft.) proceeds from the main artery at the point where it 
leaves the cesophagus, passes to the right, across and dorsal 
to the pedal ganglia, and plunges into the centre of the foot, 
giving it the appearance of a pedal nerve. The arteries, as 
I have already stated, are in section oval or flat externally 
(Pl. 5, fig. 23), and are composed of an external sheath of 
areolar and circular muscular tissue, called “tunica adven- 
titia” (/.c.m.), with an internal lining (tunica media) (l.¢.1.m.) 
of circular and longitudinal muscular tissue. This last layer 
is internally coated with a thin lining of endothelium (end.) 


Description or Conus Tura (Pl. 1, fig. 1). 


Maximum length of shell 2 in., maximum breadth 1+ in. 
Operculum oblong, small, and thinner than in C. textile. 
The foot is larger and wider, and the row of tubercles are 
not present ; body-walls thick and muscular. 


CONUS TULIPA, LINN., AND CONUS TEXTILE, LINN. 17 


The siphon is the same as in C. textile, but has not the 
invagination or elbow, and is not so deeply ridged. 

The rostrum is the same as in C. textile, having thick 
muscular walls which are corrugated and ridged throughout 
internally. There is also a single row of tentacles at the open 
end of the rostrum. This latter is in the form of a funnel 
with the constricted open end in front. The walls forming 
the base of the funnel are curved inwards to the centre, 
where they form a free and retractile proboscis ( ps.) which 
is about two thirds the length of the rostrum, the mouth (mh.) 
being situated in the centre of the free end. The rostrum, 
which is annulate, has grooves and ridges of muscle running 
circularly round it on the outside, while internally these are 
lined with strong muscle bands attached to its anterior end 
and to its base, by means of which it can be exserted or 
contracted. The cesophagus (e@.) runs from the mouth back 
towards the stomach in the centre of the ring of muscle 
bands. The muscular bands; are bound firmly to the walls 
and to the cesophagus by connective tissue. 

The cesophagus, on its emergence from the posterior end of 
the proboscis, bends at right-angles to the right side of the 
body-cavity ; it is then sharply reflected on itself (PI. 1, fig. 3), 
becomes slightly larger, and passes back to the centre of the 
body-cavity, where it is constricted, and again takes a sharp 
turn to the right and back to the left. It now rapidly 
becomes larger, and the radula-sac opens on the right side, 
and the poison duct on the same side and slightly behind. 
Here it is again constricted, and, together with the poison 
duct, is surrounded by the nerve collars. The cesophagus now 
opens out into the stomach, which has the form of a flattened 
tube running from the centre backwards to the left, lying 
under the poison duct and gland on the floor of the body-cavity. 
The canal is again constricted, and passes over the dorsal 
surface of the liver (PI. 1, fig. 1, /. /.), where it expands, and 
passes downwards and backwards over the back of the liver 
to the anus. The liver, a large brownish mass, lies on the left 
of the body and under the ctenidium (ct.). Where the stomach 

vou. 60, part 1.—NEW SERIES. 2 


18 H. O. N. SHAW. 


passes over the liver, it has the same depression as C. textile 
on its dorsal surface; in this depression lies the left end of the 
poison gland (pg.). 

The ctenidium is situated on the inside and dorsal surface 
of the mantle (ml.) ; starting in front and on the left side, it 
runs across the branchial cavity (br. c.) and then backwards, 
forming a semi-circle, lying above the liver. 

The osphradium, or “fausse branchie,” is, as in C. textile, 
trifid, being parallel to and on the right of the ctenidium, and 
runs backward about half its length. 

The eyes are placed on tubercles which are fused with the 
tentacles in the same way as in C. textile, one on each side 
of the rostrum on its dorsal surface and on the external edges. 
These tentacles are situated about =, in. behind the opening 
of the rostrum. 

The radula-sac (PI. 1, fig. 3, 7. s.) differs from that in C. 
textile in the fact that the left-hand arm of the V, which 
opens into the cesophagus (#.) through a very constricted 
passage, is much shorter than in the latter species. Moreover, 
the right arm is curved in two places in the shape of an S, 
is shorter and thicker, and the base of the V, instead of being 
expanded so as to form a kind of bulb, in C. tulipa is very 
much larger and forms a sort of triangular hood (h.). The walls 
of the radula-sac are of about the same thickness and texture 
in both species, with this exception, that the hood-like process 
in C. tulipa is much thinner than the rest of the sac. The 
right-hand arm of the radula-sac lies above and across the 
cesophagus. 

Immediately behind and above the radula-sac, and running 
across the body-cavity, are the coils of the poison duct 
(p. g.d.). These coils are much shorter and less twisted than 
in C. textile (Pl. 1, fig. 3). Behind them the poison gland 
(p.g.) lies athwart the body. The poison gland is about twice 
the diameter but shorter in length than that of C. textile. 
The poison duct enters the right extremity of the gland at 
right-angles in the same way in both species (Pl. 2, fio. 14). 
In C. textile it is constricted for about half an inch before its 


CONUS TULIPA, LINN., AND CONUS TEXTILE, LINN. 19 


junction with the gland, while in C. tulipa it is only con- 
stricted just at its entry (Pl. 1, fig. 3). The poison duct pos- 
teriorly to its opening into the cesophagus passes backwards 
through the nerve collars, and forms a large coil on the left of 
the body-cavity ; it then passes across and in front of the gland 
to the right side, where there is another coil. From thence it 
passes under the gland and enters on the right. The length 
of duct after unravelling the coils is about 115 mm., or 2,°, 
times the total length of the animal. In C. textile the 
length of duct is 270 mm., or considerably more than twice 
as long. The length of the gland is 15°5 mm., diameter 
8&3 mm., as compared to 17 mm. and 5°5 mm. in C. textile. 

The salivary gland (Pl. 1, fig. 3, s. g.) lies to the left of the 
cesophagus and opposite the radula-sac. Its two fine and 
crenulated ducts (s.d.) pass in the same way, one above and 
one below the cesophagus, and enter the hood of the radula- 
sac one on each side, and above the nervecollars. The gland 
itself is larger than in C. textile. 

This gland, called by Bouvier (5) the “glande impaire,” is in 
this species in such a macerated condition that I have been 
unable to obtain any sections sufficiently good to warrant 
description. 

There is considerable difference between the cesophagus 
and proboscis in C. textile and those in C. tulipa, as can be 
seen from the descriptions and figures of each. In C. tulipa 
the proboscis, which is conical and annulate, is two thirds the 
length of the rostrum ; the cesophagus runs down its centre, 
and the mouth is simply an opening at its anterior end. In 
C. textile the proboscis is quite short (Text-figs. 7, 8), and 
the mouth is provided with thick reflected muscular lips. The 
cesophagus is here simply a canal surrounded by thick 
muscular walls; while in C. tulipa it isa free duct surrounded 
by muscle-bands, which in turn are enclosed by the wall of the 
proboscis (PI. 1, fig. 1), the ducts and muscle-bands only being 
held together and to the internal wall of the proboscis by con- 
nective tissue. Again, the distance in C. textile between 
the mouth and the opening of the radula-sac into the cesophagus 


20 H. O. N. SHAW. 


is short (Text-fig. 7), about a quarter of an inch, and the 
passage straight, In C. tulipa (Text-fig. 9) it is about four 


Trxt-Fia. 9. 


Conus tulipa. The rostrum, proboscis, esophagus and radula- 
sac. mh. Mouth, ts. tentacles, rm. rostrum, ps. proboscis, @. @so- 
phagus, r.s. radula-sac, 0.7.s. opening of radula-sac, o. p. d. 
opening of poison duct (p. g. d.), m. ¢. nerve collars, h. hood. 
The difference will be seen between this species and C. textile, 
in the length of proboscis and length of esophagus between 
base of rostrum and opening of radula-sac into the csophagus, 
which is thin-walled. (Diagrammatic). 


times as long, orabout oneinch. The duct isalso thin walled, 
and twice sharply reflected on itself. Lastly, the constricted 
funnel opening with its very thick walls suddenly emerging 
into a large cavity, is, im C. tulipa entirely absent. The 
radula-sac (0.7.s.) and poison duct (0. p. d.) in the latter species 


CONUS TULIPA, LINN., AND CONUS TEXTILE, LINN. 21 


simply open into the straight-walled cesophagus, which is here 
rather dilated. 

Troschel (20, pl. vi, fig. 142) gives a drawing of the radula- 
sac, thick and short proboscis, and part of the cesophagus of 


TExt-FIG. 10. 


ba. 


The base of two teeth of Conus tulipa in natural position show- 
ing (ats.) attachments to wall of radula-sac. 6. a. basal attach- 
ments, 7. ¢. row of denticulations on side of teeth, ats. tooth 
attachment to sac-wall. 


C. textile, which resembles my dissection of this species, 
though the cesophagus as shown by him is of greater size than 
I found to be the case in my specimen of the same species. 
Whereas in C. textile (Pl. 1, fig. 2) the esophagus and 
stomach for the greater part of their length are large, thick- 
walled and almost circular canals, being only constricted 


A He O. N. SHAW. 


where surrounded by the nerve collars, in C. tulipa (PL & 
fig. 3) the walls are thin, the duct is three times constricted, 
and never more than a third of the width of the former. In 
©. tulipa the canal is flat instead of round, and barely one 
tenth of the diameter of that of C. textile. 

The radula-sac (r.s.) is, as in C. textile, thick-walled 
and muscular, but the length of the right arm is only 11 mm. as 
compared to 20 mm. in that species, and I have already 
given the difference in the shapes of the two sacs. 

The teeth are placed in two groups, with the barbs or free 
ends of one group in the cul-de-sac of the right arm, those of 
the other group protruding through the esophageal opening 
asin C. textile. At the base of each tooth, and firmly con- 
nected to it, is a roundish muscular attachment (Text-fig. 10) 
fixing the tooth to the wall of the radula-sac. Each attach- 
ment is generally about half the length of a tooth, and curves 
forward from its base in such a way that the other end is 
fixed to the sac wall on the side, and anterior to the base of the 
tooth. Asin C. textile, the teeth are surrounded and con- 
nected together by connective tissue, and are generally hollow, 
and of a dark yellow colour. The anterior end of each tooth 
(Text-figs. 11, 12) is provided with a lance point, while on each 
edge behind this point is a hooked barb (bs.) pointing back- 
wards, the barbs, as in C. textile, being so placed that the 
distance between the lance point and the first barb is the same 
as between the latter and the barb behind it on the other side. 
The point and barbs, though resembling those in C. textile, 
are, however, not nearly as large or stout, and the whole tooth 
is much straighter. The base (b. a.) is generally slightly 
swollen, and the chitinous walls are thicker than the rest of 
the tooth. 

The teeth seem very constant in size and shape. On one 
side of each, and on the flat surface of the lance-point, and 
therefore at right angles to each of the two barbs, is a row of 
stout hooked denticulations (r.¢.) with their points curved 
backwards and directed to the base of the tooth. These den- 
ticulations commence anteriorly about midway between the 


CONUS TULIPA, LINN., AND CONUS TEXTILE, LINN. 23 


two large barbs, and extend backwards in a row for about 
one third of the total length of the tooth. The denticulations 
vary in number from 15 to 25, and also in size in nearly every 
tooth, though 20 to 22 seems about the average number. In this 
respect the teeth of C. tulipa differ from those in C. textile, 


Trext-Fies. 1] and 12. 


12 
The tops of teeth of Conus tulipa. bs, Barbs, rt. row of 
denticulations. 


as the latter have not got this row of lateral denticulations. 
The average length of the teeth is 4 mm., or less than half 
the length of those in C. textile (8 to 10 mm.) ; while their 
diameter at the base above the attachment is the same in both 
species, viz. ‘2 mm. 


HistToLoey. 


The state of this material, though better than that of C. 
textile, was not of the best, the specimen having been in 


24. H. O. N. SHAW. 


spirit for some considerable time. The same stains were 
employed as in C. textile, but again the nuclei were only 
partially defined. 


Porson GLAND. 


As I have already stated, this gland occupies the greater 
part of the body-cavity and lies across it, with its right end 
slightly in advance. The canal runs down the centre of the 
gland, and the poison duct enters it from in front and at right 
angles at its right extremity in the same way as in C. textile 
(Pl. 2, fig. 14). 

The canal (PI. 2, fig. 7) (c. a.) is oviform, and lined through- 
out its interior with a thin layer of epithelium (J. ep.). 

The wall of the gland is very thick and muscular (PI. 2, figs. 
6, 7, 8). It is composed of a deep external layer of longitu- 
dinal muscular fibres (J. m.s.), which run parallel with the 
length of the gland and form a thick sheath. Beneath this 
layer is a very much thinner layer (c. m.s.) (about one-six- 
teenth the thickness of the former) of muscular fibres running 
round the gland. A third layer (J. m.s.), in thickness about 
one third of the outer sheath, has the same composition as the 
outer sheath of longitudinal muscular fibres, and again runs 
longitudinally. Between this layer and the lining epithelium 
of the canal is a very narrow ring of muscle-fibres (c. m. s.) 
which run circularly round the gland. When compared to 
C.textile the poison gland is twice the diameter ; the external 
sheath of longitudinal fibres is also twice as thick, the second 
layer is the same thickness, the third layer slightly deeper. 
The fourth or internal sheath of fibres running round the 
gland is about half the thickness, the epithelium is the same ; 
the canal is one quarter the diameter, and oval instead of 
round. The muscle-fibres are disposed in the same irregular 
layers (PI. 2, fig. 6) as in C. textile. 


Poison Ducr (PI. 2, fig. 11). 
The condition of these sections is not at all good, as the 
lining epithelium is very much macerated. The outer wall is 


CONUS TULIPA, LINN., AND CONUS TEXTILE, LINN. 25 


formed of a sheath of muscle-fibres (J. /. m.) which run parallel 
to the length of the duct; this is lined by a layer of fibres 
(l. c. m.) which run round the duct and are half the thickness of 
the external sheath. This last layer has an internal lining of 
epithelium (ep.cs.), which, as in C. textile, is composed of 
elongated club-shaped cells with basal nuclei. The cells, 
which are not as long or as fine as in the last species, are also 
filled with the same fine granular substance. The epithelium 
varies in length in different parts of the duct. The external 
surface of the poison duct in C. textile is circular, while in 
C. tulipa it is of a more irregular shape. 

The whole of the centre of the duct is occupied by an irre- 
gular invagination of the epithelium (inv. ca.), which hangs 
down into the duct, and is connected by a restricted attach- 
ment to the lining epithelium. There is thus formed, with 
the exception of the place of attachment, an irregular 
circular canal (ca.) between the lining epithelium and the 
invagination. Inthe right side of this latter is a deep groove 
which runs in a curve upwards and then down to the centre. 
The average external diameter of the duct is “65 mm., the 
thickness of the wall ‘05 mm. and the epithelial lining ‘02 mm. 

In U. textile there is only a slight invagination hanging 
down into the centre of the duct, but owing to the epithelial 
lining being much thicker in this species (‘2 mm.), the area of 
the canal is about one sixth of that in C. tulipa in spite ot 
the invagination in the latter. 


SromacH anp CKsopuHacus (PI. 2, fig. 13). 


The cesophagus and stomach are in section oval externally 
with thin walls. heir internal surfaces are very deeply 
crenulated and pleated, forming large folds and furrows. 
This internal ridged surface is lined with columnar epithelium 
(l. ep.) of considerable thickness, which, as can be seen in 
the figure, has become detached from the muscle and con- 
nective tissue of the wall that surrounds it, this effect being 
due to its contraction in spirit, and where they are still 


26 H. 0. N. SHAW. 


attached, the epithelium is distorted owing to its greater con- 
traction and tendency to pull away from the outer layer. As 
far as I can ascertain from the sections (and, as I have 
previously said, the material is not of the best) the epithelium 
is non-ciliated, but appears to have the striated border which 
is characteristic of columnar epithelial cells. Embedded in 
this epithelial layer are a considerable number of what appear 
to be sporozoon parasites (sp. pa.). These are large, generally 
round or oval bodies, scattered all over the epithelium and 
buried in it at different depths. Some are so large as to 
extend through the whole depth of the layer, but never 
beyond into the muscle or connective tissue. In EKhrlich’s 
hematoxylin they stain a deep purple colour, and in Heiden- 
hain’s hematoxylin a greenish-brown. It is impossible to say 
much about them with any accuracy, as they are extremely 
difficult if not impossible to define. From their appearance 
and the fact that they are clearly foreign bodies introduced 
into the epithelium, there is little doubt that they are parasites 
such as occur not uncommonly in Molluscs. The external 
layer of the cesophagus (J. c. m.) is composed of circular 
muscle-fibres which run round and ensheath it, intermingled 
with connective tissue and a certain number of longitudinal 
muscle-fibres. Beneath this layer, but very much confused 
and mixed up with it, is an imperfect sheath of longitudinal 
muscular tissue. Between this last and the epithelial lining 
of the oesophagus is a thick but irregular layer of longi- 
tudinal granular muscular tissue (. 1. m.), intermixed with 
transverse muscles and connective tissue which bind the 
epithelium to the cesophageal wall. In some places the 
epithelium has an internal lining of a granular appearance, 
while in other places it is wanting. No doubt this is a 
glandular secretion or mucus derived from the epithelial cells. 
The average external length of a transverse section is 
16 mm. with width -75 mm., the length of the cesophageal 
canal being 1°15 mm. and width °15. 

In C. textile the external length is 4 mm. and width 2°75 
mm., length of canal 2°5 mm., width 1:0 mm. It will thus 


CONUS 'ULIPA, LINN., AND CONUS TEXTILE, LINN. 27 


be seen that all the measurements are far greater in C. 
textile than in C. tulipa. 

In C. textile the wall of the cesophagus and stomach are 
thick, while in C. tulipa they are comparatively thin. In 
the former the canal is nearly circular, while in the latter it 
is oval. 

In C. textile the lining epithelium is much thinner, and 
the internal surface, instead of being deeply crenulated and 
ridged, has large curved projections of the wall into the canal. 

The eyes, as I have already mentioned, are situated on the 
internal edge of the tentacles, at the extremity of a small 
outgrowth produced by the fusion of the tentacle and tubercle. 
In both species they are identical, so I shall only describe that 
of C. tulipa, which is in the best state of preservation. The 
eyes are enclosed by a thin transparent membrane or outer 
corneal layer (Pl. 2, fig. 12, co. 0.), which on each side of the 
eye is continuous with the coat of epithelium (ep.) covering 
the tentacle. Beneath the outer membrane is an inner 
corneal layer (co. 7.), which is considerably thinner in front 
than at the back of the eye and composed of epithelium. 
This inner layer forms a hollow vesicle or eyeball, with a 
transparent cuticular lens (/e.) occupying its interior. ‘lhe 
optic nerve (0.), which is of considerable thickness, does not 
enter the retina, but expands out over the back of the eyeball 
and chiefly on the right side. As I have said, the inner layer 
of cornea is much thicker at the back of the eye, and is so 
modified that retinal cells (ret.) are formed in the epithelium, 
directed inwards to the hollow vesicle. These retinal cells 
run across the back of the eye, and round each side for a 
considerable distance, extending forward most on the right. 
These cells, which are embedded in pigment, are longest at 
the back of the eye, and gradually get smaller as they advance 
forward on each side, till they disappear posterior to the junc- 
tion of the inner and outer corneal layers. The eye, which is 
highly developed as is the case in most molluscs, contains no 
points of special interest. ‘he eye is easily discernible with 
the naked eye, appearing as a black spot on each tentacle. 


28 H. 0. N. SHAW. 


Nervous System or Conus TUtipa. 


I found that of the two specimens under discussion 
C. tulipa offered the most advantageous dissection for the 
nervous system, and I therefore described and illustrated this 
species first. As a complete description of the same system in 
C. textile would be for the most part simply a repetition of 
the former, though both have been equally carefully dis- 
sected, I shall content myself in C. textile with pointing 
out the differences which exist between the two species. 


Description oF THE NERVE CENTRES OF Conus ‘I'ULIPA 
(Pio; tis, 21), 


There are thirteen nerve centres or ganglia, a right and 
left cerebral (C.), pleural (Pl.), buccal (B.), and pedal (P., 
P. 1), aright (V. 1), left (V. 3), and median (V. 2) visceral, 
one sub-intestinal (S7.) and one supra-intestinal ganglion 
(S.). The ganglia with their nerves and connectives are of 
a whitish-yellow colour, and may be brought into prominence 
by treatment with osmic acid. The right and left cerebral 
ganglia (C.) are round in shape and fused together on their 
adjacent sides, and each is posteriorly united to the left and 
right pleural ganglia (PI.) respectively, the right pleural 
ganglion being again attached on its posterior side to the 
supra-intestinal ganglion (S.). All these ganglia are ex- 
tremely hard to define by reason of their junctions being 
indicated only by very slight constrictions and the whole 
being bound to the cesophagus by connective tissue. These 
ganglia lie on the cesophagus, where it is slightly constricted 
behind the opening of the radula-sac and poison duct. This 
position is seen in Text-fig. 9 (w. c.), but as this figure is purely 
diagrammatic, only one collar in the form of a ring, and 
without any ganglia, is shown. ‘To the right, and slightly 
posterior, are situated the pedal ganglia (P., P. 1). 

These ganglia lie acoss the right side of the body-cavity 
under the radula-sac with their anterior extremities pointing 


CONUS TULIPA, LINN., AND CONUS TEXTILE, LINN. 29 


slightly forward. Both ganglia are closely connected to- 
gether along their internal sides, being only distinguishable 
at both extremities, and more so at their anterior ends. The 
ganglia are slightly pyriform or attenuated anteriorly. 

A cerebro-pedal connective (ce. pl.) passes from the left 
side of the left cerebral ganglion under the cesophagus, and 
enters the left pedal ganglia (P.) on its dorsal and posterior 
surface. The left pleuro-pedal connective (pa. pl.) runs 
under the cesophagus and parallel to and behind the cerebro- 
pedal connective, and joins the left pedal ganglion behind 
the latter. In like manner a right pleuro-pedal and cerebro- 
pedal connective passes over the cesophagus, connecting the 
right pedal ganglion (P. 1) to the right cerebral and pleural 
ganglia. The cerebro-pedal connective is smaller than the 
pleuro-pedal. Posterior to this last connective, and having 
their origins one in the left and one in the right pleural 
ganglia, are the pleuro-subintestinal connective (d.’) and the 
zygoneurous conective (z.) respectively of the visceral com- 
missure (d.). The pleuro-subintestinal connective (d.’) issues, 
as I have stated, from the left pleural ganglion behind the 
left pleuro-pedal connective, passes in an oblique direction 
backwards and under the cesophagus, and enters the anterior 
and left extremity of the subintestinal ganglion (Sv.). The 
zy goneurous connective (z.) issues from the same place in the 
right pleural ganglion, passes backwards and over the ceso- 
phagus, and enters the right anterior extremity of the sub- 
intestinal ganglion. This ganglion is pyriform, its attenuated 
posterior end forming the origin of the right loop of the 
visceral commissure (P]. 4, fig. 19). The left pleural ganglion 
terminates posteriorly rather bluntly, being little attenuated, 
while the right ganglion, lying between the right cerebral 
and the supra-intestinal ganglia, is hardly distinguishable. 
This last ganglion, which is pyriform, is continued back- 
wards, forming the left loop of the visceral commissure. ‘This 
latter (d.), after leaving the ganglion, passes over to the left 
of the body-cavity, and through the wall where is situated 
the left visceral ganglion (V.3). This ganglion ends abruptly 


30 H. O. N. SHAW. 


posteriorly, the commissure passing backwards till it reaches 
the median visceral ganglion (V. 2), which is placed across 
the body, anterior to the recto-genital mass, and is attenuated 
at both ends, The visceral commissure leaves the right 
extremity, passing forward and to the right, where it jos 
the right visceral ganglion (V. 1). This latter is pyriform 
anteriorly, while posteriorly it is produced where connected 
to a large nerve (k.) and the visceral commissure, the latter, 
after passing forward for some considerable distance, joining 
the anterior of the right visceral ganglion, with the posterior 
extremity of the subintestinal ganglion. The commissure 1S 
shortest between the median and right visceral ganglia. 

From the anterior edges of both cerebral ganglia (PI. 3, 
fic. 18, and PI. 5, fig. 21) issue the cerebro-buccal connectives 
(c. b.). Both pass round the cesophagus, one on each side, 
and join the external edges of the right and left buccal ganglia 
(B.) respectively. These ganglia are small, circular, flat bodies 
lying under the cesophagus. They are united by two com- 
missures, but are so close together that they appear almost to 
touch. The anterior commissure is of extreme fineness, while 
the posterior, which is of considerable stoutness, is the origin 
of the large buccal nerve (s. 5) which leaves the commissure 
close to the Jeft buccal ganglion and passes backwards. 

The cesophagus and the poison gland duct (PI. 5, fig. 21) are 
completely surrounded by four nerve collars, with ganglia at 
each extremity of the collars, the most anterior being formed 
by the two cerebral and two buccal ganglia with their 
connectives. This collar is the smallest and lies close round 
the cesophagus, to which it is attached by connective tissue. 
Behind this comes the cerebro-pedal collar, and posterior again, 
the pleuro-pedal, these last two being larger and not so closely 
attached to the cesophagus, round which they lie obliquely, 
owing to the pedal ganglia being slightly behind the cerebrals 
and buccals on the right. The fourth and last nerve collar 
is the pleuro-subintestinal, and owing to the latter ganglion 
being situated still further backwards, this last collar is more 
oblique with regard to the cesophagus than the second and 


CONUS TULIPA, LINN., AND CONUS TEXTILE, LINN. 31 


third, but is smaller in circumference. The left half of this 
collar, which is the stoutest, is formed by the pleuro-sub- 
intestinal connective of the visceral loop (d.’), while the right 
half is the zygoneurous connective (z.). 

The nerves issuing from the various ganglia, with the 
description of their functions and the parts they innervate, 
etc., will be dealt with under the various headings hereafter 
mentioned. 


TENTACULAR AND Optic Nerves (PI. 3, fig. 18). 


The tentacular nerves (¢.), or, as they may be called, 
cephalo-tentacular, are the largest nerves given off from the 
cerebral ganglia (C.).. They issue from the dorsal and 
anterior surface of the ganglia, that from the right supplying 
the right half of the rostrum, while the left side is supplied 
from the left ganglon. This left nerve, after leaving its 
origin in the ganglion, passes to the right, and is then sharply 
reflected back to the left, where, after proceeding a short way, 
a stout nerve (¢. 1) is given off from its anterior side, which 
passes over to the left and is ramified, splitting up into 
numerous fine branches. Slghtly anterior to the base of this 
nerve the main tentacular nerve is bifurcated, both nerves 
passing underneath it, the left branch (¢. 3) being ramified in 
the same way as the preceding one. The right branch (¢. 2) 
is subdivided into three, of which the median is the optic 
nerve (0.), while those on either side of it are divided up into 
numerous fine nerves like the other tentacular nerves. 

The optic nerve, which, as has been shown, is only a branch 
of the tentacular nerve, is stouter than the branch on each side 
of it. When it reaches the place in the rostrum immediately 
below where the eye is situated, it turns at right angles 
through the rostrum wall and so to the eye. 

The right tentacular nerve on leaving its ganglion passes 
to the right and slightly forward, where, after a short course, 
it is bent sharply backwards and then forward again, forming 
a kink. In the left nerve this kink is close to the left 


82 H. O. N. SHAW. 


ganglion, no doubt the duty of these two kinks being to 
allow a slight contraction or extension of the thick muscular 
walls of the rostrum. After its flexion or kink a fine nerve 
is given off on the inside of the right tentacular nerve, which 
supplies the rostrum base, and anterior to this latter the main 
nerve is bifurcated, the left branch again dividing into two, 
the left of these two (¢.1) being similarly split up. Both 
these last give off fine and very numerous branches on their 
course towards the tentacles, as is also the case with the night 
branch (¢. 3) at the first bifurcation of the main nerve. The 
right branch of the second division of the main nerve (f. 2), 
after proceeding a short way, gives off a fine nerve, while the 
main nerve of these two, the left, is the optic nerve (0.), which 
proceeds to the eye in the same manner as the left optic nerve, 
and has two or three small branches issuing from it. 

In his description of Conus virgo, Bouvier (5) says that 
the left optic nerve is given off from the tentacular nerve of 
the same side sooner than is the case on the right side. In 
my dissection of C. tulipa the reverse is the case, but, hike 
Bouvier, I have been unable to ascertain whether the optic nerve 
is simply a branch of the tentacular, or whether, though both 
are in the same nerve-sheath, they really issue as two distinct 
nerves from the ganglia. The latter seems improbable, as in 
the case of the left optic nerve there are stout branches on 
each side supplying the tentacles, the optic nerve issuing from 
between them. All the tentacular nerves are extraordinarily 
ramified, splitting up and giving off very numerous fine 
branches. The ends of the nerves which supply the tentacles 
themselves, before entering the latter, are generally divided 
into at least three branches. Both the main tentacular nerves 
proceed from the ganglia under the base of the proboscis. 

A most remarkable fact in connection with the nerve system 
of this species is the following: From the posterior surface of 
the base of the kink in the right tentacular nerve a stout 
branch (¢.p.) issues; it rans backwards, and is much crenu- 
lated. This latter joins at right-angles a nerve (7.) running 
across the floor of the body-cavity, thus forming an inverted 


CONUS TULIPA, LINN., AND CONUS TEXTILE, LINN. 33 


T. The right branch is ramified, and supplies the base of the 
rostrum on the right, while the left branch proceeds to the 
dorsal surface of the right pedal ganglion, thus forming 
a connective between the tentacular nerve and the right pedal 
ganglion. This nerve, as far as I have been able to ascertain, 
has not been noticed before, and its presence is hard to 
explain. 

There is no nerve or connective of any sort given off from 
the left tentacular nerve until the latter is bifurcated, as I 
have shown, the right branch passing forward to the rostrum 
and cephalic teguments. 


Acoustic Nerves (PI. 5, fig. 21). 


These nerves are extremely fine and very hard to follow, 
on account of the great number of pedal nerves through which 
they pass. The origin of the left nerve (ac. /.) is in the left 
cerebral ganglion, immediately below the left cerebro-pedal 
connective, and between the latter and the lett pleuro-pedal 
connective. This nerve runs between these two connectives 
to the left pedal ganglion (P.) along the anterior side of the 
latter, passing between numerous pedal nerves, then slightly 
backwards, and reaches the left otocyst (ot. /.) situated some 
little distance to the right and slightly posterior to the pedal 
ganglia. The right acoustic nerve (ac. 7.) issues from the 
right cerebral ganglion at about the same place as does the 
left, but descends to the right pleuro-pedal connective, and 
runs along its anterior edge till both reach the right pedal 
ganglion (P.1). ‘lhe acoustic nerve, after following the 
posterior surface of the latter, passes along the pedal nerves 
backwards and to the right, till it reaches the right otocyst 
(ot. r.), which is situated in the muscles of the foot, and at a 
considerable distance from the pedal ganglia and behind them. 


ProgoscripEaN Nurves (PI. 4, fig. 19). 


The nerves to the proboscis, or labio-proboscidean nerves, 
are three in number on each side, and issue from the anterior 
vo. 60, pART 1.—NEW SERIES. 3 


34 H. O. N. SHAW. 


edges of the cerebral ganglia. The nerves are stout, of a 
white colour, not deeply imbedded in the muscular wall of the 
proboscis, and run parallel to the cesophagus and at about 
equal intervals round it, the central one (/. 1) being on the 
ventral surface. The two central nerves (J. 1) are the longest, 
and have the fewest nerves given off along their length, and 
are the finest and most crenulated of the three sets. They 
run to the anterior end of the proboscis, and are here broken 
up into a number of fine nerves. ‘The second pair (/. 2) are 
straighter, and the left member of the pair gives off a large 
branch about one third of its length from the ganglion, while 
in the right nerve the branch is half-way. The left nerve of 
the pair (/. 3) is trifurcated, the right one bifurcated, all 
branches being much ramified. In section these nerves are 
nearly flat and much crenulated. 

The walls of the rostrum and cephalic tegument are supplied 
by three pairs of nerves (/. 4—l. 6) as well as by branches given 
off from the tentacular nerves. As was the case with labio- 
proboscidean nerves, these nerves to the rostrum are in three 
pairs, and have their origin in the anterior part of the cerebral 
ganglia (C.). The three pairs are nearly the same size, but 
are very much finer and shorter than the three pairs of labio- 
proboscideans. As was the case in Bouvier’s dissection of 
C. virgo, so in this species, the first pair of nerves (1. 4) go 
to the base of the rostrum wall, where they are ramified and 
run forward, while the other two pairs on the right (J. 5, J. 6), 
which are slightly finer, pass from the cerebral ganglion across 
the cerebro-pedal connective, and so enter the muscular wall 
of the rostrum, where they run anterior to the first pair (1.4). 
Before they enter the muscular rostrum wall, and between here 
and their ganglia, the nerves (J. 5, /.6) are very sinuous, as is 
the case more or less with all the nerves to the rostrum and 
also the labio-proboscideans. This is due to the fact that 
nerves, being non-extensile or retractile, or nearly so, when 
they enter a wall which may be exserted or contracted, as in 
the case of the proboscis, and slightly so in the rostrum, are 
supplied with enough slack nerve between the ganglia and 


CONUS TULIPA, LINN., AND CONUS TEXTILE, LINN. 309 


their entry to allow the wall to be fully extended without 
pulling on the nerve. When fully contracted or abnormally 
sO, as in spirit, the nerves have a very sinuous appearance, 


Buccat Ganaiia AND THEIR Nerves (PI. 3, fig. 18). 


The buccal ganglia (B.) are small, circular, flat bodies which 
lie under the cesophagus, and their external edges are joined 
by a stout connective to the anterio-dorsal surface of the 
cerebral ganglia (C.). The two internal and adjacent edges 
of the buccal ganglia are connected by two sub-cesopha- 
geal commissures, but the ganglia are so close together that 
they almost touch, the commissures being about a sixteenth 
of an inch long. The anterior of the two is extremely fine, 
while the posterior one is as stout as the cerebro-buccal 
connectives (c. b.). These two ganglia, therefore, with their 
connectives, form a complete nerve collar round the cesophagus, 
the buccal ganglia being slightly posterior to the cerebrals. 
The openings of the radula-sac and poison duct into the 
cesophagus are anterior to the nerve collar. 

For the sake of comparison I have used the same letters as 
employed by Bouvier in his figures. From the posterior 
surface of the right buccal ganglion a nerve (s. 1) is given off, 
which innervates the anterior portion of the poison duct, till 
a branch is given off to the latter from the main poison-gland 
nerve. There is no corresponding nerve given off from the 
left ganglion. The buccal proboscidean nerve (s. 2), which is 
anteriorly ramified, supplies the wall of the cesophagus in the 
proboscis, the wall of the proboscis being innervated from the 
cerebral ganglia by the labio-proboscidean nerves. Here, in 
like manner, I have been unable to find any such nerve issuing 
from the left ganglion, nor was I able to find the nerve 
mentioned by Bouvier as being given off by the side and pro- 
ceeding to the cesophagus. 

Two pairs of nerves (s. 3, s. 4), one pair issuing from each 
ganglion, supply the wall of the radula-sac, those given off 
from the right ganglion supplying the dorsal surface of the 


36 H. O. N. SHAW. 


sac, while the two from the left ganglion supply the ventral 
surface. The main nerve (s.5) which innervates the poison 
gland and duct, issues from the posterior commissure, and runs 
backwards under the cesophagus and through the nerve 
collars. This nerve, which is of considerable size, after pass- 
ing backwards for some distance curves to the right, where a 
branch is given off which supplies the poison duct and takes 
the place of the first fine nerve (s.1). The main nerve now 
doubles back to the left and then again to the right, three or 
four slender nerves being thrown off to the coils of duct 
among which the main nerve runs. This nerve, after passing 
to the right, is trifurcated, the three branches being ramified 
and supplying the walls of the poison gland, the branch 
to the right entering the right extremity of the gland 
by the duct opening. After the main poison gland nerve 
has passed backwards from the buccal commissure, a fine 
nerve (s.6) is given off on the left, which runs backwards 
along the wall of the cesophagus and is much ramified. The 
existence of this nerve was indicated by Bouvier (5) p. 341, 
where he says: “ Dans un individu femelle, il me sembla qu’un 
filet gréle se rendait de ce nerf a Vcesophage, en arriére des 
colliers nerveux.” About this point he was not certain, and 
mentions the difficulty experienced in removing the connective 
tissue which completely surrounds and binds firmly together 
the nerves, the cesophagus, and coils of poison duct; this I 
also found most troublesome, but managed to free the nerve, 
and ascertained that this nerve (s. 6) does run from the main 
nerve to the cesophagus. As stated by Bouvier, the main 
nerve (s.5), on account of both its position and origin, would 
correspond to the cesophageal-aortic nerve in Buccinum, if a 
branch was given off to the cesophagus, and this I have shown 
to be the case. 


Lert Pievrat GaneLtion and Nerves Issuine (Pl. 4, fig. 19). 


This ganglion is attached to the posterior end of the left 
cerebral ganglion, the end of the one and the beginning of 


CONUS TULIPA, LINN., AND CONUS TEXTILE, LINN. 37 


the other being only distinguished by a slight constriction 
between them. The posterior extremity of this pleural gan- 
glion (Pl.) is directed slightly to the left, while the ganglion 
is very slightly pyriform and attenuated posteriorly. On the 
left, and from the external and ventral surface, a stout con- 
nective (d.’), the pleuro-subintestinal connective of the vis- 
ceral commissure, issues, turns to the right, across the floor 
of the body-cavity, runs slightly backwards and under the 
cesophagus, and enters the ventral and anterior end of the 
sub-intestinal ganglion (S7.). 

Two columellar nerves (¢. 1, 7. 2) are given off from the 
centre of the ventral surface of this pleural ganglion, while 
none are present in the corresponding right ganglion. The 
nerve on the right (7. 2) passes backwards and slightly to the 
right, being much crenulated and unattached to the body- 
cavity till it splits into four fine nerves, which diverge in the 
columellar muscle. ‘The left of this pair (¢. 1) is twice the 
size of the former, being sinuous or crenulated in the same 
way, and runs directly backwards, being unattached to the 
wall for some considerable distance. Shortly after its entry 
in the wall it is bifurcated and descends almost vertically 
into the base of the columellar muscle on the right, and 
about level with the left visceral ganglion (V. 2). Both 
nerves pass under the cesophagus, the left one running 
parallel with it, and supply the anterior and left half of the 
columellar muscle and base of the body-cavity. These two 
nerves differ from those in Conus virgo, described by 
Bouvier (5), in the following respects; they are shorter, supply 
the columellar muscle anteriorly and to the left, instead of 
posteriorly and on the right, do not pass through the nerve 
collars or go anywhere near the right pleural ganglion or the 
sub-intestinal, but are absolutely distinct and run directly 
backwards from their origin. 

Four nerves are given off from the postero-dorsal edge of 
the left pleural ganglion, two being parietal, the other two 
being the main and lesser pleuro-siphonal nerves. 

The most anterior of these four (/. 1) is the lesser pleuro- 


38 H. 0. N. SHAW. 


siphonal nerve, the second (f.), the main pleuro-siphon, the 
third and fourth (c. and c. 1) being parietal. All these four 
nerves, after leaving their origins in the ganglion, pass directly 
to the left, being suspended and quite unattached for some 
considerable distance between the ganglion and their entries 
into the base of the body-wall. The lesser pleuro-siphonal 
nerve passes through the base of the rostrum and at once 
runs downwards over the attachment of the base of the 
siphon to the base of the rostrum. When it reaches the 
siphon proper it gives off two or three small branches and 
passes forward along the left siphonal wall. The main 
siphonal nerve (f.), which is very much larger, runs parallel 
but posterior to the former till it reaches the base of the 
siphon, where it proceeds along the centre of the channel 
and gives off fine nerves which run into the walls on 
both sides. When one third of the length from the anterior 
extremity of the siphon, the main nerve is trifurcated, each 
branch ramifying in the sides and extremity. At its entry 
into the base of the siphon, a stout anastomosis (a.) connects 
it with the anterior branchial nerve (0.). 

The parietal nerve (c.) enters the floor of the body-cavity 
.on the left, just behind the main pleuro-siphonal nerve, giving 
off fine branches to the floor and wall. The other nerve of 
this pair (c. 1) issues from the ganglion behind the former, 
and after a short distance passes forward under it, and also 
beneath both the pleuro-siphonal nerves, entering the body- 
wall anterior to the foregoing. Both parietal nerves are by 
far the finest of the four nerves given off from this ganglion. 


SUPRA-INTESTINAL GANGLION AND Nerves (P]. 4. fig. 19). 


This ganglion (S.), which is closely connected to the 
posterior extremity of the right pleural ganglion, serves as 
origin for six nerves. There is hardly any constriction at 
the junction of the two ganglia, which are, therefore, hard to 
distinguish in a dissection. ‘I'he supra-intestinal ganglion is 
pyriform in shape with its posterior extremity directed slightly 


CONUS TULIPA, LINN., AND CONUS TEXTILE, LINN. 39 


to the left. Bouvier (5) mentions only four nerves as coming 
from this ganglion; of the two additional nerves that I have 
recognised, one is only a parietal nerve and of little or no 
importance ; the other will be mentioned in due course. 

There are, then,as I have said, six nerves from this ganglion, 
two being parietal, two branchial, one, the largest, the supra- 
intestinal branch of the visceral commissure, and the last 
supplying the inner edge of the branchial cavity. All these 
six nerves lie above the two columellar nerves (7. 1, 7. 2). 

All the nerves issue from the posterior extremity and 
dorsal surface of the ganglion, the most anterior being the 
main branchial nerve (b.), which is the stoutest of the six with 
the exception of the supra-intestinal branch of the visceral 
commissure. 

The main branchial nerve (b.), after leaving the ganglion, 
passes slightly backwards and to the left of the body-cavity 
where it enters the wall running forward and parallel to the 
pleuro-siphonal nerve (f.). In front of and above the anterior 
end of the liver these two nerves come closer together, and, 
as I have already stated, a stout anastomosis (a) connects 
them before the branchial nerve reaches the mantle. The 
branchial nerve is now reflected sharply to the left and passes 
under the extreme anterior edge of the osphradium and 
ctenidium, supplying fine nerves to each; the main nerve 
ramifies in front and to the left of the ctenidium, and forms a 
fine network, with numerous anastomoses, in the mantle. 
Slightly to the left of where the branchial nerve les above 
the left columellar nerve (7. 1), a stout branch (b. 1) is given 
off which runs parallel to and almost touches the main nerve 
for some considerable distance. After passing down through 
the body-wall this branch lies across the anterior lobe of the 
liver, whence it proceeds backwards and under the edge of 
the osphradium, supplying this and the central portion of the 
ctenidium. 

The posterior branchial nerve (b. 2) is much finer than the 
anterior, and crosses over to the mantle behind the branch of 
the main branchial nerve which I have just described, being 


40 H. O. N. SHAW. 


practically parallel to it throughout its length. Having 
passed under the posterior end of the osphradium and given 
off two or three fine nerves, the main portion of the posterior 
branchial nerve innervates the internal portion of the mantle 
and the posterior parts of the ctenidium. According to 
Bouvier the nerve with its fine branches and anastomoses 
supplying the anterior edge of the mantle is given off after 
the anastomosis (a.) between the anterior branchial nerve (6.) 
and the main pleuro-siphonal nerve (/.), and is thus a product 
of the pleuro-siphonal nerve. In my dissection this is not so, 
for the branchial nerve bifurcates, the right branch, which is 
the finer of the two, forms the anastomosis (a.), while the 
left branch supplies the anterior portions of the mantle in 
addition to the ctenidium and osphradium, so that this nerve, 
which in both cases supplies the anterior mantle edge, is in my 
dissection not pleuro-siphonal but branchial. The branchial 
nerve, therefore, is of much greater length than that figured 
by Bouvier. 

Another curious point and one worthy of note is the 
presence of the large branch (b. 1) issuing from the main 
branchial nerve close to its origin. At first sight I was under 
the impression that this was the posterior branchial nerve, 
but closer inspection soon proved that this was not the case, 
for it is undoubtedly simply a branch of the anterior nerve, 
the posterior branchial nerve (b. 2) being quite distinct and 
originating to the right of the former. The typical position 
for the posterior banchial nerve would be slightly behind this 
branch nerve, though not as far back as it is in this specimen. 

There are, therefore, according to the parts they innervate, 
three, and not two branchial nerves. The main and anterior 
nerve is normal; its branch forms a median branchial nerve; 
the posterior nerve proper is displaced backwards, and with 
the aforesaid branch, innervates rather more than the area 
covered by a normal posterior branchial nerve. 

There is nothing of much interest about the two parietal 
nerves (c.2,c,3). Both cross over to the left and supply the 
body-wall. The anterior (c.2) leaves its ganglion between 


CONUS TULIPA, LINN., AND CONUS TEXTILE, LINN. 41 


the anterior and posterior branchial nerves, while the posterior 
(c.3) has its origin between the visceral commissure (d.) and 
the posterior branchial nerve. Soon after its entry in the 
wall, this posterior parietal nerve expands out into a small, 
ganglionic-looking mass, from which two fine nerves run 
forward and three backwards into the body-wail. 

The four parietal nerves, two from the left pleural ganglion 
(c., c. 1) and two from the supra-intestinal ganglion (c. 2, ¢. 3) 
innervate the body-wall and side, anterior to the left visceral 
ganglion (V. 2). 

The sixth nerve (c.4), which I have mentioned, leaves the 
supra-intestinal ganglion on its right and postero-ventral 
edge, runs backwards and to the left, and plunges into the 
body-wall just in front of the left visceral ganglion, and after 
running down through the wall, emerges above and passes 
across the liver to the mantle base, where it ramifies and 
supplies the mantle wall between the hinder portion of the 
ctenidium and the base of the mantle. This nerve is not noted 
by Bouvier in C. virgo, but in the specimen under discussion 
it is as stout as the posterior branchial nerve, and from its 
position is of some interest. 

The remaining nerve, issuing from the supra-intestinal 
ganglion, is the left or supra-intestinal branch of the visceral 
commissure; this I shall discuss with the visceral ganglia. 


Tue Viscerat Commissure anp Ganeuia (PI. 4, fig. 19). 


There are three visceral ganglia. ‘The left one (V. 3) 1s 
situated in the body-wall near the anterior portion of the liver 
and above it on the right side. The right ganglion (V. T)as 
found in the wall enclosing the posterior part of the body- 
cavity and near its left extremity, while the median visceral 
ganglion (V. 2) lies to the nght of the posterior part of the 
liver, and in front of the recto-genital mass (r. g. ™.) 

The left branch of the visceral commissure (d.) has its origin 
in the posterior and dorsal surface of the supra-intestinal 
ganglion, passes obliquely backwards and to the left, and so 


42 H. O. N. SHAW. 


into the body-wall, and enters the left visceral ganglion close 
to the right edge of the liver. From the ventral surface of 
this ganglion, two nerves issue, which might equally well be 
called parietal or columellar nerves, since they supply the 
walls and columellar muscle to the right and below their 
ganglion and almost touch the left columellar nerve (t. 1) 
which issues from the left pleural ganglion. The commissure, 
after leaving the left ganglion, runs backwards and downwards 
and slightly to the right, where it meets the median visceral 
ganglion. Between these two last ganglia, and about one 
third of the distance from the left visceral, a nerve (7.) issues 
from the left side of the commissure, and runs through the 
base of the mantle to the dorsal surface of the liver. This 
hepatic nerve turns to the right and passes back through the 
dorsal and right edge of the liver and emerges on the right 
side, where it is trifurcated. The commissure itself does not 
run out in a loop over the dorsal surface of the liver, but is 
more normal, being situated in the thick tissue formed by the 
junction of the mantle to the body-wall, and is thus on the 
extreme right edge of theliver. The hepatic nerve (n.), which 
I have just described, is peculiar both from its origin and 
from the fact that it is of considerable size, and the only nerve 
I have been able to trace which supplies the liver from the 
commissure. Bouvier mentions the existence of a fine hepatic 
nerve in C. virgo. 

The median visceral ganglion is not so large as the left 
visceral. Three nerves issue from it; that to the left (m.) 
supplies the posterior lobe of the liver, while the central nerve 
(m. 1), which is the largest of the three, is the visceral nerve, 
innervating the heart as well as the genital organs and kidney. 
The nerve on the right (m.2) is the genito-rectal nerve. 
Between the median and right visceral ganglia, two parietal 
nerves are given off to the body-wall from the commissure, 
which latter, after leaving the median ganglion, passes forward 
and to the right, and so to the right visceral ganglion. One 
nerve (k.) arising from this last ganglion (V.1) is of con- 
siderable size. As stated by Bouvier, this nerve is pleural, 


CONUS TULIPA, LINN., AND CONUS TEXTILE, LINN. 43 


and issues from the posterior and right side of the ganglion, 
passes backwards and into the mantle, beneath the recto- 
genital mass and up over the back without entering it, where 
it again reaches the mantle, turns sharply to the right and is 
ramified. 

After leaving the anterior end of the right visceral ganglion, 
the visceral commissure runs forward and to the right through 
the body-wall till it reaches the posterior portion of the body- 
cavity. Here it emerges from the body-wall under the 
poison gland and duct, where it enters the hindmost part of 
the sub-intestinal ganglion. I have been unable to find any 
nerves issuing from the left visceral ganglion and passing to 
the liver as indicated by Bouvier. 


QuB-INTESTINAL GaneLion (PI. 4, fig. 19). 


This ganglion gives off five nerves, four of which innervate 
the right part of the body, while the fifth is the visceral 
commissure. All these five nerves are unattached, and lie on 
the floor of the body-cavity till they enter the posterior wall 
of the latter. 

The ganglion (Sz.) 1s pyriform and attenuated posteriorly 
and slightly to the right, the visceral commissure (d.) entering 
the posterior extremity. One nerve (¢.) issues from the left 
and ventral surface of the ganglion, passes backwards and 
slightly to the left, and enters the body-wall under the 
visceral commissure, where, after its entry, it divides up into 
numerous fine branches. The branches supply the body-wall 
as well as the posterior muscles of the latter, which eventually 
unite with the columellar muscle, so that this nerve, though 
really parietal, is in part also a columellar nerve. 

There are two true parietal nerves given off from the 
ventral and right side of the sub-intestinal ganglion, both 
being finer than the nerve just described, and lying to the 
right of it. Both these nerves (e. 1, e.2) supply the hinder 
portions of the body on the right, as also part of the right 
side, but do not extend as far back as the parietal-columellar 


4.4, H. O. N. SHAW. 


nerve. The right of these two nerves (e. 1) issues from its 
ganglion more anteriorly than does the left. Between these 
two and starting more from the right side of the ganglion is 
the largest of the four nerves, the right pleural nerve (e. 3). 
This nerve passes backwards and slightly to the right, ranning 
at no great depth through the body-wall till it reaches the 
thick muscular ridge formed by the posterior edge of the 
body-wall and bounding the anterior side of the anal channel. 
On reaching this ridge, the nerve plunges straight down and 
under the channel and to the right, where the nerve divides 
into two, the left branch running back and ramifying in the 
edge of the mantle to the right, while the right branch, curving 
forward again, supplies the under-surface of the body by its 
junction with the mantle. The zygoneurous connective 
(z.) unites the right pleural ganglion (P/.) with the sub- 
intestinal ganglion (Sz.). It is of considerable stoutness, and 
issues from the ventral and right side of the former to pass 
backwards and to the right, and is then bent abruptly 
posteriorly and enters the sub-intestinal ganglion anteriorly 
and on its dorsal and right side. This connective passes over 
the cesophagus, while the pleuro-subintestinal connective (d.’) 
of the visceral commissure is sub-cesophageal. 

There are no nerves given off from the right pleural ganglion, 
the right side of the body being innervated from the sub- 
intestinal ganglion. Bouvier, in his figures and text, has 
confused these two connectives, since he calls that between 
the left pleural ganglion and the sub-intestinal ganglion the 
zygoneurous connective, while he describes the one between 
the latter and the right pleural ganglion as the connective 
of the visceral commissure, whereas if is exactly the reverse. 


Pepa Ganeuta (PI. 4, fig. 20, and Pl. 5, fig. 21). 


These two ganglia are so closely connected together that 
they look like one irregular ganglion. On closer inspection 
it will be noticed that at both anterior and posterior extremities 
there is a slight cleft or constriction between them, which is 


CONUS TULIPA, LINN., AND CONUS TEXTILE, LINN. 45 


most distinct at their anterior end. The ganglia lie under 
the radula-sac on the floor of the body-cavity and are inclined 
so that their anterior ends are slightly lower than the posterior. 
The ganglia are so displaced that they lie across the body- 
cavity instead of their longitudinal axes running parallel to 
the foot, the anterior extremities being slightly in advance of 
the posterior, the ganglia thus lying at a tangent to the 
longitudinal body-axis and considerably to the right. From 
this it will be noticed that owing to torsion, the symmetry 
of the anterior part of the body has been entirely dis- 
placed. Of this not only the pedal ganglia but also the 
otocysts bear witness, for the latter are not really right and 
left, but anterior and posterior, the right (ot. 7.) being 
practically directly behind the left (ot. 1.), and both are 
situated at the base of the foot and on the right side. Each 
of the two pedal ganglia is connected with the cerebral and 
pleural ganglia of its own side by the cerebro-pedal and pleuro- 
pedal connectives. The cerebro-pedal connectives join the 
pedal ganglia anteriorly to the pleuro-pedals, the latter 
being the stoutest of the two pairs. These connectives with 
their ganglia form two complete and wide nerve collars round 
the cesophagus and lie obliquely round the latter, since the 
pedal gangha are to the right, and shghtly posterior to the 
cerebral and pleural ganglia. 

The nerves given off from the pedal ganglia are extra- 
ordinarily numerous, and for the most part of considerable 
size. ‘hey issue from the sides and anterior extremities of 
the ganglia. The posterior part of the foot is innervated from 
the right ganglion, while the anterior half is supplied from 
both right and left ganglia, the right ganglion sending nerves 
to the anterior and right half, the left ganglion to the left 
anterior portion. 

There are many more nerves issuing from the right gan- 
glion than from the left, the former giving off thirty, while 
only eleven proceed from the left ganglion. As I have 
already stated, most of these nerves are of considerable size, 
some—8, 9, 14, 29—equal in stoutness the tentacular nerves ; 


46 H. 0. N. SHAW. 


thus the foot and right side of the body are very highly 
innervated. 

I do not propose to describe all these forty-one pedal nerves, 
but they will be seen in Pl. 4, fig. 20, and the parts they 
supply are described in the explanation of that figure. 

The following nerves are worthy of a passing note. The 
posterior portion of the foot is supplied from the right ganglion 
by the nerves 29, 31 and 32, of which the first is the stoutest 
while the latter pair run side by side for most of their length. 
The right ganglion also innervates the central and ventral 
parts of the foot, with the exception of the one bifurcated 
nerve 23, which proceeds backwards from the left ganglion, 
while the chief nerves from the right ganglion are 25, 27 
and 28. 

There are three main nerves to the anterior region of the 
foot, of which 8 and 9 issue from the ventral and anterior 
side of the left ganglion, and supply the left side, while 14 
passes to the right side from the anterior edge of the right 
pedal ganglion. An anastomosis exists between this nerve 
and the smaller nerve 13. 

One very sinuous nerve (1) after leaving the anterior and 
ventral surface of the left ganglion, passes under the 
cesophagus and cerebro-pedal and pleuro-pedal connectives, 
directly over to the left side, and enters the floor of the body- 
cavity immediately in front of the left columellar nerve (7. 1), 
where it is trifurcated and innervates the floor under the left 
pleural ganglion. A fine nerve, 41, issues from the ventral 
surface of the right ganglion, and supplies the cavity floor 
directly beneath the pedal ganglia. 

From the foregoing account of the nervous system of this 
species, several points of interest will be noticed, some new, 
and others confirming the descriptions of previous writers on 
this genus. Among the latter is evidence of the existence of 
an oesophageal nerve given off from the poison gland nerve, 
and indicated by Bouvier. Both my dissections have been of 
females, as was Bouvier’s, and I have not yet been able to 
dissect a male of either species owing to lack of material. 


CONUS TULIPA, LINN., AND CONUS TEXTILE, LINN. 47 


Perhaps the most interesting feature about the nervous 
system of the above species is the connection between the 
right tentacular nerve with the right pedal ganglion, by 
means of a branch uniting the former to a nerve passing to 
the body-wall from the right ganglion. 


DIFFERENCES BETWEEN Nervous Systems oF CoNUs TEXTILE 
AND CONUS TULIPA. 


The nerve centres are situated in similar positions in both 
species; in C. textile both the centres and their nerves are 
of a reddish-yellow colour. The cerebral, pleural and supra- 
intestinal ganglia are more closely connected, and even harder 
to differentiate. The internal edges of both pleural ganglia 
almost meet, and the constriction is hardly noticeable between 
the cerebral ganglia, while the whole mass is covered by 
a thick sheath of connective tissue. The supra-intestinal 
ganglion is connected, as in C. tulipa, to the right pleural 
ganglion, their junction being hardly determinable, while the 
posterior extremity of the former is directed sharply to the 
left, from whence the left-hand loop of the visceral commissure 
issues. ‘This flexion of the ganglion is so abrupt as to bring 
the posterior portion immediately behind the left pleural 
ganglion, with the result that, at first sight, there appear to 
be two supra-intestinal ganglia. 

The visceral commissure and nerves issuing from the 
supra-intestinal ganglion, owing to its peculiar shape, proceed 
across the body-cavity to the left, and at right angles to its 
axis, instead of obliquely backwards as in the former species. 
The left pleural ganglion is more pyriform, but in other 
respects is the same. The positions of the ganglia lying 
around the cesophagus and also of the sub-intestinal ganglion 
are similar, but the pedal ganglia lie further back than in 
C. tulipa, and are not so attenuated anteriorly. These last 
ganglia are attached with the cerebral and pleural ganglia 
by the usual pairs of connectives, the anterior, the cerebro- 
pedal, being very much finer than the pleuro-pedal con- 


48 H. 0. N. SHAW. 


nectives. Both run parallel and close alongside of each 
other entering the sides of the pedal ganglia, and issuing 
close together from the posterior and external edges of the 
cerebral ganglia and the anterior edges of the pleural ganglia. 
‘he pleuro-subintestinal and the zygoneurous connectives 
are of about equal size. The right visceral ganglion, to which 
the former proceeds, is not pyriform, but of an oblong shape, 
and the connectives enter at the opposite extremities on the 
anterior surface. The three visceral ganglia are situated in 
about the same positions, but the commissure connecting 
them is shorter. This latter has its origin in the posterior 
and left side of the sub-intestinal ganglion behind the entry 
of the pleuro-subintestinal connective. 

The buccal ganglia are attached to the cerebral ganglia in 
the same manner, but their connectives are almost twice as 
long. The former ganglia are smaller, but globular in shape 
instead of flat, while their commissures are much longer, the 
anterior being almost as stout as the posterior. ‘I'he most 
anterior nerve collar, the cerebro-buccal, is not closely 
attached to the walls of the cesophagus, but is considerably 
larger than the latter, beng very nearly as large as the 
cerebro-pedal nerve collar. ‘his latter with the pleuro-pedal 
are the same as in C. tulipa, with the exception that they 
both lie more obliquely across the cesophagus. ‘he fourth 
collar, connecting the pleural and the sub-intestinal ganglia, 
like the second and third, is much more oblique. 

The tentacular and optic nerves are the same, with the 
exception that the former are not nearly as stout, and have 
not the kink or elbow bend, the optic nerves issuing from 
them as already described. 

The connective between the right tentacular nerve and the 
nerve issuing from the right pedal ganglion, and which, in 
the former species, was so noticeable, is absent in C. textile, 
there being no connection of any sort between the tentacular 
nerve and the pedal ganglia. 

The left or anterior otocyst is placed closer to the pedal 
ganglia, while the right one is in about the same position. 


CONUS TULIPA, LINN., AND CONUS TEXTILE, LINN. 49 


The left acoustic nerve issues above the left cerebro-pedal 
connective instead of below it and follows much the same 
course, but runs along the floor of the body-cavity under the 
anterior pedal nerves instead of through them. The right 
acoustic nerve runs along the dorsal surface of the right 
cerebro-pedal connective. 

There is no difference of note in the proboscidean nerves. 

I have already mentioned that the nerve collar round the 
cesophagus formed by the two cerebral and two buccal ganglia 
is of much greater diameter. Two nerves (s.3, s.4) issue 
from the left of these latter ganglia and proceed to the 
radula-sac in like manner, but are stouter, the fine nerve 
(s. 2) running forward; the buccal-proboscidean nerve is 
present, but only one nerve from the right ganglion to the 
radula-sac, and the small nerve (s. 1) from the same ganglion 
to the anterior part of the poison duct is also absent. The 
main poison gland nerve (s. 5) is slightly modified, as shortly 
after leaving the posterior commissure it is bifurcated, the 
right branch innervating most of the poison duct. The left 
branch, after proceeding backwards for some distance, is in 
turn divided into two, these two innervating the poison gland 
itself. 

Lastly, the fine nerve (s. 6) from the main nerve (s. 5) to the 
cesophagus, branches off from the former further back, and on 
reaching the surface of the cesophagus sends one nerve forward 
and one back over the surface of the latter. Though some- 
what modified, the function of this nerve is the same, and its 
existence is the chief point in question. 

The two columellar nerves having their origin in the left 
pleural ganglion are of equal size, and slightly stouter, and not 
so sinuous. Both lie under the cesophagus, and innervate the 
columellar muscle in much the same way. The four nerves 
issuing from this ganglion—viz. the main and lesser pleuro- 
siphonal and the two parietal nerves—differ but little, the 
anastomosis between the anterior branchial nerve and the main 
pleuro-siphonal being slightly longer. 

I have already mentioned that the posterior portion of the 

vot. 60, PART 1.—NEW SERIES. 4 


50 H. O. N. SHAW. 


supra-intestinal ganglion is reflected over to the left. There 
are only three nerves besides the left loop of the visceral 
commissure which proceed from this ganglion, namely, the 
anterior and posterior branchial nerves and one parietal nerve. 
With the exception that they pass straight across the body- 
cavity to the left they are of no particular interest. The 
branch given off from the main or anterior branchial nerve is 
again present, but much smaller, with the result that the 
anterior and posterior branchial nerves are closer together, 
and, therefore, in more normal positions. 

The sixth nerve (c.4) in C. tulipa is entirely absent in 
this species. 

The three visceral ganglia and their commissures require 
no special comment, being much the same in both species, the 
commissure being slightly shorter. 

The difference in shape of the sub-intestinal ganglion has 
already been noticed. There are only three nerves which 
issue from the ganglion besides the right loop of the visceral 
commissure. The nerve (e.) performing the functions of a 
parietal and columellar nerve is absent. 

The pedal ganglia are even more closely connected and 
ensheathed by connective tissue. The nerves leave the right 
ganglion, with one exception, from its anterior extremity only, 
while they issue from the left side as well as from the anterior 
of the left ganglion; tnus, more nerves are given off from the 
left than from the right ganglion—the reverse of what occurs 
in C. tulipa. In the latter there were 41 nerves altogether 
from both ganglia; in this species there are 39—21 from 
the left, 18 from the right ganglia, as compared to 30 and 11. 
None of these 39 pedal nerves attain so great a size as the 
chief ones in C. tulipa, and all enter the floor of the body- 
cavity sooner. The right ganglion supplies the posterior 
portion of the foot, while the central and anterior parts are 
innervated from the left ganglion. It will thus be noticed 
that the functions of the ganglia in the different species have 
been reversed, since in C. textile it is the left one that is 
the most important. The ganglia themselves do not lie at so 


CONUS TULIPA, LINN., AND CONUS TEXTILE, LINN. 51 


great a tangent to the longitudinal body-axis, but more at 
right-angles to it. 

There are no points of any great difference, or of special 
interest, between the nervous systems of both species beyond 
those already mentioned. 


Crrcunatory Sysrem or Conus Toura (Pl. 5, fig. 22, and 
Pie Gey fiers, 25): 


As in C. textile, so in this species, I have only studied 
the artery passing forward through the nerve collars. ‘This 
artery (g.), after leaving the heart, proceeds forward and 
obliquely to the right till it reaches the left branch of the 
visceral commissure (d.) anterior to the left visceral ganglion. 
For some considerable distance, both commissure and artery 
are surrounded by one sheath of muscular and connective 
tissue, the artery being on the right of the commissure. 
Owing to the artery and commissure having been cut through, 
where surrounded by this sheath, while removing the 
cesophagus from the nerve collars, I at first mistook the artery 
for a nerve given off from the commissure, and, indeed, had 
dissected it out as such, being considerably perplexed by the 
existence of such a nerve, till on having sections cut I realised 
my mistake. It is owing to this error that I have been 
enabled to work on the arterial system at all, for being then 
engaged on the nervous system, had it from the beginning 
clearly shown itself to be an artery, most probably it would 
have been removed without much comment, and some 
interesting facts remained unknown. 

Having branched off from the visceral commissure, the 
artery passes to the right almost at right angles to the longi- 
tudinal body axis, and under the cesophagus. A_ little 
distance after its separation from the commissure, a branch 
artery (j.) is given off nearly at right angles and passes 
upwards to form a loop which partly surrounds the cesophagus, 
while the end of the loop is bifurcated, one branch running 
up and the other down. 


52 H. O. N. SHAW. 


The main artery proceeds through the three nerve collars, 
pleuro-subintestinal, pleuro-pedal, and cerebro-pedal, till it 
arrives at the posterior cleft between the two pedal ganglia. 
Here a stout but very short branch (op./.) opens into the 
dorsal surface of the left pedal ganglion at its posterior edge. 
‘The artery runs down the dorsal constriction between the 
two ganglia and over the anterior edge. When about half 
way along the depression between the ganglia, the artery 1s 
slightly expanded, and two branches issue, that to the right 
(op.r.) opening into the dorsal surface of the right pedal 
ganglion, while the left branch (w.), which is much longer, 18 
directed upwards and divides into two, each being again 
bifurcated and spreading over the radula-sac. 

The branches opening into each of the ganglia are of the 
same size and length, but the left one enters the extreme 
posterior edge of its ganglion, while the right one opens into 
the centre of the dorsal surface of the right pedal ganglion. 

After leaving the anterior edge of the ganglia, the main 
artery curves to the left, becomes larger, and runs directly 
forward, two arteries branching off from it on the right side, 
then one on the left, and still more anterior to this latter, the 
main artery is bifurcated. 

The first branch (c. ft.) to be given off to the right, runs 
backwards and through some of the pedal nerves, and enters 
the centre of the foot, where it divides into two branches 
supplying the central portions of the latter. 

The second branch to the right (v. ft.) proceeds directly 
downwards into the foot to its ventral surface, where it forms 
a JT, one arm running forward and the other backward. 
Slightly anterior to this artery, the main artery sends off the 
only branch to the left (te.), which proceeds to the base of 
the cephalic integuments. The left bifurcation of the main 
artery (y. 1) runs forward through the dorsal portion of the 
foot, while the right (y.), passing downwards into the foot, 
attains the extreme right anterior basal surface of the latter. 

In section, the main and branch arteries are for the most 
part oval (Pl. 5, fig. 22), and are composed of two distinct 


CONUS TULIPA, LINN., AND CONUS TEXTILE, LINN. 59 


sheaths or layers, of which the inner (J, c. m.) consists of 
circular muscular tissue lined with endothelium (end.), while 
the outer sheath (I. m.s.), which is slightly thicker, is built 
up of areolar and muscular tissue running longitudinally. 


DIFFERENCES BETWEEN THE ARTERY IN Conus TEXTILE AND 
Conus JTULIPA. 


Although the position and functions of the artery in both 
species are roughly the same, there are many points of great 
difference between them. 

In the first place, the artery in C. textile has no connec- 
tion whatsoever with the visceral commissure, and lies entirely 
under the esophagus. In C.tulipa, for some distance, the 
artery and commissure are enclosed in the same sheath, and 
the artery traverses the nerve collars in a direct line for the 
pedal ganglia. In C. textile the main artery passes through 
the nerve collars some distance to the left of these ganglia, and 
a branch artery crosses the dorsal surface of the pedal gangha, 
but has no connection with them. In C. tulipa this branch 
is absent, but the main artery opens directly into the dorsal 
surface of each ganglion, while from the same place as the 
opening into the right ganglion a stout but short branch 
proceeds to the radula-sac. | 

The radula-sac of C. textile is supplied from a branch 
which is given off from the main artery before the latter 
passes through the pleuro-pedal nerve collar. This branch 
is bifurcated, the left arm running backwards and through 
the pleuro-subintestinal collar and» supplies the cesophagus, 
while in C. tulipa the latter has a direct supply from the 
main artery posterior to the last-named nerve collar. In this 
latter species the main artery, after passing over the pedal 
ganglia, becomes considerably larger, runs directly forward 
and divides up into five branches, which supply the base of 
the cephalic integuments, the anterior dorsal surface, the 
anterior right and basal surface, the central ventral, and the 
central portions of the foot respectively. 


o4 H. O. N. SHAW. 


In C. textile, anteriorly and to the right of the left pedal 
ganglion the main artery forms a cross, the left arm running 
to the base of the siphon, the central to the anterior and basal 
surface of the foot, while the right arm, which is twice the 
length of the other two, proceeds directly backwards to the 
extreme posterior and dorsal surface of the foot. Lastly, in 
section, the artery is of a different construction in each species. 
In C. tulipa it is composed of two layers of tissue, of which 
the outer, or tunica adventitia, which is slightly thicker than 
the inner, is built up of areolar and longitudinal muscular 
tissue. The inner layer, or tunica media, is composed of mus- 
cular tissue running circularly round the arterial canal, and 
having a smooth inner surface lined with endothelium, or 
pavement epithelium. ‘he reverse in every point, except the 
endothelial layer, is the case in C. textile. The outer layer, 
which is half the thickness of the inner, is built up of areolar 
and circular muscular tissue ; the inner layer is a confused 
mass of circular and longitudinal tissue of the same sort ; while 
the internal surface, which is lined with endothelium, is deeply 
corrugated, aud a section looks more like a vein than an artery. 
The total thickness of the arterial wall in this species is twice 
that in C. tulipa. 

Though superficially resembling one another, from the fore- 
going comparison the many and great differences between 
the arteries in the two species will be observed. 

Poli (21), pl. xlv, fig. 18, and p. xxxix has illustrated and des- 
cribed in Conus rusticus Linn. an artery which he calls 
““Paorte abdominale,” which is similar to that in C. textile 
in some respects. According to Poli’s figure, the artery runs 
forward from the ventricle of the heart beside the left loop of 
the visceral commissure, across the latter, and forward through 
the nerve collars, where it is anteriorly split up into several 
small branches. 

In concluding this paper, it is hoped that the contents on 
the anatomy of two species of the genus Conus may be of 
use to workers on this group. Since I commenced my work 
on this genus I have received from various sources a number 


CONUS TULIPA, LINN., AND CONUS TEXTILE, LINN. 55 
of different species, and I hope shortly to be able to start 
work on them with a view to further determining the ana- 
tomical variations between different species. 


BIBLIOGRAPHY OF THE CHIEF Works CONSULTED. 


1. Adams, H. and A.—‘ The Genera of Recent Mollusca,’ 1853-8, 
London. 


2. Amaudrut, A.—“ La partie antérieuse du tube digestif, et la torsion 
chez les Mollusques gastéropodes,” ‘Ann. des Sci. Nat.,’ (8), vil, 
1898, p. 1. 


3. Bergh, R.—‘“ Beitriige zur Kenntniss der Coniden,” ‘ Nova Acta der 
Ksl. Leop-Carol.,’ ‘ Deutschen Akademie der Naturforscher,’ Bd. 
Ixy, Nr. 2, 1895. 

4, Bernard, F.—* Recherches sur les organes palléaux des Gastéropodes 
prosobranches,” ‘ Ann. des Sci. Nat.,’ (7) ix, 1890, p. 88. 

5. Bouvier, E. L.—< Systeme nerveux, Morphologie générale et classifica- 
tion des Gastéropodes prosobranches,”’ ‘ Ann. des Sci. Nat.,’ (7) 


ili, 1887. 
' 6. Cuvier, G.— Régne Animal. Mollusques,’ by G. P. Deshayes, Paris, 
1836-49. 


7. Eydoux, F., et Souleyet, L.—* Voyage autour du Monde sur la 
corvette la Bonite,”’ ‘ Zoologie,’ vol. ii, pp. 632-3, pl. 45, Paris, 
1852. 

8. Fischer, P.—‘ Manuel de Conchylologie,’ Paris, 1887. 

9. de Freycinet, L.—‘* Voyage autour du Monde,” ‘ Zoologie par Quoy et 
Gaimard,’ pp. 437-440, pl. 69, Paris, 1824. 

10. Gray, J. E—* On the Division of Ctenobranchous Gasteropodous 


Mollusca into larger Groups and Families,’ ‘Ann. and Mag. Nat. 
Hist.,’ 2nd series, vol. xi, No. lxii, paper xiii, 1853. 


11. “On the Head of the Genus Conus,” ‘Ann. and Mag.,’ 2nd 
series, vol. xii, No. lxix, paper xix, 1853. 
12. ‘Guide to the Systematic Distribution of Mollusca in the 


British Museum,’ Pt. I, 1857, London. 


18. Jhering, H. von.—‘ Vergleichende Anatomie des Nervensystems 
und Phylogenie der Mollusken,’ Leipzig, 1877. 


14, “ Beitriige zur Kenntniss des Nervensystems der Amphi- 
neuren und Arthocochliden,” ‘ Morph. Jahrbuch,’ iii, 1877. 
15. “ Gibt es Orthoneuren ?” ‘ Zeit. f. wiss. Zool.,’ xlv, p. 499, 


1887. 


56 H. O. N. SHAW. 


16. Johnston, G.—‘ Introduction to Conchology,’ 1850, London. 

17. Lacaze-Duthiers.—* Otocystes ou capsules auditives des Mollusques 
(Gastéropodes),” ‘Arch. de Zool. expér.,’ (1), i, 1872. 

18. de Montfort, D—‘ Conchyliologie systématique et classification 
méthodique des Coquilles,’ vol. ii, 1810, Paris. 

19. Pelseneer, P.—‘* The Mollusca,” in a Treatise on Zoology, edited by 
E. Ray Lankester, Part v, London, 1906. 

20. Perrier, R.—* Recherches sur l’anatomie et l’histologie du rein des 
Gastéropodes prosobranches,” ‘ Ann. des Sci. Nat.,’ (7) viii, p. 61. 

21. Poli. J. X., et Delle Chiaje, S.— Testacea utriusque Sicilie, 
eorumque historia et anatome,’ iii, Parme, 1826. 

22. Quoy et Gaimard.—* Voyage de l’Astrolabe,” ‘ Zoologie,’ iui, Paris, 
1834. 

23. Simroth, H.— Gastropoda prosobranchia,’ ‘ Bronn’s Tierreichs. 
Mollusea,’ 1896-1907. 

24. Troschel, F. H.—‘ Das Gebiss der Schnecken,’ vol. ii, Berlin, 1866- 
1893. 

25. Woodward, S. P.— A Manual of the Mollusca,’ London, 1858, 1869, 
1875. 


EXPLANATION OF PLATES 1-6, 


Illustrating Mr. H. O. N. Shaw’s paper ‘On the Anatomy of 
Conus tulipa, Linn., and Conus textile, Linn.” 


LETTERING IN ALL THE FIGURES EXCEPT NUMBERS IN Fie. 20. 


a. Anastomosis between pleuro-siphonal and anterior branchial nerves. 
ac. 1. Left acoustic nerve. ac.7. Right acoustic nerve. ats. Tooth 
- attachments to sac wall. B. Buccal ganglia. 6. Anterior branchial 
nerve. 6.1. Branch off anterior branchial nerve. b. 2. Posterior 
branchial nerve. b. a. Basal attachment of radule. 6. c. Body-cavity. 
br. c. Branchial cavity. bs. Barbs of teeth. C. cerebral ganglia. c., 
c.1, c. 2, c. 3. Parietal nerves to left of body. c. 4. Main and posterior 
parietal nerve. ca. Canal. c.b. Cerebro-buccal connectives. ce. pl. 
Cerebro-pedal connectives. c. ft. Artery to centre of foot. c¢.m.s, 
Circular muscular sheath. co.z. Inner corneal layer. co. 0. Outer 
corneal layer. ct. Ctenidium. d. Visceral commissure. d.' Pleuro- 
sub-intestinal connective. dt. op. Duct opening. dts. Ducts from 
unicellular glands. e., e. 1. Parietal nerves to right of body. e.2 


« te 


Columellar nerve. e. 3. Right pleural nerve. end. Endothelium. ep. 


CONUS TULIPA, LINN., AND CONUS TEXTILE, LINN. 57 


Epithelium of tentacles. ep. cs. Epithelial cells. f. Main pleuro-siphonal 
nerve. jf.1. Lesser pleuro-siphonal nerve. ft. Foot. g. Artery from 
heart through nerve collars. g.c¢.¢. Granular contents of glands. h. 
Hood-like process of radula-sac. 7.1. Left columellar nerve. 7.2. Right 
columellar nerve. inv. Invagination into siphon. div. ca. Invagination 
into canal. y. Artery to the esophagus. k. Right pleural nerve. J/.1, 
1.2, 1.3, 1.4, 1.5, 1.6. Proboscidean nerves. J.c.l.m. Layer of inter- 
mingled circular and longitudinal muscle. J/.c.m. Layer of circular 
muscle. Je. Lens. J. ep. Lining epithelium. J/./. Left lobe of liver. 
L.d.m. Layer of longitudinal muscle. J/. m. s. Longitudinal muscular 
sheath. m. Hepatic nerve. m.1. Visceral nerve. m. 2. Recto-genital 
nerve. mh. Mouth. ml. Mantle. «. Hepatic nerve from visceral com- 
missure. n.c. Nervecollar. nw. Nuclei. o. Opticnerve. @. Gsophagus. 
o. p. d. Opening of poison duct. op. l. Artery opening into left pedal 
ganglion. op.r. Artery opening into right pedal ganglion. 0.7. s. 
Opening of radula-sac. of.1. Left otocyst. otf, r. Right otocyst. P. 
Left pedal ganglion. P.1. Right pedal ganglion. pa. pl. Pleuro-pedal 
connectives. p. ft. Artery to posterior portion of foot. p. g. Poison 
gland. p.g.d. Poison-gland duct. Pl. Pleural ganglia. ps. Proboscis. 
r. Nerve from right pedal ganglion joining tentaculo-pedal connective. 
vet. Retina. +.g.m. Recto-genital mass. 7m. Rostrum. r.s. Radula- 
sac. 7+.t. Row of denticulations. §. Supra-intestinal ganglia. s. 1. 
Nerve to anterior portion of poison duct. s. 2. Buccal proboscidean 
nerve. s.3,s.4. Nerves to radula-sac. s.5. Main nerve to poison gland 
and duct. s.6. Fine nerve to esophagus. s.d. Salivary duct.  s.g. 
Salivary gland. 8.7. Sub-intestinal ganglion. sz. Siphon. — sp. pa, 
Sporozoon parasites. st. Stomach. ¢. Main tentacular nerves. f. 1, 
t.2, t.3. Branches of main tentacular nerves. te. Artery to cephalic 
integuments. tent. mus. Muscle of tentacle. ¢. p. Tentaculo-pedal 
connective. ts. Tentacles. w.g. Unicellular salivary glands. JV. 1. 
Right visceral ganglion. V.2. Median visceral ganglion. V. 3. Left 
visceral ganglion. v. ft. Artery to ventral surface of foot. w. Artery 
to radula-sac. #. Artery to base of siphon. y. Artery to anterior base 
of foot. y. 1. Artery to anterior dorsal surface of foot. z. Zygoneur 
between right pleural ganglion and sub-intestinal ganglion. 


PLATH 1 


Fig. 1—Conus tulipa. A general view of the anatomy. The 
mantle (ml.) has been cut open, and reflected to the left showing the 
ctenidium (ct.) lying ina curve along its dorsal surface. The ospradium, 
which is situated on the inside of the ctenidium, is hidden by the base 
of the rostrum. This latter (rm.) has also been opened along its dorsal 


58 H. O. N. SHAW. 


surface, showing its thick muscular walls and tentacles (ts.) and the 
proboscis (ps.) situated at its base. The dorsal wall of the body-cavity 
has been partly removed, and the large poison gland (p. g.) will be seen 
to occupy all the posterior portion. The poison gland duct (p. g. d.) is 
coiled across and anterior to the gland, while the radula-sac (r. s.) is 
anterior and partly dorsal to the poison duct. An incision has been 
made in the base of the proboscis to show the cesophagus (@.) passing 
up the centre to the mouth (mh.). (As the various organs are situated 
in the same positions in both Conus tulipa and Conus textile, I 
have used this figure to illustrate the description of the latter species, as 
well as the one it depicts.) 

Fig. 2.—Conus textile. A view of the poison gland (p.g.) and 
duct (p. g.d.), esophagus (@.), salivary gland (s. g.), and radula-sac (7. s.) 
removed and spread out to show the coils and length of poison duct, 
stomach (st.), salivary ducts (s. d.) 

Fig. 3—Conus tulipa. A view of the poison gland (p.g.) and 
duct (p. g.d.), esophagus (@.), salivary gland (s. g.), and radula-sac (r. s.) 
removed and spread out as in fig. 2, to show the difference in size of 
poison gland, constricted opening of duct, which latter is shorter and 
less coiled than in C. textile. Hood-like process (h.) of radula-sac, 
not present in C. textile, stomach (s¢t.), salivary ducts (s. d.). 

Fig. 4.—A tooth of Conus textile. Barbs (bs.), basal attach- 
ment (0. a.). 

Fig. 5.—A tooth of Conus tulipa. Barbs (b.s.), basal attachment 
(b.a.), row of denticulations (r. t.). 


PLATE 2. 


Fig. 6.—Conus tulipa. A longitudinal section through poison 
gland, showing layers of muscle-fibres, highly magnified. 

Fig. 7.—A transverse section of same. 

Fig. 8.—A transverse section of same, highly magnified. 

Fig. 9.—Conus textile. A transverse section of poison gland, cu. 
canal, c.m.s. circular muscular sheath, J. ep. lining epithelium of canal, 
l. m.s. longitudinal muscular sheath. 

Fig. 10—Conus textile. A transverse section of poison duct, 
showing long club-shaped cells, and irregular surface of canal. 

Fig. 11—Conus tulipa. A transverse section of poison duct showing 
invagination hanging down into canal, c.a. canal, ep. cs. epithelial cells, 
inv. ca. invagination into canal, lJ. c.m. layer of circular muscle-fibres, 
1. l. m. layer of longitudinal muscle-fibres. 


CONUS TULIPA, LINN., AND CONUS TEXTILE, LINN. 99 


Fig. 12.—Conus tulipa. A transverse section through the eye, co. 7. 
Inner corneal layer, co. 0. outer corneal layer, ep. epithelium of tentacle, 
le. lens, 0. optic nerve, vet. retina, tent. mus. muscle of tentacle. 


Fig. 13—Conus tulipa. A transverse section through the 
cesophagus, showing sporozoon parasites in the lining epithelium, which 
has shrunk away from the muscle wall. ca. Canal, l. c. m. layer of circular 
muscle, /. ep. lining epithelium, J. 1. m. layer of longitudinal muscle-fibres, 
sp. pa. sporozoon parasites. 

Fig. 14.—Conus tulipa. A transverse section through the end of 
poison gland, showing disposition of layers of muscle-fibres and opening 
of poison duct (0. p. d.). 


Fig. 15.—Conus textile. A transverse section through cesophagus, 
thick muscular walls, with a layer of longitudinal and circular muscle- 
fibres (l. c. 1. m.) intermixed in the centre. 


Fig. 16.—Conus textile. A transverse section through salivary 
gland showing entry of one of the two main ducts. dts. Ducts from 
unicellular glands, s. d. salivary duct, w.g. unicellular glands. 

Fig. 17.—The above highly magnified, showing a group of five 
unicellular glands with nuclei and granular contents. g.c. ¢. Granular 
contents of cells, dt. op. duct opening, nw. nuclei. 


PLATE 3. 


Fig. 18—Conustulipa. Thenervoussystem. Buccal ganglia 
(B.) and nerves issuing therefrom, with the tentacular and optic nerves 
given off from the cerebral ganglia (C.). The two pleural and the supra- 
intestinal ganglia are not shown. 


PLATE 4. 


Fig. 19—Conus tulipa. Thenervous system.—The cerebral 
ganglia (C.), pleural ganglia (Pl.), the sub (Si.) and supra-intestinal 
ganglia (S.), with the nerves given off from each, also the visceral 
commissure (d.), and the three visceral ganglia (V.1, V.2, V.3,). The 
pedal ganglia, cerebro-pedal, and pleuro-pedal connectives are not 
represented. 

Fig. 20.—Conus tulipa. The nervous system. The pedal 
ganglia P. and P. 1 with the nerves issuing, and the cerebro-pedal (ce. pl.) 
and pleuro-pedal (pa. pl.) connectives. Parts of the foot and body to 
which the nerves (1-41) proceed. 1 to anterior floor of body-cavity in 
front of the columellar nerves (7. 1, 7. 2), 2 to centre of floor of body- 
cavity, nerve much crenulated, 3 to left and anterior portion of foot, 4 
and 5 to floor of body-cavity, 6 and 7 to left anterior side of foot, 8and 9 


60 H. O. N. SHAW. 


to left anterior extremity of foot (main nerves), 10 to central anterior 
extremity of foot, 11, 12 and 13 to right anterior extremity of foot, 14 
main nerve to right anterior extremity of foot, 15 to anterior and right 
side of foot, 16 to right side of foot, 17 to extreme light and anterior 
part of body-wall, 18 to anterior and right side of foot, 19 to extreme 
right and anterior part of body-wall, 20 to right anterior side of foot, 21 
and 22 to centre and right part of foot, 23 to ventral and central surface 
of foot, 24 to right side of body-wall, 25 and 26 to centre and lower 
surface of foot (25 main nerve), 27 and 28 to centre of foot, 29 main nerve 
to extreme posterior end and base of foot, 30 to centre of foot, 31 under 
operculum, 32 through right corner of body-wall, and to extreme end 
and base of foot, 33 to right posterior corner of body-wall, 34 to floor 
and body-wall, 35 to base of posterior wall of body-cavity, 36 to left 
central side of foot, 37 to posterior wall of body-cavity, 38 to left and 
central side of foot, 39 to base of posterior body-cavity wall, 40 and 41 
to floor of body-cavity, under cerebral ganglia. 


PLATE 5. 


Fig. 21.—Conustulipa. Thenervoussystem. Nerve centres. 
The ganglia with their connectives and nerves. Only nine nerves are 
shown issuing from the pedal ganglia (P.and P.1). The buccal ganglia 
(B.) with the cerebro-buccal connectives (c. b.), which lie parallel to the 
cerebro-pedal connectives (ce. pl.), have in this figure, for the sake of 
clearness, been brought forward. The left and right otocysts (of. L., 
ot. 7.) with their nerves (ae, l., ac. 7.) can be clearly seen, as also the 
nerve (r.) issuing from the right pedal ganglion (P.1), and united by 
means of the tentaculo-pedal connective (¢. p.) to the right tentacular 
nerve (t.). 

Fig. 22.—Conus tulipa. A transverse section of the artery to the 
pedal ganglia. 

Fig. 23.—C onus textile. A transverse section of the artery which 
passes through the nerve collars. 


Pia 6, 


Fig. 24.—Conus textile. A general view of the nerve centres 
(semi-diagrammatic), showing the position and branches of the artery 
proceeding from the heart through the nerve collars to the foot, ete. 

Fig. 25—Conus tulipa. A general view of the nerve centres (semi- 
diagrammatic), showing the position and branches of the artery which 
proceeds from the heart through the nerve collars to the pedal ganglia 
and foot, ete. 


CENTRIFUGED “EGG OF THE FROG. 


+ 


On the Relation between the Structure and the 
Development of the Centrifuged Egg of the 


Frog. 


By 


J. W. Jenkinson, M.A., D.Se., 


University Lecturer in Embryology, Oxford; Fellow of 
Exeter College. 


With Plates 7—12 and Text-figs. 1—18. 


ConTENTS. 


I. INTRODUCTORY...-~ 
Il, THe EXPERIMENTS. 
A. On the CentriPiged Egg 
(1) Method’. 
(2) Details of the Experiments 
(3) The Effect of a Centrifugal Force on— 
(a) The Egg Structure 
_ (b) Segmentation . 
— (e) Development 
B. On € ntrifuged Egg-Pulp. The Guomcey of he 
Constituents of the Cytoplasm ; 
c. Conclusions to be Drawn from the eee Experi- 
ments 
III. On THE RELATION BETWEEN THE Gennes tana Sheree 


OF THE EGG AND THE DEVELOPMENT OF THE EMBRYO IN 
GENERAL . 


I. INTRODUCTORY. 


108 
1167 
117 
130 


140 


144 


Iv is to the structure and constitution of the germ-plasm, of 
the cytoplasm as well as of the nucleus of the germ-cell, that 
the path of inquiry into the internal causes of the development 


| 
} 
J 


{ 


| 


62 J. W. JENKINSON. 


of the organism, the causes which determine in each generation 
the reproduction of the specific form, inevitably leads back. 
It is not alittle curious that a method of experimental research 
which has already given something and promises perhaps to 
contribute still more to the analysis of the factors resident in 
the egg-cytoplasm should owe its existence to those classical 
experiments of the physiologist Emil Pfliiger, which led their 
author to formulate the doctrine of the “isotropy ” of the 
ovum, and to deny to the egg-cytoplasm any structure, or at 
least any significant structure at all. 

Normally, the eggs of the frog and toad—Pfliiger used the 
ova of Bombinator—are not free to rotate inside their 
tightly adherent jelly-membranes when first they are laid. 
But shortly after and as a consequence of insemination there is 
developed between the egg and the innermost layer of the jelly 
a perivitelline space, inside which the egg then rotates until 
its axis becomes vertical with the heavier yolk-pole below. It 
will be recalled that Pfliiger succeeding in inhibiting the 
formation of this space, though not the subsequent segmenta- 
tion and development, by giving each eg’; only a minimal drop 
of sperm-containing water. Such eggs ould be kept forcibly 
inverted in any position in which the ex:perimenter chose to 
place them, with the original axis making any angle with the 
vertical, and, except when the inversion was absolutely complete, 
were capable of normal segmentation and development. In 
cleavage it was found by Pfliiger that the firs two divisions 
were vertical and the third horizontal, just as in the undis- 
turbed egg the first two (meridional) and third (latitudinal) 
are also respectively vertical and horizontal, while later on the 
dorsal lip of the blastopore appeared just below the actual 
equator, and travelled downto the actual lower pole in the plane 
including the actual (vertical) and the original (inclined) egg- 
axes. The conclusion drawn from these facts—that the embryo 
may be developed from any part of the egg, that the egg is 
consequently isotropic, and only undergoes a specific develop- 
ment because it is always brought under the same external 
conditions—was not allowed to remain unchallenged for long. 


CENTRIFUGED EGG OF THE FROG, 63 


First Roux showed that the constant directive influence of 
gravity might be easily eliminated by keeping the eggs in a 
perpetual state of slov rotation, under which circumstances 
their segmentation and development bear the same relation to 
the original axis and polar structure as is ordinarily the case. 
And secondly, Born’s examination of sections showed at once 
that in these forcibly inverted eggs there takes place a re- 
distribution of material, the hghter liquid plasma ascending, 
the heavier yolk-particles descending until there is conferred 
upon the ovum a new structure with a new axis, which is of 
course vertical, and has unlike animal or protoplasmic and 
vegetative or yolk poles. The pigment, however, is not wholly 
shifted, though some is carried up into the lighter plasma. 
The upturned white pole remains white, or at most becomes 
greyish. To this new structure—which may have any relation, 
make any angle with the original structure—the cleavage and 
development of the egg has the same relation as normally. 
With regard to the new axis the first and second furrows 
are meridional, the third latitudinal, while the head of the 
embryo appears near the new animal pole. The median plane 
of the embryo is that plane in which both the original and 
the new axes lie, for the streaming up and down of the plasma 
and yolk take place symmetrically about that plane, and the 
bilaterality so conferred upon the egg-contents persists as 
that of the embryo. 

The very experiments, therefore, which were vainly imagined 
to prove the isotropy of the cytoplasm have only succeeded in 
emphasising the significance of the structure of the ovum for 
the development of the embryo. 

This experiment can be performed in what is, I think, a 
more convenient manner by substituting for gravity a cen- 
trifugal force, greater than gravity, but not too large. As 
soon as the eggs have been inseminated they are placed with 
sufficient water to allow the jelly to swell and the perivitelline 
fluid to be exuded, in the tube of the centrifuge, and centri- 
fuged at once for a short time. The eggs, of course, lie hap- 
hazard in the tube, with their axes making any angle with the 


64, J. W. \JENKINSON. 


direction of the force, and sihce they cannot move until the 
perivitelline space has been de veloped their contents are re-dis- 
tributed as in Pfliiger’s experimewt. Taey are then removed 
from the machine. It wiil be found that they do not turn 
over, that the first and second cleavages are parallel, the 
third at right angles to the direction of the force, and that the 
dorsal lip of the blastopore appears just on the centrifugal 
side of the (new) equator of the egg. 

Wetzel and Hertwig have recently employed this method in 
the study of the particular case of complete inversion, which 
does not apparently prevent the formation of the embryo, as 
stated by Pfliiger. 

The credit of subjecting the eggs of the frog to a still 
higher centrifugal force belongs to O. Hertwig. He found 
that the sezmentation of such eggs was meroblastic. A cap 
of small cells or blastoderm was formed resting upon an 
undivided though nucleated yolk, and these yolk-nuclei were 
large and irregular, resembling the giant nuclei found in the 
large-yolked eggs of fishes and other forms. If removed 
from the centrifuge in time these eggs developed, though 
monstrosities (spina bifida) were frequent. 

More recently Morgan, Konopacka and McClendon have 
all made contributions to the problem of the relation between 
this disturbance of the egg-structure aud the development of 
the embryo. 

Morgan has shown that with a still higher speed (1600 revo- 
lutions per minute for seven minutes, R = 5 in., f. = 370 g.) 
the egg of the American species Rana sylvatica develops 
a grey patch round the animal pole owing to the heavy 
pigment being driven centrifugally into the interior of the 
egg like a plate. In these eggs the first and second furrows 
are approximately meridional, but those of the third phase 
are abnormal in being again meridional. The dorsal lip of 
the blastopore is in the normal position, but the yolk plug 
may be pigmented. The tadpole is antero-ventrally unpig- 
mented, but this defect may be made good in later stages. 
Morgan refers to but does not follow up nor adequately describe 


CENTRIFUGED EGG OF THE FROG. 65 


an interesting abnormality which he says is not uncommon. 
In this the front end of the nervous system is malformed, the 
neural folds each terminating in a knobbed extremity. With 
still longer exposures the yolk fails to segment, but ‘‘ the more 
fundamental questions relating to the distribution of the 
materials of the egg, and the interpretation as to whether 
these visible substances are organ-forming or organ-deter- 
mining have not been discussed. It is evident that the black 
pigment has no such function, but further experiments will be 
necessary in order to determine what value the other sub- 
stances in the egg may have ””—a conclusion with which we 
may cordially agree. 

The principal object of the work of Konopacka (on Rana 
fusca) is to discover whether there is any alteration in the 
sensitiveness of the egg to this disturbing agent during the 
early stages of fertilisation and segmentation, and it does 
indeed appear that the number of abnormal embryos is greater 
when the eggs are centrifuged during the early cleavages than 
when the operation is performed upon unfertilised intra-uterine 
eggs, or on eggs in process of fertilisation. Apart from this, 
however, the paper contains a description and figure, the 
first published, of the alteration in the structure of the egg- 
cytoplasm, as well as an account of some of the monstrosities 
produced. With short exposures to a fairly high acceleration 
(f = 228 g.) the grey area of Morgan appears, and is seen in 
sections as a layer of yellow vacuolated hyaloplasm sur- 
mounting the inwardly driven pigment and the yolk. With 
longer exposures a white hyaloplasmatic layer is interpolated 
between the vacuolated and pigmented layers, while another 
vacuolated layer appears between the white layer of hyalo- 
plasm and the yolk while the first vacuolated layer becomes 
folded. No attempt is made, however, to investigate the 
gradual genesis of these layers, nor to ascertain the chemical 
nature of the substances composing them. Amongst the 
abnormalities described are numerous half-embryos, a portion 
of the egg having remained unsegmented, embryos with 
persistent blastopores, and tailed but headless monsters. 


~ 


voL. 60, part 1.—NEW SERIES. 9) 


66 J. W. JENKINSON. 


The last, as I hope to show, are of particular. interest, but the 
author figures only the external appearance of one, whose 
development has been very much arrested, and gives no 
detailed description of the anatomy of either this or other 
stages of the malformation. 

The most recent work of all, that of McClendon, marks a 
very great advance, for here we have for the first time a 
chemical analysis of the various layers into which the egg 
cytoplasm is separated by the centrifuge. By the use of a 
considerable force for a short time (f= 2771 g., for five 
minutes) the eggs of Rana pipiens and Acris gryllus 
become divided into zones. In the first there are three 
zones, A, a yellow centripetal cap (corresponding to the grey 
area of Morgan and the vacuolated hyaloplasm of Konopacka) 
which consists of globules—soluble in oils and ether—which 
are mostly fat but partly lecithin, inasmuch as there is some 
phosphorus in the alcohol extract of this layer. The water 
content of this layer is 50 per cent. The second layer, B, is 
grey, of translucent protoplasm, 82 per cent. of which is water ; 
there is also some fatand lecithin. The third layer, ¢, is black, 
it consists of the yolk and pigment, contains 42 per cent. of 
water, some fat, a good deal of lecithin, and a large quantity of 
protein which is rich in phosphorus and therefore supposed to 
be a nucleo-protein. In Acris the first layer consists of a 
yellow cap and a white ring, while the pigmented portion of 
the third is divided into three rings. 

It goes without saying that the chemical composition 
of the several layers could not have been ascertained from 
any investigation of individual eggs. For this purpose a 
whole mass of eggs was centrifuged, the layers separated 
and analysed. The method was to take, not, of course, the 
laid egg with its coating of jelly, but ripe ovarian eggs. In 
actual practice the whole ovary was removed, including not 
only the fully grown ova, but all the young ones as well, 
washed and squeezed through bolting cloth to get rid of the 
stroma. The pulp so obtained was then centrifuged. This 
appears to me to introduce a certain error. I must also point 


CENTRIFUGED EGG OF THE FROG. 67 
out that McClendon has determined only the phosphorus 
content of the ethereal and alcoholic extracts and of the 
solid protein-containing residues, and has given no further 
proof of the presence of lecithin and nucleo-protein. ‘These 
points, therefore, and of course many others remain for investi- 
gation, but McClendon’s work is certainly a beginning, and 
the priority belongs to him. 

I myself had frequently had occasion to examine, somewhat 
cursorily, the development of the centrifuged eggs of the 
frog, and it had occurred to me quite independently, as indeed 
it would naturally occur to anyone with such eggs under his 
eyes, and before I became acquainted with McClendon’s papers, 
that the layers might be obtained in sufficient quantity for 
chemical investigation if a mass of egg-pulp were centrifuged. 
Moreover, if the abnormal development of the embryo really is 
a consequence of the derangement of the materials of the 
cytoplasm, it ought to be possible to relate a certain degree 
of malformation with a certain degree of derangement by com- 
paring on the one hand the composition of the several layers 
in masses of egg-pulp centrifuged at different speeds, and on 
the other the development of embryos from eggs centrifuged 
at corresponding speeds. Such an investigation, though it 
would not certainly lead to great results, yet seemed well 
worth undertaking. But before that could even be possible 
there was a preliminary question to answer. As a result of 
similar experiments on the ova of various Invertebrates it has 
been seriously suggested that the polarity of the egg, to 
which the structure of the embryo bears such a very detinite 
relation, is-not determined by the disposition of the various 
visible and separable constituents of the cytoplasm, which, 
indeed, so it is maintained, may be driven by the centrifuge 
to any region of the egg, leaving the original polarity intact, 
and without prejudice to the normality of development. 

It became necessary, therefore, first of all to inquire into the 
structure of the embryos produced from such ova, and the 
relation of that to the derangement of the egg materials. 
For only if it should turn out that the polar structure of the egg 


68 J. W. JENKINSON. 


to which the normal development of the embryo is related is 
determined, in part at least, by a certain arrangement of the 
visible materials which can be actually separated by the 
centrifuge, would any attempt to ascertain their chemical 
nature be of the slightest value, however successful. It was 
with these objects in view that the experiments which are now 
to be described were undertaken in March and April last. 

I shall deal first with the structure and development of the 
centrifuged eggs (which in any case have not yet been com- 
pletely elucidated), and afterwards give such observations as 
I have made on the chemical composition of the various con- 
stituents of the cytoplasm. 

A general discussion of these results and of their bearing 
on the large problem of the nature and origin of the polarity 
of the ovum and its relation to the development of the 
embryo will be found in a final chapter. 


T will only add here that I am under very great obligations 
to my friend Dr. Ramsden, not only for permitting me to 
work in the Laboratory of Physiological Chemistry, but also 
for giving me the most valuable advice and much personal 
assistance. 

I am also very much indebted to Professor Dreyer for 
allowing me the use of the centrifuge in his laboratory, and 
to Dr. Scott for the loan of a freezing microtome. 


Il. THE EXPERIMENTS. 


(A) THe Structure anD DEVELOPMENT OF THE CENTRIFUGED 
Kea or THE F Rog. 


(1) Methods.—I have used the eggs of the common 
English frog. The eggs were taken from the uterus, in- 
seminated, and allowed to remain in water for about an hour 
until the jelly had swollen and the perivitelline fluid been 
exuded. The eggs turned into the normal position with the 
axis vertical and the white pole below. They were then 


CENTRIFUGED EGG OF ‘THE FROG. 69 


placed on the centrifuge, which was at first set in gentle 
motion to turn the axes of the eggs into the direction of the 
force, and then more rapidly to bring about the desired effect. 
The eggs were thus centrifuged in the direction of their axes 
with the animal pole centripetal. 

Various speeds were used, but I am unable to state in any 
case the precise number of revolutions per minute. 

In series G the electrically driven machine was used, and at 
a speed of about 1500 revolutions. In the others a water- 
driven machine was employed, at speeds varying from about 
1100 to 3000 revolutions a minute. These speeds, which [ 
eall I, II, III and IV, I being the lowest, were determined by 
the angle through which the tap of the apparatus was turned, 
Owing, however, to the inconstancy of the water-pressure 
they probably varied in the course of the experiments. 

The radius in both machines was about three inches. 

Different exposures were used, from five to thirty minutes. 

After the treatment the ova were removed to vessels of 
clean water. 


(2) Details of the Experiments. 


G. 


Centrifuged 27 : iii: 13 on the electrically driven machine at bottom 
speed. 

1. One hour after insemination, centrifuged for 5 minutes. 

A grey patch appears round the animal pole. 

The first two segmentation furrows are normal. 

28 : iii: °13.—The eggs have segmented normally. 

The grey patch is no longer visible. 

8:iv:°13.—The tadpoles have hatched out, but some are abnormal, 
the yolk-sac being swollen, and are inert. 

All preserved in picric acid. 

Of these tadpoles 41 are apparently normal, i.e. like the controls | 
12 abnormal. Of the latter 10 are of type (a), 2 of type (pb). 

G.1. 8:iv:’13 (a) (Text-fig. 1, Pl. 11, fig. 19)—The operculum is 
growing back over the external gills, the tail and fin well-developed. 
The head is somewhat warty and wrinkled. 

Sections show that the anterior head ectoderm is vacuolated, as are 
also the olfactory pits, the brain, the Gasserian ganglia, the suckers 


70 J. W. JENKINSON. 


and the wandering mesoderm cells. The anterior end of the brain is 
single. 

The optic cup is some distance from the ectoderm ; its cavity is small 
and encloses a blood-vessel (vitreous body). 

There is no lens. The corneal mesoderm is present, but the con- 
junctiva is not cleared. The infundibulum is present and the pituitary 
body. ‘The auditory vesicles are well formed. 

The notochord begins behind the infundibulum: its structure isnormal. 


TExt-FIG. 1. 


G.1. 8:iv:’13 (a). The lens-less eye. There is a blood-vessel in 
the small cavity of the optic cup. Corneal mesoderm and con- 
junctival ectoderm, choroid and sclerotic mesoderm are shown. 


Differentiation of the skull and visceral arches has begun, the 
labial cartilages, Meckel’s cartilage, the trabecule with the anterior 
trabecular plate and cornua, the quadrate, the parachordals, the 
hyoid and branchial arches being all present. 

The thyroid is already separated, the trachea and lungs have been 
formed. There are external gills, and the internal gills are beginning 
to appear. 

The heart is normal. Aortz and cardinal veins present. 

The pronephros has the usual three funnels on each side and a 
glomus; the ducts open to the cloaca. 


CENTRIFUGED EGG OF THE FROG. yA 


The gut is normal. 

There is a blastema for the pelvis and hind limbs. 

Germ-cells are present at the root of the mesentery. 

The tadpole is clearly abnormal only in the vacuolation of the anterior 
ectoderm and its derivatives and in the absence of the lens. 

G.1. 8:iv:°13 (b) (Text-fig. 2, ad, Text-fig. 4, Pl. 11, figs. 20, 44a, 
45, 47).—The tail isas well-developed as in the last, but the anterior end 


TExT-FIG 2a. 


Section of the series through the tadpole, G. 1, 8: iv: °13 (b). 
Through the degenerate brain represented by a mass of pigment- 
cells (br.). The rudimentary cranium (er.), the hyoid (hy.) and 
parts of the branchial arches are alsoshown. Between the parts 
of the skeleton are bundles of myoblasts (my.). The section 
also shows gill-cleft, external gills, blood-vessels, a large lym- 
phatie ventrally, and dorsally thickened and pitted ectoderm. 


is seriously affected. The head appears to be abruptly truncated, the 
gills far forwards. The body is swollen. There are no suckers. 
Sections show the ectoderm of the head to be folded and much 
vacuolated. 
Olfactory pits, fore-brain, eyes, mid-brain, and part of the hind- 
brain are all absent, or represented only by masses of degenerating cells. 


72 J. W. JENKINSON. 


In the front part of the head is found an aggregation of pigmented and 
vacuolated cells, some still containing yolk-granules. These cells are 
the disintegrated residue of the anterior part of the nervous system, 
for some have the large pale nuclei characteristic of neuroblasts, others 
the smaller, darker nuclei of spongioblasts. Groups of cells with nerve- 
fibres proceeding from them are the ganglia of the fifth and seventh 


TExtT-FIG. 2 b. 


Section of the series through the tadpole, G. 1, 8: iv : 13 (b). 
Hind-brain, with folded roof, auditory vesicles and ganglia, 
cesophagus, trachea, heart, blood-vessels (aorte and cardinals). 
No notochord. 


nerves. One or two small vesicles of doubtful significance are also 
found. In addition to the intact cells are cell-fragments, and chromatic 
spherules, the remains of nuclear degeneration. 

In and around this accumulation are mesoderm cells, some vacuo- 
lated, others deeply pigmented (chromatophores). 

The auditory vesicles are large and apparently normal; at their level 
begins that part of the central nervous system which has escaped des- 


CENTRIFUGED EGG OF THE FROG. 73 


truction, namely the medulla, continued behind into the spinal cord. 
Even this, however, is not normal. In outline it is circular, its lumen 
crescentic. The thin roof is excessively folded, while in the thick floor 
there is a continuous mass of white matter across the middle line below ; 
next to this comes a layer of neuroblasts and next the lumen a layer of 


TExtT-FIG. 2c. 


Section of the series through the tadpole, G. 1, 8: iv: °15 (b). 
Pronephros (one funnel on right), stomach, liver, intestine, 
aorte, cardinals. Fusion of myotomes below medulla. No 
notochord. 


spongioblasts. The spinal cord presents the same histological char- 
acters, the lateral tracts of white matter having been apparently forced 
down into a median ventral position. 

The ganglia of the eighth and ninth and tenth nerves are present. 
The spinal ganglia are abnormal in position, being united ventrally 
below the cord, an effect, presumably, of the same cause to which the 
peculiar structure of the cord is due. 


74 J. W.: JENKINSON. 


The skull has also suffered, being represented only by a small bilateral 
plate ventral to the degenerate brain, possibly the anterior trabecular 
plate. Of the visceral skeleton there is a wedge-shaped piece in front 
of the gut, which may be interpreted as first branchial arch, with 


Text-FIG. 2 d. 


Section of the series through the tadpole, G. 1,8: iv : 13 (D). 
Pronephric ducts, primordial germ-cells, intestine. Spinal cord 
abnormal, with white matter in a continuous ventral band, 
and ganglia fused below. Myotomes fused. No notochord. 


possibly mandibular and hyoid elements included in it. and behind this, 
underneath the throat, a basi-branchial hearing two pairs of arches; 
the relation of these to the gill-slits proves them to be the second and 
third of the series. Attached to these skeletal elements are masses of 
myoblasts. 


CENTRIFUGED EGG OF THE FROG. 75 


There is no stomodeum, and the fore-gut is in communication with 
the exterior only by the gill-slits, of which the usual four are present, 
the last three being open. There isarudimentary sixth cleft (pharyngeal 
outgrowth). External gills are borne on the first three branchial 
arches. j 

The thyroid, still connected with the throat, lies in a depression of the 
basi-branchial. The trachea and lungs have been formed. The heart 
is normal, aorta and cardinal veins present. In front of the pericardium 


TEXT-FIG. 3a. 


i 


i 


ease 
Vy 
th 


‘ 

iB 

Wi 
<= 


tele) 
90 

~ 
(eXe) 
OO % 


Ks 

(ola 
ie 

O50 
2) 


SE Res 
7 Ah “ eS 
se) 
CESS RHA 


OX 


0 
(e) 


) 
fe) 
oO 
BS 
BS 
= 
% 


fe) 


Ly, 


oe. 
oS 
pegeton Si 


ce G, ereere 9 
sremaasceseuans ceo 


Optic vesicles and fore-brain. The pharynx below. i Dieisaviecks: (e): 
Normal. 


is a large sinus, lymphatic, partially divided by a septum. It causes 
the body-wall to protrude. The gut is normal. 

The pronephros is well-formed, but on one side there are only two 
nephrostomes. The ducts open to the cloaca. There is a glomus. 
The notochord is very imperfectly developed. The myotomes bend 
round and fuse in a median mass of cells (myoblasts) below the spinal 
cord; in this mass notochordal tissue may be here and there distin- 
guished, but it is feebly differentiated. Even when present it is not 
immediately below the spinal cord, but separated from the latter by 
myoblasts. Germ-cells are found at the root of the mesentery. 

2. One hour after insemination, centrifuged for 10 minutes. 

The grey patch round the animal pole is more conspicuous. 


76 J. W. JENKINSON. 


Segmentation, the first two furrows, may be normal, but is sometimes 
abnormal. 

28 : iii: °15.—Segmentation has been normally completed. The grey 
patch is still visible. 

8:iv:'13.—Many of the tadpoles are retarded in development and 
have failed to hatch. Those which have hatched are inert. 

All were preserved in picric acid. 

Of these tadpoles 28 are normal in appearance, 15 abnormal. 


TEXT-FIG. 3 D. 


O 
= > Na 


5} 


2 0 


2% 
x Seeeraberc cis 


I. 1,3:iv:13 (a). Optic vesicles and fore-brain. Lumen reduced. 
The pharynx below. 


Of the latter 6 of type (a) resemble externally G.1. 8: iv: 713 (a), 4 of 
type (b) resemble G.1. 8:iv:°13 (b), while types (c), (d), (e), (f) and (g) 
are represented by one each. Suckers and mouth are absent. 

G.2. 8:iv:’13 (a) (Pl. 12, fig. 46)—The head ectoderm is con- 
siderably vacuolated, and so are the mesodermal cells. 

There are no olfactory pits, but two masses of vacuolated cells are 
found, with nerve-fibres differentiated, each hollow, and connected with 
the ectoderm by cell-strands. 

The fore- and mid-brains and eyes are represented by paired and 


CENTRIFUGED EGG OF THE FROG. (Wh 


median masses of pigmented vacuolated cells. Some of the nuclei have 
the characters of those of neuroblasts, and nerve-fibres can be seen. 
The remains of the fifth and seventh ganglia and nerves are also distin- 
guishable. 

The brain begins as a solid mass with longitudinal and transverse 
fibres at this level, and immediately behind, opposite the auditory 
vesicles, a lumen appears. The hind brain and spinal cord have the 
structure seen already in G.1. 8:iv:713(b). The spinal ganglia are 


TEXT-FIG. 3c. 


A 
TEER 


L. 1, 6:iv:13 (a). Optic vesicles and fore-brain. Very much 
reduced. The pharynx below. The ectoderm (epidermis) is 
thickened and pitted. 


pressed down and may meet below. The ganglia of viii and ix and x 
are present. The auditory vesicles are large, the cells vacuolated. 

There is no trace of the trabeculz, but the visceral skeleton is repre- 
sented by a median plate below the throat bearing three pairs of 
branchial arches—better developed on one side than on the other—and 
produced in front into a ring-shaped cartilage which seems to represent 
the quadrate. To this skeleton are attached bundles of myoblasts, 
especially anteriorly. Of gill-clefts, the hyo-mandibular and first two 
branchials (the second open) are present on one side, the first three 
branchials on the other (the second and third open). There are external 
gills. 

The trachea and lungs have been formed, the heart is small but 


78 J. W. JENKINSON. 


twisted. Aortze and cardinal veins are present. There are three pro- 
nephric funnels on one side, but only two on the other. The ducts open 
into the cloaca. The glomus is absent. 

The gut is normal, but the intestine is not yet coiled. There are 
germ-cells at the root of the mesentery. The myotomes bend down and 
meet below the spinal cord. In this mass is the notochord, imperfectly 
differentiated in front, more posteriorly vacuolated. More posteriorly 
still the notochord is free and the myotomes separate. 

The whole body is very edematous; the connective tissue and the 


TEXT-FIG. 4. 


Reconstruction of the cranium (er.) and visceral skeleton (hy. br. 1, 
br. 2, br. 3) of G. 1, 8 : iv: 18 (b). 


posterior cardinal vein round the pronephros suffer especially from this 
accumulation of fluid. 

G.2. 8:iv:’138 (b) (Text-figs. 8, 11)—Anteriorly the ectoderm is 
highly vacuolated; so also are the mesodermal cells. 

Ventrally the ectoderm is also folded for some little way back. 

There are no suckers and no mouth. The gills are far forwards. 
Olfactory pits, eyes, fore- and mid-brains represented only by a heap of 
pigmented vacuolated cells, with traces of fibres. 

The auditory vesicles with ductus endolymphatici. At this level the 
brain begins. The hind-brain and the spinal cord have the same 
characters as in (a), but the posterior part of the spinal cord is normal. 
The auditory, vagus, and spinal ganglia are present, the spinal ganglia 
being united ventrally. 


CENTRIFUGED EGG OF THE FROG. 79 


The visceral skeleton—the only skeleton developed—consists of a 
median plate bearing three arches, the first, second and third, and an 
anterior prolongation, situated in the front wall of the pharynx, which 
represents hyoid and perhaps quadrate elements. 

The hyomandibular cleft is absent. The first, second, third, and 
fourth branchial clefts are present on both sides, the fourth being just 
open on one side, while on the other the second and fourth are open. 


TEXT-FIG. 5. 


SA Ke 
Mm ee 


Hes. : 2:iv: 13 (b). Not very abnormal front end (section 
oblique). Brain, neural crest, pharynx, one sucker cut. 


There are external gills. 

The trachea and lungs are formed, but the trachea is solid. 

The heart is also solid, small, and hardly twisted. Aorte and cardinal 
veins are formed. The pronephros has three funnels on each side, and 
a glomus. The ducts open to the cloaca. 

The intestine is not differentiated into regions, and in the yolk-cells 
is a central mass where there are no cell-divisions and the granules are 
fused into a coagulable liquid. 


80 J. W. JENKINSON. 


Germ-cells are found at the root of the mesentery. The myotomes 
are fused below the spinal cord. The notochord is only distinguishable 
behind the pronephros. Here it lies immediately ventral to the spinal 
cord. 


TEXT-FIG. 6. 


G.2. 8:iv:'13(g). Degenerate brain (pigmented cells). The 
auditory vesicles united (au.); below them the auditory and 
vagus ganglia (vii, w) also united. (Esophagus, lungs, liver, 
pronephros (one funnel cut on right). idematous connective 
tissue. 


In the ceelom and in the posterior cardinal vein round the pronephros 
is an accumulation of fluid. 


G2. 8:iv:’18(c) (Text-fig. 10, Pl. 11, fig. 21)—Body much 


CENTRIFUGED EGG OF THE FROG. 8] 


swollen; tail well developed. Neither mouth nor suckers. The head 
ectoderm is vacuolated and the mesoderm. The mesoderm s edematous. 
Olfactory sacs, eyes and front part of brain represented by an aggrega- 
tion of pigmented and vacuolated cells, with traces of the ganglia of 
v and vii, but no nerve-fibres. Hind brain, spinal cord and spinal 
ganglia as in (a); the spinal cord abnormal to the end. 

The auditory vesicles and the auditory and vagus ganglia are 
present. 

A small nodule of perichondrium in the front wall of the pharynx is 


TEXxtT-FIG. 7. 


Il. 4:iv:13(d). Auditory invaginations. The solid wedge 
of cells in the middle of the larger endoderm cells is the de- 
generating fore- and mid-brains. Ectoderm thick and pitted 
(cut somewhat tangentially on the ventral side). 


the only representative of the visceral skeleton, There is no trace of 
the cranium. 

The pharynx has an irregular cavity, but there are no gill-clefts. 

The trachea and lungs are present, but the latter are short. 

There is no heart. 

The pronephros has only two funnels on each side, and there is no 
glomus. The ducts are open to the cloaca. 

The stomach and duodenum are very small. 

Germ-cells are present in the mesentery. 


von. 60, part 1.—NEW SERIES. 6 


82 J. W. JENKINSON. 


The myotomes, which are fused below the spinal cord, are poorly 
developed in front. 

There is no notochord except in the tail. 

There is great edema of the connective-tissue spaces, of the posterior 
cardinal veins round the pronephros, of the celom, and of spaces 
(blood-vessels or lymphatics) around the stomach and duodenum. 

The ectoderm is distended and flattened. 


TExtT-FIG. 8. 


G.2. 8 iv:13(b). Hind-end, normal, except for central mass 
of undivided yolk in intestine and absence of lumen. 


1G.2. 8:iv:’13 (d) (Text-fig. 14 a, b, Pl. 11, fig. 22).—The body 
is very stunted; there is a short tail. The blastopore is widely open, 
and the yolk-plug protrudes. 

At the front end the ectoderm is folded, pitted and highly vacuolated. 

Internally the front end is occupied by a large cavity with a lining 
of flattened mesoderm cells. The cavity is probably persistent blasto- 
cel. The hinder wall of this space is occupied by a mass of undiffe- 
rentiated mesodermal and yolk-cells. Further back are traces of celom 
and blood-vessels. 

There is no sign of nervous system, notochord, or myotomes. 


CENTRIFUGED EGG OF THE FROG. 83 
G.2. 8: iv: 713 (e) (Text-figs. 12, 13, Pl. 11, fig. 23, Pl. 12, fig. 44 b). 
—Body short; short tail; persistent large yolk-plug. No mouth; no 


suckers. 
Ectoderm of head folded and wrinkled ; highly vacuolated. 


No olfactory pits nor eyes; no fore- nor mid-brain. These structures 


are represented by a mass of pigmented cells. 
Ganglia of v and vii not to be found. The auditory and vagus 


ganglion pairs each united across the middle line. 


TEXT-FIG. 9, 


eo 


ane 
T 


= PSU 
() X a by, 
ues 
pO Sine 
ai 


secu: 


I.1. 4:iv:’13(d). Thick and pitted ectoderm. Spinal cord 
solid. Notochord not properly differentiated. Pronephric 


ridge. 
Auditory vesicles well formed, each constricted into two cavities. 
The brain begins at the level of the auditory vesicles; it is solid, with 


fibres ventrally. 

The spinal cord has similar characters, but a lumen appears in it 
here and there. The spinal cord does not reach the hind end of the 
body. The spinal ganglia are united below. 

Fore-gut, but no gill-slits. No lungs. 


84 J. W. JENKINSON. 


Gut not much differentiated. In the yolk-cells a central mass without 
cell-boundaries ; here the yolk-granules are fused. 

Blood-corpuscles are being formed from the yolk-cells ventro-laterally. 

A pericardium is present, but the heart is represented only by a solid 
cell-mass projecting into this cavity from above. Vitelline veins and 
cardinal veins are present. 

The pronephros has three funnels on each side, and a small glomus. 
The ducts end blindly. 

The pronephric tubules are enormously swollen. 


TrExt-FIG. 10. 


G.2. 8:iv:’13(c). Pronephros, one funnel cut on right. 
Abnormal medulla. Fusion of myotomes; no notochord. Cso- 
phagus. Lungs. Liver. Enlargement of posterior cardinal 
vein and of lymphatics round liver. Cidematous connective 
tissue. 


The myotomes are united across the middle line by a mass of fusiform 
myoblasts, some of which are vacuolated. No notochord, except for an 
occasional vacuolation. 

The peritoneal cavity is well developed. 

G.2. 8:iv:’13(f)—Body very short and undifferentiated. At 
one end, presumably anterior, the ectoderm is highly vacuolated and 
wrinkled. At this end is what appears to be a yolk-plug, but is in 
reality a yolk-burst. 


CENTRIFUGED EGG OF THE FROG. 85 


There is a small archenteron posteriorly opening by the blastopore, 
a rudimentary pericardium, and the mesoderm is to some extent diffe- 
rentiated—as myoblasts, connective tissue and blood-corpuscles. 

No trace of nervous system, nor of sense-organs, nor of notochord, 
nor of pronephros. 

G.2. 8:iv:°13 (g) (Text-fig. 6)—Body short, neither mouth nor 


Text-FIeG. 11. 


cr 
ae ‘| ar SS 


L SerAatg 
7 


G.2. 8:iv:’13(b). Pronephros, one funnel cut on right, 
glomus. Cclom much enlarged. Myotomes fused below 
medulla. Ventral ectoderm folded. 


suckers. Tail with dorsal and ventral fins. There is a side branch to 
the tail but this does not receive any spinal cord. 

The anterior ectoderm is much vacuolated. Neither olfactory pits 
nor eyes. Fore- and mid-brains represented by a mass of vacuolated 
pigmented cells. Ganglia of v and vii present, with the nerves 
developing. 

Auditory and vagus ganglion pairs each united below the brain. 

The hind-brain, with fibres and neuroblasts stretching across the 


86 J. W. JENKINSON. 


thick floor, begins behind the auditory vesicle; the thin roof is much 
folded. The spinal cord, which has similar characters and a small 
lumen, is continued into the tail. The ganglia are united below the 
cord. 

The auditory vesicles are large and united, with their cavities also in 
communication, above the hind-brain. 


TrExt-Fic. 12. 


G.2. 8:iv:°13 (e). Spinal cord nearly solid. Pronephric 
tubules much swollen, especially on left. 


There are external gills, and gill-clefts, two on one side, probably the 
first and second branchial, the second being open, and three on the other 
sides, probably the first, second and third branchial, the last two being 
open. 

The visceral skeleton is represented by two pieces: one, anterior to 
the pharynx and even dorsal to it, probably represents the hyoid and 
first branchial arches. Muscles (myoblasts) are attached to it. The 
other is below the pharynx and bears a pair of arches, probably the 
second branchial. 


CENTRIFUGED EGG OF THE FROG. 87 


Trachea and lungs. 

Heart and pericardium, the heart straight. 

Pronephros with three funnels on each side, the ducts open. The 
glomus much enlarged. 

Stomach and intestine differentiated, liver present, proctodeum 
open. A central mass of fused granules in the yolk of the intestine. 

The myotomes are united below the medullary tube by elongated 
myoblasts, but there is no trace of a notochord. 

Germ-cells at the root of the mesentery. 

The connective tissue is edematous and the posterior cardinal veins 
much distended. 

3 One and a half hours after insemination, centrifuged for 174 
minutes. 

A whitish-yellow ring appears round the grey patch. 

Segmentation in early stages is normal, except for the inequality of 
the first or second division, or both. 

98 :iii:°13.—A grey and yellow patch is still visible. The animal 
hemisphere is segmented in some, but not in all. The vegetative 
hemisphere is segmented in none. 

8: iv:°13—All the eggs are dead, undeveloped. 

4. One anda half hours after insemination, centrifuged for 28 minutes. 

As the last; folds appear in the grey patch. 

Segmentation is more abnormal than in the last. 

28: iii :’13.—Like the last. 

8: iv :’13.—All the ova dead, undeveloped. 


H, 


Centrifuged, 28 : iii : 13, on the water-driven machine, about one hour 
after insemination. 

1. For 10 minutes at speed III. 

A grey and yellow patch, with folds, appears round the animal pole. 

29 : iii :°13.—The yellowish patch is still present. The animal hemi- 
sphere is segmented, but the vegetative is not; it is blotched with 
pigment. 

30 : iii :’13.—No blastopore has appeared. 

2. For 10 minutes at speed II. 

There is a grey patch surrounded by a yellowish ring. 

29 : iii: °13—The animal hemisphere, which still shows the grey patch. 
is segmented, but the vegetative is not; it is blotched with pigment. 

30: iii: °13.—There is no blastopore. 

3. For 10 minutes at speed I. 

A grey patch appears round the animal pole, but there is no yellowish 
margin. 

29 :iii:’13,—The grey patch is still present. 


88 J. W. JENKINSON. 


The animal hemisphere is well segmented. The vegetative hemisphere 
is also certainly segmented in some of the ova. It is blotchy. 

30 : iii: °138.—A semicircular blastopore is present. 

2: iv : 13.—Four of the embryos are dead in an early stage. Of the 
others, some are normal, some abnormal, and of two types, (a) and (b). All 
these embryos were preserved in picric acid. 

H. 3. 2:iv:’18.—The normal embryos. 


Trxt-Fie. 13. 


G.2. 8:iv:18(e). Heart small, fore-gut solid. Auditory 
vesicles well-formed. Ectoderm very thick and pitted. 


There is a short tail stump. Stomodzeumand proctodeum. Suckers. 
Sections show olfactory pits, optic vesicles, lens thickenings, auditory 
vesicles in an early stage of invagination, endothelial cells of heart, 
paired pericardium, pronephric ridge, sclerotome, notochord and sub- 
notochordal rod. The only abnormality is the mass of fused granules 
in the centre of the yolk-cells of the gut. 

H.3. 2:iv:’18 (a) (Pl. 11, fig. 24)—The front end is normal, 
with olfactory pits, optic vesicle, lens thickening, auditory vesicle, 
stomodeum, pituitary body, suckers, paired pericardium, endocardial 


CENTRIFUGED EGG OF THE FROG. 89 


cells, pronephric ridge, sclerotome, notochord and subnotochordal rod. 
In the centre of the yolk is a mass of fused granules. Posteriorly there 
is an exposed yolk-plug bounded above by the tail—containing spinal 
cord, notochord and mesoderm—and in front and on the left by a small 
knob which may be a second tail, inasmuch as it contains two rounded 
cell masses and some mesoderm. 

H.3. 2:iv:’13 (b) (Text-fig. 5, Pl. 11, fig. 25)—The anterior end is 
abnormal. The ectoderm here is highly vacuolated, the olfactory pits 


TextT-Fic. 14a. 


G. 2. 8:iv’13(d). Section of front end through the en- 
larged head-vesicle lined by a thin layer of mesoderm and con- 
taining some scattered mesoderm cells. This is the persistent 
segmentation cavity. 


very shallow, the fore- and mid-brain represented by a solid mass of 
vacuolated cells, and there is no sign of the infundibulum nor of the optic 
vesicles, The neural crest goes forwards into this region. The lumen of 
the medullary tube appears first in the hind-brain just in front of the 
auditory vesicles, and is continued into the spinal cord, which is normal. 

The auditory vesicles are just invaginated, ductus endolymphatici 
being present. 

There is no stomodzeum ; the suckers are very far forwards. 


90 J. W. JENKINSON. 


Fore-gut with incipient gill outgrowths, but no clefts. 

Heart-tube between the two sides of the paired pericardium. Slight 
peritoneal cavity, pronephric ridge, sclerotome, notochord and subnoto- 
chordal rod. 

What appears to be an enlarged liver outgrowth extends forwards 
below the heart. 

A mass of fused yolk-granules in gut. Hind-end normal with tail gut 
and neurenteric passage (a streak of pigmented cells). 


TExt-FIc. 14D. 


Posterior end of the same as Text-fig. 14a, through the yolk- 
plug, with deep blastoporic involution. Some myoblasts and 
blood-vessels have been differentiated. 


ib 


Centrifuged 31 : iii : 13 on the water-driven machine at speed IV. 

1. Fifty minutes after insemination, centrifuged for 10 minutes. 

A grey patch appears round the animal pole; it is surrounded by a 
yellowish-white ring, and in its centre is a yellowish spot. Segmentation 
is normal. 

2:iv:13.—The grey patch is still visible. There is a crescentic 
blastopore. 


CENTRIFUGED EGG OF THE FROG. 9] 


3:iv: 13.—The grey patch is still present. 

Medullary folds formed, blastopore widely open and yolk-plug pro- 
truding. Three embryos preserved, (a), (b) and (c). 

I.1. 3:iv:’13 (a) (Text-fig. 36, Pl. 11, fig. 26).—The anterior 
ectoderm thickened and vacuolated. The brain-cells are also vacuo- 
lated. The optic vesicles are abnormally thick-walled, and have an 
abnormally narrow lumen. 

Nerve crest, sense-plate and gill-plate present. The notochord is 
distinct. 

In the yolk isa central mass of fused granules. The yolk-plug is 
lateral (the blastopore has closed on the right, and the medullary tube 
has grown back). 

I.1. 8:iv:’18 (b) (Pl. 11, fig. 27)—Anterior ectoderm thick- 
ened and vacuolated. The medullary groove shallow, deeper in front, 
but the optic vesicles have not yet been evaginated. There is a neural 
crest. The notochord is hardly distinct from the mesoderm. 

There is a central mass of fused yolk-granules in the gut. 

I.1. 3:iv:718 (c) (Text-fig. 3 a).—The anterior ectoderm and brain 
vacuolated as in (a). The optic vesicles have a lumen of nearly normal 
size. There is a neural crest, and the notochord is beginning to be 
vacuolated. The vertebral plate is beginning to be separated from the 
lateral plate. 

There is a central mass of fused yolk-granules in the gut. 

4: iv :°13.—Some are normal or nearly so, but many have large, 
persistent yolk-plugs. 

All these embryos were preserved. 

Three are apparently normal, (a), ten abnormal. Of the latter there 
are seven of type (b), and one each of types (c), (d) and (e). 

I.1. 4:iv:’18 (a)—Normal except for the vacuolation of the 
head ectoderm and the fusion of yolk-granules in the centre of the yolk- 
cell mass. 

Olfactory pits, optic vesicles, slight lens thickening, auditory invagina- 
tions, stomodeum, pituitary body, notochord, pharynx, gill-outgrowths, 
pericardium and heart-tube, suckers, liver, pronephric ridge, slight 
peritoneal cavity, subnotochordal rod, proctodeum and neurenteric 
streak of pigmented cells. 

I.1. 4:iv:’18 (b) (Pl. U1, fig. 28)—Like (a), except for the 
yolk-plug. 

I.1. 4:iv:’18 (ce) (Pl. 11, fig. 29)—Head ectoderm vacuolated. 
The front part of the brain degenerate, but the optic vesicles can be 
distinguished as small, almost solid outgrowths. 

Shallow olfactory pit, auditory invaginations. Suckers absent. 

There are indications of a pericardium, but there is no heart. 

The notochord is distinct. There is a pronephric ridge. The hind- 


92 J. W. JENKINSON. 


end is fairly normal except for the large yolk-plug. The spinal cord 
ends in the middle line, it does not pass into the caudal swellings. 

There is fusion of yolk-granules in the centre of the yolk-cells. 

I.1. 4:iv:’18 (d) (Text-figs. 7, 9).—The head ectoderm is highly 
vacuolated, thickened and pitted. 

The fore- and mid-brains are represented only by a mass of pigmented 
cells which are mingled with an aggregation of yolk-containing cells, 
the front wall of the fore-gut. 

In this region are found the two auditory ingrowths. 

The medullary tube begins behind the latter asa solid cord, but 
further back a lumen appears, nearer the dorsal than the ventral side. 


TExt-FIG. 15. 


L.1. 6:iv:713(c). Spinal cord very small. Notochord not 
differentiated from the mesoderm. The mesoderm of the 
vertebral plates united across the middle line by cells which 
are beginning to elongate. 


The spinal cord is similar, but anteriorly the lumen is only present here 
and there; at the hind end it is better developed. 

Suckers, heart, and pericardium are absent. 

The neural crest is lateral, not ventral. 

The pronephric ridge is well developed. There is no notochord; the 
myotomes are united across the middle line by horizontally elongated 
myoblasts. ‘This median cell-mass is, like the myotomes, segmented. 

The gut hasa lumen. There is a mass of fused yolk-granules. The 
yolk-plug protrudes. 

Ti. 4:iv-*13\(e) (Text-fig. 16, Pl. 11, fig. 30).—The antenor 
ectoderm is very thick, vacuolated and pitted. 


CENTRIFUGED EGG OF THE FROG. 93 


No fore-brain nor mid-brain, nor even auditory vesicles. There are 
no suckers. The medullary tube is narrow, almost solid, and passes 
into one lip of the widely open blastopore, round one side and there 
ends. There is no notochord. Dorsal and ventral mesoderm are 
present, but the former is not differentiated into vertebral and lateral 
plates. 

Neither heart nor pericardium are found, nor a pronephric ridge. 

At the ventral lip of the blastopore is a short ectodermal involution, 
presumably the proctodeum. 

The gut has a narrow crescentic lumen. 

2. Seventy-five minutes after insemination, centrifuged for 30 minutes. 

There appears a yellow-grey patch round the animal pole; it is much 
folded. Round it is a double whitish ring, and round this again a grey 
ring. 

Segmentation is abnormal. The first and the second divisions are not 
meridional, and the grey patch may be cut off as a separate “ cell.” 

Some fail to segment. 

2: iv: °13.—Only one or two have segmented and developed further ; 
the blastopore, if present, is a wide semicircle. 

3: iv: ’13.—No further development. Cellular disintegration setting 
in. 

The zones are still distinguishable. 

I. Controls, 2 : iv : °13.—Semicircular blastopore. 


J. 


Centrifuged, 1: iv :°13, on the water-driven machine, at speed I, 
50 minutes after insemination. 

1. For 10 minutes (PI. 7, fig. 7.) 

2. For 20 minutes. 

The results are similar in the two cases. A faint grey patch appears 
round the animal pole with a margin of a rather lighter colour. 

Segmentation is perfectly normal. 

8:iv:°13.—The tadpoles are ready to hatch, and as normal as the 
controls. 


K. 


Centrifuged 2:iv:’13 on the water-driven machine at speed II, 70 
minutes after insemination. 

1. For 10 minutes (P1. 9, fig. 13). 

A grey patch appears round the animal pole with central, or sometimes 
excentric, yellowish-white spots, and surrounded by a lighter marginal 
ring. 

Segmentation is normal. 


94. J. W. JENKINSON. 


2: iv: °13.—The grey patch is still visible. The ova appear to be as 


well segmented as the controls. 
3: iv :°13.—The grey patch is faintly visible. The blastopore is three 


quarters of a circle. 
4: iv :°13.—The medullary folds are closing. Fairly normal. 


6 : iv : °13,—Fairly normal. 


Text-Fic. 16. 


‘ 
Haas 


(/) TA 
& 


\ 


I 


I.1. 4:iv 718 (e). Spinal cord reduced to a minimum. Gut 
lumen very narrow. Dorsal and ventral mesoderm differen- 
tiated, but no notochord. Ectoderm thickened and pitted. 
Central undivided yolk-mass. 


8: iv: °138.—Fairly normal, with stomodeum, suckers, nostril, gills, 
tail and fin. 

These tadpoles were not preserved. 

2. For 20 minutes. 

The grey patch becomes folded. The marginal ring is more marked, 
and may be confluent with the spots inside the grey patch. 

Segmentation is not normal, the first and second furrows not being 
meridional, and the rate of division is retarded. 


CENTRIFUGED EGG OF THE FROG. 95 


2 : iv : 13.—The animal hemisphere alone is segmented. It is streaked 
with white or grey. The vegetative hemisphere is blotched with 
pigment. 

3:iv:’13.—The blastopore has not yet appeared. 

4:iv:°13.—The embryos resemble those of I. 1. 3:iv:°13. The 
medullary folds are developed, the yolk-plug is exposed. The grey 
patch is still to be seen at the anterior end. 

6 : iv : °13.—Dead in an early stage. 

These embryos were not preserved. 


L. 


Centrifuged 1 : iv : 13 on the water-driven machine at speed III. 

1. Forty minutes after insemination, centrifuged for 5 minutes. 

The grey patch round the animal pole has a whitish spot in the 
centre and is surrounded by a faint marginal ring. 

Segmentation is normal in form but slightly retarded on the controls. 

2:iv:°13—The grey patch and whitish spot are still visible. 
Segmentation has proceeded not quite as far as in the controls. 

3: iv: 13.—The grey patch is still present. In some eggs the dorsal 
lip of the blastopore has appeared. 

4: iv : °13.—The blastopore is circular, the yolk-plug large and pro- 
truding. 

6: iv: °13.—There is a tail stump. The blastopore is open in many. 

Twelve of these embryos were preserved. 

Of these, 2 are normal, with nostrils, suckers, stomodzeum and in- 
cipient gill-clefts, 9 are monstrous, and 1 undeveloped. Of the 9 
monstrous embryos, | is of type (a), 1 of type (b), 1 of type (c), 2 of 
type (d), 1 of type (e), 1 of type (f), and 2 of type (g). 

L.1. 6:iv:’13 (a) (Pl. 11, fig. 31, Text-fig. 3c)—Anteriorly the 
ectoderm is very thick and vacuolated. A solid in-growth of this 
represents the fore-brain, produced into minute optic vesicles. 

There are traces of olfactory pits. 

The in-growth of ectoderm becomes grooved behind this point, and 
the groove deepens: this is the mid-brain. Then the groove closes, in 
the region of the hind-brain. 

The auditory vesicles are in an early stage of invagination. 

There are no suckers. 

The heart and pericardium are not formed yet. 

The notochord is normal, slightly vacuolated. 

The mesoderm is also normal, with somites, lateral plate and pro- 
nephric ridge. 

There is a small lumen to the gut, open at the blastopore. There is 
no proctodeum. 


96 J. W. JENKINSON. 


The medullary tube and notochord pass to one side of the protruding 
yolk-plug. Neural crests small. In the yolk-cells there is a central 
mass of fused granules. 

L.1. 6:iv: 13 (b)—The vacuolated anterior ctoderm is very 
thick and crinkled. 

Neither nostrils nor suckers are present. 

There is hardly any medullary tube in front of the auditory invagina- 
tions, and that is solid. 

The cells of the hind-brain are vacuolated, the spinal cord is normal. 
The neural crests are small. 

There are no gill-clefts. 

There is a pericardium but no heart. 


TExT-FIG. 17. 


L.1. 6:iv:715(g) (1). No nervous system. No differentia- 
tion of the notochord from the dorsal mesoderm. 


Somites, lateral plate and pronephric ridge are present. 

The notochord goes as far forwards as the auditory vesicle. 

The proctodzum is open. 

There is a mass of fused yolk-granules in the yolk-cells; the large 
yolk-plug is one-sided. 

L.1. 6:iv:’13 (ce) (Pl. 11, fig. 32, Text-fig. 15)—Anterior ecto- 
derm thick and vacuolated. 

No olfactory pits, no auditory vesicles, no suckers. 

The medullary tube very asymmetrical in front; this is presumably 
the hind-brain, but in the absence of the auditory vesicles it is im- 
possible to say with certainty. 

The lumen of the medullary tube is very small and absent in places. 
Behind is an open medullary groove. 

The notochord is not distinguishable in front, barely so behind. 

There is a pericardium, but no heart. 


CENTRIFUGED EGG OF THE FROG. 97 


The pronephric ridge is indicated. 

The fore-gut is much folded and crumpled. 

The gut lumen disappears behind. 

There is the usual mass of fused granules in the middle of the yolk- 
cells, and a large yolk-plug. 

L. 1. 6:iv:’13(d) (1) (Pl. 11, fig. 33)—Medullary tube and 
notochord both absent. On one side of the large yolk-plug is a slight 
protrusion: this is probably dorsal and represents the tail. The meso- 
derm in this tail is continued forwards into the body; ventral meso- 
derm is being differentiated. 

There is no celom. 

There are no suckers. 

The anterior ectoderm is thickened, pitted and vacuolated, and there 
is a central fusion of yolk-granules. The gut is limited to the small 
archenteric cavity round the yolk-plug. 

L.1. 6:iv:’13(d) (2).—This is similar to the last. In the dorsal 
mesoderm there are indications of the differentiation of a median tract, 
the notochord. 

L.1. 6:iv:’13 (e) (Pl. 11, fig. 34)—Like (d), except that the 
blastopore is closed, and the tail protuberance bilobed. 

Of the two pits seen at the base of this tail one is quite superficial, a 
mere fold of ectoderm, while the other is a deep involution which 
passes in some way to end blindly. It is the proctodeum. 

The vacuolated ectoderm is at the other (anterior) end. 

L.1. 6:iv:’13 (f) (Pl. 11, fig. 35)—This is similar to (e) except 
that the rudiment of the tail is not bilobed. Of the two depressions at 
its base, one is a mere superficial folding, the other a deep passage 
leading into a cavity lying behind the yolk-mass, a rudimentary archen- 
teron. 

L.1. 6:iv:’18 (g) (1) Text-fig. 17)—This resembles (d) (2). 
There is an indication of the differentiation of the notochord—as a 
median band of smaller cells—from the dorsal mesoderm. There is a 
rudimentary archenteron opening under the dorsal lip of the blasto- 
pore. 

L.1. 6:iv:’13 (g) (2).—Similar to the last and to (d) (1). There 
is no trace of a separation of the notochord. 

8: iv: 13.—The remainder were preserved. 

One is undeveloped, 2 are normal in form with the tail longer than 
before, and 9 are monstrous. Among these tadpoles types (a), (b) and 
(e), (f) and (g) are represented each by one, types (c), (d), each by two 
specimens. 

L.1. 8:iv:’13 (a) (PI. 11, fig. 36)—The anterior ectoderm is very 
thick and pitted, and highly vacuolated. 

There is no nervous system nor sense-organs. From the dorsal lip of 

vot. 60, parr 1.—NEW SERIES. | 


98 J. W. JENKINSON. 


the blastopore a thick band of mesoderm stretches forwards ; in front 
this becomes thinner. There is no trace of a notochord in it. 
There is ventral mesoderm and the beginning of a pericardium. 


TExT-FIG. 18; 


Sez “i 


Wy] 
Wh 
ae, 


Oo 
rr Al 
sae 


L.1. 8:iv:°13(g). Both tails are cut. Signs of segmentation 
of the mesoderm in them. In the body mesoderm has been 
differentiated from the yolk-cells, in which is a central undi- 
vided mass of fused yolk-granules. 


There is no sucker. 


The blastopore is widely open; on one side is a deep involution, but 
except for this no gut. The yolk-plug is large. In the centre of the 
yolk there is a mass of fused granules. 


L.1. 8:iv:’18(b) (Pl. 11, fig. 37).—The head ectoderm is 


CENTRIFUGED EGG OF THE FROG. 99 


vacuolated and thick. The fore-brain, with the optic vesicles and 
infundibulum, is present. The mid-brain is also present, and the hind- 
brain and spinal cord. All parts of the brain show cell-vacuolation. 

Auditory vesicles, and ganglia, vagus ganglia, and ganglia of v and 
vii all formed, and neural crest. The suckers are present. 

The notochord, normally vacuolated, extends to in front of the auditory 
vesicles. Both it and the spinal cord pass into the tail. 

There is no stomodeum. The gill-slit outgrowths are solid. 

The pericardium is large, but the heart rudiment is solid. The peri- 
toneal cavity is still small. The pronephric tubules are developing, but 
are still without lumina and funnels. There is a liver diverticulum, but 
the lumen of the gut is obliterated. 

There is fusion of yolk-granules in the yolk-mass, and a large yolk- 
plug. 

L1. 8:iv:’13(c) (1).—This resembles (a), but there is a deep 
involution underneath each lateral blastoporic lip. Otherwise there is 
no gut. 

There are signs of the differentiation of a notochord, 

There is no pericardium in the ventral mesoderm. The nervous system 
is absent, the longitudinal pleats looking like medullary folds being 
merely ectodermal ridges. 

L.1. 8:iv:’13 (ce) (2) (Pl. 11, fig. 38)—Like the last, but a 
rudimentary nervous system is present. A solid median mass of ecto- 
dermal cells with paired appendages represents the brain and ganglia. 
Imbedded in this are two small auditory vesicles. Behind this point a 
solid spinal cord runs back and enters one of the caudal swellings. 

The dorsal mesoderm is a thick plate ; it shows no signs of a notochord. 
The ventral mesoderm is present, but neither pericardium nor heart. 
The pronephric ridge is indicated on one side. The gut has no lumen. 
In the yolk-cells a central fusion of yolk-granules. 

L.1. 8:iv:’18 (d) (1).—The anterior end can only be determined 
by the vacuolation of the ectoderm. 

The segmentation cavity persists in front and dorsally. Below it is 
the mass of yolk-cells; its roof isformed of a band of mesoderm cells 
which passes backinto the dorsal lip. This band stops before the front end 
is reached so that there the segmentation cavity is roofed only by 
ectoderm. 

Ventral mesoderm is being differentiated. There is an open blasto- 
pore and persistent yolk-plug; the archenteric cavity is very small. 
There is neither notochord nor nervous system. 

L.1. 8:iv:’13(d) (2).—The segmentation cavity is obliterated 
and there is an archenteron, narrow, but longer than inthe last. Other- 
wise the same. 

L.1. 8:iv:’13(e) (Pl. 11, fig. 39).—Body rounded, bearing a 


100 J. W. JENKINSON. 


short tail. No nervous system nor sense-organs. Dorsal mesoderm, 
but no notochord. Dorso-laterally the mesoderm is thickened and these 
two thickenings pass back into the tail. In each is a cavity (ccelom). 

The ventral mesoderm is developed, but slight. There is an archen- 
teron on the dorsal side with a postero-ventral extension. The latter is 
in communication with the exterior by an ectodermal involution (proc- 
todzeum). 

The anterior ectoderm is vacuolated. 

L.1. 8:iv:’13(f) (Pl. 11, fig. 40)—The body consists of a solid 
mass of yolk-cells (in the centre of which is a region in which the 
granules are fused together), enclosed by a vacuolated ectoderm. On one 
side of the large yolk-plug is a tail. In this tail are paired segmented 
somites, with indications of myocel. There is, however, no notochord. 
On the opposite side of the blastopore is a much shorter tail-like structure. 
The nervous system is completely absent. Dorsally and ventrally there 
are thin layers of mesoderm. 

L.1. 8:iv:’18(g) (Text-fig. 18, Pl. 11, fig. 41)—Body stumpy, 
with two tail-like structures, one on each side of the large blastopore. 

Each of these tails contains mesoderm, which shows signs of being 
divided into pairs of somites. Otherwise this embryo resembles the 
last. 

The pit at the anterior end is a mere infolding of the ectoderm. 

2. Forty minutes after insemination, centrifuged for 10 minutes. 

The central yellowish spot is surrounded by folds of the grey patch, 
the marginal white ring broader. 

The first two furrows are normal in form, but retarded, not only on 
the controls, but slightly on L. 1. 

2: iv: °13—The grey patch and yellowish spot are still visible. The 
animal hemisphere is as well segmented as in L. 1; the vegetative 
hemisphere is, however, divided into only a few large cells. 

' 3:iv:’13—The grey patch is still present. The dorsal lip of the 
blastopore has not yet appeared. 

4: iv : °13.—There are indications of a blastoporic groove at the edge 
of the pigmented area. 

6: iv: °13.—The yolk is still exposed. 

3. Seventy-five minutes after insemination, centrifuged for 16 
minutes. 

The folding of the grey patch round the yellowish spot is more 
marked ; the spot appears to sink in. The marginal ring is more 
conspicuous and flecked with dark spots. 

The first and second segmentation furrows may be normal in form, 
but there are many irregularities; the furrows of the third phase are 
also irregular, for instance, parallel to the first. These meridional 
furrows fail to reach the vegetative pole. 


CENTRIFUGED EGG OF THE FROG. 101 


2: iv : °13.—The grey patch and yellowish-white ring are still visible, 
but the foldings seem to have been smoothed out. 

The animal hemisphere imperfectly segmented, the vegetative not at 
all. 

3: iv: ’13—No change. 

4: iv : °13.—There is still no sign of a blastopore. 

4. Seventy-five minutes after insemination, centrifuged for 25 
minutes. 

The marginal zone becomes broader still. 

Segmentation as in 3, or even more irregular. The yellowish spot 
may become cut off as a separate “cell,” the folds being incorporated 
in the divisions or else entirely smoothed out. 

2: iv : ’13.—The white ring is still visible. 

Segmented in the animal hemisphere. 

3:iv:’13.—As 3. 

4: iv : °13.—There is no sign of a blastopore. 

6:iv:°13.—The white ring is now very faint. No other change 
(Pl. 9, fig. 14). 


M. 

Centrifuged 2 : iv : 13 one hour after insemination at speed II on the 
water-driven machine. 

1. For 10 minutes. 

A circular grey patch appears round the animal pole, containing 
sometimes a yellowish spot; it is surrounded by a paler border (PI. 7, 
fig. 8a). 

The first and second furrows are meridional as normally, but some- 
times either of these divisions is unequal. They appear later than in 
the controls. 

The furrows of the third phase are abnormal in being meridional. 
All these furrows reach the vegetative pole. 

3: iv: °13.—Segmentation is completed ; the cells of the vegetative 
hemisphere are rather large. 

The grey patch has disappeared. 

4:iv:°13.—There is a blastopore, three quarters of a circle or 
circular ; it is rather large. 

6:iv:°13.—The medullary folds are closing (when they have been 
formed). The blastopore is still open in some cases. 

Six embryos were preserved at this stage, one each of types (a) and 
(b), two each of types (c) and (d). 

M.1. 6: iv: ’13(a).—Rounded body, yolk-plug still exposed. A 
slight back-growth of the body on the dorsal side of the blastopore is 
the beginning of a tail. The archenteron is short and does not extend 
in front of the middle of the body. 


102 J. W. JENKINSON. 


Fusion of yolk-granules occurs. 

There is no medullary plate. The dorsal mesoderm is hardly differ- 
entiated from the roof of the archenteron. There is no distinct 
notochord ; ventral mesoderm is present. 

Anteriorly the ectoderm is thickened but not so remarkably as in 
later stages. It is slightly pitted. 

M.1. 6:iv:’13 (b).—The anterior ectoderm is vacuolated. There 
are no suckers. 

Neither olfactory pits nor eyes are present. The auditory vesicles 
are curved plates, slightly detached from the ectoderm. 

A mass of pigmented cells in front probably represents the fore- and 
mid-brains. Similar lateral masses would be the ganglia of this region, 
possibly also the optic vesicles. The median mass is continued into a 
fold, and this passes behind into a solid cord which presently enlarges. 
The lumen of the hind-brain appears in this. The lumen is dorsally 
situated. The hind-brain is continued in turn into the spinal cord, in 
which the lumen is also excentrically dorsal ; posteriorly, however, itis 
in the normal position. 

The neural crests come fairly far down but do not meet below the 
cord. 

The vertebral plates are united across the middle line by horizontally 
elongated cells. There is no notochord. 

A pronephric ridge is present and there are indications of a 
splanchnoceel in the lateral plate. 

Ventral mesoderm is present but there is as yet no pericardium. 

The enteron is narrow but extends far forwards. 

There is a central fusion of yolk-granules. 

There is a proctodeum, and a streak of pigment-cells indicates the 
neurenteric passage. 

The yolk-plug has been withdrawn. 

M.1. 6:iv:’13 (c).—As (a), but there is no back-growth above the 
yolk-plug, and the archenteron is shorter. 

M.1. 6:iv:’13 (d).—As (b), but the medullary tube is smaller and 
the lumen minute. 

There is a yolk-plug. 

8:iv:’13.—A tail is developed in most. The head is deficient. 

Seven embryos were preserved, two each of types (a) and (c) and one 
each of types (b), (d) and (e). 

M.1. 8:iv:’13 (a).—Anteriorly the ectoderm is vacuolated. Here 
there is a large cavity bounded by a layer of flattened cells. The cavity 
is the persistent segmentation cavity, the lining cells mesodermal 
elements which have been differentiated from the yolk-cells and passed 


into the cavity, while the yolk-cells have remained in their original 
position. 


CENTRIFUGED EGG OF THE FROG. 103 


Posteriorly the mesoderm cells pass into the main mass of yolk- 
cells in which is the enteron. 

There is a central mass of fused yolk-granules. 

Dorsal mesoderm is present and differentiated into vertebral and 
lateral plates. A median notochordal tract is distinguishable from the 
vertebral plate in places. 

Pronephric tubules are in course of formation. 

Ventral mesoderm not well developed. 

There is no nervous system. What appears from the outside to be 
such is a median fold of ectoderm enclosing some mesodermal elements. 
Proctodeum large but not open into enteron. 

M.1. 8:iv:’13 (b) (Pl. U1, fig. 42)—Anterior ectoderm vacuo- 
lated, enclosing a persistent segmentation cavity. In front this is 
bounded by the ectoderm alone, but further back a thin layer of cells 
appears lying inside the ectoderm. This layer is continued into the 
mass of yolk-cells behind. In the latter a central mass of fused 
granules. 

The dorsal mesoderm passes back into the tail, and is there segmented. 
There is no notochord. Ventrally the mesoderm is poorly developed. 

The cavity of the enteron is small and short. The proctodzeum opens 
into it. 

There is no nervous system. 

M.1. 8:iv:’13 (c) (Pl. 11, fig. 43)—As in (a) and (c) there is 
anteriorly a persistent segmentation cavity. It is not, however, lined in 
whole orin part by a sheet of mesoderm, but merely includes a certain 
number of mesodermal cells. 

The yolk plug is exposed, and there is a short archenteron opening by 
the blastopore. 

The tail springs from a thick mass of mesoderm lying beneath the 
lip of the blastopore on one side; there is a similar mass of mesoderm 
on the opposite side, and the two become confluent in front. 

Correct orientation of this embryo is very difficult as there is neither 
nervous system nor notochord, but assuming that the mesodermal 
thickenings are lateral the tail is either right or left. The side on 
which the two masses meet may then be dorsal. 

M.1. 8: iv: ’13 (d).—The anterior vacuolated ectoderm encloses a 
solid mass of yolk-cells, with peripheral mesoderm. The segmentation 
cavity does not persist. 

The mesoderm becomes concentrated posteriorly into three masses 
which project wedge-like into the yolk-cells. At the hinder end these 
three pass into the blastoporic lip, where they become continuous with 
small cells on the surface of the yolk-plug. Enteron, nervous system 
and notochord are all absent. 


M.1. 8:iv:’13(e).—The anterior vacuolated ectoderm encloses a 


104 J. W. JENKINSON. 


solid mass of yolk-cells. Peripherally mesoderm is differentiated. 
Posteriorly there is a short tail springing from what appears to be the 
dorsal side, since the archenteron is on this and the proctodeum on 
the opposite side. 

The proctodeum ends blindly. 

The tail contains mesoderm which springs from a dorsal median, 
partly also from a lateral concentration of mesoderm. 

Nervous system and notochord both absent. 

2. For 20 minutes. 

The grey patchislarger. A grey ring appears outside the white ring. 

Segmentation as in 1 (PI. 7, fig. 8, b). 

3: iv: 713.—The grey patch is still to be seen. 

The vegetative hemisphere is imperfectly segmented. 

4: iv :’13.—No blastopore has appeared. 

3. For 30 minutes. 

There is a yellowish central spot inside the grey patch. 

Segmentation more irregular. The second furrow, for instance, may 
be parallel to the first, and the divisions do not reach the vegetative pole. 

3: iv : °13.—The grey patch is still visible. The vegetative hemisphere 
is not segmented. 

4: iv: °13.—No blastopore has yet been developed. 


INE 

Centrifuged 2:iv:°13 on the water-driven machine one and a half 
hours after insemination at speed IV. 

1. For 10 minutes. 

The circular grey patch contains a yellowish spot surrounded by folds. 
Round it is a yellowish-white margin. 

By the time segmentation has begun the folds have disappeared and 
the central spot is not very distinct. 

The first furrow may be normal butis notalwaysso. The second also 
may be meridional, but is not so always. Later furrows are irregular, 
sometimes meridional, sometimes circular—that is, circumscribing a 
small area in the animal hemisphere. The meridional furrows do not 
reach the vegetative pole. 

3: iv:713.—The animal hemisphere alone is segmented, and that 
incompletely. 

4: iv :’13.—No development has occurred. 

2. For 20 minutes. 

There is a grey ring outside the white margin. 

The first two furrows may be meridional but are not necessarily so. 
They do not reach the vegetative pole. Later furrows are very irregular. 

3: iv: 13.—As 1. 


CENTRIFUGED EGG. OF THE FROG. 105 


4: iv : 13.—No development. 

3. For 50 minutes. 

The grey patch is surrounded by a groove. Outside this is first a 
white, then a grey ring. Segmentation is very seldom regular, even in 
the earliest stages. Divisions are often parallel to a meridian, and 
circular furrows are common. The vegetative hemisphere is not seg- 
mented. 

ofiv: 13. j eo 

ae ae 
O. 


Centrifuged 8 : iv : 13 on the water-driven machine at speed IV. 

All these eggs were preserved in formol immediately or in early 
cleavage stages. 

1. One hour after insemination, centrifuged for 5 minutes (Pls. 7, 
8, 10, figs. 1, 9, 16). 

A circular grey patch appears round the animal pole; in it are radial 
dark strive converging towards a central yellowish spot. The patch is 
separated by a groove from the pigmented area. 

The patch is sometimes grooved or folded. The first and second 
furrows may be meridional, but are often parallel to a meridian. Later 
furrows are very irregular. Some, but not all of the meridional furrows 
reach the vegetative pole. 

2. One hour after insemination, centrifuged for 10 minutes (PI. 7, 
fig. 2). 

The central spot is slightly depressed and surrounded by folds of the 
grey patch. The patch is marked by radial striz, and surrounded by a 
groove. It has a narrow, faint whitish border. 

Segmentation as in 1. 

3. One hour after insemination, centrifuged for 15 minutes (Pls. 7, 
8, 10, figs. 3, 10, 17). 

The central spot is sinking in and being covered over by the folds. 
The radial striz are still present. The white borderis broader. Outside 
the groove is a grey ring. 

Segmentation is more irregular than in 2. The furrows hardly 
_ reach the vegetative pole. 

4. One and a quarter hours after insemination, centrifuged for 
20 minutes (Pls. 7 and 8, figs. 4, 11). 

The central spot is no longer visible, the folds having grown over it. 
The grey patch is still radially striated. External to the marginal 
groove is a second white ring, just above the grey ring and derived 
from it. 

The first furrow may reach the vegetative pole, the others do not. 
Circular furrows are seen. 


106 J. W. JENKINSON. 


5. One and a quarter hours after insemination, centrifuged for 
25 minutes. 

The folds have nearly met and fused. Internal to the white border 
of the grey patch is a yellowish ring. 

The furrows, when meridional, do not pass beyond the pigmented 
area (Pl. 7, fig. 5). 

6. One and a half hours after insemination, centrifuged for 30 
minutes (Pls. 7, 8, and 10, figs. 6, 12, 18). 

The grey patch is now homogeneous, the radial strize having dis- 
appeared. It is smaller, and immediately surrounded by a groove. 

Outside this groove is a wide white ring which appears to be due to 
the confluence of the two white rings of the earlier stage. 

Then comes the grey ring subdivided into three zones, of which the 
middle is darker. 


Segmentation very irregular. The first furrow may begin at the side 
instead of at the animal pole. Circular furrows are found. 


(3) The Effect produced by the Centrifuge on the 
Structure, Segmentation, and Development of 
the Egg. 


From the details of the experiments that have been given 
in the preceding section it will be clear that the effect 
produced by the centrifuge upon the structure of the ovum, 
and upon its segmentation and development, varies with the 
degree of acceleration and the length of exposure. 

A strict comparison is, of course, only possible between 
eggs of the same batch, centrifuged on the same machine, at 
the same time, at different speeds and exposures. Obviously, 
however, these conditions cannot always be observed, and 
then, owing, as already explained, to diurnal, if not hourly 
inconstancies in the water pressure, as well as to variations 
in the eggs themselves, it may well happen that on some 
occasion an exposure to a higher will not effect a greater 
change, will effect, perhaps, even a less change than an’ 
exposure of the same length to a lower acceleration. 

Fortunately, in my series of experiments, this has not 
occurred, except, perhaps, once. 

The accompanying table gives a résumé of the experi- 
ments performed with the water-driven centrifuge : 


CENTRIFUGED EGG OF THE FROG, 107 


Exposure in minutes. 
15 2¢ 


Speed. 5 10. 5. 20. 25. 30. 
if Beef PEL ON a _— — 
— . Jil — J.2 -- — 

II — . He2 a a a a 
= ed 0) re “2 — — 

ee, Mid = Me 2 — M.3 

BET Seen i de, — — — a 
1 iva ie L. 3 L. 4 — — 
Mesos ple bs —_—_ . — — i 2 
— . N.1 Sp Nes 3 ae NGS 

Ope. 10.2 On3si 22  OSAR En FOs oy ONG 


On referring back to the preceding section it will be found 
that in nearly all cases the same acceleration and exposure 
produce the same, or nearly the same, effect upon the egg. 

Thus in H. 3 and J. 1 there is a grey patch only, without a 
markedly white border. In H. 2, K. 1 and M. 1 the patch is 
surrounded by a lighter ring. In K. 2 and M. 2 the patch is 
folded round a central lighter spot, but in M. 2 there is an 
additional grey ring, which in K. 2 isabsent. In H. 1 and 
L. 2 the patch is deeply infolded. In I. 1, N.1,and0O. 2 there 
is a central yellowish spot in the grey patch, and the latter 
has a lighter border. In N. 2 and O.4 the grey patch is 
deeply infolded, the white border marked, while in I. 4, N. 3, 
and O. 6 the grey patch is surrounded by a broad white 
border, and this by a composite grey ring. When the same 
speed is employed the effect is always increased by pro- 
longing the exposure. Conversely, when the exposure is 
constant the effect ought to vary with the acceleration, but 
this has not proved to be the case invariably. ‘hus the 
change in J. 2 is less than in K. 2, and less in K. 2 than in L. 4, 
and similarly in the series J.1, K.1, L.2, and, again, less in 
M. 3 than in O. 6, but H.1 and L. 2 are very like O. 4, L. 4 like 
O. 6, so that here the case is realised of a lower acceleration 
having caused a greater alteration than a higher in two 
different lots of eggs, the exposure being the same. 

The same discrepancy appears in the subsequent develop- 


108 J. W. JENKINSON. 


ment, the embryos of I. 1 being as well if not better developed 
than those of L. 1 and M. 1 (which are muchalike). Develop- 
ment only occurs in the series J.1 and J.2, H.3, I.1, K.1, 
L. 1, M.1, and G.1 and G.2. The ova of the series H. 2 did 
not develop though it would have been expected that they 
should. O.1,N.1 and O. 2 would very possibly have developed 
if they had been kept. 

Of these J. 1 and J.2 develop best, then H.3, then G. 1, 
then G. 2, then I. 1, L. 1 and M. 1. The position of the 
embryos of K. 1 in the list is uncertain since they were 
unfortunately not preserved. There is, however, sufficient 
evidence that in the main the capacity to develop is progres- 
sively decreased as the acceleration is raised and the exposure 
prolonged. The experiments of the G series can only be 
assigned a position amongst the others by the extent to which 
the egg-structure and development are altered. In G. 1 and 
2 there is a simple grey patch, and development is much 
better than in any except J. 1 and J. 2. They probably lie 
very near H.3. G.3 and G. 4 have the ring round the grey 
patch and fail to develop. 

We may now proceed to discuss in order the effects which 
the operation produces upon the structure of the egg, its 
segmentation and its development. 


a. The Effect Produced upon the Egg-structure. 


The Structure of the Normal Egg.—As is well 
known, the spherical egg of the frog has a radially sym- 
metrical structure about an axis, which axis has unlike poles. 
The polarity is determined first by the distribution of the 
superficial pigment, which only occupies about two thirds 
of the egg-surface, and second, by the disposition of the plasma 
(protoplasm) and yolk, the plasma being mainly, though not 
exclusively, situated in the pigmented region, the yolk-granules 
larger and more abundant in the unpigmented region, though 
found, of course, in the plasmatic portion as well. The line 
drawn through the centre (at the surface) of the pigmented 


CENTRIFUGED EGG OF THE FROG. 109 


and plasmatic region, the centre of the egg itself, and the centre 
(at the surface) of the unpigmented yolky region is the axis, 
and its unlike plasmatic and yolky poles respectively the animal 
and vegetative. Internally there is also diffuse pigment, 
aggregated a little more intensely in the axis and in the animal 
hemisphere. The nucleus of the full-grown but immature 
oocyte hes, of course, axially, but excentrically, in the animal 
hemisphere, and it is in the same place that the pronuclei 
meet in fertilisation. All this is, of course, a common place 
of embryology, but the distribution of certain other substances 
in the ovum has not, as far as I am aware, ever been described. 
I allude to the fat (including lecithin, which, as I shall have 
occasion to show later, is present) and the glycogen. The fat 
appears to be uniformly distributed in the shape of small 
globules, variable in size, lying in the plasma between the 
yolk. These globules are easily demonstrated in sections of 
eggs, preserved in formalin and frozen, by means of the fat stain 
Sudan III. The yolk-granules remain uncoloured by the dye. 

In respect of the glycogen the egg is, however, polarised, 
for this substance, present in the form of small spherules, is 
more abundant in the plasma of the animal than in that of the 
vegetative region, the spherules being larger and more 
numerous in the former, while in the latter they are excessively 
minute and very scanty (Pl. 10, fig. 15). 

The glycogen, therefore, like the plasma in which it is 
embedded, gradually decreases in concentration from the 
animal to the vegetative pole, while the concentration of the 
yolk increases in the opposite direction! 

This polar structure is seriously affected when the egg is 
centrifuged, in my experiments initsaxis. Roughly speaking 
what happens is that the lighter fat comes to the centripetal, 
that is, the animal pole; the next lightest, the plasma with the 


1 The distribution of glycogen in the course of development has not 
yet been worked out. I have only observed that in tadpoles in which 
the operculum is closing this substance is found only in the myotomes, 
in the tubules and duct of the pronephros and in the roof of the 
medulla. 


110 J. W. JENKINSON. 


glycogen, forms a layer next to the fat ; while the heavy yolk 
and pigment are driven to the centrifugal, the vegetative pole. 
Hence the various zones or strata into which these eggs 
become divided. The somewhat complex details of the 
changes are most readily made out in a series of eggs centri- 
fuged at the same acceleration with successively longer 
exposures, as, for example, the seriesO. The acceleration was 
here considerable: the exposures were 5, 10, 15, 20, 25 and 
30 minutes. The ova were preserved in formol; from this 
material good sections are obtainable, either by the paraffin 
method, or after freezing. The latter are necessary for the 
study of the fat. 

In O. 1 (Pl. 7, fig. 1, a, 6), there is developed round the 
animal pole a circular grey patch radially striated with dark 
lines converging towards a central yellowish spot. The patch 
is separated by a groove from the surrounding pigment. The 
radial striz are the lines along which the pigment is streaming 
away from the centripetal pole, while the lighter fat and 
plasma is moving in the opposite direction. 

The yellowish spot may be excentric, and the grey patch 
may be folded. 

A meridional section shows, beginning at the animal pole 
(Pls. 8 and 10, figs. 9, 16), (1) a superficial layer of rather 
finely vacuolated plasma, staining violet with hematoxylin. 
In it there is an occasional yolk-granule. Some of the 
original pigment remains in this layer, being disposed in (a) 
a denser sheet at the surface, (3) more diffusely below. 

This layer is the grey patch seen from the surface. 

(2) A more coarsely vacuolated layer of the same violet 
staining plasmatic substance. In it is some pigment and a 
few yolk-granules. At the level of this layer is one enormous 
vacuole (v.) which presses out layer 1 on the centripetal side. 
The floor of this vacuole is formed of the next layer. Both 
the large vacuole and the smaller vacuoles of layers 1 and 
2 contain fat, staining vividly with Sudan III, and it is the 
large vacuole which is seen from the outside, shining through 
the first layer, as the yellowish spot. 


CENTRIFUGED EGG OF THE FROG. 1g 


(3) A layer of a violet-staining plasmatic substance which 
appears to have an alveolar structure (though this may be the 
effect of one of the reagents used). his layer is composite, 
there being denser sheets with a good deal of pigment running 
through lighter patches which contain little pigment. Fat 
vacuoles, singly and in groups, occur in this layer, and a good 
deal of yolk remains in it. 

Neither the second nor the third layer comes to the surface, 
the first layer being here co-terminous with the fourth. 

(4) The yolk and pigment. 

Immediately below layer 3 is a sheet of pigment from 
which dense streamers depend into the yolk below. The ori- 
ginal sheet of superficial pigment remains at the surface. 

(5) The unpigmented region round the vegetative pole 
forms a fifth layer, but this, of course, was not produced by 
the centrifuge. 

A good deal of fat and plasma remain still in the yolk. 

In O, 2 the central spot is slightly depressed, and a white 
border is beginning to appear around the edge of the grey 
patch (Pl. 7, fig. 2, a, b). 

Sections show the following layers : 

(1) With the same characters as in O. 1. Below it a large 
fat-vacuole. 

(2) With the same characters as O. 1, but with still less 
pigment, hence the appearance of a white border round the 
grey patch. This layer has probably been reinforced by 
some plasma from the third layer seen in O. 1 from which the 
yolk-granules have been drawn away. 

(3) A broad band with numerous small yolk-granules—those 
found in the third layer of O. 1—and some vacuolated plasma. 
The latter has presumably just come up from the yolk below. 
This layer lies immediately centrifugal to the layer of 
pigment. 

(4) The rest of the yolk with large granules. The super- 
ficial pigment is unaltered. As layer— 

(5) may be distinguished the original unpigmented area. 

In O. 3 the central spot begins to sink in below the sur- 


112 J. W. JENKINSON. 


rounding folds of the grey patch, while the white border to 
the latter is more distinct. Immediately outside this there is 
in the pigmented area a grey ring (PI. 7, fig. 3, a). 

The section shows the following layers (Pls. 8 and 10, figs. 
VOGT): 

(1) The coarsely vacuolated pigmented layer in which the 
large vacuoles are embedded. This is the grey patch which 
is folded. The folds appear to arise by the accumulation of 
large globules or vacuoles of fat. As these are forced more and © 
more centripetally the folds which are caused by them pass 
over the central depressed portion of the grey patch, which 
itself appears yellow owing to the accumulation of fat there. 

In some of the vacuoles of this layer there is a coagulated 
fluid, which is stained pink with eosin. This appears to be a 
protein distinct from the material of which the walls of the 
vacuoles are composed. Whether the same vacuoles also 
contain fat is not certain ; in that case this might be a protein 
material accompanying the fat—the thin envelope, perhaps, 
of the fatty globules. On the other hand, the eosinophilous 
material may be normally associated with the other proteins of 
the plasma, and only dissociated from it by a certain degree 
of centrifuging. We shall see later that there is evidence for 
the existence of more than one kind of protein in the plasma. 

(2) A layer which appears homogeneous under a low, 
finely alveolar under a high power. It stains violet with 
hematoxylin. In it are a few fat-vacuoles, but it contains no 
yolk-granules. Pigment is scattered through it, and the 
remains of the superficial pigment is seen at its external 
surface. It is the white border of the grey patch. This is 
identical with layer 2 in O. 2, but is thicker, and has lost 
nearly all its fat. 

(3) Below this stretches a thin sheet of pigment, ana 
immediately below this a broad vacuolated layer. The walls 
of the vacuoles are formed of plasma—staining with hema- 
toxylin—with yolk-granules embedded in it; the cavities 
of the vacuoles are occupied by fat. Pigment is scattered 
through this layer, but sparsely, and at the external surface the 


CENTRIFUGED EGG OF THE FROG. 113 


denseness of the original pigment sheet is much diminished. 
Externally this appears as the grey ring. 

This layer is evidently derived from layer 3 of O.2. What 
has happened is that more fat-globules have been driven 
centripetally and accumulated here in vacuoles underneath 
the dense plasma of layer 2, through which at present they 
are unable to pass. 

In this layer are found local accumulations of deeply 
staining (violet) plasma. The plasma is, of course, being 
driven centripetally. There are also vacuoles containing 
eosinophilous coagulum. 

(4) The pigmented region of the yolk. The lower edge of 
the pigment sheet is driven inwards. 

(5) The pigment-free yolk. 

O. 4. (Pl. 7, fig. 4, a, b). The folds of the grey patch 
have nearly met. A second white ring, derived from the 
grey ring, lies just external to the white ring of the grey 
patch. 

The section (Pl. 8, fig. 11) reveals the same layers as in O. 
3. The third layer—of fatty vacuoles with yolk-granules inthe 
walls—is emancipating itself from the sheet of pigment, and 
sending streamers into the plasmatic layer (layer 2) above. 
The yolk-granules are evidently adherent—for the moment— 
to the fat-globules, and caught up in the centripetal move- 
ment. This gives the white division of the grey ring. Layer 
2 is now practically devoid of pigment. The other layers are 
as in O. 3. 

O. 5 (PI. 7, fig. 5, a) presents very little change beyond the 
close approximation of the lips of the folds of the grey patch, 
but in— 

O. 6 great alterations have occurred. 

The grey patch—no longer radially striated—is imme- 
diately surrounded by a deep groove. Outside this lies a very 
broad white ring, and then a grey ring, compounded of three 
zones, the middle of which is darker than the others (Pl. 7, 
fig. 6, a). 

Sections (Pls. 8 and 10, figs. 12, 18) show that (1) the grey 

vou. 60, part 1.—NEW SERIES. 8 


114 J. W. JENKINSON. 


patch is composed of the same coarsely vacuolated pigmented 
material as before, and encloses a large fat-vacuole; (2) out- 
side the groove which bounds this patch is another layer of 
vacuolated material, but this includes but little pigment. The 
vacuoles contain fat. 

(3) There follows the finely alveolar plasmatic layer, almost 
devoid of pigment, with a few yolk-granules, and here and 
there a fat-vacuole. With the transitional condition O. 4 before 
us it seems easy to understand what has happened. The fat 
of layer 3, which was there beginning to penetrate the 
plasmatic layer 2, has completely passed through the latter 
and given rise to the present layer 2—a layer which, as we 
should expect, contains but little pigment. In so doing it 
has shaken off the yolk-granules, which have passed back in 
the opposite direction, although a few remain entangled in 
the third layer. The second and third layers then form 
together the broad white wing. In this process the original 
groove external to the white ring (layer 2 of O. 3 and O. 4) 
disappears, while a fresh groove is formed round that portion 
of the fatty layer in which a considerable quantity of pigment 
is still entangled, namely, in the grey patch. 

(4) Underneath the homogeneous plasma layer is a sheet 
—the uppermost sheet of the yolk—from which a good deal 
of the pigment has disappeared. ‘This is the grey ring. 

In it a finely vacuolated layer can be distinguished above 
from a layer which is not so vacuolated. The former is the 
paler upper zone of the grey ring, and is due to a fresh 
agglomeration of fat-globules in vacuoles below the plasma 
layer. In other words, the process seen in O. 4 is about to 
be repeated. It must be remembered that there is still a 
good deal of fat left in the yolk. With Sudan III the yolk 
stains a faint orange, and proper examination reveals the 
fat-globules between the yolk-granules (PI. 10, fig. 18 ¢). 

I have not succeeded in finding in the sections the lower 
pale zone of the grey ring. 

The other layers are as before. 

We can now form some idea of the changes that take 


CENTRIFUGED EGG OF THE FROG. 115 


place when the egg is centrifuged at this high acceleration 
for progressively longer periods. 

First the pigment and yolk are driven centrifugally while 
the fat and plasma are urged in the reverse direction. 
The fat is, however, lighter than the plasma, so that the 
former occupies the most centripetal position, At the 
same time the globules become agglomerated into larger 
masses, some of which are enormous; a certain amount of 
pigment remains obstinately adherent to the fat. Hence the 
grey patch and yellow spot. The plasma forms a layer next 
to this, and gradually rids itself of fat and yolk, in opposite 
directions. When it is free from pigment it appears as the 
white border. At the same time the pigment is spread out 
in a third layer, which sends streamers into the yolk below. 
This fact, coupled with the frequent mottling of the original 
unpigmented area, suggests that the pigment is perhaps 
heavier than the yolk. 

The supply of fat and plasma in the vegetative region is, 
however, by no means yet exhausted, and a fresh accumulation 
is soon spread out beneath the barrier, at present impenetrable, 
of the plasma layer. But eventually this gives way, the fat 
passes through, dragging at first some yolk-granules with it, 
but these are quickly discarded and driven back. ‘The con- 
dition seen in O. 6 is thus reached, though this is by no means 
the final stage, since there are evident preparations for a 
fresh conglomeration of fat. What the end would be is clear 
enough. The fat—with adherent pigment and plasma—would 
be centripetally disposed, the yolk and pigment centrifugally, 
while the plasma, including, I may perhaps now say, the 
glycogen, would lie between. A discussion of the chemistry 
of the components of these layers must, however, be reserved 
for a later chapter. 

The same kind of effect is produced with a smaller accelera- 
tion, but, as a rule, the white ring appears before the central 
spot—as, for example, in series M (PI. 7, fig. 8). Probably 
considerable force is required to make the fat-globules cohere 
in one or more large masses. 


116 J. W. JENKINSON. 


A section of K. 1 (PI. 9, fig. 13) shows the grey patch as a 
coarsely vacuolated layer, with pigment, the pale border as 
a plasma layer, with some pigment, and below this a sheet of 
pigment and the yolk. 

When the acceleration is smaller still, as in series J (PI. 7, 
fig. 7), the only alteration revealed by the sections is an 
immigration of pigment round the animal pole. This causes 
the faint grey patch seen in these eggs. 

As a rule the various zones and rings become somewhat 
confused, and the folds of the grey patch disappear after a 
short interval; but a grey or blotched patch usually persists 
for a considerable time, and may often be seen at the anterior 
end of the embryo. 

There is, in fact, a slight re-diffusion of the disarranged 
materials through one another, but the normal arrangement 
is never regained—not at least in eggs centrifuged so soon 
before segmentation. 


b. The Effect produced upon the Segmentation of 
the Egg. 


With the lower acceleration the segmentation of the egg is 
normal except for a slight retardation, as for example in L. 1, 
L. 2, M. 1 and M. 2 (PI. 7, fig. 8), but when greater force is 
applied irregularities appear (Pl. 7, figs. 1 c, 2 c,d, 3b, 4c, d, 
5 b,66,c). Even though the first two furrows are meridional, 
those of the third phase may be meridional instead of latitu- 
dinal (as in L. 4) or parallel to the first. The first and second 
are often parallel toa meridian, and with greater accelerations 
the normal sequence is almost entirely abandoned. The first 
furrow may begin at the side instead of at the pole (O. 6), and 
circular furrows, cutting off a small region of the animal hemi- 
sphere completely or incompletely, are frequently observed 
(O. 4, O. 6, L. 4). Many or all of the meridional furrows fail 
to reach the vegetative pole, and as a result the segmentation 
becomes more or less meroblastic ; in extreme cases, indeed, 


‘ 


CENTRIFUGED EGG OF THE FROG. 1 Ly 


it ends in the formation of a blastoderm resting upon an unseg- 
mented yolk, as Oscar Hertwig pointed out; as mentioned 
by the same author the nuclei in the yolk become enlarged, 
irregular in shape and highly chromatic, so resembling the 
yolk nuclei of Elasmobranchs and Teleostei, bodies which are 
concerned in the elaboration of the yolk alone, and do not 
play any part in the formation of the embryo. 

A section (Pl. 9, fig. 14) through one of these eggs (L. 4, 
6:iv:’13) shows that though five days have elapsed since 
the operation no differentiation has occurred. The egg has 
merely continued to segment slowly. 

Some of the strata can still be recognised. 

Round the animal pole is the grey patch, lightly pigmented 
and vacuolated. This region alone is properly segmented, and 
even here it is not possible to distinguish cell-boundaries 
between the nuclei of the deeper layers. There are three or 
four layers of nuclei in all. ‘ 

The grey patch passes into the remains of the plasmatic 
layer, but this contains now masses of yolk-granules and a good 
deal of pigment, both of which have evidently returned from 
the inferior position to which they were driven. In this 
layer are numerous large vacuoles, some clear (these evidently 
contained fat), and some filled with a coagulated liquid which 
stains with the plasma dye (picro-indigo-carmine). The fat- 
globules are presumably due to the breaking up of the larger 
vacuoles of the centrifuged egg—again a return of a substance 
towards its original position. This layer contains nuclei, 
many of which are homogeneous and highly chromatic, while 
others are very large and irregularly amoeboid. 

Below is the yolk, in the upper (centripetal) region of 
which are a few nuclei, of a large but not excessive size. 


Ge The Effect produced upon the Development of 
the Embryo. 


We have at our disposal embryos and larvee obtained from 
the series J, H. 3, G. 1, G. 2, I. 1, L. 1 and M. 1. 


118 J. W. JENKINSON. 


In the series J the acceleration was small and the tadpoles 
were normal. 

In the series G. 1 a larger acceleration was (probably) 
employed, and out of 53 tadpoles 12 were abnormal. ‘These 
tadpoles were twelve days old when preserved. The H. 3 ova 
were subjected to about the same force as those of G. 1, but 
the embryos were killed at an earlier stage when five days 
old. In G. 2 the same force was used as in G. 1 but the 
exposure was longer: out of 43 larve (killed at 12 days) 15 
were abnormal. The force employed and the effect produced 
in the remaining series were greater, the degree of abnor- 
mality being progressively larger in I. 1 (10 abnormal out of 
13), L. 1 (17 abnormal out of 21), and M. 1 (all abnormal). 
The numbers are, however, too small to be genuinely signifi- 
cant, and all the embryos of these three series may be placed 
in one class. Those of I. 1 were preserved on the third and 
fourth days, those of L. 1 on the fifth and seventh days, and 
those of M. 1 on the fourth and sixth days after the operation. 

The available material should, therefore, provide an oppor- 
tunity for the study of the genesis of these aberrations of 
development. 

The distortion of development produced by the centrifuge 
is of a very striking kind, and is, moreover, one which cannot 
be induced as far as Iam aware by any other method. It 
consists essentially of, first, a disturbance at the anterior end, 
which may manifest itself merely by a vacuolation of the 
ectoderm and of the nervous system and other structures in 
that region, but more usually takes the form of a total disin- 
tegration of the front part of the head: the olfactory pits, the 
fore-brain and eyes, and the mid-brain, the skull and the 
mouth, all disappear as such, and the nervous system begins 
in the region of the medulla. Secondly, the yolk is affected. 
The only sign of any derangement may bea tract of undivided 
yolk in which the granules have become fused into one mass, 
but the closure of the blastopore is often prevented, or at 
least delayed, and there is a more or less persistent yolk-plug. 
When it is remembered that the yolk-plug is derived from 


CENTRIFUGED EGG OF THE FROG. 119 


material situated in the vegetative region of the egg while 
the head end of the embryo is developed near the animal pole, 
the significance of the relation between these malformations 
and the structural derangement produced by the centrifuge 
along the axis of the ovum will be sufficiently obvious. While 
these abnormalities are proceeding in the head and round the 
blastopore the middle region may be developing normally, 
such structures as the auditory vesicles, medulla and spinal 
cord, pharynx and gill-clefts, branchial skeleton (in part), 
lungs, heart and blood-vessels, alimentary canal, pronephros, 
germ-cells and tail may all be of ordinary appearance. (The 
tail, I may remind the reader, though, of course, posterior, 1s 
developed from the lateral lips of the blastopore—that is, from 
material placed originally in the equatorial region of the egg.) 
There is, however, one curious malformation in this region, 
which may best be described as a fusion of paired structures 
in the middle line. It is seen in the abnormal arrangement 
of neuroblasts in the medulla and spinal cord, in the fusion 
below the spinal cord of the paired spinal ganglia, sometimes 
of the posterior cranial ganglia too, and in the fusion below 
this, again, of the mesodermal somites, or rather of the myo- 
tomes. The last leads to the obliteration of the notochord. 
Exceptionally the auditory vesicles may unite above the 
medulla. All these changes occur when the ova have been 
subjected to a moderate degree of force (as in the G. series) : 
when a greater force is employed, the derangement is more 
serious; the whole of the nervous system and the organs of 
the middle region of the body may disappear, leaving an 
embryo with perhaps a very short archenteron and an un- 
differentiated mesoderm. A longer or shorter tail may, 
however, grow out ; sometimes it is double, and the mesoderm 
contained in it may show traces of segmentation. 

These monsters occur mainly in the L. 1 and M. 1 series; 
that they are not merely retarded embryos of an early stage, 
which might possibly have developed further, is indicated by 
their age—four to seven days—and by the extreme irregu- 
larity of their appearance. 


120 J. W. JENKINSON. 


The Vacuolation of the Anterior Ectoderm (Pl. 12, 
fig. 44). 


This alteration occurs at the front end, extending back 
some little way on both dorsal and ventral sides, but never 
reaches quite to the posterior end though the ectoderm is 
here often folded. Where the centrifuging has been more 
severe it is more extensive. 

In slight cases the ectoderm remains two-layered and only 
the epidermal layer is affected (asin G. 1,8: iv: 713 [a and 
b]), but after a more violent operation the ectoderm becomes 
thickened byan increase in the number of layers and at the same 
same time folded and pittedand the cellsinall layersareaffevted. 
At the same time the vacuoles are larger (as in G. 2, and 
the embryosof the I.,L.and M.series). There canbe no doubt 
that in the fresh condition these vacuoles were full of the fat 
forced by the centrifuge to the animal pole. Other structures 
at the anterior end may be similarly vacuolated ; the olfactory 
pits (G,1, 8: iv 2713 [a]), the brain (G. 1, 8: iv -°loiee 
J. 1,3: iv: 713 [a, band c]), the optic vesicles (I. 1,3: iv: 
713 [a and c]), the auditory vesicles (G. 2, 8: iv: 713 [a]), 
the ganglia of cranial nerves, and mesoderm cells (G. 1, 8: 


iv: 18 [b], G. 2, 8: iv: 18 [a}). 


The Degeneration of the Front Part of the Head. 


InG. 1, 8:iv:713 (a) (PL. 11, fig. 19) the brain and skull are 
normal and there isa mouth. The only abnormality, indeed, 
to be noticed is the absence of a lens. The optic cup has a 
narrow mouth and is some little distance from the ectoderm 
(Text-fig. 1). 

In such early stages as I. 1, 3 : iv : 713 (a, b, and oc), 
H. 3, 2: iv: 718 (a) (PI. 11, figs. 24, 26, 27), both fore-brain 
and mid-brain are present and apparently normal; while 
normal olfactory pits are found in H. 3, 2: iv : 713 (a), 
I. 1, 4: iv: 718 (a and b), and optic vesicles in H. 3, 2: iv: 
715 (ay ki oe: LS (b): 


CENTRIFUGED EGG OF THE FROG. 121 


As a rule, however, these structures, together with the 
front part of the hind-brain, suffer disintegration by the time 
such stages asG. 1, 8: iv: 713 (b), G. 2, 8: iv: 718 (a, b, c, e, 
and g) (Pl. 11, figs. 20, 21, 23) are reached. All that remains 
of them is then a mass of pigmented vacuolated cells lying in 
the front of the head (‘l'ext-fig. 2, a), some of which have the 
large pale nuclei characteristic of neuroblasts (PI. 12, fig. 45), 
and from some of which nerve-fibres proceed. Mingled with 
the cells is a débris of cell and nuclear fragments (the 
latter having undergone chromatolytic degeneration into 
deeply staining spherules), yolk- and pigment-granules. 

The ganglia of this region (v and vii) often preserve 
their individuality more or less completely (G. 1, 8: iv :715 
[b], G. 2, 8 : iv: 713 [a, c, g]); they appear as groups of 
vacuolated pigmented neuroblasts, and the appropriate nerves 
can often be traced (Pl. 12, fig. 46). In one case (G. 2, 8: 
iv : 713 [a]) the olfactory sacs. remain as two vacuolated 
fibrous masses, still retaining their connection with the ecto- 
derm. No trace of the eyes is found in these embryos. 

At an earlier stage the rudiments of these organs were laid 
down; their degeneration seems to set in almost at once. 
Mimsin H: 3, 2: iv: 713 (b), L. 1, 6:1v: 718 (a), the. front 
part of the brain is a solid wedge of cells (the mid-brain 
region in the last-mentioned is an open groove), in L. 1,6: iv : 
°13 (b) the brain is very small, the olfactory pits are very 
shallow in H. 3, 2: iv : 718 (b), I. 1, 4: iv: 713 (c), and the 
optic vesicles thick-walled (I. 1, 3: iv : ’13 [a]) or much 
reduced (L.1, 6: iv: 713 [a], I.1, 4: iv: 18 [c]) (‘Text-fig. 3). 
Associated with débris of the nervous system and sense 
organs are of course mesodermal elements, such as wandering 
connective tissue cellsand chromatophores. Mesodermal cells 
may suffer vacuolation (G. 1, 8: iv : 713 [b] (PI. 12, fig. 
45 b,c). 

The mouth is almost always absent. It is found in the 
nearly normal G. 1,8: iv : 713 (a), and in early stages a 
stomodeum is present (H. 3,2: iv: 713 [a] I. 1,4: iv: 713 
{a and b]). The pharynx then communicates with the exterior 


122 J. W. JENKINSON. 


only by the gill-clefts, which are often well developed 
(Text-fig. 2, a). 

The head skeleton is also profoundly modified (except in 
G.1, 8: iv: 713 [a]). Properly speaking, the cranium is 
entirely absent, with the exception of a small plate of cartilage 
situated below the débris of the brain, which appears to 
represent some part of the cranial floor—perhaps the anterior 
trabecular plate—in G. 1,8 :iv :’13(b). The branchial skeleton 
is, however, better developed in a few cases. In G.1, 8: iv: 
’13 (b) it consists of a wedge-shaped piece embedded in the 
anterior wall of the pharynx, and a triangular plate placed 
below the pharynx, with a median depression, in which the 
thyroid is lodged, a backwardly directed apex, vertical lateral 
edges, and two pairs of branchial arches. From their rela- 
tion to the gills and gill-clefts these appear to be the second 
and third branchial arches. ‘The anterior piece then represents 
the first branchial and the hyoid, with, perhaps, an admixture 
of mandibular (quadrate) elements. Bundles of myoblasts 
connect these two pieces to one another and the rudiment 
of the skull (Text-fig. 4). 

In G. 2,8: iv:713 (a) there is a plate under the throat 
bearing branchial arches, not symmetrically nor fully de- 
veloped: on one side is a large third branchial and small 
fourth, second and first branchials, and a hyoid, the last two 
being united a little way from the median plate ; on the other 
side are long second and third, short fourth and _ first 
branchials, and a short hyoid. In front the plate is produced 
into two curved pieces, which may be extensions of the hyoid. 
Bundles of myoblasts pass from this anterior to the more 
posterior parts of the apparatus. 

In G. 2, 8: iv: 713 (b) there is a plate bearing three pairs 
of short branchial arches, the first and second being united 
on each side. In front, embedded in the anterior wall of 
the pharynx, is a wedge-shaped piece, resembling that of G.1, 
8:iv:713 (b), but here united to the main plate. It appears 
to be hyoidean. 

In G. 2, 8: iv: 713 (g) there is a similar anterior piece, 


CENTRIFUGED EGG OF THE FROG. 123 


with, however, irregular anterior prolongation. Behind, and 
below the pharynx, is a plate with one pair of irregular 
arches. lastly, in G. 2, 8: iv: 713 (c) the branchial skeleton 
is reduced to a small nodule in the front wall of the pharynx. 
Gill-slits are absent in this larva. 

Suckers are almost invariably absent. They are found, 
indeed, only in G. 1, 8: iv: 713 (a), L.1,8:iv: 718 (b) (PI. 11, 
fig. 37), and in comparatively early stages, such as H. 3, 
2:iv:713 (aand b), I. 1, 4: iv: 713 (a and b) (Text-fig. 5). 


The Changes at the Yolk and Blastopore End. 

SaoG- 2,8: iv:713 (b) (Text-fig. 8 
which are as much or more malformed, a central region is 
observable in the mass of yolk-cells in which cell divisions 
are absent and the yolk-granules fused together. ‘This would 
seem to be due to withdrawal of plasma. Another con- 
sequence of the operation is the delay in the closure of the 
blastopore, seen, for instance, in G. 2, 8 : iv: 713 (d, e), H. 3, 
mee 13 (b), I. 1,.3:1v 2713 (a, b, c), 1. 1,4:1v2 718 (b, 
c, d, e), and in most of the very severely affected embryos of 
the Land M series (PI. 11, figs. 22, 23, 25, 26, 27, 28, 29, 30). 
io, 1, 6: iv : 713 (e and f), in L. 1, 8: iv : 713 (e), m M. 1, 
6: iv: 713 (b) and in M. 1,8: iv: 138 (b and e) the yolk-plug 
is, however, withdrawn in spite of the very serious arrest of 


development (PI. 11, figs. 34, 35, 39, 42). 


), and in all embryos 


/ 


Fusion of Paired Structures in the Middle Line; 
the Hind-brain and Spinal Cord, the Myotomes 
and Notochord. 


In those embryos in which the front part of the nervous 
system has disintegrated, the brain begins at the level of tlie 
auditory vesicles or sometimes a little in front of or behind 
this point (Text-fig. 2b). Inasmuch as the ganglia of v and 
vii, or the remains of them, are found with the rest of the 
débris in the head, we must suppose that part only of the 
hind-brain has escaped destruction. 


124 J. W. JENKINSON. 


This part has the structure of the medulla with a thin roof 
and a thick floor, but is still not quite normal since the floor 
is excessively thick, often projecting into the lumen, while a 
mass of white matter occupies the whole of its ventral side, 
uninterrupted by any cells. The cells lie above this, the 
spongioblasts next the lumen, the neuroblasts above the 
fibres. The lumen is itself further dorsal than it should be. 
The roof is excessively folded. 

In this region are found the auditory and vagus ganglia; 
they often approach one another in the middle line and may 
meet (G. 2,8: iv: 713 [e and g]) (Text-fig. 6). 

The auditory vesicles (Text-fig. 2) are well developed. In 
G. 2,8: iv: 13 (e) each is constricted into two distinct parts, 
and in G. 2,8: iv: 713 (g) the vesicles of opposite sides are 
united in front of the hind-brain (Text-fig. 6). They are 
found, of course, in earlier stages (H. 3, 2: iv: 713; 1.1, 
4:iv:713 [d]) (Text-fig. 7). The spinal cord (PI. 12, fig. 47) 
(Text-fig. 2d) has the same defects as the medulla—that is to 
say, the lumen is driven dorsally by an excessive thickening 
of the floor across which runs a continuous band of fibres; 
next to this comes a layer of neuroblasts, and then the spongio- 
blasts next the lumen (G. 1, 8: iv: 713 [b], G. 2,8: iv: 713 
[a, b,c, e, g], M.1,6:iv:713 [b]). Posteriorly, however, 
it may be normal (G. 2, 8: iv: 713 [b], M. 1, 6: iv: 713 [b]) 
(Text-fig. 8), and may be normal throughout (L. 1, 6: iv : 713 
[a]; 8:iv:713[b]). The abnormality is interesting ; it is as 
though the lateral tracts of fibres had come down, forced the 
neuroblasts of the ventral cornua upwards towards the canal, 
and then met in a continuous ventral band. In other words, 
there has been a median fusion of paired structures, and the 
same phenomenon is seen in the fusion below the cord of the 
spinal gangha (G. 1, 8: iv : 713 [b], G. 2,8: iv +718 fa, 
partial], [b, in front], [c, e, g]). The ganglia, however, 
still retain their proper relations with the cord in that the 
fibres of the dorsal root pass upwards from their cells to enter 
the cord at the side. 


Like the ganglia the myotomes unite below the cord 


CENTRIFUGED EGG OF THE FROG. 125 


(Text-figs. 2, c, d, 10, 11) in a mass of fusiform, horizontally 
placed myoblasts (G. 1, 8: iv: 713 [b], G. 2, 8: iv: 713 [a, 
not posteriorly], [b, c, e, g], I. 1, 4: iv: 713 [d],M.1,6:1v: 
713 [b, d]). There is little doubt that these latter cells by 
which the junction is effected are cells which should have, but 
have not, given rise to the notochord ; for first the notochord 
is absent in these cases or only represented by an occasional 
vacuolation (as in G.1,8: iv: 713 [b], G. 2, 8: iv: 713 [eand 
o]),except where some part of the original material has been 
saved (as in G. 2, 8: iv : 7183 [b]), and a notochord is seen 
lying above the median conjunctive mass, or where there has 
been no fusion (as in the hind-end of G.2,8:iv; 713 [a]). In 
the second place the beginning of the process is seen in 
such early arrested stages as L. 1,6: iv :713 (c) (Text-fig. 15), 
M. 1, 6: iv: 713 (b, d), where the vertebral plate mesoderm 
of the two sides is continuous across the middle line in a 
median mass of cells which are already beginning to elongate. 

In early stages (of slightly centrifugalised eggs) there is a 
normal notochord (H. 3, 2: iv:’13 [b]), and sometimes in 
feel. 1, 62iv :713 [a,b]; 8:iv:713 [b] 5 1. 1,4:1v:713 
fa. b; c]). 

Lastly, in one case, already referred to, the two auditory 
vesicles are united, while a fusion of the auditory and vagus 
ganglia may also occur. 

I am unable to suggest any explanation of this curious 
change. 


The Organs of the Middle Region of the Body. 


With the exceptions already noted, the posterior part of the 
nervous system, the auditory vesicles, and the muscles of the 
back are well developed. The same may be generally stated 
of the pharynx, lungs, alimentary canal, heart and blood- 
vessels, pronephros, germ-cells and tail. 

The pharynx is provided with gill-evaginations, some of 
which may be open (Text-fig. 2,a). InG. 1, 8:iv:713 (b) 
all five pairs are present, and the last three (second, third and 
fourth branchials) are open. In G. 2. 8: iv: 713 (a) the first 


126 J. W. JENKINSON. 


three branchials are present on one side (the second and third 
open) and on the other the second, third and fourth branchials 
(the last two open). In G. 2, 8:iv:7’13 (b) the first, second, 
third and fourth branchials are present on both sides (the 
last two open on one side, the last hardly open on the other), 
and in G.2, 8: iv :713 (g) the first, second and third branchials 
are present on one side (the second and third open), while on 
the other are the first and second branchials (the second open). 
Thus the hyomandibular evagination is found only in the 
first tadpole; in the remainder branchial clefts in numbers 
which differ in individuals, and on the two sides of the same 
individual. 

The embryos of Series H and I are too young to show 
gill-clefts. In L. 1, 8:iv:’18 (b) there are solid outgrowths, 
but no perforations. 

As already pointed out, there is a correlation between the 
presence of clefts and the development of a branchial skeleton, 
as would, of course, be expected. 

External gills are found in G. 1, 8: iv: 713 (b), G. 2, 8: iv: 
°13 (a, band g). They are placed very far forwards (PI. 11, 
fie.20). In G.1, 8:iv:713 (b) there is a small opercular fold. 

The trachea and lungs (Text-figs. 2b, 6, 10) are found 
in the older embryos (G. 1, 8: iv: 713 [b]. G. 2, 8: iv:718 [a, 
b, c, g, but not e]), whose development is not too much 
arrested. The pericardium and heart (Text-figs. 2b, 13), 
with the principal blood-vessels, aortee and cardinal and 
vitelline veins, are found in the better developed embryos 
and are usually well formed (G. 1, 8:iv:713 [b], G. 2, 8:iv: 
13 °{a, b} 6, g|, lL. 1, 8:iv: 713 [b]), but in G. 2,8: 1y2 tae 
the heart is small, in (g) not twisted, and in (e) and L. 1,8: 
iv :713 (b) very small and solid. Aorte and cardinal veins 
may be present when the heart is absent (G. 2, 8: iv: 713 [e]). 

In G. 2, 8:iv:’13 (d and e) there are irregular blood- 
vessels (l'ext-fig, 146) and a structure, which is possibly 
the heart, lying on one side of the pericardial cavity. In the 
younger but fairly normal embryos (H. 3, 2:iv:718 [a, b], 
Tai 215 )a,eb)-c), lL. 1, 6:iv::713 [b, c]))-thegimee 


CENTRIFUGED EGG OF THE FROG. 127 


cardium is still small, and the heart in either a very early 
stage or quite undeveloped. 

The peritoneal cavity is frequently well formed (Text-figs. 
feo. 8) (as 1n-G. 1,8: iv :7t3 [b], G. 2, dziv : 713 
[a, b, c, e, g]), and in communication with the exterior by the 
pronephros (Text-figs. 2c, 6,10, 11). The full number of 
pronephric tubules and funnels is found in G. 2, 8:iv:715 (a, 
b, e, g), but in G. 1, 8:iv:’18 (b) there are only two 
funnels on one side, in G. 2,8: iv:’13 (c) two only on both 
sides. The tubules are bathed, as usually, in the capillaries of 
the posterior cardinal vein. The glomus is found (except in 
G. 2, 8:iv:’13 [a andc]). The ducts open into the cloaca 
(except in G. 2, 8: iv: 713 [e]). 

In the younger embryos (H, I, L. 1, 6:iv:713 [a, b, ¢], 
M. 1, 6:iv:’13 [b, d]), only the pronephric ridge is found 
(Text-fic. 9). In L. 1, 8:iv:’13 (b) differentiation of the 
tubules has not gone very far. 

The gut is well differentiated (with stomach, liver and 
intestine) (Text-fig. 2c); in G. 1,8:iv:'18 (b), G. 2,.8:1iv: 
713 (c, g), less differentiated in G. 2, 8:iv:713 (a, b, e), L. 
1,6:iv:’13 (c). Inthe H and I series little differentiation, 
has occurred beyond the formation of the liver diverticulum, 
and the embryos L. 1, 6: iv:713 (a), 8: iv: 718 (b), M. 1, 6: 
iv:713 (b, d) are in the same early, probably arrested, 
condition. 

A proctodzum, not open to the gut in the early or arrested 
stages, is found always except where the blastopore persists. 

Primordial germ-cells are found at the root of the mesentery 
Gl, S:iv:713(b), G. 2, 8:iv:713 (a, b, ¢, g) (Text-fig. 
2d). 

A tail is found, provided with a fin, in G. 1, 8:1iv:’13 (b), 
G. 2, 8: iv : 7138 (a, b, c) (Pl. 11, figs. 20, 21).. In the young 
embryos it is of course only a short stump. 


(Edema. 


Many of these tadpoles suffer from cedema or an accumu- 
lation of fluid in the connective-tissue inter-cellular spaces, 


128 J. W. JENKINSON. 


or in cavities (Text-figs. 6, 10, 11, 12). Such an accumu- 
lation is seen in the connective-tissue in G. 2, 8:iv:713 (a 
and c), in the ccelom (G.2, 8: iv:’13 [b, c, g], in the pos- 
terior cardinal vein round the pronephros (G. 2, 8:iv: 713 
[a, g]),in the lymphatics or blood-vessels round the gut and 
liver (G. 2, 8:iv: 713 [c]), and in the pronephric tubules 
(G. 2, 8:iv:713 [e]). In G. 1, 8:iv:718 (b) there is a large 
ventral cavity, in front of, but quite independent of, the 
pericardium, partially divided by a median septum, which 
is probably due to the same causes. In several cases where 
the development has been more seriously interfered with, the 
segmentation cavity persists. This, perhaps, belongs to the 
same category. The persistent blastoccel may contain a 
few scattered mesodermal cells, but otherwise retain its 
original character, its front wall being formed of the small 
ectodermal cells of the animal hemisphere, its hind wall of the 
large yolk-cells (M. 1, 8:iv:°13 [c]), or the mesodermal cells 
which have advanced into it may give it a lining of its own, 
incomplete (M..1, 8: iv :713 [b], L. 1, 8:iv:713 fa ae 
or complete (M. 1, 8:iv:’13 [a], G. 2, 8:iv:713 [d]) (Text- 
fig. 14a). The cavity in question has no communication 
with the alimentary canal, which is, indeed, in most of these 
cases very small, or restricted to the blastoporic groove. 


The Changes after Severer Treatment. 


While it is thus possible, if the centrifugal force applied 
be not too great, to obtain tadpoles which, though deformed 
anteriorly and in the region of the blastopore, are yet 
more or less normal in the middle portion of the body 
and in the tail, in individuals which have suffered more 
seriously there is witnessed a gradual loss of structure, until 
eventually no more differentiation occurs than is involved 
in the production of some dorsal and ventral mesoderm 
and a slight blastoporic overgrowth. 

The heart usually goes before the pericardium, or, to express 
it in a better way, the latter may be developed while the former 


CENTRIFUGED EGG OF THE FROG. 129 


is not. Thus in L.1,8:iv:’13 (b) the heart is solid, the peri- 
cardium large, and in I.1,4:iv:713 (c), L.1, 6:iv:713 (b,c), 
the former is absent while the latter is present. In the more 
seriously injured embryos neither is found, though blood- 
vessels may be (G.2, 8:iv:’13 [d]). The pronephros holds 
out perhaps a little longer, as it is formed not only in the em- 
bryos just mentioned, but also in M.1,6:iv:713 (b,d), which 
possess no pericardial cavity. The peritoneal cavity appears 
to persist in these same embryos, but is not found in the more 
degenerate individuals (except possibly in L. 1,8 :1v: “LS ifel}). 
The absence of the heart and pericardium in L. 1, 6: iv :’13 (a), 
while the nervous system and auditory vesicles are present, may 
indicate that in general the former organs are affected before 
the latter. The auditory vesicles are found almost as long as 
the central nervous system persists. Indeed, in G. 2, 8:iv: 
’13 (e) they are well developed, while the hind-brain and spinal 
cord are reduced. On the other hand, in IJ.1,4:1iv:713 (e) 
and in L. 1,6:iv: 713 (c) they have disappeared, while the 
hind-brain and spinal cord have remained. 

Thesuccessive steps in the degeneration of what is left of the 
nervous system are easy to follow. The hind-brain has a very 
small lumen in I. 1,4: iv : 718 (d), is nearly solid in I. 1, 
4: iv :’13 (a, b, c, e), L.1,6:iv:713 (a), M. 1, 6:iv:713(b, d), 
and quite solid in G. 2, 8: iv :’13 (e), L.1, 8 :iv:18 (e) (2). The 
spinal cord is solid in the last-mentioned and nearly so in the 
others; in L. 1, 6:iv: 713 (c) it is still open behind (Text-figs. 
15, 16). In all other cases there is no sign of the nervous 
system at all (Text-fig. 17). The gut may remain—even if 
only as an archenteron—when the other organs have dis- 
appeared. ‘Thus it is found asa narrow but fairly long cavity 
in L. 1,6: iv :713 (c), M.1, 6:iv: 718 (b, d), which still possess 
a nervous system (Text-fig. 16). In G. 2, 8: iv : 718 (f), 
L. 1, 6:iv:718 (f) (Pl. 11, fig. 35), L. 1, 8: iv: 713 (d, e) (PL. 11, 
fig. 39), M. 1, 6: iv:’13 (a,c), and M. 1,8: iv: 713 (a, b,c, d, e) 
(Pl. 11, figs. 42, 43)—none of which have any nervous system— 
it is a very short cavity, opening by the proctodeeum or by the 
blastopore. In G.2,8:iv:713 (d) (Pl. 11, fig. 22), L. 1, 6: iv: 

von. 60, PART 1.—NEW SERIES. 9 


130 J. W. JENKINSON. 


13 (d), 1, 8:iverls (a, b)~(Pl. 11, figs.'36, 37), (cytes 
(Pl. 11, figs. 40, 41), itis reduced to the blastoporic involution 
(Text-fig. 14), while in L. 1, 6: iv: 713 (e) it is altogether 
absent, though a proctodzum is present. 

he tail remains in many of these extremely stunted forms 
as a longer or shorter stump, as in G, 2, 8: iv: 713 (d), 
li. 1, 6: iv : 713,(f), lu. 1, 8: iv: 718 (a, b, e); M. Eibe ayaa 
(a,b, d), M.1,8:iv: 713 (a, b,c, e), and it is sometimes double 
(L. 1, 6: iv:’18[e], L.1,8:iv:’18 [f,g]). The double rudi- 
ment of the tail is seen of course in those cases where it is 
represented only by two caudal swellings (I. 1, 4: iv:718 [ce], 
L. 1,6:iv: 713 [a], L. 1, 8: iv: 718 [e]) (figs. 2955 aah 
These tails, though devoid of nervous system and notochord, 
may yet display traces of a metameric segmentation of the 
mesoderm (as in L. 1, 8: iv: ’13[f, g], M.1, 8: iv: 713 [b]) 
(Text-fig. 18). 

In the rest there is not evena tail. Dorsal and ventral 
mesoderm are, however, always developed. The dorsal meso- 
derm may give indications of the differentiation of a median 
notochordal tract (L. 1, 6: iv: 713 [g], 8: iv: 713 [e], [1]). 


(ps) On tHE CHemicaL NatTuRE oF THE SUBSTANCES IN THE 
Froac’s Eaa WHICH MAY BE SEPARATED FROM ONE ANOTHER 
BY THE CENTRIFUGE. 


The inquiry into the chemical nature of the substances 
in the several layers which appear in the centrifuged egg can 
only be prosecuted with any hope of success by centrifuging 
a large quantity of egg-material in vitro. For this purpose 
the eggs must be obtained free from their coating of mucin- 
jelly, that is, before they have entered the oviduct. 

Such eggs are not, strictly speaking, in quite the same 
physiological condition as the fertilised eggs, imasmuch as 
in them the maturation divisions have not yet occurred. 
If, however, the eggs be taken after their release from the 
ovary—while they are still in the peritoneal cavity—or 
immediately before that release, it will be found that the 
germinal vesicle has already broken down and dispersed its 


CENTRIFUGED EGG OF THE FROG. 131 


contents into the cytoplasm, while the first polar spindle has 
come to the surface of the egg. In their cytoplasm such 
eggs are probably not so very different from those that have 
become completely mature. 

My first idea was to employ only ccelomic ova, but I soon 
found that I could not obtain anything like a sufficient 
quantity of material in this way, since at any one moment 
but few eggs are found in the body-cavity, some being still 
retained in the ovary, while others are already in the oviduct. 
I, therefore, adopted the plan of removing the ovaries, wash- 
ing them well in Ringer’s solution to remove blood and 
peritoneal fluid, and then leaving them in a quantity of the 
same solution at a low temperature until the eggs dropped 
out. I was able to get considerable numbers of eggs by this 
method, which seems to me to be preferable to that employed 
by McClendon; this, as already pointed out, involves the 
inclusion of all the young ova with the old ones. 

From the eggs so obtained the Ringer’s solution was poured 
away as completely as possible. The ova were then ground 
to a fine pulp in a mortar, and this egg-pulp was centrifuged. 
In the first experiments a fairly high velocity was used (about 
3200 revolutions a minute) for about twenty minutes; in later 
experiments, to which I shall refer presently, different and 
lower accelerations and exposures were employed and their 
effects compared. 

The centrifuged mass becomes separated into three distinct 
layers : 

1. A centripetal dark yellowish-grey layer. 

2. A middle light grey or opalescent layer. 

3. A very thick black centrifugal layer. 

The first of these consists of fatty substances and some 
protein, the second of proteins and glycogen, and the third 
of pigment and yolk with an admixture of fatty substances. 

Layer 1. 

(i) This layer is slimy and viscid, hardened at its surface into a 


thin crust. 
(ii) Microscopical examination: Pigment-granules, refringent 


12 J. W. JENKINSON. 


globules entangled into angular masses in some other 
material, and some fiuid. 

The refringent globules are soluble in chloroform and 
acetone, and blacken with osmic acid. They appear, 
therefore, to consist of a fatty material. The material in 
which the globules are entangled and by which they are 
partly obscured is soluble in 1 per cent. NaOH. The 
globules then become more evident. 

Some of the globules are nearly or quite colourless, others 
of a yellow colour. There are no yolk-granules in this 
layer. 

(iii) When this stuff is boiled for some time in alcoholic potash 
a solution of soap is formed which may be salted out with 
NaCl. By CaCl, the soap is precipitated, and on the 
addition of acetic a layer of fatty acid rises to the surface. 

(iv) A portion of the material was mixed with °75 per cent. 
NaCl and filtered. 

A. The filtrate— 

(1) Filters quickly. 

(2) Is opalescent. 

(3) Gives Millon’s reaction slightly. 

(4) Gives Heller’s reaction slightly. 

(5) Gives the xanthoproteic reaction slightly. 

(6) Gives a slight heat coagulum. 

(7) Gives a slight glycogen reaction with iodine. 
B. The residue— 

(1) Gives Millon’s reaction. 

(2) Gives the xanthoproteic reaction. 

(3) Dried to a black colour and washed well withether. The 

ether becomes yellow. The residue is now grey. 
(a) This grey residue gives the xanthoproteic reaction. 
(8) This residue was now washed with hot alcohol, dried 
and incinerated. Theashes were dissolved in dilute 
HNO.,. On the addition of ammonium molybdate 
yellow crystals of ammonium phosphomolybdate 
appear. 

(v) Another portion of this material was placed in alcohol. 

a. A flocculent coagulum appears at once. The coagulum gives 
Millon’s reaction. 

b. The material was then boiled repeatedly in alcohol. The 
alcohol becomes yellow, and several large yellow globules 
fall to the bottom without dissolving. 

Filtered hot— 

A. (i) The filtrate, which becomes cloudy on cooling, was now boiled 


CENTRIFUGED EGG OF THE FROG. 133 


for half an hour with BaOH, the baryta soap filtered off, 
the barium removed by passing CO,, the BaCO, filtered off 
and the filtrate evaporated to dryness. 

A minute fragment of the evaporate placed on a slide 
under a cover-glass in IKI showed clouds of black globules 
and then rectangular black crystals. This is choline ennea- 
iodide, and proves that part at least of the fatty material is 
lecithin. 

(ii) On evaporating a portion of the alcoholic filtrate a yellow- 
brown residue is left. This residue is soluble in ether, but 
not in acetone. A small piece placed in water slowly swells 
and puts out finger-shaped processes. 

When incinerated and dissolved in dilute HNO, it gives 
crystals of ammonio-phosphomolybdate with ammonium- 
molybdate and coffin-lid crystals of ammonio-magnesio- 
phosphate with magnesia mixture. 

B. The residue washed with ether. The ether becomes yellow. 

(i) While acetone gives no precipitate with this yellow filtrate, 
the choline reaction can nevertheless be obtained from the 
dried residue. 

(ii) The second residue, after washing with ether, was dried and 
placed in 1 per cent. NaOH, in which it dissolves. The 
solution— 

(a) Gives the xanthoproteic reaction. 

(3) Gives Millon’s reaction. 

(y) Gives a pink biuret reaction. 

(8) Does not give the iodine reaction for glycogen. 


The centripetal layer therefore contains fatty substance, 
protein and a little glycogen. Part of the fatty substance is 
lecithin, which can be precipitated by acetone from alcoholic 
but not from ethereal solution, which will give choline, and 
from which phosphorus may be obtained. Part seems to be 
fat, as some of the globules are soluble in acetone. 

The proteins seem to include a globulin, but others are 
possibly present ; for instance, the solid protein in which the 
fat-globules are embedded, which may be the same as the 
eosinophilous coagulum seen in the vacuoles of the egg- 
cytoplasm. The phosphorus obtained from the protein-con- 
taining residues may be due to inorganic phosphates or 
perhaps to a phospho-protein ora nucleo-proteid. No purine 


134 J. W. JENKINSON. 


bases have, however, so far been satisfactorily demonstrated 
in any constituent of this layer. 

This layer obviously corresponds to the grey patch in the 
highly centrifugalised eggs. 

The admixture of a good deal of melanin pigment is an 
additional difficulty in the investigation of the proteins of this 
layer. 


Layer 2. 
A. (i) This layer is opalescent and liquid. 

(ii) The greyish colour is not due to pigment, but to angular 
masses enclosing refringent globules of fatty material. 
These masses are not very numerous. They are soluble in 
1 per cent. NaOH. 

(iii) The liquid plasma is coagulated by alcohol, the coagulum 
being finely granular. 

(iv) Distilled water produces a finely granular precipitate, the 
granules being arranged in a reticulum in which the angular 
masses are emmeshed. The greater part of the plasma 
remains liquid. 

(v) The plasma may be coagulated by heat. 

(vi) It may be precipitated by sublimate, and by 2 per cent. 
acetic. 

(vii) HNO, gives a white precipitate. This becomes yellow on 
heating, and with NH, an intense yellow. 

B. A portion of this layer is dissolved in distilled water and 
filtered. 

(i) The filtrate— 

(a) Is perfectly clear under the microscope. 

(b) Is slightly alkaline. 

(c) Is heat coagulable. 

(d) HNO, produces a slight opalescence. 

(e) Gives the xanthoproteic reaction. 

(f) Gives Millon’s reaction. 

(g) Gives a purple biuret reaction. 

(h) Gives a slight glyoxylic reaction (Adamkiewicz). 

(¢) Boiled with 40 per cent. NaOH and treated with lead 
acetate gives no sulphur reaction. 

(ii) The residue, incinerated, gives the ammonium molybdate 

phosphorus reaction. 
c. A portion of this layer is dissolved in ‘75 per cent. NaCl and 
filtered ; it filters slowly. 


CENTRIFUGED EGG OF THE FROG. 135 


(i) The filtrate— 
(a) Is opalescent. 
(b) Is alkaline. 
(c) Gives a heat coagulum. 
(d) Gives no precipitate with 2 per cent. acetic. 
(e) Gives Heller’s HNO, reaction. 
(f) Gives the xanthoproteic reaction. 
(g) Gives Millon’s reaction. 
(hk) When boiled with H,SO, the vapours (of furfural) turn 
anilin acetate red. 
(<) Acidified with acetic, iodine gives an abundant red colour, 
which disappears on heating but reappears on cooling. 
(j) This glycogen reaction is not given after the liquid has 
been digested with saliva. 
(k) After incineration gives the ammonio-phosphomolyb- 
date and ammonio-magnesio-phosphate reactions. 
(ii) The residue, dried and washed with ether— 
(a) Gives the xanthoproteic and— 
(b) Millon’s reactions. 
(c) Washed again, in hot alcohol, the residue, incinerated 
gives the two phosphorus reactions. 


It appears, therefore, that the second layer contains pro- 
teins, a good deal of glycogen, presumably in solution, and 
a small quantity of fatty substance, of which a part may be 
lecithin. 

The evidence, so far as it goes, seems to point to the 
existence of at least two heat coagulable proteins, of which 
one is soluble in water, the other not. There is also the solid 
protein, in which the fat-globules are entangled. The phos- 
phorus may be due to inorganic phosphates, or to phospho- 
proteius or to nucleo-proteins, but no satisfactory proof of the 
existence of the last has been obtained. 

This layer is represented in the eggs by the white circle 
outside the grey patch. 


Layer 3 
A. (i) Is thick and pasty. 
(ii) Microscopically examined, yolk-granules, pigment and fat- 
globules are seen. 


136 J. W. JENKINSON. 


The yolk-granules are soluble in— 

(a) NaCl 10 per cent., 5 per cent., 23 per cent., but in 14 per 
cent. only swell a little and lose their refringency with- 
out disappearing. In saturated solution the yolk- 
granules swell and become less refringent, but remain 
distinct. 

(b) (NH,).SO, 2 saturated solution, but not in 3 saturated or 
saturated solution. 

(ec) NaOH 1 per cent. 

(d) Na,CO, 1 per cent., 2 per cent. 

B. A watery extract is made and filtered. 
The filtrate— 

(a) Is opalescent. 

(b) Is alkaline. 

(ce) Is coagulable by alcohol. 

(2d) When acidified and boiled gives a heat coagulum. 

(e) Gives Heller’s reaction. 

(7) Gives the xanthoproteic reaction. 

(g) Gives Millon’s reaction. 

(hk) Gives a purple biuret reaction. 

(it) Re-filtered, the filtrate still gives a heat coagulum, and 
Heller’s, Millon’s, and the xanthoproteic reactions. 

(k) Gives no glycogen reaction with iodine. 

c. An extract is made in °75 per cent. NaCl and filtered. 
The filtrate— 

(a) Is opalescent. 

(b) Is alkaline. 

(c) Gives a heat coagulum when acidified and boiled. 

(d) With HNO, gives a cloudy precipitate, which partially 
clears on boiling and reappears on cooling. 

(e) This turns yellow on addition of NH,. 

(f) Gives Millon’s reaction. 

(g) Gives a purple biuret reaction. 

(h) Is precipitated by distilled water. 

(z) Is precipitated by alcohol. 

(j) Does not give the anilin acetate reaction. 

p. An extract is made with 1 per cent. NaOH and filtered. 
(i) The filtrate gives a precipitate with 2 per cent. acetic, com- 
pletely when the liquid has been neutralised. 
(ii) Re-filtered, the filtrate— 

(a) Is coagulated by alcohol. 

(b) Gives the xanthoproteic reaction. 

(c) Gives Millon’s reaction. 

(d) Gives a purple biuret reaction. 


CENTRIFUGED EGG OF THE FROG. LS 


gE. A quantity of the material is ground up in a mortar and washed 
with ether. The ether becomes yellow. 

It is then boiled in alcohol, which extracts still more fatty 
substance. 

The residue— 

(a) Dissolved in 10 per cent. trichloracetic acid, and filtered. 
The filtrate gives no phosphorus reaction. 

(b) Incinerated and the ashes dissolved in dilute HNO, ; 

(1) Ammonium-molybdate gives crystals of ammonio- 
phosphomolyhdate. 

(2) The crystals washed in water and dissolved in NH;. 
Magnesia mixture gives crystals of triple phosphate. 

(c) Boiled with H,SO;. The fumes do not give the anilin 
acetate reaction. 

F. A quantity of the stuff is treated with ether and hot alcohol 
until the fat and lecithin has been extracted. 

It is then subjected to the following treatment, borrowed with 
slight modification from Fridericia, in order to see whether 
purine bases are present : 

[((1) Hydrolysed by boiling for fifteen hours with 1 per 
cent. H,SO,. 

(2) NH, added till alkaline. 

(3) Boiled till no alkaline vapours are given off. 

(4) Two per cent. acetic is added till the reaction is acid, 
and the liquid is boiled. 

(5) The coagulated proteins are filtered off. 

(6) Equal parts of 40 per cent. sodium bisulphite and 10 
per cent. copper sulphate are added. The whole is 
boiled for three minutes. 

(7) Filtered. The residue washed repeatedly with boiling 
water until the water is no longer blue. 

(8) Filter-paper and residue are now put back into the same 
flask that was used for the copper precipitation, water 
is added, the whole brought to the boiling-point, and 
excess of sodium sulphide added. 

(9) Acidified with acetic and the H,S boiled off; filtered. 
The residue washed with boiling water which is added 
to the filtrate. 

In this filtrate are the purine bases if any. } 

The copper precipitate obtained from the material of layer 5 
was a bright Indian red colour. The final filtrate (9) gave the 
following reactions : 

(a2) NH,AgNO,added. A gelatinous precipitate came down 
atonce. This is the silver compound of the purine base. 


138 J. W. JENKINSON. 


(b) The precipitate of (a@) was washed and dissolved in hot 33 
per cent. HNO;. <A crystalline precipitate slowly settled. 

(c) Evaporated with HNO, to dryness it became yellow. 
The addition of NaOH turned this a bright red, 
which colour heat converted to a brownish-red. This 
points to xanthine. 

(d) Chlorine water and a drop of HNO, was added. The 
whole is evaporated. The residue was white. The 
addition of NH, turned this first yellow then dark 
red on heating. This again points to xanthine. 

(e) It was precipitated by ammoniacal basic lead acetate but 
not by basic lead acetate alone. 

Jf) It was precipitated by NaOH; the precipitate was 
soluble in excess. 

(g) With HCl a white crystalline precipitate was given. 

(h) With Zn and HCl no red colour was given. 

While, therefore, there is no doubt that a purine base may be 
extracted from the yolk-granules, it is not certain which one 
is present. The reactions appear to point to x anthine. 


The third layer consists, therefore, of pigment and yolk, 
and some fat. There is no glycogen. There may be a little 
plasma left between the yolk-granules, which would give the 
proteins found in the watery extract. 

The solubilities of the yolk-granules point to their protein 
being a globulin or a nucleo-protein. They certainly consist 
of or contain a nucleo-protein, from which a purine base, 
probably xanthine, can be obtained. 

The third layer obviously corresponds to the pigmented 
yolk region of the centrifuged egg. 

These data do not, of course, pretend to be in any sense a 
complete account of the chemistry of the frog’s ovum, but 
they are, I hope, a beginning which may serve as the basis of 
future work. Even as they stand they may throw some light 
on the relation between the stratification of the egg-cyto- 
plasm and the abnormal development of the embryo. 

Before concluding this section I will refer briefly to the 
results of one ortwo other experiments in which the thickness of 
the layers was determined in egg-pulp centrifuged at different 


CENTRIFUGED EGG OF THE FROG. 139 


speeds, and at the same time, and for the same length of time, 
as the eggs of the series H. 1, 2 and 3 and I. 1 and 2. 
These results are embodied in the accompanying table : 


Thickness in millimetres : 


Layer 1. Layer 2. Layer 3. 

H.1 (10 minutes ——. 

at speed ITI) : 8 : 24. 
H. 2 (10 minutes ——e 

at speed IT) + 26 
H.3 (ten minutes ——_ 

at speed I) 2 22 
I. 1 (10 minutes 

at speed IV) : 1 24 274 
I. 2 (80 minutes 

at speed IV)  . 1 5) 31 


It was not possible in the H series to distinguish accurately 
the boundary between the first and second layers. 

In the next table the relative volumes of layers 1 and 2 
taken together and of layer 3 are given, and the ratios of the 
volume of the third to that of the others.! 


Relative volumes. 


Layers 1 and 2. Layer 3. Ratio eee and. 
H. 1 8 14 2-7 
H. 2 + 234 : a9 
H. 3 2 194 : 9.75 
ed D4 25 ; FLA 
| ey 6 284 3 4°75 


The relative volume is a measure of the disturbance. Since 
the embryos of H 3 develop nearly normally, while those of 
I, 1 develop, but not so normally, but the remainder not at all, 
it may be said that centrifuged eggs are only able to develop 
when in a mass of egg-pulp centrifuged at the same accelera- 
tion and for the same time the combined volumes of the 


' In calculating the volume allowance must, of course, be made for 
the hemispherical bottom of the tube. 


140 J. W. JENKINSON. 


layers of fat and plasma are not more than one seventh of 
the volume of the layer of yolk. 

I have only to add that the few quantitative determinations 
I was able to make show, as we should expect, that the water 
content and the fat content of the third layer diminish 
with increased centrifuging. The fat was extracted in a 
Soxhlet apparatus. I give the figures in tabular form : 


Fat in layer 3 
given as a 
percentage of the 


Percentage of 
dry substance 


an layer: dry substance. 
Eel 32°6 : ‘ 22°4 
HH. 2 ol 24-4. 
H.3 29-9 a 
ie ; : : 32°4 . . Ph 
12 . : ; ! 34°9 : : =n 
Egg-pulp not centrifuged 30°4 : : 95 


That there is a smaller percentage of dry substance in the 
third layer of H. 3 than in the uncentrifuged egg-pulp must 
be attributed simply to the fact that different lots of eggs 
were used in these two determinations : 


(c) Conctusions 10 BE Drawn FRoM THE ForeGoING EXxpERt- 
MEN'S. 


The egg of the frog has a structure which depends on a 
certain arrangement of visible materials—the plasma, the 
glycogen, the yolk and the pigment—this arrangement being 
such that the plasma and the glycogen gradually decrease in 
concentration from one point on the surface towards the dia- 
metrically opposite point, while the yolk gradually increases 
in concentration in the same direction. The pigment lies in 
a superficial sheet in the plasmatic two thirds of the egg. In 
respect of the distribution of these substances, therefore, the 
egg has a polar structure, or polarity, or is radially sym- 
metrical about an axis with unlike poles. As a result of 
fertilisation this polar is replaced by a bilateral structure 


CENTRIFUGED EGG OF THE FROG. 141 


when the grey crescent appears on one side of the egg at the 
margin of the pigmented area. 

To this original polar and superimposed bilateral structure 
the structure and symmetry of the embryo are definitely 
related in ordinary development in such a way that the head 
of the embryo is formed near that pole towards which the 
concentration of the plasma is increasing—the animal pole— 
the blastopore closes near the vegetative pole, that towards 
which the concentration of the yolk is increasing, while the 
dorsal side is that on which the grey crescent appeared. 
Ventral, right and left sides are therefore also simultaneously 
determined. 

From Pfliiger’s experiment, and the variant of it im which 
a centrifugal force is substituted for gravity, it appears that 
a new polarity, and, indeed, a bilateral symmetry, may be con- 
ferred upon the egg merely by redistributing the heavier and 
lighter constituents of the cytoplasm. ‘lhe new polar struc- 
ture resembles the old in being determined by a siniilar 
graduated arrangement of materials, and to it the segmenta- 
- tion of the egg and the symmetry of the embryo bear precisely 
the same relation as do the cleavage furrows and the parts of 
the embryo to the structure of the egg in normal develop- 
ment. 

We might reasonably suppose, therefore, that any dis- 
arrangement of these materials, severe enough to alter their 
normal graduation, would bring about a distortion of develop- 
ment or might even prevent it altogether. And this, as we 
have seen, actually occurs. 

By the centrifuge the materials of the egg are driven past 
one another in opposite directions, the fat towards the 
centripetal (animal) pole, with some entangled pigment and 
some plasma, the yolk and pigment towards the centrifugal 
(vegetative) pole, while the movement of the plasma is 
opposite at one end to that of the fat, at the other to that of 
the yolk. 

_ Where the separation of materials is nearly or quite complete 
no development is possible nor even a normal cleavage. In 


142 J. W. JENKINSON. 


less severely centrifuged eggs, while there must in any case be 
an excess of fatty substances round the animal pole, and a 
deficiency of plasma, glycogen and yolk, an excess of yolk 
and a deficiency of plasma and fat round the vegetative pole, 
it is possible that in the equatorial region the relative con- 
centration of the different stuffs, their graduation, may 
remain normal. Such eggs may develop, but monstrously. 

The irregularities of cleavage and development may be 
directly traced to this disturbance. 

As far as cleavage goes it appears that yolk deprived of its 
plasma cannot divide, as, of course, is seen in normally mero- 
blastic eggs, though nuclei may be present. In the enlarge- 
ment of these yolk-nuclei, their excess of chromatin and their 
apparently amceboid movements, there is an interesting 
analogy with the yolk-nuclei of Elasmobranch and Teleostean 
ova. The distortions of development are seen primarily at 
the head and blastopore ends—that is, in the parts developed 
from materials situated at the animal (centripetal) and vege- 
tative (centrifugal) poles. In front there is a fatty vacuola- 
tion of the ectoderm and of other organs, which increases in 
the more severely handled eggs. This is obviously due to 
the accumulation of fat around the animal pole. In the yolk 
towards the other end there is seen a mass where there are no 
cell-divisions and the yolk-granules have fused. This is to be 
attributed to the deficiency of plasma. Where the derange- 
ment of materials has been greater the front part of the 
head is degenerate, and the blastopore fails to close, the 
yolk-plug remaining exposed. The first of these malforma- 
tions must be assigned to the excess of fat and lecithin, the 
deficiency of plasma and possibly of glycogen, and probably 
also of yolk, since the yolk contains a store of nucleo-protein 
which is presumably normally used in the production of fresh 
nuclear material. ‘The persistence of the yolk-plug, on the 
other hand, is caused ultimately by a deficiency of plasma. 
In the normal process of the closure of the blastopore the 
yolk-cells are not merely passively pushed beneath the over- 
growing blastoporic lip, but creep actively inwards, as Kopsch 


CENTRIFUGED EGG OF THE FROG. 143 


showed long since. Anything which diminishes their vitality, 
as, for example, in experiments of another kind, heat or 
solutions of salt and other substances, will inhibit the process 
and the yolk-plug will persist. In the present case it is the 
deficiency of the plasma; the cells are too large and over- 
loaded with yolk. But while at the animal and vegetative 
poles, or head and blastopore ends, the development is being 
distorted in this way, the middle region of the body may 
remain fairly normal. As already pointed out, the distribu- 
tion of materials may remain approximately normal in the 
equatorial tract while altered at the poles, and the develop- 
ment of the middle part be regular for this reason. With 
more centrifuging, however, the usual distribution of stuffs in 
this region also is lost and then the power of development 
disappears ; the heart, pronephros, auditory vesicles, nervous 
system and alimentary canal ail go, and we are left with an 
embryo in which no development has occurred beyond the 
differentiation of some mesoderm. The significance of the 
experimental results obtained by Konopacka is now apparent. 
When the egg is centrifuged in an early stage (prior to 
fertilisation), sufficient time (several hours) elapses before 
cleavage begins, a return of the materials to their original 
position is possible, and development is normal. But when 
the disturbance takes place immediately before cleavage, or 
during its early phases, there is no time for recuperation and 
there are many abnormalities. That such a return does take 
place, or at least begins to do so, is shown by the blurring 
of the zones and rings, even in highly centrifuged eggs, by 
the time cleavage begins. 

A certain arrangement of most of the visible materials of 
the cytoplasm appears therefore to be a condition of normal 
development: the concentration of the plasma, glycogen and 
yolk must be graduated in a certain way, that of the fat must 
be uniform. The position of the pigment is alone inessential, 
as Morgan has pointed out. But with this exception the factors 
to which the egg owes its visible poiar structure are also the 
causes of the production from that egg of a normal organism. 


144 J. W. JENKINSON. 


[Though it does not fall within the scope of these experi- 
ments, it seems probable that the relation between the 
bilateral structure of the ovum and the bilateral symmetry of 
the embryo might be expressed in similar terms. It seems 
likely that at the time of the formation of the grey crescent 
there is some redistribution of the materials in the cytoplasm : 
it is, indeed, known that the grey crescent is due to an 
immigration of pigment. | 

The visible materials of the egg are partly living, the 
plasma, partly not, the yolk, glycogen and fat. Though the 
latter are not properly to be termed organo-genetic, yet they 
condition development in the manner described. In the 
living plasma, on the other hand, there may be distinct organo- 
genetic bodies, arranged in a certain way, and each causally 
related to the formation of some particular organ, and the 
disarrangement of these would necessarily involve malforma- 
tion. No evidence for their existence is, however, brought 
forward by these experiments, and we must postpone the dis- 
cussion of this problem until we have inquired how far in other 
cases also a certain arrangement of materials not only bestows 
a polarity upon the egg, but is also a condition of the 
normal development of the embryo. 


(III) ON THE RELATION BETWEEN THE CYTO- 
PLASMIC STRUCTURE OF THE EGG AND THE 
DEVELOPMENT OF THE EMBRYO IN GENERAL. 


The effect upon development of deranging the cytoplasmic 
materials of the egg has now been investigated in the ova 
of several Invertebrates. 

We turn first to the experiments carried out by Lyon, and 
by Morgon and Spooner, upon the eggs of the sea-urchin, 
Arbacia, in which there is a diffuse red pigment. 

If the ripe but unfertilised ovum be strongly centrifuged 
(f = 6400 g) four strata appear. The red pigment passes to 
the centrifugal pole; next to this is a grey granular layer, 
blackened by osmic acid, then a fluid hyaline layer in which 


CENTRIFUGED EGG OF THE FROG. 145 


lies the nucleus, while the centripetal pole is occupied by a 
cap of opaque white material. The new axis of stratification 
which is thus produced by the operation may make any angle 
with the original axis as determined by the micropyle. When 
removed from the centrifuge the strata begin to re-mingle, 
but the first and fourth return to their original positions very 
slowly, if at all. The second and third layers, on the other 
hand, intermingle with one another rapidly, and it is 
apparently necessary that they should do so before segmenta- 
tion and development can occur, for if the egg be broken into 
two portions between them, then neither portion can be 
fertilised (Lyon). 

In segmentation it is the axis of stratification which deter- 
mines the direction of the furrows, since the first three, which 
are at right angles to one another as in the normal egg, either 
include or are at right angles to this axis, or to put it in 
another way, the axis of stratification coincides with one line of 
intersection between some two of these three divisions. At the 
next division the micromeres are formed at that intersection 
of two furrows which is at the anti-micropylar pole or nearest 
to it (when the axis of stratification is oblique to the original 
egg-axis). It appears therefore that some invisible polarity 
of the egg has remained unaffected by the centrifugal force, 
and that this determines, or at least helps to determine, the 
symmetry of the embryo, since the micromere pole becomes 
the blastopore pole and the original egg-axis the gastrula axis, 
or as nearly as possible (Morgan). The pigment is found in 
any part of the larva, right or left, dorsal or ventral, anterior 
or posterior. It is not, therefore, essential to the development 
of any particular part. 

It may be added here that the yellow pigment band of 
Strongylocentrotus is equally unnecessary. Normally it 
is subequatorial and passes into the archenteron, but it may 
be meridional or oblique to the egg-axis and so become incor- 
porated wholly or partly in the ectoderm (Garbowski). 

Experiments of alike kind on other eggs have yielded very 
similar results, for while the existence of an invisible structure 

vou. 60, part 1.—NEW SERIES. 10 


146 J. W. JENKINSON. 


has been revealed, a structure which is not disturbed by the 
centrifuge and is definitely related to the subsequent differen- 
tiation, that differentiation has been shown to be independent 
of some at least of the visible constituents of the cytoplasm. 

Thus Lillie, by centrifuging the egg of Chetopterus 
during the first maturation division, produced in it three 
layers—a small grey cap at the centripetal pole, a clear layer, 
and a yellow granular hemisphere (on the centrifugal side). 
These strata, it was found, might occupy any position with 
regard to the egg-axis (as defined by the polar bodies), yet 
in fertilisation the sperm always entered at the vegetative 
pole and cleavage was always normally related to that axis. 
The grey cap is derived from the contents of the germinal 
vesicle, the clear band from the microsomes of the endoplasm, 
and the yellow granules from the coarser endoplasmic con- 
stituents. It would be interesting to know something of the 
further fate of these centrifuged eggs. 

This we do know in some other cases. 

The ovum of the Lamellibranch Cumingia contains a 
red pigment and an oily green material, both scattered 
through the cytoplasm. When the egg is centrifuged during 
the first polar division (Morgan) these go to opposite poles, 
the red pigment to the centrifugal, the green oil to the 
centripetal. Between the two is a broad hyaline layer. 
Maturation proceeds and the polar bodies are extruded. With 
the egg-axis, as so determined, the axis of stratification may 
make any angle. Fertilisation occurs, and in the subsequent 
cleavage the planes of division bear the normal relation to 
the axis of the egg. ‘The strata persist so that the red 


pigment may be in the AB or the CD cell and so on, the — 


green oil on the opposite side. Development follows, and 
these two coloured materials are found in the trochophore 
larva in any position opposite to one another. The structure 
of the trochophore has the normal relation to the cleavage. 
system and therefore to the original axis. There does, 
however, appear to be some tendency for the green stuff to 
redistribute itself. 


CENTRIFUGED EGG OF THE FROG. 14.7 


So,again,in Pulmonateeggs(Physa, Planorbis, Limnea) 
Conklin has by the same means produced three strata 


a grey 
finely granular zone at the centripetal pole, a narrow clear 
band, and a yellow granular centrifugal hemisphere. When 
segmentation and development take place the strata make 
any angle with the first and subsequent furrows, and any angle 
with the principal planes of the embryo. Conklin has, how- 
ever, added the important observation that the possibility of 
obtaining a normal development is largely dependent on the 
redistribution to or towards their original positions of some 
of the disturbed cytoplasmic materials, for it is only when 
the operation is performed prior to maturation or during its 
earliest stages—only, that is, when some time elapses between 
the operation and cleavage—that development is afterwards 
normal. Eggs centrifuged during the extrusion of the first 
polar body or later either die or give rise to monstrous em- 
bryos. It appears, further, that during the interval the clear 
substance disappears into the grey orthe yellow layer or both— 
a readjustment which cannot occur unless sufficient time be 
allowed. 

In another molluse, Crepidula, on the other hand, as well 
as in the Ascidian Cynthia, Conklin has found it possible, 
by prolonged centrifuging, to shift the original polar axis 
(the position of which in the experiments just quoted is left 
unaltered) without prejudice to normal development. ‘The 
symmetry of cleavage and of differentiation are, it seems, 
determined by the new polarity, as in the frog’s ege. 

In Cyclops (Spooner) the centrifuge separates the cyto- 
plasm into three similar zones—a greenish-white layer at the 
centripetal pole, a middle clear stratum, and the blue. yolk- 
granules. These eggs are stated to develop normally, even 
when continuously centrifuged. 

From Hegner’s observations it appears that in centrifuged 
beetle eges the embryo is developed at the centripetal end, 
whether this be the morphologically anterior or posterior end ; 
also that when the eggs are centrifuged violently enough to 
produce well-marked layers the development is abnormal. 

VoL. 60, PART 1,.—NEW SERIKS. 10§ 


148 J. W. JENKINSON. 


In the Rotifer Hydatina (Whitney) the polar zones are 
pink and grey, the middle clear as in the foregoing instances. 
The stratification may have any relation to the original axis, 
and the first cleavage is, as in the normal egg, transverse to 
this axis. Normal young are produced, become mature and 
reproduce in their turn. 

Lastly, in the centrifuged egg of Ascaris similar strata 
appear. The normal egg is telolecithal. here are in the 
cytoplasm also some clear spherules and pigment-granules. 
The layers that appear after centrifuging are a sheet of yolk 
at the centripetal end (the yolk is here lighter than the cyto- 
plasm), a layer of clear spherules, clear plasma, and finally at 
the centrifugal pole a cap of brown pigment. When strongly 
centrifuged the ege becomes flattened against the slide on 
which it is placed (the centrifugal force is perpendicular to 
the slide), and, if still subjected to the action of the force, 
the fertilisation spindle places itself at right angles to the 
direction of the latter—that is, parallel to the stratification 
(and in the clear zone), and the division is meridional. If, 
on the other hand, the egg is removed from the machine (or 
is less strongly centrifuged), it resumes or keeps its spherical 
shape, and the spindle returns more or less completely to its 
proper axial position, and the first division is equatorial (or 
oblique). 

It is suggested by Boveri and Miss Hogue (to whom the 
experiment is due) that there is an invisible polarised struc- 
ture in the cytoplasm which is not affected by the operation 
and with the axis of which the stratification of the movable 
substances can make any angle. Into the axis of this 
invisible polarity the spindle is supposed to return if and 
when the egg is allowed to resume its spherical shape. The 
facts do not appear to necessitate this view, for when placed 
on the machine the whole egg rotates inside its shell until 
the heavier animal pole is centrifugal, and then the stratifica- 
tion of the cytoplasmic materials begins. As long as the 
force is operating the spindle is compelled to place itself 
parallel to the stratification, but when released from the force, 


CENTRIFUGED EGG OF THE FROG. 149 


returns or attempts to return to its usual position, namely, in 
the egg-axis, that is, in the stratification axis. The obliquity 
of the spindle in those cases where the return to the normal 
position is incomplete would then be the result of two ten- 
dencies, at right angles to one another, urging the spindle to 
place itself the one perpendicular, the other parallel to the 
stratification. 

When the spindle returns (more or less completely) to its 
normal situation the division is equatorial or oblique, and a 
normal embryo is developed, in spite of the stratification. 
When, however, the spindle remains parallel to the stratifica- 
tion the first division is meridional, and each cell behaves as 
the P, (vegetative) cell of an entire ovum. (The greater part 
of the pigment cap is usually extruded from these eggs at the 
centrifugal pole as a “ball.’) Hach half blastomere divides 
into two, which can be recognised as H. M. St. (endo-meso- 
stomoblast) and P, (meso-gonoblast), by the chromosomes 
being diminished in the one, intact in the other, and by their 
subsequent behaviour, and so gives rise to what is essentially 
a blastula without ectoderm. (Ihe ectoderm is normally 
derived from the absent Sj, the sister ceil of P,.) It might 
be imagined that the ectodermal material had been extruded 
with the “ ball,” but apparently this is not so, since develop- 
ment is the same when (as may happen) no ball is extruded. 

Now, when we survey the results of these experiments, it 
certainly does seem at first sight as though the most incon- 
trovertible proof had been brought forward for the existence 
of a polar structure in the cytoplasm, invisible, not to be 
shifted by the centrifuge, and not therefore dependent on any 
graduation or stratification of heavier and lighter materials 
in opposite directions, a structure, moreover, to which the 
symmetry of the embryo is definitely related, while it bears 
no such relation to the visible fat and yolk and pigment, 
which, indeed, may be driven to any part of the egg without 
prejudice to the eventual normality of development. 

The evidence is, however, not altogether flawless. For, in 
the first place, we notice that in certain cases (Arbacia, the 


150 J. W. JENKINSON. 


Pulmonates) a return of the materials towards their original 


positions is described—a return which, moreover, appears 
certainly in the Pulmonates and possibly in the sea-urchin 
to be a condition of normal development as in the frog. In 
the second place, it does not follow that because the position 
of the egg-axis has not been hitherto affected by the centri- 
fugal foree employed it never can be so moved. Conklin, 
indeed, has succeeded in shifting it in Crepidula and 
Cynthia. The polar structure to which it belongs may there- 
fore eventually prove to be dependent, to quote Conklin’s 
own words, on “the heteropolar arrangement of certain 
odplasmic substances,” though these are indistinguishable 
to the eye, and need not, of course, be of sufficiently different 
specific gravities to allow the force applied to them to over- 
come their viscosities. When it is remembered that a solute 
can be centrifugally separated from its solvent, there is no 
need to despair of the possibility of still further separating 
from one another the materials of the egg cytoplasm. 

Thirdly, the development of these embryos and larve has 
not always been followed either sufficiently closely or suffi- 
ciently long to vindicate the claim that is made for their 
normality. 

In the present state of our knowledge, therefore, it is still 
possible that that polarity of egg-structure to which the 
development of the embryo is so intimately related is in these 
cases also itself dependent on a graduated concentration of 
different materials, and renewed investigation may justify the 
generalisation of this conception. That would probably enable 
us at once to express the polarity of the developing and the 
polarity of the regenerating organism in common terms. For 
the “ gradation of materials,” a formula originally proposed by 
Morgan himself, is certainly the best hypothesis yet put 
forward to explain one of the most constant features of 
regeneration—the formation of a regenerated structure with 
the same characters in the same direction as the original. 
Polarity may, of course, be reversed, but the reversal can 
usually be accounted for by supposing that the concentration 


CENTRIFUGED EGG OF THE FROG. 151 


of a particular substance at a certain point has overcome the 
direction of gradation of the other substance (as, for 
example, when two Hydras are grafted together by their 
head ends, one cut off after the union is complete close to 
the graft plane, and a head instead of a foot regenerated 
from the cut surface). 

Now though the polarity of an organism is often compared 
to the polarity of a crystal, experiment has made _ it 
abundantly clear that there is no real parallel between 
the two phenomena. Any fragment of a crystal has, of 
course, the same optical properties with the same orientation 
as the whole from which it was taken, no matter from 
what part of the larger piece it came; and any such 
fragment will regenerate the whole when suspended in the 
mother liqnor. But with organisms it is not so. In certain 
‘eases (Protozoa, Hydroids, Planarians, some Oligochets) 
any piece that is not too small, or at least any piece 
transversely cut out, will give rise to a complete organism, 
‘and the original polarity is observed in the process, but the 
proportions are not necessarily the same as in the original, 
and differ with the level or region of the whole from which 
the part was removed ; and, indeed, the influence of region 
is frequently so great that in certain parts of the body the 
original structure cannot be replaced at all, as in the earth- 
worm, where a head is never regenerated at an anterior cut 
‘surface behind a certain level, but a tail instead—another 
instance of region, or substance, overcoming polarity or the 
graduation of another substance. The hypothesis, of course, 
‘though the best available, remains a pure speculation, for 
the graduation of materials cannot be seen here as it can in 
the ovum. 

It is not, however, only by means of the centrifuge that the 
causal relation between the polarity of the egg-cytoplasm, 
a certain arrangement whether of its visibly different or of 
its indistinguishable materials, and the formation of the 
organs of the embryo can be demonstrated. ‘There is a 
whole series of experiments of another kind in which some 


ay J. W. JENKINSON. 


definite part of the ovum being removed, some definite 
organ or system of organs of the embryo is defective or 
lacking. Not only, therefore, does the polarity of the egg- 
cytoplasm determine the orientation of the parts of the 
embryo, but certain materials in it are causally associated 
with the eventual appearance of certain organs; the 
materials in question may, but need not be, situated in that 
region in which the organ will be developed, as in Denta- 
lium, where factors on which the formation of both the 
anterior sense-organ and the trunk of the larva depend are 
both resident in the vegetative polar lobe of the egg. 

In the cytoplasm, therefore, are placed, on both lines of 
evidence, the causes on which the primary differentiation of 
the embryo depends. Not that it is proved or even necessary 
to suppose that every separately inheritable organ or cha- 
racter of an organ is represented by some distinct material, 
for there is evidence that fresh structures may be developed 
by the interactions of already existing parts on one another. 
Given, indeed, some heterogeneity to start with, and the rest, 
however complex it may be, will follow in a regular and 
orderly sequence. 

And this primary heterogeneity resides in the cytoplasm, 
which, therefore, is a vehicle of inheritable characters. The 
characters which the cytoplasm of the egg-cell thus trans- 
mits appear to be those which give the organism its rough 
outline, the large features which place it in its proper phylum, 
class, order, family, perhaps, which make it an Echinoderm 
and not a Vertebrate or a Mollusc, an Echinoid and not an 
Asteroid, and such general characters are probably trans- 
mitted only by the cytoplasm, and, therefore, only by the 
female cell, if we may trust the evidence of heteregeneous 
hybridisations. ‘The smaller characters, of course—generic, 
specific, varietal—which might be more correctly described 
as so many modifications of the larger ones, are obviously 
transmissible equally by the germ-cells of the two sexes; 
that is, are carried by the chromosomes or smaller chromatic 
elements of their nuclei, But even the different chromatic 


CENTRIFUGED EGG OF THE FROG. 153 


elements require a difference in the cytoplasm before they can 
exert each their appropriate activity, and call forth each the 
character to which each is beforehand appropriate. For every 
cell in the body contains in its nucleus a complete set—indeed, 
in sexually produced organisms two complete sets—of the 
chromatin elements or determinants characteristic of the 
species, since nuclear division is quantitative always, with 
the exception of the maturation division of the germ-cells. 
It is hard to conceive of the several chromatic determinants 
coming into operation, each at the right place and time, save 
in a heterogeneous medium. Given such a medium, then, this 
nuclear element will become active here, and that there, this 
will produce a specific modification of one character, that of 
another. And it is in the cytoplasm that this primitive and 
necessary heterogeneity is found, and from the cytoplasm 
that the prime factors of differentiation may, we hope, be 
isolated, and perhaps by means of the centrifuge. 


OXFORD ; 
October, 1913. 


LITERATURE. 


Born, G.—‘* Ueber den Einfluss der Schwere auf das Froschei,” ‘ Arch. 
mikr. Anat.,’ xxiv, 1885. 
Boveri, T.—‘ Ueber die Teilung centrifugierter Eier von Ascaris 
megalocephala,’ ‘ Arch. Ent.-Mech.,’ xxx, 2, 1910. 
Child, C. M—‘“ An Analysis of Form-regulation in Tubularia,” 
* Arch. Ent.-Mech.,’ xxiii, xxiv, 1907. 
“Studies on the Dynamics of Morphogenesis and Inheritance 
in Experimental Reproduction,” I, ‘ Journ. Exp. Zool.,’ x, 1911. 
Conklin, E. G.—* The Application of Experiment to the Study of the 
Organisation and Early Differentiation of the Egg,” ‘ Anat. 
Rec.,’ iii, 1909. 
‘The Effects of Centrifugal Force upon the Organisation and 
Development of the Eggs of Fresh-water Pulmonates.’ 
Dimon, A. C.—*The Regeneration of a Heteromorphic Tail in 
Allolobophora fwtida,” ‘Journ. Exp. Zool.,’ i, 1904. 


154 J. W. JENKINSON. 


Fridericia, L. 8.—* Untersuchungen iiber die Harnsiiureproduktion 
und die Nucleoproteidneubildung beim Hiithnerembryo,” ‘Skan- 
dinav. Arch. Physiol.,’ xxvi, 1912. 

Garbowski, M. T.—* Ueber die Polaritit des Seeigeleies,” ‘ Bull. 
Internat. Acad. Sci. de Cracovie,’ 1905. 

Hammarsten, O.—‘ Text-book of Physiological Chemistry,’ tr. by J. A. 
Mandel, New York, 1911. 

Hegner, R. W.—“ The Effects of Centrifugal Force upon the Eggs of 
some Chrysomelid Beetles,” ‘Journ. Exp. Zool.,’ vi, 1909. 
Hertwig, O.—* Ueber einige am befruchteten Froschei durch Centri- 
fugalkraft hervorgerufene Mechanomorphosen,”’ ‘S.-B. Konig. 

preuss. Akad. Wiss. Berlin,’ 1897. 

— “Weitere Versuche iiber den Einfluss der Zentrifugalkraft auf 
die Entwicklung tierischer Hier,” ‘ Arch. mikr. Anat.,’ xiii, 1904. 

Hogue, M. J.—‘ Ueber die Wirkung der Centrifugalkraft auf die Kier 
von Ascaris megalocephala,” ‘ Arch. Ent.-Mech.,’ xxix, 1910. 

Jenkinson, J. W.— Experimental Embryology,’ Oxford, 1909. 

King, H. D.—* Observations and Experiments on Regeneration in 
Hydra viridis,” ‘Arch. Ent.-Mech.,’ xiii, 1902. 

Konopacka, B.-— Die Gestaltungsvorgiinge der in verschiedenen 
Entwicklungsstadien centrifugierten Froschkeime,” ‘ Bull. 
Internat. Acad. Se. Cracovie,’ 1908. 

Lillie, F. R.—‘‘ Observations and Experiments concerning the Elemen- 
tary Phenomena of Development in Chietopterus,” ‘ Journ. Exp. 
Zool.,’ iii, 1906. 

Lyon, E. P.—“ Results of Centrifugalising Eggs,” ‘ Arch. Ent.-Mech.,’ 
xxiii, 1507. 

McClendon, J. F.—‘* Cytological and Chemical Studies of Centrifuged 
Frog Eggs,” *‘ Arch. Ent.-Mech.,’ xxvii, 1909. 

— “On the Nucleo-albumin in the Yolk-platelets of the Frog’s 
Egg,” ‘ Amer. Journ. Phys.,’ xxv, 1909. 

Morgan, T. H.—* The Relation between Normal and Abnormal Develop- 
ment of the Embryo of the Frog as Determined by Injury to 
the Yolk Portion of the Egg,” ‘Arch. Ent.-Mech.,’ xv, 1903. 

—— “The Influence of a Strong Centrifugal Force on the Frog's 

Egg,” ‘ Arch. Ent.-Mech.,’ xxii, 1906. 

and G. B. Spooner.—* The Polarity of the Centrifuged Egg,” 

‘ Arch. Ent.-Mech.,’ xxviii, 1909. 

—— “Cytological Studies of Centrifuged Eggs: I. Cumingia,” 
‘Journ. Exp. Zool.,’ ix, 1910. 


CENTRIFUGED EGG OF THE FROG. 155 


Morgan, T. H.—*‘ Polarity’ considered as a Phenomenon of Gradation 
of Materials,” ‘ Journ. Exp. Zool.,’ ii, 1905. 


——— “ Regeneration in Planarians,” ‘ Arch. Ent.-Mech.,’ x, 1900. 
— “Regeneration in Bipaliam kewense,” ‘ Arch. Ent.-Mech.,’ 
ix, 1900. 

“The Internal Influences that Determine the Relative Size of 


Double Structures in Planaria lugubris,’ ‘Biol. Bull.,’ iii, 
1902. 


—— “Growth and Regeneration in Planaria lugubris,” ‘ Arch. 
Ent.-Mech.,’ xiii, 1902. 
—— “Regeneration of Heteromorphic Tails in Posterior Pieces of 
Planaria simplicissima,” ‘Journ. Exp. Zool.,’ i, 1904. 
— “The Control of Heteromorphosis in Planaria maculata,” 
* Arch. Ent.-Mech.,’ xvii, 1904. 

Pfliiger, E —‘‘ Ueber den Hinfluss der Schwerkraft auf die Teilung der 
Zellen,” ‘ Pfluger’s Arch.,’ xxxi, xxxii, xxxiv, 1883. 

Roux, W.—‘ Ueber die Entwicklung des Froscheies bei Aufhebung der 
richtenden Wirkung der Schwere,” ‘ Breslau artz. Zeitschr.,’ 1884, 
also ‘ Ges. Abh.,’ 19. 

Spooner, G. B.—* Embryological Studies with the Centrifuge,” ‘ Journ. 
Exp. Zool.,’ x, 1911. 

Wetzel, G.—“ Zentrifugierversuche au unbefruchteten Hiern von Rana 
fusca,” ‘ Arch. mikr. Anat.,’ lxiii, 1904. 

Whitney, D. D.—* The Effect of a Centrifugal Force upon the Develop- 


ment and Sex of Parthenogenetic Eggs of Hydatina senta,” 
‘Journ. Exp. Zool.,’ vi, 1909. 


EXPLANATION OF PLATES 7-12, 


Illustrating Dr. J. W. Jenkinson’s paper “On the Relation 
between the Structure and the Development of the 
Centrifuged Egg of the Frog.” 


PLATE 7. 


Fig. 1—0O.1. a, from the equator; b, from the animal pole; c, seg- 
mentation. 


Fig. 2.—O. 2. a, from the equator; b, from the animal pole; ¢, d 
segmentation. 


b 


Fig. 3.—O. 3. a, from the animal pole ; b, segmentation. 


156 J. W. JENKINSON. 


Fig. 4—O. 4. a, from the equator; b, from the animal pole; ¢, d, 
cleavage, from the equator and from the animal pole. 

Fig. 5.—O.5. a, from the animal pole; }, cleavage. 

Fig. 6.—O.6. a, fromthe equator; b, cleavage from the equator ; 
ce, cleavage from the animal pole. 

Fig. 7.—J.1. From the animal pole. 

Fig. 8a.—M.1. 4cells. 8b—M.2. 4 cells. Both from the 
equator. 

PLATE 8. 
Figs. 9-13.—Meridional sections of centrifuged eggs. 

Fig. 9—O.1. 1-5, the strata or layers (see text) ; v., enormous fat- 
vacuole. 

Fig. 10.—O. 3. 1-5, the layers described in the text ; v., vacuole. 

Fig. 11—O.4. Figures and letters as before. 

Fig. 12—O.6. 1-4, the layers described in the text; 5, yolk with 
pigment; 6, the original unpigmented portion of the yolk. (5 and 6 cor- 
respond to 4 and 5 of the previous figures) ; in 3 there is no fat; 5 and 
6 have not yet been exhausted of their fat.) 


PLATE 9. 


Fig. 13.—K. 1. 1, vacuolated fatty layer with some pigment; 2, 
plasma; 3, yolk and pigment; 4, unpigmented yolk. 

Fig. 14.—Meridional section through the meroblastically segmented 
egg, L. 4,6:iv:°13. There is a cap of cells (bl.) lying on a vacuolated 
unsegmented plasmatic mass in which are patches of yolk-granules and 
some pigment. The nuclei here are large, chromatic and irregular. 
Below is the yolk, with nuclei only in its more animal region. Of the 
vacuoles some are filled with a (protein) coagulum, while others (clear) 
presumably contained fat. 


PLATE 110. 


Fig. 15.—Glycogen spherules amongst the yolk-granules : a, from the 
animal; 6b, from the vegetative region of the normal egg. 


Figs. 16-18.—Minute structure of the layers in the centrifuged eggs. 

Fig. 16.—a, layer1; 6, layer 2 and ¢, layer 3 (with some yolk-granules) 
oO... 

Fig. 17.—O. 3. a, Kosinophilous coagulum in vacuoles of layer 1; 
b, plasma of layer 2, finely alveolar, with some pigment and a few 
vacuoles; c, the vacuoles of layer 3, with yolk-granules in the walls; 
d, lamp of plasma from layer 3. 


CENTRIFUGED EGG OF THE FROG. Baye 


Fig. 18.—O. 6. a, layers 2 (above) and 3 (below); a few yolk-granules 
in 3; b, the upper vacuolated sheet of the grey ring (layer 4); c, fat- 
globules (stained with Sudan III) amongst the yolk-granules of the 
vegetative moiety of the egg. 


PEATE, UT. 


Figs. 19-43.—Development of the centrifuged eggs. 
- Fig. 19.—G. 1. 8:iv: 13 (a). 


Fig. 20.—G.1. 8:iv: 13 (b). 
Fig. 21—G. 2. 8:iv: 713 (c). 
Fig. 22—G.2. 8:iv: 713 (d). 
Fig. 23.—G. 2. 8:iv: ‘13 (e). 
Fig. 24.—H. 3. 2:iv: ‘18 (a). 
Fig. 25.—H. 3. 2:iv: 713 (b). 
Fig. 26—I.1. 3:iv: ’13 (a). 

Fig. 27.—I. 1. 3:iv: 13 (b). 

Fig. 28.—I.1. 4:iv: 713 (b). 

Fig. 29.—I. 1. 4:iv: 18 (ce). 

Fig. 30.—L 1. 4:iv: 718 (e). 

Fig. 31—L.1. 6:iv: 18 (a). 
Fig. 32—L.1. 6:iv: ’18 (ce). 
Fig. 33.—L. 1 :iv: 713 (d). 
Fig. 34.—L. 1 :iv: 713 (e). 
Fig. 35.—L. 1 civ) 1S.(): 
Fig. 36.—L. 1 :iv: 713 (a). 


iv: 13‘ (b). 
: 713 (c) (2). 
:iv: 713 (e). 
civ: 13 (6). 
:iv: 13 (g). 


ey 

Lol 

og 

ey) 

oO 

Is" 
ieee Ee t ; 
aomm nmnanmDmnnmDnnann 

< 


PLATE 12. 


Fig. 44.—Vacuolated anterior ectoderm, in b much thickened and 
pitted. In a the vacuoles are confined to the epidermal layer. a, G.1, 
8:iv: 13 (b); b, G. 2, 8: iv: 13 (e). 

voL. 60, PART 1.—NEW SERIES. 11 


158 J. W. JENKINSON. 


Fig. 45.—Cells and cell-débris from the anterior end of G. 1, 8:iv: 
“13 (b). a, from the degenerate remains of the brain. The figure includes 
vacuolated pigmented cells, one with the large pale nucleus of a neuro- 
blast, the other, containing also yolk-granules, probably a spongioblast, 
a spongioblast without vacuoles but with yolk-granules, fragments of 
cells with spherules of chromatin from chromatolytic nuclei, and free 
spherules of chromatin and bits of cytoplasm; b, deeply pigmented 
cells, probably mesodermal; c, vacuolated mesodermal elements. 

Fig. 46.—G. 2. 8:iv:°13 (a). Group of vacuolated and pigmented 
neuroblasts, with nerve-fibres. Probably the ganglion of the fifth 
nerve. 

Fig. 47.—G. 1. 8:iv:’13 (b). Section of the abnormal spinal cord 
and ventrally fused ganglia. In the cord the lumen is pushed dorsally 
by the extension of the white matter on to the ventral side, where it 
forms a continuous band. Above this a layer of neuroblasts, with pale 
nuclei, and then round the lumen the spongioblasts with darker nuclei. 
There are yolk-granules in both spongioblasts and neuroblasts and a 
few vacuoles in the neuroblasts. The fibres of the dorsal root are seen 
on one side (the right) growing in from the cells of the ganglion. 


SPOROGONY AND SYSTEMATIC POSITION OF AGGREGATID®. 159 


The Sporogony and Systematic Position of the 
Aggregatide. 


By 
Helen L. M. Pixell-Goodrich, B.Sc., 
Beit Memorial Research Fellow. 


With Plate 15. 


THE very noticeable parasites infesting the alimentary 
canal of many Cephalopods were recorded by Eberth (8) so 
long ago as 1862. The genus comprising these forms has 
suffered many changes of nomenclature. In 1875 it was 
named Benedenia by Schneider (21, Note xiii, p. xl) to 
include B. octopiana from the octopus, and this generic name 
was employed so late as 1900 by Schaudinn (20, p. 203). In 
Schneider’s second paper (22) in 1883 the parasite was 
referred to throughout as Klossia instead of Benedenia, as 
the author stated that he was convinced that the two genera 
could not be distinguished. In this Siedlecki followed him 
in his two papers of 1898 (28 and 24) on the parasites of 
Sepia, which he considered to be identical with those of the 
octopus, although he found only three or four sporozoites in 
a sporocyst, whereas Schneider had described eight to fifteen 
in Klossia octopiana, and Labbé had already given the 
specific name of K. eberthi (7) to the species inhabiting 
Sepia. In a footnote to his second paper, however, Siedlecki 
(24, p. 800) stated that after seeing Laveran’s work on K. 
helicina he thought the original name of Benedenia should 
be readopted. It was then discovered that this name had in 
1858 been appropriated for a Trematode genus, therefore 
Blanchard, in 1900, altered it to Légeria. Jacquement (5) 

vot. 60, pART 1.—NEW SERIES. 12 


160 HELEN L. M. PIXELL-GOODRICH. 


in 1903 pointed out that this had been already applied to a 
polycystid Gregarine, and again changed the name to 
Légerina. This, however, was never used; for, in the 
previous year, after Siedlecki had stated that, like the so- 
called Eugregarines, these parasites did not appear to undergo 
Schizogony, the genus had been called Eucoccidium by 
Lithe (11). By this generic name the parasites were generally 
known until, in 1906, the excellent researches of Leger and 
Duboseq (§ and 9) showed that the schizogony, which had 
been supposed to be absent from the life-history of the genus, 
took place in the Crab, Portunus, where Frenzel (4) had 
described it in 1885 as a Gregarine under the name of 
Aggregata; this name is therefore now universally applied 
to the genus which I hope to prove is undoubtedly Coccidian 
in character as assumed by Siedlecki. Moroff (15) in 1906 
claimed to have shown that the fertilisation was of the 
gregarine type; that is, that the macrogametocyte gave rise 
before fertilisation to a number of small cells, and that these 
were the true macrogametes, each of which gave, after ferti- 
lisation, a single spore. In 1908, in a later work of great 
length, Moroff (16) admitted that he might have been 
mistaken in his description of the process of fertilisation, and 
no longer insisted on these parasites being Gregarines, stating 
that further investigations were necessary. It was for this 
reason that, at Prof. Minchin’s suggestion, I undertook this 
study of Aggregata, which has extended over some years. 
During this time I have examined Cephalopods, chiefly Sepia, 
from Plymouth, the Solent, and more recently from the 
Mediterranean in the neighbourhood of Banyuls-sur-Mer. 
To the authorities and staff, especially to the Director, Prof. 
Pruvot, of the Laboratoire Arago, I should like to express 
my thanks for their hospitality and assistance ; also to Prof. 
Minchin, not only for suggesting such an interesting subject 
for investigation, but for much kind help given me during 
the years that I have been privileged to work in his Depart- 
ment at the Lister Institute. Since last October the work 
has been carried on in Oxford, and I am much indebted to 


SPOROGONY AND SYSTEMATIC POSITION OF AGGREGATIDA. 161 


Prof. G. C. Bourne for the kind interest he has shown in the 
progress of the investigations, and for allowing me the use of 
a laboratory in his department. My husband, Mr. E. 38. 
Goodrich, has also given me much advice and assistance. 

As pointed out above, Aggregata octopiana of the 
octopus and A. eberthi of Sepia have long been known, but 
in addition to these, Moroff (16, p. 144) claims to have dis- 
tinguished some eight or nine more species in Octopus and 
five in Sepia. The distinguishing characteristics given, being 
chiefly those of size, are not at all conclusive, and, so far as 
my observations go, I have been unable to corroborate their 
existence. 

According to Léger and Duboscq (10) A. eberthi can 
undergo its schizogony in Portunus arcuatus and P. hol- 
satus as well asin P.depurator (10, p. 90 et seq.), but not in 
P.puber. L. Their attempts to infect other Crustacea, such 
as Inachus dorsettensis, Carcinus maenas, etc. (10, 
p. 95), failed, although some of these, for instance, Inachus 
dorsettensis, are sometimes found in nature infected with a 
species of Aggregata (Smith, 19). Possibly the natural cepha- 
lopod host of the latter is an octopus; it does not, at any rate, 
seem to be Sepia officinalis. Many experiments such as 
these appear to me to be necessary before we can hope con- 
clusively to distinguish different species of Aggregata, the 
possible existence of several of which Tam not prepared to deny. 


MareERIAL AND TECHNIQUE. 

It is very essential to have the material quite fresh, and in 
this connection a method of killing Cephalopods so as to 
eliminate the difficulty of dealing with the emission of ink 
may be of interest. This is to slip a loop of fine, though 
strong string round the head just above the mantle, and 
suddenly to pull the ends very tightly. In this way the 
animal is immediately strangled—the general paralysis being 
too quick to allow time for any ink to be thrown out, and at 
the same time quantities of blood are left in the heart and may 
be drawn off to be used as a mounting medium if required. 


162 HELEN L. M. PIXELL-GOODRICH. 


All Sepia officinalis, whether from Plymouth, the Solent 
or Banyuls, have been found to be infected: but none of the 
numerous specimens of Sepia elegans, which occurred in 
quantities near Banyuls-sur-Mer, had any infection during 
September, neither had the Eledone and Loligo that were 
examined. 

Both Sepia officinalis and Eledone from Banyuls were 
almost universally infected with the Cestode Scolex poly- 
morphus, the occurrence of which has already been recorded 
by Monticelli (14) from Naples. In Sepia officinalis the 
region of the alimentary canal most infected was the spiral 
stomach between the entrance into the muscular stomach and 
the longitudinal ridges which run up into the rectum. The 
latter chiefly contained large cysts, which, being somewhat 
over one millimetre in diameter, were quite visible to the naked 
eye. For examination in the fresh state some of the tissue 
from the spiral stomach was teased out, in sea-water asa general 
rule, though sometimes intestinal fluid or blood were found 
to be preferable. Films and portions of the spiral stomach 
were fixed in Schaudinn’s fluid, ordinary corrosive sublimate 
and acetic acid, or mixtures of the latter with equal quantities 
of 4 per cent. formalin—this last perhaps giving on the whole 
the best results. Iron hematoxylin was by far the most 
satisfactory nuclear stain. Ehrlich’s and Delafield’s hema- 
toxylin mixtures were also used, as well as paracarmine, borax 
carmine and many cytoplasmic stains, such as Lichtgriin and 
picric acid, eosin or orange G. 


VEGETATIVE STAGES. 


Since Léger and Duboscq (10), Moroff (16), and others 
have already given excellent figures of merozoites, tropho- 
zoites, and sporozoites, we can begin our study with the adult 
trophozoite ripe for reproduction. At this stage there is 
certainly a delicate limiting membrane present—a fact denied 
by Siedlecki (24, p. 804),—also a large vesicular nucleus, 
containing, as a rule, a single large karyosome. In some 


SPOROGONY AND SYSTEMATIC POSITION OF AGGREGATIDA. 163 


eases the cytoplasm contains large irregular masses, which 
stain intensely with iron-hematoxylin and other chromatin 
stains. These would seem to be metabolic products of the 
cytoplasm, such as Comes (1) has described in Stenophora, 
rather than of nuclear origin, and to bear no obvious relation 
to the processes of reproduction described below. Smaller 
particles of more uniform size, probably of similar nature, 
are usually scattered throughout the cytoplasm. The ripe 
trophozoites vary enormously in size, the longer diameter of 
some reaching 400 u, while others may be only about 50 p ; 
and, since all intermediate dimensions can be found, it does not 
seem possible to distinguish several species, either in this or 
subsequent stages, by size alone. Therefore I cannot support 
Moroff’s conclusions on this point (16, pp. 146 and 147). 


FORMATION OF THE GAMETES. 


While the distinction between males and females becomes 
very pronounced Jater on, it is extremely difficult to trace 
the differences back to the trophozoite stage. However, 
individuals which are to become females appear, on the whole, 
less dense, aud have their cytoplasm more vacuolated than 
those destined to be males. ‘I'he females also as a rule grow 
to a larger size. After a considerable resting period at the 
completion of growth the early stages leading up to the 
formation of the gametes are gone through very rapidly. 

In the female there is, first, a migration of the nucleus 
towards the surface, from which, however, it remaius 
separated by a plug of dense, finely granular protoplasm, 
which gives off radiations, spreading in all directions through 
the cytoplasm (fig. 1). The central part of this protoplasmic 
plug rises up on the surface, forming a cone of attraction, 
and, after entrance of the microgamete, appears to sink in 
again, carrying the male chromatin with it (figs. 2 and 3). 

In the male the nucleus also approaches the surface and 
rapidly divides by a peculiar method of mitosis. A centro- 
some with distinct astral radiations apparently emerges from 


164. HELEN L. M. PIXELL-GOODRICH. 


the nucleus at the beginning of the process, and undergoes | 
repeated division before the nucleus becomes subdivided. 
These centrosomes, with their conspicuous asters, spread over 
the surface of the microgametocyte, giving rise to such 
polymitotic figures as those given by Moroff (16, fig. 50, pl. vi). 
In the meantime the nucleus becomes irregularly lobed, the 
lobes growing out in various directions towards the asters. 
The peripheral lobes eventually become constricted off as 
separate nuclei, into which the greater part of the chromatin 
has migrated. Ata later stage the nuclei, containing deeply 
staining chromatic networks, are found evenly distributed 
over the whole surface. ‘Throughout this process the centro- 
somes and asters take up a position between the nuclei 
and the surface of the gametocyte. The peripheral nuclei 
continue to undergo repeated binary division, preceded in 
every case by the division of the centrosome and aster. As 
the nuclei become more numerous the surface at their dis- 
posal is increased by the formation of rounded lobes, which 
gradually become nipped off as separate spheres, covered 
with developing microgametes (fig. 13). A fully developed 
male generally consists of several such spheres, sometimes as 
many as six or seven. The nuclear divisions described above 
are very similar to those which take place in the formation of 
the sporoblasts in the zygote, and this fact may have con- 
tributed in some measure to Moroff’s unfortunate mistake in 
thinking that these sporoblasts represented the macro- 
gametes and were separately fertilised. It is not clear 
which process some of his figures are intended to represent ; 
but his figs. 53 and 55 (16, pl. vii) appear to me to be stages in 
the formation of microgametes. The differences to be 
observed between them will be again referred to (p. 168) 
after the description of the formation of the sporocytes. 
Siedlecki described the nuclear division of the male as 
taking place by a process of chromidial fragmentation (24, 
p- 815), such as has been described in other Coccidia. This 
error may, I think, be partly ascribed to the fact that he 
worked chiefly with films, which he stated (24, p. 801) that 


SPUROGONY AND SYSTEMATIC POSITION OF AGGREGATID®. 165 


he found to be better than sections. I, on the contrary, have 
found that, owing to the large size and great density, espect- 
ally of the male gamete, it is practically impossible to make 
out the very rapid nuclear changes which take place at the 
beginning of the process without studying an extensive series 
of sections. 

The chromatin of the developing microgametes becomes 
more and more concentrated until the nucleus is reduced to a 
small, densely staining mass. This, surrounded by a certain 
amount of cytoplasm, projects on the surface of the micro- 
gametocyte and becomes elongated, carrying the centrosome 
at its distal extremity. The microgamete now becomes more 
and more elongated, the nucleus lengthens out, and the stalk 
connecting it with the residual sphere of cytoplasm becomes 
drawn out to form the tail region. Two flagella make their 
appearance at the free end, and the microgamete gradually 
assumes its fully developed form. 

The mature microgametes during life measure about 40 
in length, which is very much the same as the fixed ones 
(fig. 11). These lengths do not, of course, include the 
flagella, which are generally considerably more than half the 
length of the body of the microgamete. ‘he latter is strap- 
shaped, with a more or less cork-screw nucleus running 
through the anterior half. This, at any rate, appears to be 
the form that the nucleus takes during the phase at which it 
emerges from the microgametocyst. 

It may be of interest to recall that a spiral nucleus has also 
been recorded in Trypanosoma brucei (Robertson 18, p. 
236) and in the so-called male forms of ‘I’. lewisi described 
and tigured by Prowazek (17, figs. 35 and 36). In some 
cases I have observed a faint wavy line extending along 
the tail, but, except for this, there is no sign of an un- 
dulating membrane. In front of the nucleus and at the 
base of the flagella are situated the granules, which I take to 
be blepharoplasts. In deeply stained specimens they tfre- 
quently appear as one large granule. These basal granules 
are apparently derived from the original centrosome of the 


166 HELEN IL. M. PIXELL-GOODRICH. 


younger stage. During this time the limiting membrane 
surrounding the male gametocyte becomes considerably 
thicker. The male individual therefore becomes enclosed 
in a distinct cyst, which is, however, thinner than that in the 
female. At an early stage the gametocyte shrinks away from 
the cyst-wall, becoming separated from it by fluid in which 
it is suspended, while the microgametes develop on its surface 
(fig. 13). The cyst is sufficiently resistant to be dissected out 
of the host tissue, and the microgametes, after the develop- 
ment of their flagella, have been observed actively swimming 
about inside. Upon the rupture of this cyst they swarm out, 
but I did not succeed in making them retain their activity 
for any length of time, although many different media were 
employed for mounting them, including Sepia, blood and 
intestinal fluid, as well as ordinary sea-water and saline of 
varying densities. On this account it was not easy to make 
out the details of their movement, but flagella could be seen 
being vigorously lashed in front, and, in addition, the flattened, 
somewhat expanded tail was used in propulsion. 


FERTILISATION. 

In the fresh state numerous microgametes have been 
observed crowded round a macrogamete, but only one 
appears to enter. ‘This it does by the cone of attraction, and 
on reaching and penetrating the female nucleus its chromatin 
rounds itself off into granules, as shown in figs. 2 and 3. 
Moroff’s figures (16, pl. x1, fig. 91, and text-figs. M, aud Mg) 
doubtless represent this stage, although he failed to dis- 
cover the true meaning, as he was convinced that fertilisation 
was of the Gregarine type. The macrogamete nucleus has 
assumed an oval or pear-shaped form by the time the micro- 
gamete enters. A network of spindle threads appears 
running irregularly along the long axis of the nucleus from 
the point of entrance of the microgamete to the opposite 
pole. Meanwhile the karyosome undergoes changes. ‘The 
chromatic substance emerges from the micropyle as a stream 
of spherules which arrange themselves along the spindle 


SPOROGONY AND SYSTEMATIC POSITION OF AGGREGATIDA. 167 


threads. In some cases strings of granules can be seen 
inside the karyosome and coming from it; possibly the 
spindle threads themselves arise in this way. Somewhat 
similar irregular spindles have been described in other 
Coccidia; for example, m Coccidium proprium by 
Siedlecki (25) and in Orcheobius herpobdellez by Kunze 
(6), and no doubt the nuclear fibres figured by Moroff (16, 
pl. 1, fig. 16, and text-fig. N. 2) are of the same nature. 

The granules of male chromatin which have become 
lodged at the superficial end of the nucleus stream inwards, 
and presumably mingle with the granules of female cliromatin 
on the spindle threads: fertilisation is thus accomplished. 

Directly after the entrance of the microgamete a resistant 
cyst appears round the zygote, which thereupon contracts 
and comes to lie freely suspended in fluid. A similar con- 
traction has been described as taking place in Coccidium 
proprium (25), but in this species the oocyst is formed 
quite early, and a micropyle is left for the entrance of the 
microgamete. 


SPOROGONY. 


After fertilisation, upon the break-up of the spindle, the 
nucleus of the zygote becomes reconstituted, and this 
synkaryon undergoes mitosis. A centrosome and aster make 
their appearance on the outer side of the nucleus: they 
divide, and their daughter-centrosomes and asters separate, 
accompanied by lobes of the synkaryon, into which passes 
the chromatin, arranged on spindle fibres in the form of 
irregular loops (fig. 10). The first few nuclear divisions 
follow one another in such quick succession that, before the 
chromatin threads of one spindle have completely separated, 
the centrosomes have again divided, and the next pair of 
spindles are being formed approximately at right angles to 
the preceding one, as shown in fig. 10, where the large asters 
so generally to be seen are also figured. In this way an 
extensively branching nucleus is formed, the central region 
only disappearing at a later stage. 


168 HELEN L. M. PIXELL-GOODRICH. 


This peculiar form of mitosis in the fertilised female may 
be considered as intermediate between ordinary mitosis and 
the polymitosis described above in the male. In the latter 
the nucleus lags so far behind the centrosomes in its division, 
and so disturbs the sequence of events, that the connection 
between the division of the two organs is obscured or lost. 
Moroff (16) has already given numerous beautiful figures of 
these stages. 

After repeated division of the centrosomes and asters large 
daughter-nuclei with numerous small faintly staining granules 
are constructed off near the surface (fig. 12). The stage with 
about twenty to fifty of these is a commonly occurring one, 
and would seem to indicate a somewhat prolonged resting 
period. It can generally only be distinguished from the 
corresponding resting stage in the formation of the micro- 
gametes by the larger size and lesser affinity for chromatin 
stains exhibited by its nuclei. The large early sporoblast 
nuclei contain, as a rule, small evenly distributed granules of 
chromatin only, whereas the small early male nuclei have 
large deeply staining masses, or threads, of chromatin. 

This resemblance between the early stages of division in 
the microgametocyte and those in the zygote is of consider- 
able interest, since if seems to support the view that the two 
processes correspond to one another, and that not only is the 
macrogamete of the female homologous with the micro- 
gametocyte of the male, but also that the microgamete of the 
male corresponds to the sporoblast of the female. 

Further, whereas in Gregarines the adult female gives rise 
to small cells (macrogametes) which become separately 
fertilised and then form the zygotes (sporoblasts), in the 
Coccidia the adult female gives rise to the small cells (sporo- 
blasts) after fertilisation. Thus the chief difference between 
the processes of sporogony in the two groups is that fertilisa- 
‘tion takes place at an earlier stage in Coccidia than in 
Gregarines. ‘This could be shown in tabular form as follows, 
the arrow representing fertilisation : 


SPOROGONY AND SYSTEMATIC POSITION OF AGGREGATIDAS. 169 


Coccidia. 
2 trophozoite — macrogamete x n sporoblasts — nsporocysts * mn 


sporozoites. aaa 


—=— 
¢ trophozoite — microgametocyte X n microgametes. 


Gregarines. 
2 trophozoite — macrogaietocyte X n macrogametes — n sporocysts 
x mn sporozoites. 
¢ trophozoite — microgametocyte x n microgametes. 


Cuénot, in 1901 (2, p. 632), remarked that fertilisation was 
tardy in Gregarines and precocious in Coccidia, but refused 
to commit himself to a homology of the various steps. The 
evidence I have given above is, however, in agreement with 
the theory put forward by Minchin (12, pp. 271-274) in 1905 
as to the homologies that exist between the gametes of 
Coceidia and those of Gregarines, namely, that the sporo- 
blasts of Coccidia (formed by the division of the zygote 
after fertilisation) “may be compared to those of Gregarines 
by supposing that the process of multiplication, by which the 
gametocyte of the Coccidia gave rise primitively to a number 
of female gametes, has not been completely suppressed, but 
merely deferred until after the process of zygosis” (12, 
p. 272). The same theory is again referred to in 1912 (18, 
p- 304). 

During the resting stage in sporoblast formation, as in the 
same stage of the male, very distinct centrosomes and asters 
are generally apparent, being situated on pointed processes 
projecting from the surface. The next stage is marked by 
the repeated division of these centrosomes and asters, followed 
by the division of the nuclei. During this time the whole 
body of the zygote becomes more or less folded, thus 
increasing the area of the surface on which the nuclei are 
situated, and giving room for their increasing number. ‘This 
process corresponds to the breaking up of the microgameto- 
cyte into separate spheres at the stage when its nuclei are 
rapidly increasing in number. 


170 HELEN L. M. PIXELL-GOODRICH. 


One very usual method of folding in the dividing zygote is 
for one pole to become invaginated, as in a gastrula; this 
was noticed by Schneider (22, pl. ix, fig. 24). It is clear 
that when this cup-like body is cut in one direction an 
arc of cytoplasm surrounded by nuclei will result such as 
described by Moroff as characteristic of his A. arcuata 
(16, pp. 118 and 146), just as, when cut through a plane at 
right angles to the former direction, a ring may result. 

The effect of this folding and increase of nuclei is to 
diminish the relative amount of cytoplasm compared to nuclei, 
so that when the sporoblasts round off there is never any great 
quantity of residuum, and in most cases none at all. At this 
stage, therefore, the oocyst is generally crowded with sporo- 
blasts. In these the centrosome and aster persist for some 
time, forming a projection on the surface (fig. 4), but I have not 
been able to trace their division preceding nuclear division to 
form the sporozoites. Before this happens the whole pro- 
jection sinks down and the nucleus assumes a resting form 
(fig. 5). On resuming its activity the nucleus divides into 
two, but though a spindle is formed, no asters are visible. 
During this first division the sporocyst makes its appearance 
(fig. 6). In the vast majority of cases only one of these 
secondary nuclei divides again (fig. 7), and in this way are 
formed the nuclei of the three sporozoites (figs. 8 and 9) 
which are normally present. Only occasionally do both of 
the secondary nuclei divide again, forming four sporozoites. 
Kach sporozoite has at its posterior end an elongated deeply 
staining nucleus in which no karyosome can be distinguished 
(fig. 9). 

Sometimes streams of sporocysts are to be observed pass- 
ing along-blood spaces, although, in the Sepia, no tendency 
can be seen on their part to open in order that auto-infection 
may take place. The whole oocyst is generally said to pass 
into the lumen of the Cephalopod’s alimentary canal whence 
it reaches the exterior. So far no definite evidence of the 
escape of unbroken oocysts has been obtained, and many of 
them are very deeply embeddea in the thick connective tissue 


SPOROGONY AND SYSIYEMATIC POSITION OF AGGREGATIDA. 17] 


of the wall far from the lining epithelium. Possibly it is by 
eating the cysts in dead Cephalopods that crabs generally 
become infected. About the process of schizogony in the 
crab nothing further need be said, for it has been beautifully 
described by Leger and Duboseq (10). 


SYSTEMATIC PosIrion. 


In having this alternation of hosts corresponding to its 
alternation of generations, the genus Aggregata differs from 
all other Coccidia whose life-histories are at present known, 
and on this account would constitute a distinct family, the 
Aggregatide making a third family of Section A of the 
Eucoccidia in the classifications drawn up by Minchin 
(18, p. 852). The family could not easily find a place in the 
classification given by Leger (10a, p. 86), except that it 
would belong to the Himeridea, which corresponds to Minchin’s 
Section A. The subdivisions of Leger’s two sections are 
based on the number of sporozoites in the oocyst, which 
Minchin (18, p. 353) has suggested is not a natural method of 
classification. This suggestion is supported by the fact that 
in the Aggregatidz, now to be included in Coccidia, there is 
no constancy in the number of sporozoites, A. eberthi having 
only three or four in each of its numerous sporocysts, while 
A. octopiana, according to Moroff (16), has sixteen, and 
other species have from eight to twenty-four. 


SUMMARY. 


(1) The genus Aggregata is undoubtedly to be included 
among the Coccidia—the fertilisation being not Gregarine in 
character as described by Moroff (15 and 16), but typically 
Coccidian. In Aggregata eberthi, here described, from 
Sepia officinalis, the large female gamete is fertilised by 
the small, active, biflagellate microgamete, and the zygote so 
formed gives rise to a large number of sporoblasts. 


172 HELEN L. M. PIXELL-GOODRICH. 


(2) The polymitotie nuclear divisions giving rise to the 
microgametes are so similar to those giving rise to the sporo- 
blasts that they afford some evidence in favour of the view 
that these stages are homologous. It is remarkable also that 
the microgametes further resemble the sporoblasts in being 
enclosed in a distinct cyst. 


REFERENCES, 
1. Comes, S.—* Untersuchungen iiber den Chromidialapparat der 
Gregarinen,” ‘Archiv. Protistenkunde,’ x, 1907, pp. 416-4389. 
2. Cuénot, L.— Recherches sur l’évolution et la conjugaison des 


Gregarines,” ‘Archiv. de Biol.,’ xvii, 1901, pp. 581-652. 

3. Eberth, C. J.— Uber die Psorospermienschliuche der Cephalo- 
poden,” ‘ Zeitschr. wiss. Zool.,’ Bd. xi, 1862, pp. 397-399. 

4. Frenzel, J.—‘‘ Ubereinige in seetieren lebende Gregarinen,” ‘Arch. 
Mikr. Anat.,’ Bd. xxiv, 1885, pp. 546-588. 

5. Jacquement, M.—* Sur la systématique des Coccidies des Cephalo- 
podes,” ‘Archiv. Protistenkunde,’ ii, 1903. 

6. Kunze, W.—‘Orcheobius herpobdelle,” ‘Archiv. Protisten- 
kunde,’ ix, 1907, p. 382. 

7. Labbé, Alphonse.—* Recherches Zoologiques, cytologiques et biolo- 
giques sur les Coccidies,’ ‘Arch. Zool. expér.’ (3), iv, 1896, pp. 
515-654. 

8. Léger, L.,and Duboscq, O.—*Sur l’évolution des Gregarines gymno- 
sporées des Crustacées,” ‘C. R. Acad. Sc. Paris,’ vol. exlii, 1906, 
pp. 1225-1226. 


9. “ L’évolution d’une Aggregata de la Seiche chez le Portunus 
depurator,” ‘C. R. Se. Soc. de Biol.,’ vol. lx, 1906, pp. 1001-1003. 
10. “T/évolution schizogonique de l’Aggregata (Eucoc- 
cidium) eberthi (Labbé),” ‘Archiv. Protistenkunde,’ xii, 1908, 

pp. 44-108. 
10a. Léger, L.—* Caryospora simplex . . . et la classification des Coc- 


cidies,” ‘Archiv. Protistenkunde,’ xxii, 1911, pp. 71-89. 

11. Liihe, M—‘“ Uber Geltung und Bedeutung der Gattungsnamen 
Eimeria und Coccidium,” ‘ Centralbl. f. Bakt. v. Parasit.,’ Abt. 
xxx1, 1902. 

12. Minchin, E. AA—‘The Sporozoa: a Treatise on Zoology ’ (Lankester). 
London: A. & C. Black, 1903. 

‘An Introduction to the Study of the Protozoa.’ London: 

Edward Arnold, 1912. 


13. 


SPOROGONY AND SYSTEMATIC POSITION OF AGGREGATID®. 173 


14, Monticelli, F. S.—* Fauna elmintologica Golfe Napoli,’ ‘ Mitt. 
Stat. Zool. Neapel.,’ viii, 1888, pp. 85-152. 

15. Moroff, Theodor.—* Sur l’évolution des prétendues Coccidies des 
Céphalopodes,” ‘C. R. Ac. Se. Paris,’ vol. exlii, 1906, pp. 652-654. 

“Die bei den Cephalopoden vorkommenden Aggregataarten 
als Grundlage einer kritischen Studie iiber die Physiologie des 
Zellkernes,’ ‘Archiv. Protistenkunde,’ xi, 1908, pp. 1-224. 

17. Prowazek, 8.—“ Studien iiber Saugetier trypanosomen, I,” ‘ Arb. 
Kais. Gesundheitsamte Berlin,’ xxii, 1905, pp. 351-395. 

18. Robertson, Muriel.‘ Notes on certain Blood-inhabiting Protozoa,” 
‘Proc. R. Phys. Soc. Edin.,’ xvi, 1906, pp. 232-247, 

19. Smith, Geoffrey.—“‘ Note on a Gregarine (Aggregata inachi n. 
sp.), ete.,” ‘ Mitt. Zool. Stat. Neapel,’ xvii, 1905, pp. 406-410. 


16. 


20. Schaudinn, F.—‘‘ Untersuchungen tiber den Generations-wechsel 
bei Cocciden,” ‘ Zool. Jahrb. abt. Anat.,’ Bd. xiii, 1900, pp. 197- 
292. 


21. Schneider, Aimé.—“ Note sur la Psorospermie oviforme du poulpe,” 
‘Arch. Zool. expér.,’ vol. iv, Note xiii, 1875, pp. xl-xlv. 


22. “Nouvelles observations sur la Sporulation du Klossia 
octopiana, ‘Arch. Zool. expér.,’ 1885, pp. 77-104. 

23. Siedlecki, M.—‘ Reproduction sexuée . . . de la Coccidie de la 
Seiche (Klossia octopiana),’ ‘C. R. Soe. Biol. Paris’ (10), v, 
1, 1898, pp. 540-543. 

24. “Etude cytologique . . . de la Coccidie de la Seiche,” ‘Ann. 
Inst. Pasteur,’ vol. xii, 1898, pp. 799-836. 

25. “ Reproduction sexuée .. . chez Coccidium proprium,” 


‘C. R. Soe. Biol. Paris’ (10), v, 1898, pp. 664-666. 


EXPLANATION OF PLATE 18, 


Ilustrating Mrs. Helen L. M. Pixell-Goodrich’s paper on 
“The Sporogony and Systematic Position of the Agere- 
gatide.” 


| Preparations stained iron-hematoxylin and drawn with a 2 mm. lens 
and compensating ocular 12, giving approximately a magnification of 
2000, unless otherwise stated. | 

Fig. 1.—Part of a section through a macrogamete ready for fertilisa- 
tion, with cone of attraction (co.) and radiating protoplasmic plug (pp.) 


174 HELEN Ll. M. PIXELL-GOODRICH. 


separating the nucleus (n.) from the surface. Stained Ehrlich’s hema- 
toxylin. x 800. 

Fig. 2.—Fertilisation. The microgamete (m.) has just entered, and 
its chromatin in the form of granules (ch.) is passing along the spindle 
threads (sp.). ca. Large karyosome also giving off chromatin granules. 
_ From a section stained iron-hematoxylin and orange G. 


Fig. 3—Complete section through a specimen soon after fertilisation 
showing nuclear spindle (sp.). x 1500. 

Fig. 4.—Sporoblast with centrosome and aster. 

Fig. 5.—Sporoblast during resting stage. 

Fig. 6—Sporocyst with dividing nucleus. Stained Ehrlich’s hema- 
toxylin. 

Fig. 7.—Sporocyst during its second nuclear division. Stained 
Ehrlich’s hematoxylin. 


Fig. 8.—Sporocyst with three nuclei. 
Fig. 9—Sporocyst containing three sporozoites with terminal nuclei. 


Fig. 10.—Portion of a section of a dividing zygote showing one 
nuclear division nearly complete and secondary spindles already forming 
with conspicuous centrosomes (c.) and asters. The loops of chromatin 
granules (ch.) are already beginning to arrange themselves on the secon- 
dary spindles. Preparation stained with iron-hematoxylin and Licht- 
griin and picric acid. x 3000. 

Fig. 11.—Microgamete with more or less corkscrew nucleus (n.) and 
a blepharoplast (b.) at base of each flagellum (f.). From a film stained 
with Delafield’s hematoxylin. x 1500. 


Fig. 12—Portion of a section of one lobe of a folded and dividing 
zygote. to show developing sporoblasts with large faintly staining 
nuclei (n.) each accompanied by an aster and centrosome (¢.). From 
same preparation as fig. 15. x 1000. 


Fig. 13.—Section through a microgametocyte showing cyst (cy.) con- 
taining three spheres surrounded by the nuclei (n.) of developing micro- 
gametes. The centrosomes (c.) and asters are situated on processes of 
the cytoplasm projecting from the surface. Stained Ehrlich’s hema- 
toxylin. x 1000. 


BLASTOCYST AND PLACENTA OF 'T'HE BEAVER. 75 


The Blastocyst and Placenta of the Beaver. 


By 


Arthur Willey, F.R.S., 
Strathcona Professor of Zoology in McGill University, Montreal. 


With Plates 14 to 21 and 6 Text-figures. 


ConreENTs. 
PAGE 
I. INTRODUCTION , : : ; 5 Alzps 
Il. Sex or THE Fetus : 1 . 5 ¢he: 
III. THE PREPLACENTAL BLASTOCYST  . ; 2 A180 
IV. Susstace A E F : : 2) 184: 
V. Suspstace B : : 5 . 188 
VI. SuBstTaGcE C : : : : . inh 
VII. Susstace D : : : = = 2.00 
VIII. Susstace E : : : 2 5 PIB: 
IX. Supstace F p : : F a 206 
X. MATERNAL TROPHOSPONGIA 3 : 5 Ales 
XI. DiscopLAcENTAL ADHESION } ; 218 
XII. Tor EstTaBLisHeD PLACENTA ‘ ; 5 PAS 
XIIl. INTERMEDIATE STAGES ; ; ; 29 
XIV. THE ConNECTING MEMBRANE : ; 5 BBY 
XV. SPECIAL CONSIDERATIONS . , : . 236 
XVI. BrsLioGRAPHY é ! , : » 247 
XVII. EXPLANATION OF THE PLATES : : ; 249 


I. Inrropuction. 


THE material upon which this contribution is based was 
collected by me during the seasons of 1911, 1912, and 1913. 
In the first season, at the Laurentides National Park in the 

VOL. 60, PART 2.—NEW SERIES. 13 


176 ARTHUR WILLEY. 


province of Quebec, the stage obtained between April 18th 
and April 27th was that of the submature naked foetus. In 
the second season, at the Aloonquin Park in the province of 
Ontario, from May Ist to May 14th, the mature fur-covered 
foetus was procured. In the third season, from February 21st 
to March 11th, in the Lake Edward district between Quebec 
and Lake St. John, the stages of the early blastocyst and the 
young placenta were acquired. The result of my first visit 
to the woods of eastern Canada was published in the second 
volume of Prof. Spengel’s ‘ Festschrift’ (1912). In that 
paper there will be found the description of a strong extra- 
placental or periplacental connecting membrane, the “um- 
bilico-uterine ligament,” by which the large foetal sac is 
suspended from the wall of the gestation sac. This was one 
of the points that stood in need of further elucidation from 
new material. 

In February, 1913, through the good offices of Profs. 
J.G. Adami and P. E. Nobbs, of this University, an oppor- 
tunity for making an effort to secure earlier stages in the 
development of the beaver was presented to me. ‘l’o Prof. 
Adami I am indebted for his kindness in apprising me of 
the fortunate occasion; while Prof. Nobbs, in the most 
generous manner, gave me the temporary freedom of a game 
reservation in which he was interested, together with the use 
of a hunting-lodge, compactly built of rough-hewn logs, 
situated on the shore of Lac Gagnon, near Pearl Lake, which 
is a few miles south of Lake Edward, on the line between 
the city of Quebec and Lake St. John. I spent eighteen 
days on this reservation, accompanied by a French Canadian 
gardien, skilled in woodcraft, named David Fournier, anda 
French Indian tracker named Joe Moreau. The severity of 
the season may be judged from the fact that the temperature 
ranged from —35° Fahr. (February 25th, at 6.380 a.m.) to 
+ 26° Fahr. (March 10th, at 1 p.m.). 

We succeeded in getting some very young gestation sacs, 
which were preserved in 10 per cent. formalin. In the 
summer I brought this material to Europe, and, being 


BLASTOCYST AND PLACENTA OF THE BEAVER. 177 
desirous of working upon it without undue delay, I wrote to 
Prof. Hubrecht, who promptly assured me of a welcome 
at Utrecht, so that I was able to enjoy the privilege of 
continuing the research in the Embryological Laboratory 
of the University of Utrecht for nie weeks, from May 
30th to July 31st inclusive. I am deeply grateful to Prof. 
Hubrecht for his friendship, for the use of his laboratory 
with all its excellent facilities, for access to the library 
and to his collection of reprints, and particularly for his 
criticism upon various questions raised by the preparations, 
which at the beginning presented puzzling and unique 

appearances. 

_ The material having been transferred from formalin to 
alcohol, the several gestation sacs were stained in toto 
with hemalaun, serial sections cut, mounted, and counter- 
stained by Mr. J. G. de Groot, Conservator in the Embryo- 
logical Laboratory at Utrecht. The preparations have been 
deposited in the collection of the Institut International 
d’Embryologie, which was established at Utrecht by 
Prof. Hubrecht in 1911. Most of the drawings illustrating 
this paper were made with the assistance of EHdinger’s 
“ Zeichenapparat.” Some of the more finished drawings 
have been executed by Mr. John Prijs, artist at the Institute, 
with the kind permission of Prof. Hubrecht. 

‘In addition to the two youngest broods, which were the 
principal fruit of my third journey, and the only material 
collected before I left the reservation, later stages were 
forthcoming by arrangement with the gardien, who 
forwarded two subsequent consignments to me after my 
return to Montreal. These included the stage of the half- 
grown placenta and that of the two-thirds-grown placenta. 
I made drawings of the more advanced stages in Montreal, 
but the study of the youngest stages was carried out, as 
mentioned, in Utrecht. 

Finally, I must not omit to acknowledge the assistance in 
procuring material from the Algonquin Park in 1912, with 
which I was favoured by Mr. G. W. Bartlett, the Super- 


178 ARTHUR WILLEY. 


intendent, and by the park rangers stationed at Joe Lake, 
Messrs. James Bartlett and Mark Robinson. 


IJ. Sex or tHe Fetus. 


Before entering upon the description of the early stages, 
reference may be made to the question of the sex of the 
advanced foetus. In 1911, of two beavers with young, one 
contained four foetuses in the right uterus, the other carried 
two in the right uterus. There was none in the left horn of 
the uterus in either case, and all six foetuses happened to be 
males. In 1912 I was informed that one gravid beaver had 
been caught, skinned, and buried several days before my 
arrival at Joe Lake. When we exhumed the carcase it was 
evident that the cool sandy loam had kept it perfectly fresh. 
It yielded five foetuses, the only foetal material obtained during 
the trip. There were three fcetuses in the right horn of the 
uterus, and these were all males. Of the two remaining 
foetuses in the left horn, the one near the oviducal end was a 
female, the other near the cervical end of the uterus was a 
male. The male sex of the three right foetuses thus happens 
to be consistent with the previous observations, but I am not 
disposed to lay undue stress upon the coincidence. Still it is 
well to put such data on record in the presumption that in 
future years some other observer may similarly rescue trapped 
beavers from their sandy sepulchre and determine the sex of 
the unborn young. Out of a total of eleven foetal beavers 
whose sex has been determined, ten were males, nine of 
which were lodged in the right uterus, which has not yielded 
a female according to present information. 

It seems doubtful whether the determination of foetal sex 
in a large number of cases would justify the time and expense 
which would have to be devoted to it. Such data can only 
be accumulated by recording casual observations of no 
immediate value in themselves, but likely to be useful for 
future comparisons. Perhaps it is because occasional 
determinations seem to be so trivial that there is such a 


BLASTOCYST AND PLACENTA OF THE BEAVER. 179 


paucity of references to the curious distribution of the sexes. 
The records would possibly have an additional interest if the 
age of the parents could be ascertained, but this is practically 
impossible in the case of the beaver. The isolated position 
of the beaver amongst rodents perhaps renders it worthy of 
special consideration in regard to this matter of the sex of 
the foetus which are lodged in the right and left horns of the 
uterus respectively. As for the apparent preponderance of 
males, it is known how misleading even impressive numbers 
may be in regard to the proportions of the sexes of various 
animals. Thus, in 1911, six females were captured and six 
foetal males extracted from two of them. In 1912 beavers 
were being taken alive in muffled traps and the sex was not 
always ascertained. My notes contain records of seven 
females and ten males, including male and female ‘ kittens,” 
two-year-olds and adults. These small numbers, so far as 
they go, would indicate an approximate equality of the 
SEXES. 

With regard to the remarkable multiple embryos of 
Armadillos, M. Fernandez (1909), investigating the case of 
the “Mulita” (Tatusia hybrida), where the embryos are 
in packets of about eight (varying in number from seven to 
twelve) contained within a common chorion and derived from 
a single ovum, noted that out of nine sets, three were all males, 
six all females; but as the material belonged to different 
regions and years it remained doubtful whether the excess of 
females had any significance. Newman and Patterson (1909, 
1910), working independently about the same time upon the 
nine-banded Armadillo (Tatusia novemcincta), where the 
embryos occur as quadruplets, found that out of thirty-eight 
embryonic vesicles exactly half were male and half female. 
“All embryos in a vesicle are of the same sex,” as was 
originally recorded for the ‘‘ Mulita” by H. von Jhering in 
1885, and only one set is produced at a time. 

Newman and Patterson (1910) state that it has been known 
to the natives of Paraguay and of Argentina for over a 
century that the “ Mulita” brings forth at a birth a litter of 


180 ARTHUR WILLEY. 


young all of one sex. The nine-banded Armadillo does the 
same; and for this species Newman and Patterson have found 
that ‘‘ there is no correlation between the sex of the embryos 
and the dextrality or sinistrality of the functional ovary. Out 
of twenty cases in which the right ovary contained the corpus 
luteum, the sex of the embryos was male in seven and female 
in thirteen; while out of thirteen cases in which the left ovary 
held the corpus luteum, the sex was male eight times and 
female five. Evidently, then, the position of the functional 
ovary has no determining influence on sex.” 

I will conclude this chapter with a couple of miscellaneous 
notes relating to the beaver. On May 2nd, 1912, a large 
pregnant female was caught alive by the toes in a muffled 
trap. On the following day she was taken in a sack to head- 
quarters at Algonquin Park, and on May 10th gave birth to 
two kittens about 9 a.m.; at 10.30 a.m. it was noticed that 
their eyes were open. ‘Towards the end of the day two more 
were born. The successful rearing of the kittens was ardently 
hoped for, but was frustrated by their death within two or three 
days. On May 14th a large male was trapped, measuring 
42 in. from tip to tip. It had lost both of its fore-limbs at 
two previous trappings. The scars of the amputated members 
had healed up completely. This serious maiming would 
prevent the beaver from doing its share of work at the 
lodge and dam, but would not destroy its swimming and 
feeding powers. It need scarcely be stated that the traps 
which are set for beaver are concealed as far as_ possible, 
and much care is devoted to placing them in the most likely 
position from the trapper’s point of view, so that it says 
nothing for or against the intelligence of the beaver that 
after two sad experiences the third proved fatal. 


III. Tae Preptacentat Briasrocysr. 


The most conspicuous feature of the preplacental blasto- 
cyst of the beaver is the presence of a deep keel along the 
length of its embryonic side. 


BLASTOCYST AND PLACENTA OF THE BEAVER. 181 


The Rodentia, as was first made known by the pioneer 
researches of Bischoff, show a great diversity in the ontogeny 
of the feetal sac, though the end-result is fairly uniform, and 
the discoplacenta is always mesometric in position. The 
divergent behaviour of the early blastocyst within the limits 
of the order involves different methods of amniogenesis. 
“Schizamnion ” is the term used by R. Bonnet for the amnion 
which arises by dehiscence in a solid embryonic knob. He 
contrasts it with the more familiar “ Faltenamnion,” which | 
will translate into pleuramnion. 

Whereas Mus and Cavia possess an inverted blastocyst 
and a schizamnion, Lepus and Sciurus havea plano-convex 
blastocyst and a pleuramnion ; and it may be said that Castor 
has an everted blastocyst and pleuramnion, in the sense that 
the embryonic area and adjacent parts project as a sharp- 
edged keel which floats in the mesometric cavum. ‘The 
shape of the blastocyst conforms to that of the cavity of the 
gestation sac or cavum uteri. Reference to Pl. 14, fig. 1, 
will make this point clear. Here we see the highly charac- 
teristic appearance of the walls and cavity of one of the 
uterine swellings containing a blastocyst in transverse section. 
There is a round wide portion occupying the antimesometric 
division and a narrow deep portion in the mesometric division 
of the gestation sac: the former is the omphaloid cavum, 
the latter the mesometric or placental cavum. 

Lodged securely in the omphaloid cavum is the balloon- 
shaped omphalon, and projecting freely into the placental 
cavum is the keel. The longitudinal diameter of the blasto- 
cyst is greater than its transverse diameter. Omphalon is a 
convenient term introduced by A. J. Resink (1904) in place of 
“Dottersack ” or ‘‘ yolk-sac,” as applied to the blastocyst of 
placental mammals. 

The keeled blastocyst is attached to the obplacental wall of 
the omphaloid cavum by means of trophoblastic implantations 
which substitute themselves in places for the uterine epithe- 
lium. There is no constant symmetry in the distribution of 
the areas of fixation, but a certain degree of regularity may 


182 ARTHUR WILLEY. 


be detected. Sometimes a single trophoblast cell will take 
the place of a uterine epithelial cell as shown in Pliage 
fig. 2, where the implantation takes the form ofa festoon-like 
arrangement. In other places the substitution is more exten- 
sive, as in P]. 14, fig. 3; and sometimes there is a broad are 
of adhesion involving the greater part of the dome of the 
blastocyst, but sparing the mouths of the glands, as is shown 
in P]. 14, fig. 4. 

The cells which constitute the obplacental trophoblast, 
surrounding the dome of the blastocyst, are megalokaryocytes? 
(large cells with large nuclei) of two intergrading kinds. 
_ This trophoblast, as we shall see hereafter, is engaged in the 
ingestion and absorption of three visible maternal products: 
(1) The albuminous fluid secreted by the uterine glands, 
which appears as a dense coagulum in the cavum uterl; 
(2) darkly staining granules, which are the lobed nuclei of 
leucocytes ; (3) red blood-corpuscles. 

The cells which ingest the leucocytes with their character- 
istic iobular nuclei may be named chromatophile megalo- 
karyocytes, or more simply, chromatophile trophoblast. This 
is usually quite free from the uterine wall, the coagulum inter- 
vening; and at these places the uterine epithelium remains 
intact or but slightly altered. ‘The normal megalokaryo- 
cytes which ingest maternal erythrocytes are those which 
attack the uterine epithelium and take its place upon the 
surface of the mucosa. Inthe placental or carinal hemisphere 
of the blastocyst, the megalokaryocytes pass quite gradually 
into the cubical ectoderm of the keel. 

As stated above, there is some regularity in the disposition 
of the different kinds of trophoblast, and this necessitates the 
employment of descriptive topographical terms. The semi- 
diagrammatic fig. 5 on Pl. 14 shows the completest differ- 
entiation of regions of the trophoblast and of the mucosa 
with which itis associated. At the summit of the dome 
there is an area of trophoblastic implantation marking out 
the coronal region. On either side of this we have a tract 

1 J. W. Jenkinson (1902). 


BLASTOCYST AND PLACENTA OF 'THE BEAVER. 183 


of free chromatophile trophoblast occupying the peri- 
coronal region of the blastocyst; between this and the 
uterine epithelium, here intact, there is a narrow peri- 
coronal cavum into which glands open. Beyond the well- 
marked pericoronal tract there follows a zone which can 
sometimes be differentiated into two portions: the part next 
to the pericoronal tract may be distinguished as the equatorial 
or periomphaloid tract, and the subequatorial festooned 
implantation may be:known, by scientific license, as the 
pericarinal tract. Between the latter and the keel there 
extends a nearly horizontal omphalopleuric membrane of very 
constant formation, the adcarinal membrane. Lastly 
follows the keel or carinal region, which includes the 
embryonic shield and extra-embryonic keel. 

The gestation sacs which contained preplacental blastocysts 
were six in number, and were extracted from one female on 
March 10th, 1913, three in each horn of the uterus. The 
swellings varied slightly in size, and the differences proved 
subsequently to be connected with the state of development. 
The average diameter of the swellings was 10 mm. The 
lowest swelling in the left horn measured 11 mm., the middle 
and upper about 10 mm. The lowest right swelling measured 
°10 mm., the middle right 10 mm., the upper right (i. e. nearest 
the Fallopian tube) 8 mm. 

On March 4th I had obtained two gestation sacs 25 mm. in 
diameter, and on March 10th four more of similar dimensions. 
These dates seem to indicate that the typical stage at this 
part of the season is that of the young established placenta, 
but that the reproduction of the beaver does not take place 
with absolute precision between fixed dates, although the 
pairing season is known to be relatively constant. ‘To the 
range of fluctuation that exists I owe the fortunate acquisi- 
tion of the preplacental blastocysts. No two of these blasto- 
cysts are exactly alike, the six individuals representing as 
many substages which exhibit the characteristics of three 
‘main phases in the growth of the tridermic blastocyst: (a) 
with solid mesoblast; (b) with exoccelom; (c) with incipient 


184 ARTHUR WILLEY. 


notochord. I will now proceed to the more detailed descrip- 
tion of the successive substages of the preplacental blasto- 
cyst. Three of them are well-defined, the other three are 
more transitional. 


IV. Supstace A. 


This stage is represented by a series of sections, cut from 
before backwards relatively to the embryo, transverse to the 
longitudinal axis of the uterus, numbered VIB in the Utrecht 
catalogue. It is a critical stage, inasmuch as it affords the 
key to the comprehension of subsequent stages. The primi- 
tive streak occupies nearly the entire axis of the blastodise, 
the mesoblast is solid throughout, ané@ there is no notochordal 
contact between endoderm and medullary plate. It is the 
smallest of the swellings, having the following sectional 


dimensions : 
Millimetres. 

Total height of gestation sac or uterine swell- 
ing, from root of mesometrium . : . 890 
Total diameter of gestation sac. ‘ 2) pod 

Total height of cavum uteri or cavity of 
gestation sac. ; : ; P » 4950 
Depthof mesometric groove orplacentalcavum 3°00 
Diameter of ditto. : : ‘ 0:25 to 0°50 
ss omphaloid cavum : : . 100 


Thickness of placental mucosa, between root 
of mesometrium and bottom of placental 
erooye : : : » » B08 
Thickness of plecetital trophospongia or 
maternal proliferation : 1-25 
Thickness of periplacental wall or feral wall 
of cavum uteri at each side of the pla- 
cental groove . : : . he 
Thickness of obplacental w af of adeen sac 1:50 
The thickness of the placental mucosa is given in the above 
table in order to exhibit its contrast with that of the obpla- 
cental wall; it includes the layer of circular muscles which 


BLASTOCYST AND PLACENTA OF 'THE BEAVER. 185 


passes across the placental wall of the gestation sac between 
the root of the mesometrium (in which the main trunks of the 
uterine vessels occur) and the placental trophospongia. 
These transverse fibres of the circular muscles in the pla- 
cental region become interrupted and oblique to a varying 
extent in later stages. The diameter of the omphaloid 
cavum is identical with that of the omphaloid portion of the 
blastocyst, or, in one word, the omphalon, because the 
trophoblastic wall of the omphalon is implanted upon the 
inner surface of the uterine mucosa, which becomes denuded 
of its epithelium at the points and areas of implantation. 

The anterior pole of the blastocyst is free, and the 
embryonic area begins immediately behind the pole. The 
area itself is keel-shaped, since the embryonic shield of 
formative epiblast bends round the free edge of the keel. 
The formative epiblast which is thus exposed at the surface 
of the blastocyst is distinguished from the peripheral tropho- 
blast by its nuclear constitution, by the fact that the nuclei 
occur at different levels, and also by the presence of scattered 
vesicular cells, i.e. embryonic cells in a vesicular phase. 
The transition from formative epiblast to peripheral tropho- 
blast is singularly abrupt, there being no gradual transition 
from one to the other at the level where they meet. 

The hypoblast dips into the keel, and, at a very short 
distance from the anterior end, we find in section the solid 
U-shaped tip of the mesoblast embracing the floor of the 
hypoblastic groove, with a few ceils migrating out from it 
along the hypoblast (Pl. 14, fig. 6). The free end of the 
mesoblast has but slight longitudinal extension, since the 
primitive streak begins immediately behind it, so that we 
find in section a perfect U-shaped, solid mesoblast, united 
to the epiblast by the primitive streak. The position of the 
primitive streak marks the strictly embryonic side of the 
keel. The peripheral limits of the formative epiblast may be 
seen clearly under Zeiss oc. 2, obj. A; the edge of the shield 
is placed at a much higher level on the embryonic side of 
the keel than on the other side (Pl. 14, fig. 7). 


186 ARTHUR WILLEY. 


From this point backwards the primitive streak is con- 
tinuous, and presents various incidents of growth at different 
parts of its length. The wedge-like proliferation at Pla 
fig. 8, followed by the chink-like cleft in Pl. 14, fig. 9, hes 
near the anterior end. From the V-shaped mass which 
proceeds from the primitive streak and embraces the hypo- 
blastic groove a single line of mesoblast cells extends on 
each side along the hypoblast, about as far as the equatorial 
region of the blastocyst at this early stage. Later, the 
mesoblast has a much more restricted peripheral distribution. 
The juxtaposed walls of the epiblastic keel distad of the 
primitive streak become closely welded together. 

The lateral wings of the mesoblast are in close contact 
with the hypoblast, but are separated from the epiblast 
except at the primitive streak. This is especially obvious on 
the anti-embryonic side of the keel (Pl. 14, fig. i0). The 
embryonic shield becomes narrower as it is traced backwards, 
so that its peripheral limits are much lower than in front ; 
on the anti-embryonic side the edge of the formative epi- 
blast reaches the level of the welded walls of the keel, and 
finally the apex. A small pit appears intermittently at the 
surface of the primitive streak, but there is no continuous 
primitive groove (Pl. 14, fig. 11). The mesoblastic wings 
become unequal as we near the posterior end of the primitive 
streak, that on the embryonic side being longer than the part 
which bends round to the anti-embryonic side (Pl. 14, figs. 12 
and 13). In this situation it is easy to recognise that the 
auti-embryonic wall of the keel consists of cubical tropho- 
blast, the embryonic wall of formative epiblast. The keel is 
thus partly embryonic and partly extra-embryonic in the 
same transverse section. 

Posteriorly the primitive streak contact widens and the 
distal border of the keel consists of cubical trophoblast only 
(Pl. 14, fig. 14). In this region we see very clearly a broad 
primitive streak with mitoses, and no other formative epiblast. 
One limb of the mesoblast becomes very slender, the other 
limb dwindles to zero, whilst the bulk becomes massed about 


BLASTOCYST AND PLACENTA OF THE BEAVER. 187 


the primitive streak, partly blocking up the cavity of the 
trophoblastic keel. Shortly behind this region the primitive 
streak ceases and the mesoblast rapidly diminishes until only 
a single line of cells remains on one side of the hypoblast as 
mel. 14, fir. 15. 

Finally, the mesoblast terminates and the keel becomes 
didermic. At the same time the two layers are badly folded 
in the sections owing to a collapse which apparently resulted 
from rupture of the adcarinal membrane. ‘This fact, added 
to the great local differences in the character of the epithelial 
walls of the blastocyst, seems to point to the existence of a 
considerable amount of elasticity in the inner and outer mem- 
branes. In this case the rupture seems to have been due to 
an extraordinary agglutination of the adcarinal trophoblast on 
one side to the edge of the placental groove. This adhesion 
was retained whilst the rest of the blastocyst followed the 
contraction of the coagulum contained within the omphalon ; 
except for this unusual attachment, the blastocyst would be 
free at this level. A little farther back the hypoblast with- 
draws, leaving a trophoblastic keel with its walls welded 
together distally. ‘hus we have a trophoblastic keel near 
the posterior pole of this blastocyst, but not near the anterior 
pole where the keel is embryonic. The keel formation 
gradually dwindles away as the posterior pole is reached. 

I will postpone the history of the uterine mucosa until the 
other substages have been dealt with, but a few words are 
necessary concerning the fixation of the blastocyst. ‘The 
antericr pole which lies in front of the primitive streak pro- 
jects freely into the cavum uteri; the posterior pole is fixed 
with a broad obplacental implantation. Between the free 
anterior pole and the middle of the blastocyst there is a region 
corresponding approximately with the middle of the primitive 
streak where obplacental arcades become established, i.e. 
intermittent areas of trophoblastic implantation with destruc- 
tion of uterine epithelium at the points of contact. About 
the level of the posterior end of the primitive streak, which 
is about the middle of the blastocyst, the obplacental contact 


188 ARTHUR WILLEY. 


is given up except at the two pericoronal corners, where a 
very narrow implantation is retained as in PI. 14, fig. 18. 
Between these two angles there extends the free coronal 
trophoblast, whose cells contain chromatic granules identical 
with those which occur in the intervening coronal cavum 
and in the mucosa. I have determined them to be the nuclei 
of leucocytes. The leucocytes migrate across the uterine 
epithelium into the cavum uteri and are ingested by the 
chromatophile megalokaryocytes. 

Towards the posterior pole of the blastocyst the active 
phagocytic attack is resumed, the megalokaryocytes broadly 
replacing the uterine epithelium, with formation of arcades, 
skipping the crypt-like mouths of the obplacental glands. In 
the coronal region there is no visible sign of hypoblast which 
appears to stop short about the equatorial zone of the blasto- 
cyst. The hypoblast comes to anend posteriorly coincidently 
with the termination of the trophoblastic keel, so that the 
extreme posterior pole of the blastocyst is monodermic, its 
wall consisting entirely of flattened cells with distant nuclei, 
appearing subfusiform in section. 

At no stage is the hyploblast conspicuous in the coronal 
region, but sometimes straggling cells can be found. 


V. Susstace B. 


This is represented by a series of transverse sections cut 
from before backwards and numbered VI D in the Utrecht 
Catalogue. It is characterised by the first indication of a 
linear exoccelom. 


Millimetres. 
Height of gestation sac p ; : s 0:00 
Diameter of gestation sac. . : ~~ ae 
Height of cavum uteri.  . : 4 . 608 


Depth of placental groove. ; : . 4:00 
Diameter of placental groove near bottom . 0°30 
z os »  aboutthecentre 0°90 


. omphaloid cavum . : . og 


BLASTOCYST AND PLACENTA OF THE BEAVER. 189 


Millimetres. 
Thickness of placental mucosa. ; sno OO) 
a placental trophospongia . about 1°75 
- periplacental wall . ‘ a) ~or00 
By obplacental wall ; ‘ Bae EL 


These measurements, especially as regards the diameter of 
the omphaloid cavum, show that this substage is intermediate 
between that described above and that which follows below. 
The blastocyst covers thirteen slides, with twenty-four sections, 
10 « thick, on each slide; hence the length of the preserved 
blastocyst is 3°12 mm., its transverse diameter being that of 
the omphaloid cavum, viz. two millimetres. In front of the 
blastocyst the omphaloid cavum contains a homogeneous 
coagulum which has shrunken away from the epithelium upon 
which it was moulded; in it are scattered cellular débris 
and chromatic (leucocytic) granules. 

The anterior pole of the blastocyst is free ; the trophoblast 
here consists exclusively of megalokaryocytes, and a few 
hypoblast cells appear against its mesometric wall, not forming 
-a closed hypoblastic sac. The megalokaryocytes exhibit 
coarsely alveolar cytoplasm and contain a few chromatic 
granules. The blastocyst appears in section quite free in the 
centre of the uterine coagulum ; its own internal coagulum is 
reticulated and paler than the surrounding coagulum. Near 
to the anterior pole the blastocyst is pyriform in section, 
the narrow portion only being didermic (Pl. 15, fig. 16). 
Farther back, the blastocyst still appearing free in section, 
a long keel arises, somewhat folded in the preparation, com- 
posed of cubical trophoblast. The hypoblast is now seen 
to pass uninterruptedly over the cavity of the keel (PI. 15, 
fig. 17). About this same region, where the free anterior end 
of the blastocyst possesses the long trophoblastic keel, as 
many as nine glandular crypts were counted, opening into 
the omphaloid cavum. 

The plane of the blastocyst represented in fig. 17 shows 
well the two kinds of trophoblast composing the outer wall, 
namely, the omphaloid trophoblast with its megalo- 


190 ARTHUR WILLEY. 


karyocytes, and the carinal trophoblast with its cubical 
cells. Shortly behind this plane the omphaloid trophoblast 
begins to attack the uterine epithelium at intervals, Here 
and there the direct passage of leucocytes with their lobed 
chromatic nuclei may be exceptionally observed to take place 
from the mucosa into the coronal trophoblast. As we proceed 
backwards we come upon conspicuous obplacental attacks 
by megalokaryocytes between the mouths of the glands, and 
a broad pericarinal attack on one side only. The pericarinal 
zone is doubled by hypoblast which does not cover the obpla- 
cental half of the omphalon in this plane. The pericarinal 
adhesion is characterised by the formation of festoons; at the 
periomphaloid zone, just beyond the pericarinal festoons, 
there has taken place an ingestion of erythrocytes by the 
trophoblast. This also occurs in the coronal region, where 
there arises a broad adhesion bounded on each side by a free 
belt of pericoronal chromatophile trophoblast with the megalo- 
karyocytes full of granules. ‘The attacking trophoblast 
abuts, in a manner which is as clear as it is surprising, upon 
the columnar epithelium of the glandular orifice, two or three’ 
trophoblast cells bridging over the gaps as shown in PI. 14, 
fig. 4. 

Upon approaching the embryonic region we find, in this 
series, a confused mass of hypoblast and mesoblast collapsed 
at the foot of the keel, almost defying interpretation, until 
we come to the region of the primitive streak. There we see 
an extension of the process which commenced in the preceding 
stage, namely, the filling up of the keel by a mass of meso- 
blast derived from the primitive streak. Into this massive 
carinal mesoblast the hypoblastic groove extends as shown 
in Pl. 15, fig. 18. In this figure it may be noted that the 
formative epiblast which lies above the primitive groove is 
thicker than that which bends round the edge of the keel. 
This illustrates the stretching of the formative epiblast, which 
reduces its thickness and causes it to approximate to the 
condition of the peripheral trophoblast. 

Above the massive mesoblast on the anti-embryonic side of 


BLASTOCYST AND PLACENTA OF THE BEAVER. 191 


the keel (i.e. the side opposite to the primitive streak) there 
is an indication of the linear splitting which precedes the 
formation of the exocclom. On the embryonic side the 
mesoblast band is thicker, its nuclei lie at about three levels, 
and there is no distinct sign of the future splitting. The 
massive carinal mesoblast persists through some fifty sections 
of 10 4. At last it comes to an end, and leaves the elongated 
apical region of the keel empty. The two mesoblastic bands 
then bend continuously into one another round the bottom of 
the hypoblastic groove (FI. 15, fig. 19). They are generally 
two cell-layers in thickness with potential ccelom between 
the layers. More posteriorly there is a little interrupted 
splitting, giving rise to discontinuous spaces of incipient 
exoccelom. 

Soon the mesoblast ceases and the hypoblast withdraws 
from the keel. The latter is now in section a very long and 
narrow trophoblastic process, composed of cubical epiblast, 
with its opposite walls touching each other in places, hanging 
freely into the placental groove. This posterior portion of 
the keel attains a depth of about 0°75 mm. Thus in this 
case there is a distinct trophoblastic keel-formation both in 
front and behind. In other words the two poles of the 
blastocyst are essentially alike; but there is this difference, 
that the obplacental adhesion is continued nearly to the 
extremity of the posterior pole. 


VI. Susstace C. 


This was cut from behind forwards, and numbered VIA 
in the Utrecht Catalogue. It was the first of the set of six 
early gestation sacs to be sectioned, and the preservation, 
staining, and mounting left little to be desired. But the 
series presented the apparent anomaly that whereas there 
was an exoccelom as well as a quantity of massive mesoblast, 
there was no obvious embryonic shield. Comparison with the 
other substages, the recognition of the primitive streak, and 
the excellent state of preservation showed that the apparent 

VoL. 60, PART 2.—NEW SERIES. 14, 


192 ARTHUR WILLEY. 


absence of the formative epiblast which constitutes the 
embryonic shield was due to the general distension of the 
blastocyst and the elongation of the keel, whereby the 
thickened epiblast was reduced to a cubical epithelium. _ 
Millimetres. 
Height of gestation sac (i.e. total height of 
section, the mesometrium having been 


trimmed off). : ; : : - 10:08 
Diameter of gestation sac. : : -. Sao 
Height of cavumuteri. : : 7 Spam 
Depth of placental groove . ; : . . 2°00 
Diameter of placental groove ; . 0°50 to 0°75 

is omphaloidcavum. - . 2380 
Thickness of placental mucosa. . about 3°00 
Bf placental trophospongia . . e550 
s periplacental wall . ; . sae 
os obplacental wall . : «ie 


The anterior didermic extremity of the blastocyst is free, 
the posterior didermic pole is fixed by an abundant ob- 
placental phagocytic contact with destruction of uterine 
epithelium. This occurs extensively also in the middle 
region of the blastocyst. Mutoses occur in the hypoblast and 
carinal trophoblast, but not in the obplacental trophoblast, 
where there are signs of amitosis. Here and there, embedded 
in the substance of the megalokaryocytes, aspherical mass of 
protoplasm is segregated from the surrounding cytoplasm 
and contains about a dozen small nuclei arranged in couples. 
In other cases the small nuclei appear to be budding from a 
meganucleus. Ido not know the precise significance of these 
nests of small"nuclei in the obplacental trophoblast, but they 
are not to be confused with the megalokaryocytes which are 
engaged in the ingestion of erythrocytes. Perhaps it is 
simply a means of adding, by amitosis, to the number of 
nuclei in the individual megalokaryocytes. 

There is an extraordinary luxuriance of maternal capillaries 
at the denuded obplacental wall of the gestation sac, opening 
upon the trophoblast. There is no extravasation of blood 


BLASTOCYST AND PLACENTA OF THE BEAVER. 193 


into the cavum uteri, but the capillaries actually abut upon 
the implanted megalokaryocytes and discharge their erythro- 
eytes directly into the trophoblast. Consequently the megalo- 
karyocytes are found commonly charged with half-dissolved 
blood-corpuscles. As mentioned there is no hemorrhage ; 
the extravasated erythrocytes are all contained in megalo- 
karyocytes. In the posterior region of this blastocyst the 
whole of the upper third of the dome is implanted upon the 
uterine wall. Nevertheless the epithelial crypts at the 
mouths of the obplacental glands are not attacked by the 
trophoblast; phagocytic adhesion only occurs between 
neighbouring crypts, so that the obplacental union is a system 
of arcades, arching freely over the crypts. These are not 
like the pericarinal festoons which are miniature arcades not 
related to crypts. 

Near the posterior pole the sections present a typical 
picture of a blastocyst with a deep trophoblastic keel, the 
hypoblast passing directly across the cavity without bending 
into it, precisely as already shown in PI. 15, fig. 17, with the 
difference that in the case now under consideration the 
pole of the blastocyst does not lie freely in the cavum 
uteri, but is attached by a general obplacental implanta- 
tion, skipping the mouths of the glands. Proceeding for- 
wards we soon find the hypoblast dipping deeply into the 
keel, the distal walls of which are agglutinated together. 
At this level the blastocyst is didermic, there being no 
mesoblast in sight (Pl. 15, fig. 20). 

As we approach the posterior V-shaped extremity of the 
mesoblast, we observe that the hypoblastic cells become 
loosened from their epithelial contiguity in the wall of the 
groove, but they do not escape into the cavity of the tropho- 
blastic keel. This loosening of the carinal hypoblast occurs 
variously in all the substages ; it would naturally be ascribed 
to defective preservation, but whether this would be correct 
or not, it certainly denotes a peculiar physiological modifica- 
tion of the hypoblast in this region (Pl. 15, fig. 21). Passing 
on, the mesoblast dehisces to form a spacious exoccelom, the 


194, ARTHUR WILLEY. 


main cavity of which is lodged in the keel, while the narrow 
bands terminate in a solid proliferation or sinus terminalis 
on each side of the keel (Pl. 15, fig. 22). 

About the plane of the posterior end of the mesoblast in 
the carinal region, the obplacental trophoblast changes its 
character in the coronal region. The megalokaryocytes are 
no longer implanted upon the mucosa, but present a free 
outer surface to the cavum uteri (PI. 15, figs. 23 and 24). 
The large, evenly disposed cells contain numerous chromatic 
granules which they have ingested ; similar granules occur in 
the uterine coagulum, and others are seen traversing the 
uterine epithelium. ‘The special function of these modified 
megalokaryocytes is to ingest leucocytes whose nuclei yield 
and are composed of the granules in question. They occur 
over a continuous shield-like area of the coronal region of the 
blastocyst, and spread out forwards as the two pericoronal 
bands already mentioned (cf. Pl. 14, fig. 5). In some places 
the chromatophilous cells assume a flattened form as in PI. 15, 
fig. 25. 

The posterior exoccelom now requires a little extra con- 
sideration. It persists through some thirty sections and then 
gives way to massive tissue in the apical region of the keel, 
with cubical hypoblast flanked by narrow exoccelom dipping 
into it. In previous substages we have seen the commence- 
ment of the massive tissue of the keel at the posterior end of 
the primitive streak. In that part of its course where it is 
occupied by massive mesoblast I propose to designate the 
keel-formation, which appears like a stalk in section, by the 
term exostyle. When necessary the posterior exoccelom 
may be referred to briefly as the post-stylar celom. It 
is not merely post-embryonic in position, because the exostyle 
itself is post-embryonic. 

As we follow the series forwards the exoccelom becomes 
greatly reduced, almost obsolete, and the massive keel actually 
adheres by its apex to one wall of the placental groove; this 
wall happens to be anti-embryonic, i.e. on the side away from 
the primitive streak. From the tip of the exostyle a cord of 


BLASTOCYST AND PLACENTA OF THE BEAVER. 195 


slime is continued towards the bottom of the placental groove. 
It is noteworthy that where the hypoblastic groove of the 
omphalon penetrates into the carinal mesoblast, the epithe- 
lium retains a cubical character; whereas between this situa- 
tion and the adcarinal membrane the hypoblast is loose. 
Another detail to be mentioned is that the uterine epithelium 
is not specially modified at the point of agglutination of the 
tip of the exostyle. It can hardly be doubted that the 
exostyle possesses a highly viscid surface, and the particular 
adhesion here described does not represent the position of 
the ultimate placental union (Pl. 16, fig. 26). Im this con- 
nection it may be called to mind that in his monograph 
entitled ‘ Affen Ostindiens’ (Wiesbaden, 1891, see p. 201), 
E. Selenka found to his cost that the chorionic ectoderm 
(trophoblast) of Semnopithecus possesses a high degree of 
viscidity. 

At the level of fig. 26 the chromatophagous layer of 
trophoblast still forms an uninterrupted calotte over the 
dome of the omphalon, and the corresponding portion of 
the uterine epithelium is intact. In the section from which 
the figure was drawn, two obplacental glands open into the 
coronal cavum. ‘lhe granules which have been ingested by 
the trophoblast are commonly surrounded by what looks like 
a food-vacuole ; this effect may be in part assignable to the 
bodies of the leucocytes (PI. 15, fig. 25). A broad peripheral 
festooned zone of implantation occupies the widest portion of 
the omphalon, and uterine fluid can circulate through the 
trophoblastic arcades. 

Twenty-six sections of 10 4 intervene between figs. 26 and 
27. The nuclei of the exostyle present the appearance of 
being arranged in vertical rows or sheets. Mitoses are to be 
observed in the midst of the tissue as well as at the surface. 
The exostyle is marked off from the rest of the blastocyst by 
a geniculate bend due to increased resistance of its massive 
walls. It is penetrated throughout the greater part of its 
length by a central lumen continuous with the omphalon ; the 
lining of this groove is not distinct from the rest of the stalk- 


196 ARTHUR WILLEY. 


tissue except at its proximal or omphalad end. At the distal 
end there is a distinct trophoblastic space forming the cavity 
of the foot-like extremity which adjoins the uterine wall 
below the opening of a gland, the neck of which js under- 
going occlusion by epithelial mitotic proliferation. Near the 
bottom of the placental groove the lumina of the glands cease 
before reaching the surface of the mucosa, the necks being 
blocked by a mixed tissue composed of its own degenerating 
elements and proliferating epithelial cells. The cubical 
trophoblast stops quite sharply at the top of the stalk in the 
figure. On the right of the figure a piece of the wall of the 
stalk is torn away and lies near the corresponding uterine 
wall without any jagged edge; nor is any jagged surface left 
upon the stalk where the detached piece belonged—a fact that 
accords with the sheet-like arrangement of the stalk cells 
mentioned above (Pl. 16, fig. 27). 

The central lumen of the stalk (or exostyle) becomes inter- 
rupted in places as we trace it forwards, the tissue appearing 
solid in its central third, with a narrow chink in its distal 
third (Text-fig. 1). In this figure we see the swollen peri- 
pheral edges of the mesoblast, forecasting the sinus term1- 
nalis, and on one side only, namely on the embryonic side, 
a clear narrow exoccelom not extending into the stalk. The 
cubical trophoblast of the adcarinal wall of the omphalon 
stops short at the proximal bend of the stalk. <A little farther 
on the stalk is solid throughout its central half, with proximal 
cavity continuous with the omphalon and a distal linear lumen 
terminated by a solid apex (PI. 16, fig. 28). 

The chromatophile trophoblast is now present at the sides 
of the dome, i.e. in the pericoronal region. In the centre of 
the dome there isa broad phagocytic contact with the denuded 
mucosa. The columnar uterine epithelium stands out with 
great prominence on either side of the denuded coronal area, 
while the broad equatorial zone of the omphalon shows very 
beautifully the pericarinal festoons (PI. 14, fig. 5). It is a 
constant surprise to see normal uterine epithelium contiguous 
with a belt of implanted megalokaryocytes. 


BLASTOCYST AND PLACENTA OF THE BEAVER. 197 


Between Pl. 16, figs. 28 and 29 there are fifty-four sections 
of 10 u. The central extension of the umbilical cavity has 
again become continuous throughout the greater part of the 


TEext-Fie. 1. 


Section through mid-region of keel showing nearly solid exostyle. 
1. Sinus terminalis. 2. Exoccelom (on embryonic side). 3. 
Unsplit mesoblast (on anti-embryonic side). 4. Exostyle. 
5. Trophoblastic apical cavity. 


length of the exostyle. That part of the primordial sinus 
terminalis which is on the anti-embryonic side of the keel 
is now at the level of the geniculate bend. As before, the 


198 ARTHUR WILLEY. 


cubical trophoblast ceases near the proximal end of the stalk. 
The nature of the cellular membrane which limits the surface 
of the exostyle is shown very clearly in fig. 29 near the distal 
end, where it is partly separated from the massive tissue. 
The solid mesoblast on the embryonic side of the central 
cavity of the exostyle is thicker than that on the other 
side. Below each section of the sinus terminalis a 


narrow exoccelom, very short on the anti-embryonic side, is 
visible. ; 

On the anti-mesometric side of the cavum uteri the uterine 
epithelium at this level is complete except for a narrow area 
in the centre of the coronal region; the variously depressed 
chromatophile megalokaryocytes are stretched over the dome 
of the omphalon as far as the pericarinal zone where phago- 
cytic contact is maintained. 

The small coelomic cleft noted on the anti-embryonic side 
below the sinus terminalis in fig. 29 is a local dehiscence 
in the mesoblast at that point. The mesoblast on this side 
becomes thinner, about two layers deep, as compared with six 
sheets of nuclei on the embryonic side. Great pieces become 
torn out of the embryonic side of the stalk, apparently owing 
to their adhesion to that side of the placental groove before 
preservation ; the stalk as a whole is often much lacerated. 
The uterine muscles had presumably relaxed and so widened 
the groove after death. ‘he blastocyst becomes quite free 
from the mucosa except at the two pericarinal angles. After 
about two dozen sections forwards from fig. 29, the exoccelom 
on the embryonic side begins to extend well into the stalk, 
none being present on the anti-embryonic side. 

With the extension of the exoccelom into the exostyle on 
the embryonic side of the central hypoblastic cavity, we 
begin to recognise the position of the primitive streak at the 
distal end of the linear exoccelom; and it continues through 
some twenty sections. Close to the plane of the posterior 
end of the primitive streak a groove appears at the proximal 
face of the keel on the embryonic side. This groove persists 
through about ninety sections, and marks the peripheral 


BLASTOCYST AND PLACENTA OF THE BEAVER. 199 


limit of the embryonic shield. At the same level a coelomic 
space appears in the apical region of the keel beyond the 
distal end of the hypoblast. ‘This is the pre-stylar exoccelom 
(Pl. 16, fig. 30). 

We have now followed the massive exostyle throughout its 
entire extent between the post-stylar and the pre-stylar 
exoccelomic spaces. We have seen that it is never quite solid, 
there being always a trace of a linear hypoblastic axis. The 
position of the primitive streak marks the posterior boundary 
of the embryonic shield, which at first appeared strangely 
inconspicuous in this otherwise excellent series. The entire 
blastocyst is so well extended by the pressure of its internal 
fluid that the formative epiblast has become stretched, and 
so has suffered a diminution of thickness. This explana- 
tion is supported by the disposition of the nuclei, which are 
much more crowded on the embryonic than on the anti- 
embryonic side of the keel. The embryonic ectoderm con- 
sists of narrow cubical cells, the nuclei thus occurring in 
close juxtaposition; on the anti-embryonic side the tropho- 
blastic ectoderm consists of flattened cells with distant nuclei 
pel 16, fig. 30). 

The anti-embryonic section of the primordial sinus ter- 
minalis passes definitely into the keel, becoming elongated 
in section (PI. 16, fig. 31). Farther forwards it rounds the 
distal border of the carinal hypoblast, and finally meets its 
companion, which has meanwhile descended into the keel. 
In proportion as the two sections of the sinus terminalis 
approach their confluence, the coelom dwindles to extinction 
(Pl. 16, fig. 32). After this the mesoblast ceases. 

In front of the primitive streak it becomes difficult to 
determine with precision the distal limit of the embryonic 
shield, and its anterior border cannot be defined. In front 
of the mesoblast there are signs of ectodermal proliferation, 
especially in the vicinity of the proximal groove, which 
becomes deeper anteriorly. ‘This is probably a proliferation 
of the amniogenic trophoblast. The growth of the obplacental 
trophoblast is accompanied by amitotic nuclear division, that 


200 ARTHUR WILLEY. 


of the carinal trophoblast is effected by nuclear mitosis 
(Pl. 16, fig. 38, and Pl. 17, fig. 34). 

In this anterior region the keel is didermic. Below the 
hypoblast there is a long narrow monodermic extension of 
the keel. The abrupt transition from cubical carinal hypo- 
blast to flattened adcarinal hypoblast, shown on the anti- 
embryonic side in figs. 32 and 34, is a characteristic feature. 
It should also be mentioned, though not shown in the figures, 
that the flattened hypoblast of the adcarinal omphalopleure 
is separated from the adjacent trophoblast by a well-defined 
basement membrane, which stains strongly with orange G. 
In front of the region represented by these figures the 
proximal groove of the keel soon flattens out; the blasto- 
cyst appears free in the cavum uteri, and the anterior pole 
is reached. 


VII. Susstace D. 


This substage is rather poorly represented by a series of 
transverse sections cut from behind forwards and numbered 
Vie inthe Utrecht catalogue. It may be considered to be 
characterised by the initiation of the lateral amniotic folds. 


Millimetvres. 

Height of gestation sac ; : : . LEG 

Diameter ae A : : 3 . S08 

Height of cavumuteri : , . oe 

Depth of placental groove . 3 : > 2a 

Diameter of placental groove in middle . O@s 
a re » near omphaloid 

end . ; : : 5 : oOo 

Diameter of praanateid cavum. : ) ee 

Thickness of placental mucosa. : . 4°00 

- oe trophospongia . . | Bee 

cs periplacental wall . : . 4°00 

i obplacental wall. : 1:00 


Both poles of the blastocyst are fixed to ae wall of the 
omphaloid cavum. ‘Towards the posterior pole the adcarinal 


BLASTOCYST AND PLACENTA OF THE BEAVER. 201 


omphalopleure was greatly folded, while the folding of the 
carinal trophoblast indicates the existence of a long keel like 
that shown in PI. 15, fig. 17. As in the preceding substage, 
we first meet the posterior U-shaped solid end of the meso- 
blast, immediately in front of it the post-stylar coelom, and 
then the massive exostylar tissue (Pl. 17, fig. 35). 

This series is very useful for the migration of leucocytes 
from the mucosa, through the uterine epithelium and across 
the cavum uteri into the chromatophilous megalokaryocytes, 
as well as for the ingestion of erythrocytes by the phagocytic 
megalokaryocytes. The mechanism of ingestion of the leuco- 
cytes seems to consist of an active immigration, the megalo- 
karyocytes, so long as they retain a free, unattached outer 
surface, remaining apparently passive. On the other hand 
the phagocytic activity of the implanted megalokaryocytes 
is often very obvious. Such cells, when engaged in the 
ingestion of red blood-corpuscles, commonly possess several 
large nuclei. The differentiation of the obplacental tropho- 
blast into free leucocytophagous and attached erythrocyto- 
phagous cells is a very singular and constant feature of the 
beaver’s preplacental blastocyst. 

In front of the region represented in Pl. 17, fig. 35, the 
carinal epiblast is ruptured, and this has led, over a certain 
extent of the series, to almost unaccountable confusion of the 
layers. It can be made out that the mesoblast becomes 
massed around the hypoblastic groove which dips into the 
keel as usual and is lined by cubical epithelium. The 
exoccelom becomes reduced to a linear cleft above the solid 
exostylar mass, and its walls may be in contact so that no 
Open ccelom appears in the section. ‘The carinal epiblast 
becomes greatly flattened, and there is an abrupt transition 
to the peripheral or adcarinal cubical epiblast (Pl. 17, 
fig. 36). 

The rupture of the carinal epiblast noted in preceding 
sections culminates forwards in a notch at the apex of the 
keel. If such a section is examined apart from the rest, the 
bilobed apex of the keel might appear to have a special 


202 * ARTHUR WILLEY. 


bearing, whereas it is really an artefact. After a score of 
sections from the last one figured, the primitive streak can 
be recognised on one side of the keel not far from the free 
distal edge. Its position is approximately in the centre of 
the thickened formative epiblast which bends round the edge 
of the keel (Pl. 17, fig. 37). On the left side of the figure 
the formative epiblast is seen blending with the cubical 
adcarinal trophoblast; on the right it passes abruptly into 
what is left of the flattened epiblast of the keel. 

The mesoblast continues to be massive on either side of 
the primitive streak. At the proximal end of the keel on 
the left side of the figure there is a small space limited 
externally by cubical mesoderm, internally by columnar © 
mesoderm. I interpret this as part of the pericardial pri- 
mordium, as will be expiained more clearly below. As we 
pass forwards from this level the massive mesoblast gives way 
to open ccelom. In the anterior region of the embryonic shield 
the sections present the general appearance illustrated in Pl. 
17, fig. 38. The keel is now embryonic; at its edges are 
seen incipient amniotic folds. The hypoblastic groove is cut 
tangentially near its anterior termination, and its carinal part 
appears separated from the omphaloidean hypoblast. Another 
part of the pericardial primordium is here seen in the meso- 
blast distad of the hypoblastic groove. 

In front of the embryonic shield the anterior terminal 
portion of the exoccelom extends characteristically into the 
proximal half of the keel, leaving the distal half free (Pl. 17, 
fig. 39). Thus there is a very definite keel-formation near the 
anterior pole of the blastocyst, well in front of the embryonic 
shield. 

The anterior and posterior trophoblastic portions of the 
keel with their exoccelomic spaces correspond in position 
with the ovate areas on the mature foetal sac which I described 
(1912, p. 203) as fenestre pyriformes or umbilico- 
placental areas. Between the pre-stylar and _post-stylar 
keels occurs the exostyle ; between the anterior and posterior 
pyritorm areas occurs the placenta. 


BLASTOCYST AND PLACENTA OF THE BEAVER. 203 


VIII. Susstrace E. 


The series is cut transversely from before backwards, and 
is numbered VIF in the Utrecht catalogue. There is noto- 
chordal contact between hypoblast and formative epiblast in 
front of the primitive streak. 


Millimetres. 
Height of gestation sac : : . about 10-00 
Diameter a = : : : 727560 
Height of cavum uteri. ' ; Sar tor 00) 
Depth of placental groove . ; 2°79 
Diameter “4 e : : 6: 50 to 0°80 
Pe omphaloid cavum : : S) s;00 
Thickness of placental mucosa. - about 3:00 
s s trophospongia . > dred 
53 periplacental wall . : ae) 
sé obplacental wall : ' 1:00 


The anterior pole of the blastocyst exhibits phagocytic 
adhesion to the wall of the omphaloid cavum; the posterior 
pole is free. Beginning from the front end, we find a long 
keel composed of cubical epiblast with a hypoblastic groove 
dipping into the proximal part of its cavity. Distally the 
walls of the keel are agglutinated together so that the cavity 
is occluded. After about twenty sections of the formation 
just described, the mesoblast appears with a ccelomic cavity 
embracing the sides and bottom of the hypoblastic groove, 
with a Jong epiblastic keel beyond it. The hypoblastic 
groove is lined by cubical epithelium with large round nuclei. 
There is general phagocytic adhesion of the entire obplacental 
hemisphere in the anterior region. 

At length the anterior end of the folded embryonic shield 
appears on the right side of the keel when the latter points 
towards the observer under the microscope (Pl. 17, fig. 40). 
Somewhat before this level is reached the coronal tropho- 
blast begims to show patches of chromatophile cells alter- 
nating with phagocytic megalokaryocytes. This leads on to 
the condition of a continuous coronal chromatophilous calotte 


204. ARTHUR WILLEY. 


such as has been described in previous substages, the leuco- 
cyte granules being extremely abundant. The mesoblast in 
the region of the embryonic shield is nearly solid, with a 
linear cleft representing the pericardial primordium, 

Proceeding backwards with the sections, the solid meso- 
blastic bands become separated at a certain spot, and the 
embryonic hypoblast there comes into apposition with the 
formative epiblast, its cells at the same time acquiring a 
typical columnar form. It is the notochordal primordium 
(Pl. 17, fig. 41). This figure shows the solid peripheral 
swellings (sinus terminalis) of the mesoblastic bands 
placed at slightly different levels and a little exoccelom 
occurring between the swellings and the mass of the bands, 
The notochordal primordium extends through about a dozen 
sections of 10 u, and then it passes into the primitive streak 
(Pl. 17, figs. 42 and 43). At this level of the blastocyst there 
is a very extensive megalocytic attack in the coronal region, 
flanked on either side by a pericoronal chromatophile band ; 
beyond this there is a periomphaloid attack followed by peri- 
carinal festoons — all in the typical manner previously 
described (cf. Pl. 14, fig. 5). 

At the posterior end of the primitive streak the massive 
exostylar tissue commences and the mesoblastic bands enter 
into close contiguity with the carinal epiblast over a wide 
area (Pl. 17, fig. 44). The significance of the exostylar 
portion of the keel with its massive mesoblast and superjacent 
trophoblast (carinal epiblast) will be discussed below under 
“Special Considerations.” In the mid-region of the exostyle 
in this series the keel is damaged in nearly every section in 
consequence of its having been torn from its viscid adhesion 
to the uterine wall. The figures are further complicated by 
the peculiar folding of the keel, which continues for some 
distance behind the embryonic shield (Pl. 17, fig. 45). 
Towards the posterior end of the exostyle the fold straightens 
out and we have the appearance shown in PI. 17, fig. 46, 
where the massive tissue is bordered by a simple keel with 
juxtaposed walls. After this the post-stylar exoccelom opens 


ig 


BLASTOCYST AND PLACENTA OF THE BEAVER. 208 


out in a typical manner (PI. 18, fig. 47). Finally, the two 
sections of the primordial sinus terminalis meet together 
(Pl. 18, fig. 48), and thereafter the mesoblast ceases. 
Between the last two figures there are thirty-nine sections 
of 10 wu. 

Behind the posterior limit of the mesoblast we find for 
some distance a didermic keel with trophoblastic extension 
(Pl. 18, fig. 49). The balloon-shaped blastocyst is here free 
from all adhesion to the uterine wall. Farther back, the 
hypoblast withdraws from the keel (Pl. 18, fig. 50), and 
eventually the section is exactly similar to that shown in 
Pl. 15, fig. 17, which relates to the anterior pole of the 
blastocyst in Substage B. The two poles of the blastocyst, 
though differing in the extent of their adhesion to the uterine 
wall, are naturally similar in other respects. 

Apart from the broad difference between a free pole of the 
blastocyst and a fixed pole, there is another more subtle 
distinction which has been referred to incidentally and may 
be exemplified once more in this series, namely, the difference 
between a free and an implanted coronal disc. It may be 
noted in passing that the term “ coronal” is not synonymous 
with ‘ obplacental.’”” The whole of the wall of the omphaloid 
cavum to which the omphaloid hemisphere of the blastocyst 
may be attached is obplacental. At the level of Pl. 17, 
fig. 44, the coronal region of the blastocyst in transverse 
section exactly equals the diameter of the field of the micro- 
scope when viewed under Zeiss oc. 4, obj. A; throughout 
this extent the denuded mucosa is lined by a continuous 
pseudo-epithelial sheet of megalokaryocytes closely attached 
by their amoeboid peripheral ends to the decidual surface, 
without any arcades. At one edge of this implanted area, a 
gland opens into the pericoronal cavum at a true epithelial 
surface ; at the other edge the pseudo-epithelium adjoins the 
true cylindrical epithelium of the pericoronal cavum of that 
side. ‘The implanted trophoblast does not sink into the 
mucosa; in general the line of implantation is straight, 
corresponding to the original basement membrane of the 


206 ARTHUR WILLEY. 


uterine epithelium. The subepithelial capillaries, at frequent 
intervals, where they come into contact with the trophoblast, 
actually open and discharge their red corpuscles directly 
into individual megalokaryocytes. The obplacental glands 
almost always open at an epithelial surface subtended by a 
trophoblastic arcade, but in a few rare instances a gland is 
seen to open directly upon the trophoblast, into which it 
discharges a mass of intrusive chromatic granules. At the 
level of fig. 47, instead of the implanted pseudo-epithehal 
coronal trophoblast, we find the coronal region of the blasto- 
cyst entirely free, composed of chromatophilous megalo- 
karyocytes, as represented diagrammatically in PI. 15, fig. 24. 


IX. Susstace F. 


This excellent series shows very clearly the embryonic 
shield with primordia of medullary groove, notochord, peri- 
cardium and amniotic fold. It is cut from before backwards 
and is numbered VI c in the Utrecht catalogue. Both poles 
of the blastocyst are fixed to the wall of the omphaloid 
cavum. 


Millimetres. 

Height of gestation sac : : : . Looe 
Diameter of gestation sac. : , . 8:00 
Height of cavum uteri . : : . ' oe 
Depth of placental groove . ; : > Oe 
Diameter of placental groove : . 0°75 to Ae 
Gs omphaloid cavum. : . 9°00 
Thickness of placental mucosa. : .- OG 
= a trophospongia . - ee 

of periplacental wall. 3°50 on one side, 

4-00 on the other. 
a obplacental wall =. . 1:00 to 1°85 


These measurements, taken in conjunction with Pl. 18, 
fig. 51, exhibit a placental groove which is considerably 
wider and shallower than anything that has gone before. 
This progressive expansion of the cavum uteri accords with 


BLASTOCYST AND PLACENTA OF THE BEAVER. 207 


the fact that the blastodise is farther advanced in develop- 
ment, though the exostyle is still far from its destination on 
the placental trophospongia. The estimated length of the 
blastocyst is nearly 4 mm. 

At the anterior pole the blastocyst is monodermic, and is 
attached by the omphaloid hemisphere in broad arcades. 
Then the hypoblast appears a short distance removed from 
the pole; at first it stretches straight across the adcarinal 
plane and a little further back dips down into the proximal 
half of the keel, the distal half of which remains monodermic. 
In previous substages it has been observed that in places, 
the epithelium which lines the hypoblastic groove, i.e. the 
carinal hypoblast, tends to break up into rounded cells that 
he loosely in the cavity of the groove, but never escape into 
the distal cavity of the keel. The extent and position of the 
hypoblastic groove can always be made out, however dis- 
connected its cells may be. This fact is due to the peculiar 
behaviour of the basement membrane which exists between 
the adcarinal epiblast and hypoblast. 

At a certain point near the mouth of the keel the 
adcarinal basement membrane can be seen to cross the 
narrow interval that separates the epiblast from the hypo- 
blast, and to attach itself to the latter, which it accompanies 
into the keel. The trophoblastic cavity of the keel is thus 
shut off from the omphalon by a delicate wrinkled membrane. 
The keel-cavity contains a finely granular coagulum, which 
differs slightly in its staining properties from the adjacent 
coagulum in the adcarinal region, so that there is a sharp 
contrast in the character of the coagulum on either side of 
this structureless membrane (Text-fig. 2). The membrane is 
inconspicuous and easily overlooked. 

The anterior border of the mesoblast appears on one side 
of the keel, viz. the embryonic side, as shown by subsequent 
sections. In this region there is aremarkably deep epiblastic 
keel extending far beyond hypoblast and mesoblast (PI. 18, 
fig. 52). The mesoblast rapidly increases in volume as we 
trace it backwards, and presently the anterior ccelom opens 

VoL. 60, PART 2,—NEW SERIES. 15 


208 ARTHUR WILLEY. 


out with a deep extension into the keel. Produced beyond 
this carinal ccelom there isa rather long and somewhat folded 
epiblastic keel with its walls agglutinated so as to occlude 
the potential trophoblastic cavity. 


TEXT-FIG. 2. 


Anterior region of keel, showing relation of basement membrane 
to carinal hypoblast. 1. Interval between hypoblast and epi- 
blast containing adecarinal coagulum. 2. Basement membrane 
crossing the interval to join the carinal hypoblast. 3. Tropho- 
blastic cavity of keel. 4. Hypoblastic groove. 


The epiblastic keel soon begins to shorten, the carinal 
coelom becomes reduced, and the cells of the hypoblastic 
groove appear as a continuous epithelium on the embryonic 
side. In the diagram (Text-fig. 3) I have reconstructed the 


BLASTOCYST AND PLACENTA OF T'HE BEAVER. 209 


hypoblast on the anti-embryonic side of the keel, where it is 
evanescent in the actual section. Between the condition 


itisp-crsinives 6}. 


Appearance of the keel behind the anterior polar region. The 
index numbers are placed on the embryonic side of the keel. 
1. Exocelom. 2. Hypoblastic groove. 3. Shortened epiblastic 
keel. 


shown in Pl. 18, fig. 52, and that in Text-fig. 3 there are 
thirty-nine sections of 10 . At the region of the blasto- 
cyst which we have now reached the coronal disc is composed 


210 ARTHUR WILLEY. 


of large chromatophile cells freely subtending the coronal 
cavum, which is filled with coagulum containing many 
chromatic granules. Two or three sections after Text-fig. 3, 
the front part of the embryonic shield begins to be cut 
tangentially. Its abrupt appearance in section indicates a 
much sharper definition than in previous substages. 

Under the Zeiss microscope, when the keel appears to 
point towards the observer, the embryonic shield is placed 
on the right side of the keel. The plate-drawings were 
reversed under the Edinger apparatus. For descriptive 
purposes we may regard the edge of the keel as pointing 
downwards towards the observer, although it really points 
upwards towards the mesometrium. Above the embryonic 
shield in the figure (Pl. 18, fig. 553) there is a somatopleuric 
fold which looks like a lateral amniotic fold and is so inter- 
preted. Below the shield in the figure there is the epiblastic 
keel into the neck of which the exoccelom of that side is 
produced, so that the material for the other amniotic fold is 
in continuity with the keel. Underneath the thick formative 
epiblast of the shield we find a thick-walled mesodermic 
saccule, interpreted as the pericardial primordium of the 
embryonic cceelom. The only ccelom at present existing in 
the embryo is the pericardial ccelom. Against this the 
embryonic hypoblast consists of a continuous cubical epithe- 
lium, while on the opposite side the hypoblast is evanescent 
and granular. Below the distal border of the hypoblast a 
thin sheet of mesoblast intervenes between it and the epiblast 
(Pl. 18, fig. 53). 

Proceeding backwards, the pericardial saccule shortly 
separates into two moieties. The portion which lies sub- 
centrally under the thickened epiblast shows a character- 
istic thicker inner wall of columnar cells and a thinner 
outer wall of cubical cells. Between the two parts of the 
‘pericardial ccelom the mesoblast thins out, and so the cubical 
hypoblast comes nearer to the formative epiblast at this 
point (PI. 18, fig.54). The next figure shows an accentuation 
of the preceding features, and is introduced to exhibit the 


BLASTOCYSI AND PLACENTA OF THE BEAVER. Past bl: 


two diverging limbs of the pericardial ccelom which now 
appear as antimeres (PI. 18, fig. 55). 

As we approach the posterior quarter of the blastodisc, the 
embryonic shield becomes larger, and a medullary groove 
with subjacent notochordal plate appears in the position 
shown in PI. i9, fig. 56. After another half-score sections 
we have the structure represented in Pl. 19, fig. 57. The 
medullary groove and the distal pericardial process have 
ceased, and the lower part of the keel begins to be occupied 
by massive mesoblast with a hypoblastic diverticulum extend- 
ing into it. At the foot of the keel there is an open hollow 
space containing coagulum ; it is possible that it is an incident 
of growth rather than that it has any special significance, but 
this may remain an open question. In this figure and the 
preceding, the sinus terminalis is seen low down on the 
anti-embryonic side of the keel; on the embryonic side it 
occurs near the junction of the adcarinal and carinal regions. 
This unequal behaviour of the parts of the sinus termi- 
nalis has been observed in ali substages. 

At the posterior end of the blastodisc, the primitive streak 
in this substage is reduced both in extent and definition, and 
it is evident that it would soon be given up. Its position and 
the wings of massive mesoblast proceeding from it are shown 
in Pl. 19, fig. 58. The hypoblastic groove lined by cubical 
epithelium intruding into the distal mesoblast is conspicuous, 
and the transition from cubical to flattened epithelium is 
remarkably abrupt. 

We are now rapidly nearing the region of the massive 
exostyle which always follows behind the primitive streak. 
In this substage the exostyle exhibits in section a broad distal 
base with a small trophoblastic cavity and hypoblast pene- 
trating into the massive mesoblast. ‘The epiblast on one side 
of the keel is peculiarly folded. The distal lumen of the 
hypoblastic groove is occluded, the occlusion taking place by 
opposition of the walls accompanied by intense proliferation 
of the massive mesoblast, whereby the axial hypoblast merges 
imperceptibly with the surrounding mesoblast, without the 


212 ARTHUR WILLEY. 


intervention of a basement membrane (Pl. 19, fig. 59). The 
large clear nuclei of the intrusive tongue of hypoblast contrast 
with the more darkly stained mesoblastic nuclei in their 
normal aspect, but they appear to blend with the latter at the 
sides and apex. On the anti-embryonic side of the keel there 
is unsplit mesoblast; on the embryonic side an exoccelom 
whose epithelium is partly cubical. The anti-embryonic part 
of the sinus terminalis lies low down, midway between 
the foot of the keel and the position of the sinus on the 
embryonic side. The massive tissue of the exostyle is con- 
centrated within somewhat narrow bounds, covering about a 
dozen sections of 10. This concentration and delimitation of 
the exostyle are what must necessarily precede the initial 
discoplacental adhesion. 

A few sections behind PI. 19, fig. 59, the axial cavity re- 
appears in the hypoblast and this marks the transition from 
the exostyle to the post-stylar region. A deep and narrow 
hypoblastic groove occupies the centre of the mesoblast ; 
above the massive tissue the exoccelom opens out on both 
sides to form a wide cavity. Below the mesoblast there is a 
trophoblastic extension of the keel (Pl. 19, fig. 60). 

At the region of the blastocyst corresponding approximately 
with the course of the exostyle, the obplacental wall exhibits 
the condition of a continuous megalocytic attack accompanied 
by ingestion of red corpuscles in the coronal area, and a wide 
pericoronal arcade subtending a pericoronal cavum on each 
side. As already explained, the term arcade in this connec- 
tion expresses typically a single bridge of chromatophilous 
trophoblast with free outer surface washed by the uterine 
fluid. The particular function of the obplacental arcades, in 
addition to the ingestion of leucocytes by the chromatophile 
cells, is to afford spaces for the openings of the persistent 
obplacental glands. 

Posteriorly the two halves of the exoccelom unite to form a 
common cavity, the post-stylar coelom, which extends some way 
into the distal cavity of the keel (Pl. 19, fig. 61). After this 
the trophoblastic keel lengthens enormously (Pl. 19, fig. 62). 


BLASTOCYST AND PLACENTA OF THE BEAVER. 213 


About this region of the blastocyst the sections show demon- 
stratively the ingestion of erythrocytes by the coronal megalo- 
karyocytes and the ingestion of leucocytes with their lobed 
chromatic nuclei by the pericoronal chromatocytes. In 
addition to the coronal implantation and the pericoronal cava, 
there is a typical periomphaloid adhesion and _ pericarinal 
festoons. 

The two sections of the sinus terminalis descend deeper 
into the keel, and finally come near to each other at the bottom 
of the carinal hypoblast as shown in PI. 19, fig. 63, which is 
thirty-nine sections behind the preceding figure. At the same 
time the post-stylar ccelom is reduced, and the trophoblastic 
extension of the keel is devoid of a continuous axiallumen. In 
Pl. 19, fig. 64, the two parts of the sinus nearly touch; and 
fig. 65 shows the sigmoid confluence of the sinus, marking the 
posterior limit of ccelom and mesoblast. 

Next follows the posterior didermic region of the blasto- 
eyst. The trophoblastic keel shortens, but the hypoblast 
retains its carinal extension until near the posterior pole. 
These features, as well as the relation of the keel to the 
adcarinal omphalopleure, are portrayed in Pl. 20, fig. 66. 
(Quite at the posterior pole we find in section a round didermic 
blastocyst free on its mesometric side, adhering intermittently 
elsewhere. 


X. MATERNAL T'ROPHOSPONGIA. 


Without undertaking a detailed discussion of the histology 
of the mucosa, attention may be invited to certain interesting 
changes which take place in the uterine mucosa of the 
gestation sac, preliminary to placentation. ‘The perusal of 
the foregoing pages will doubtless have left a vivid impres- 
sion of the vital importance to the beaver’s blastocyst 
of its obplacental implantation. One might have expected, 
from the analogy of the rabbit, that this would involve 
the degeneration of the obplacental uterine glands. On 
the contrary, the obplacental glands retain their full func- 


214 ARTHUR WILLEY. 


tion throughout the period dealt with, and the trophoblast 
gives free way to their crypt-like openings. It is only 
in the walls of the mesometric or placental groove, which 
have not yet entered into very close association with the 
blastocyst, that the glands have already become atrophic. 

The mucosa may be regarded as comprising two classes of 
tissue: (1) Uterine epithelium and glands; (2) dermatic 
connective tissue and capillaries. The fixation of the blasto- 
cyst resolves itself into two periods: (1) Preplacental period ; 
(2) euplacental period. The size of the gestation sac is 
determined by two factors: (1) Dermatic proliferation; (2) 
pressure of blastocyst. 

In one of the uterine cornua belonging to a later phase of 
gestation which will be described below, there was a slightly 
swollen segment of the uterus, which looked as if it might 
contain a very young blastocyst. After it had been sectioned, 
it was found to be in a perfectly healthy condition, but non- 
gravid. This series is numbered VII x in the Utrecht cata- 
logue. The cavum uteri is a rather narrow lumen 
bifurcated towards the mesometrium and lned by a high 
columnar epithelium; opening into it on all sides there is a 
profusion of glands (PI. 20, fig. 67). It contains a coagulum 
with some dark-stained granules diffused throughout it. The 
lumen is invested outside the epithelium by a narrow uni- 
form zone of intense dermatic proliferation riddled with 
capillaries; in this zone (not indicated in the figure) the 
connective-tissue cells are more numerous and closer together 
than in the deeper parts of the mucosa. On the mesometric 
side the Jumen branches into two grooves, separated by a 
ridge or dermatic cone covered by epithelium. Numerous 
large vessels occur between the circular and longitudinal 
muscles of the mesometric region; the circular muscle-ring 
is broad and entire, except where penetrated by the vessels. 

In contrast with the initial condition of uniform dermatic 
proliferation and intact glands described above in a non- 
gravid segment, we find in a gravid gestation sac belonging 
to the preplacental period that all the mesometric and peri- 


BLASTOCYST AND PLACENTA OF THE BEAVER. Palla, 


placental glands are undergoing necrosis, being met and 
vanquished by an epithelial proliferation operating centri- 
fugally.- The dermatic proliferation acts centripetally. Mean- 
while the dermatic proliferation in the future placental 
region has outstripped the remainder, and now constitutes 
what has been referred to as the placental trophospongia, 
the term “‘trophospongia” having been introduced by 
Hubrecht in 1889 to denote a vascular maternal proliferation. 
In the beaver such a proliferation takes place all round the 
cavum uteri, and is at first of uniform thickness. 

The epithelial proliferation is a phenomenon of substitu- 
tion accompanying the necrosis of the glands. It is paralleled 
by a corresponding phenomenon described by Hubrecht 
(1893) in the case of the shrew (Sorex), where the epithelial 
proliferation leads to the formation of secondary crypts. In 
the beaver’s gestation sac during the preplacental period, 
the placental mucosa, towards the bottom of the placental 
groove, exhibits, sometimes very clearly, radiating bands of 
necrotic glands, with dark-stained shrunken nuclei (PI. 20, 
fig. 68). These glands are to be replaced by an extensive 
epithelial proliferation, which grows centrifugally into the 
substance of the placental trophospongia in the form of two 
lobulate wings, which correspond in their position with the 
grooves on either side of the dermatic cone in the non-gravid 
uterus (Pl. 20, fig. 67). Thus the wing-like epithelial pro- 
liferations are preceded by necrotic belts. The cells multiply 
by mitosis. In the niches between the lobes are found 
subepithelial capillaries (Pl. 20, fig. 69). 

In the obplacental region, where the subepithelial capil- 
laries are also excessively abundant, the capillaries are con- 
veyed by the centripetal dermatic proliferation. This happens 
in the placental region as well, but in addition the centri- 
fugal epithelial proliferation descends to meet the capillaries 
and embrace them, thereby preparing a nidus for the true 
placental implantation. At the beginning, the proliferating 
cells retain their cell-boundaries (Pl. 20, fig. 70). Even- 
tually the proliferation will become syncytial. At the level 


216 ARTHUR WILLEY. 


of the anterior part of the embryonic shield in substage F, 
the capillaries which adjoin the inner borders of the wing-like 
mesometric proliferations are greatly dilated, but there is no 
trace of any endothelial proliferation (Pl. 20, fig. 71). 

At the distal borders of the two principal mesometric pro- 
liferations there are to be seen traces of the glands which 
they have supplanted. There are no signs of glands between 
the wings; the latter cause the more laterally placed meso- 
metric glands to diverge widely, arching round the meso- 
metric area, appearing healthy in their deep-lying portions, 
but failing to reach the surface, their neck-portions having 
been killed. Sometimes a transverse section of a glandular 
tube is seen, the centripetal half of which is necrotic, the 
distal half normal. In this way, centripetal glandular 
degeneration and centrifugal epithelial proliferation take 
place simultaneously. In the obplacental region the glands 
persist and the uterine epithelium is largely replaced by 
trophoblast ; in the placental region the glands degenerate 
and are replaced by epithelial proliferations which are to 
some extent moulded upon the pre-existing glands. At the 
sides of the placental groove the necks of the glands some- 
times widen out into large ampulliform dilatations with ill- 
defined walls and wrinkled nuclei; such ampulle fail to open 
into the cavum uteri. They may nearly open, but their 
mouths are blocked, partly by their own degenerate cells and 
partly by epithelial cells (Pl. 20, fig. 72). 

There is not much to add to what has been said already 
regarding the uterine epithelium. Where the trophoblast is 
free and the epithelium consequently intact, the character of 
the latter varies at different points without any regard to 
symmetry. In one instance my notes record an epithelium so 
flattened as to become a pavement epithelium, covering the 
left side only of the omphaloid cavum. The cytoplasm of the 
uterine epithelium stains dark yellow with orange G, and 
thus offers a marked contrast to the adjoining trophoblast, 
whose cytoplasm remains pale. J. W. Jenkinson (1902, p. 
132) has noted the same distinction in the case of the mouse: 


BLASTOCYST AND PLACENTA OF THE BEAVER. Pilea 


“The cytoplasm [of the trophoblast cells] does not stain 
intensely with acid stains, and may in this way be readily 
distinguished from the cytoplasm of the maternal cells.” 
Again, on p. 159, referring to epithelial cells which have 
been ingested by the trophoblast, he says: “As for the cyto- 
plasm, it stains brilliantly with plasma stains, and offers in 
this respect a marked contrast to the cytoplasm of the 
trophoblast by which it has been ingested.” This last is true 
of the erythrocytes ingested by the megalokaryocytes in the 
beaver. 

There are signs that in the beaver the uterine epithelium 
may not only disappear by direct substitution of the tropho- 
blast in situ, but that, in a manner not unlike what P. Nolf 
(1896) described in the bat, it may first be reduced to a 
cubical and then to a pavement epithelium; or finally it may, 
in places, be shredded off into the cavum uteri. 

The migration of leucocytes from the mucosa across the 
uterine epithelium into the omphaloid cavum and thence into 
the chromatophilous megalokaryocytes, or, more briefly, 
chromatocytes, is a phenomenon that requires further com- 
ment. Granules staining darkly with hemalaun, occurring 
commonly in triads and tetrads, are found deep down in the 
stroma of the obplacental mucosa. From the deeper zone 
they scatter through the intermediate stratum and are again 
met with in great numbers in a subepithelial position, from 
whence they pass into the epithelium. At first I supposed that 
they might be derived from the nuclei of the decidual cells, 
but was unable to find transitional forms. Not only do they 
traverse the uterine epithelium, but they also pass in smailer 
quantities into the lumina of the glands, and are then dis- 
charged with the glandular secretion into the cavum uteri. 
In much less numbers are they observed in the capillaries. 

In places the obplacental mucosa showed an abundant 
infiltration of undoubted leucocytes whose lobed nuclei were 
these same chromatic granules of whose nature I had been 
uncertain. When they reach the bases of the epithelial cells 
and penetrate through the latter, their cytoplasm does not 


218 ARTHUR WILLEY. 


stain so well as it does in the middle of the mucosa, but 
remains clear so that the granules often appear to be con- 
tained in vacuoles. When they arrive in the uterine fluid, 
which is represented in the sections by a coagulum staining 
deeply with orange G, the cell-body is often invisible though 
sometimes quite obvious. The marked tendency which they 
exhibit to discharge themselves in volleys through the uterine 
epithelium corresponds with their deep-seated gregarious 
habit in the mucosa. ‘This infiltration of leucocytes is com- 
parable with that observed by Nolf in the epithelioid tissue 
and hypertrophied venous endothelium of the bat’s tropho- 
spongia, although their fate is different. 


XI. DiscopLaceNntaL ADHESION. 


The stage of the incipient placentation is unfortunately 
lacking in the material collected by me. In the keel of the 
preplacental blastocyst four parts with independent destinies 
may be distinguished: the anterior keel with anterior exo- 
coelom, the embryonic keel, the exostyle, and the posterior 
keel with post-stylar exoceelom. The first and last will be 
employed in the formation of the umbilico-placental mem- 
branes or fenestre pyriformes, consisting of diplotropho- 
blast, i.e. trophoblast doubled by the somatic mesoblast of 
the exocelom. The embryo will become folded off, and will 
sink with its amnion into the exoceelom. ‘The exostyle will 
furnish the material for the placental labyrinth. 

But the chapter recording these events in their actuality 
cannot yet be written. The groove at the foot of the keel in 
Pl. 19, fig. 59, is ready to receive the dermatic cone of the 
maternal trophospongia with its syncytial caps (cf. Pl. 20, 
fios. 69-71), but we are not at present privileged to witness 
the act. 


XII. Tue EstaslisHep PLACENTA. 


The material upon which the following description is 
based consisted of six spherical gestation sacs about 25 mm. 


BLASTOCYST AND PLACENTA OF THE BEAVER. 219 


in diameter, obtained on two separate occasions—March 4th 
and March 10th, 1913. It was on the latter date that I also 
obtained the six youngest gestation sacs. Ina female beaver, 
taken on March 4th, there were two spherical swellings, each 
25 mm. in diameter, in the left uterus only, and two con- 
tiguous corpora lutea in the left ovary. Above the upper 
swelling there was a smaller one of 12°5 mm., upon which 
ereat hopes were set; but it proved barren. The axillary 
teats of the mother were distinct, but the post-axillary pair 
was only found after the skin had been removed. The 
trapper said it was a young female that would have been 
three years old in the coming spring. In another female, 
taken on March 10th, there was a single spherical sac in the 
left uterus, and three in the right. From these data it may 
be gathered that whilst the beaver produces a litter only once 
a year in the spring, there isa considerable range of variation 
with regard to the stages which may be observed upona given 
date during the season of reproduction. 

I opened one of these sacs in Prof. Hubrecht’s presence by 
slicing away the antimesometric wall with a cartilage knife. 
We were surprised to find that no liquid exuded when the 
cavity of the omphalon was exposed; on the contrary it was 
entirely filled by a translucent coagulum which cut like cheese 
and showed no tendency to break away. One slice, about 
2mm. thick, was cut through the middle of the gestation 
sac; the coagulum, stretched from wall to wall like a sheet 
of gelatine, retained its position exactly, and the cut surfaces 
remained flat. 

In the mesometric half the embryo could be seen through 
the coagulum lying on one side, its left side being presented 
towards the omphalon. The embryo lies in the exoccelom 
between the placenta and the umbilical membrane. The 
minimum thickness of the uterine wall exceeded 4 mm. ; the 
maximum was nearly 6 mm.; the diameter of the cavum 
uteri, and consequently of the omphalon that occupied it, 
was about 12 mm. (PI. 20, figs. 73 and 74). The placenta, 
measured in section, had a length of 4°25 mm., and a height 


220 ARTHUR WILLEY. 


of 225 mm. ‘The latter measurement signifies that the some- 
what mushroom-shaped placenta projected freely to that 
extent into the exocelom. After being reduced to sections 
the body of the embryo was found to have shrunken rather 
badly owing to the treatment it had received. When the 
destination of such early stages is the microtome they should 
not be cut open previously. Fortunately there was no lack 
of good material. 

In its shape and projection the placenta may be likened to 
a cushion or a mushroom ora dome. At its edge it overlaps 
the base of insertion, so that in sections passing tangentially 
through the margin it appears detached from the uterine 
wall. Such tangential sections are useful for determining 
the essential structure of the placental labyrinth. There isa 
straight trophoblastic base continuous with the rest of the 
trophoblastic wall of the blastocyst, from which centripetal 
trabecule ascend into the allantoic mesoblast. These tropho- 
blastic trabecule are excavated by canals which carry 
maternal blood, and are what M. Duval called sanguimaternal 
lacunz in the ectoplacenta. The allantoic mesoblast conveys 
foetal capillaries to the placental labyrinth. The allantoic 
tissue appears in the form of centrifugal vill interlacing with 
the centripetal trabecule, together establishing the placental 
labyrinth (Text-fig. 4). Neither the trophoblast nor the 
allantoic mesoblast remains passive in the growing placenta, 
but they grow in opposite directions, the trophoblast centri- 
petally, the allantoic mesoblast centrifugally. 

After passing the point where the placenta is inserted into 
and united with the uterine wall, the definite basal layer of 
trophoblast with its sharp-demarcation disappears. Wherever 
the trophoblast touches allantoic tissue it is composed of a 
cellular layer corresponding to what E. van Beneden (1888) 
called the cytoblast, subsequently lengthened into cyto- 
trophoblast by J. H. Vernhout (1894). Towards maternal 
blood and trophospongia the trophoblast presents a plasmodial 
layer termed “ plasmodiblast” or ‘ plasmoditrophoblast.” 
For the sanguimaternal lacune in the trophoblastic trabeculae 


DID) | 


BLASTOCYST AND PLACENTA OF THE BEAVER. 


the plasmodiblast furnishes a pseudo-endothelium. The solid 
buds produced at the growing edge of the placenta consist of 
eytoblast which ascends into the allantoic tissue. The cyto- 
blastic buds anastomose and so enclose allantoic islands. he 
eytoblast is the formative layer of the trophoblast ; it may be 
compared with the Malpighian layer of the skin, but instead 
of giving rise to horny cells on its outer surface, it produces 
plasmodiblast which is rooted in the vascular trophospongia. 
The cytoblast itself is rooted in the vascular allantoic tissue ; 


TEXT-FIG. 4. 


Section through growing margin of placenta. Trophoblast 
black, mesoblast dotted. The asterisks indicate the line where 
the allantoic and somatic tissues meet; outside the asterisks 
the double membrane is diplotrophoblast. 


it does not give off any buds towards the maternal tissue. 
The cytoblast is a growing tissue, the plasmodiblast is a 
feeding organ, and the trophospongia is the nidus supplying 
the nutriment. 

The foregoing aphorisms are indited with special reference 
to the conditions observed in the beaver, but they derive 
support from the observations of others upon different 
animals. In the rabbit, H. Schoenfeld (1903), confirming 
Maximow, states that the plasmodiblast is supplied with 
elements from the cytoblast, and adds that it has no growth 
of its own; it moves about upon the surface of the cytoblast 


222 ARTHUR WILLEY. 


as described by A.Maximow (1900). Inthe squirrel (Sciurus), 
F. Muller (1905, p. 560) says that the growth of the placenta 
does not take place merely by the substitution of foetal for 
maternal tissues, but by the progressive centripetal growth 
of the ectoplacental mass, which thus surrounds the allantoic 
villi more and more; nevertheless the greater part of the 
uterine mucosa is supplanted by the placenta, because a 
continual process of degeneration and resorption of maternal 
tissue is taking place. 

In the bat, P. Nolf (1896, p. 610) says that the increase in 
thickness of the placenta is not the result of peripheral 
growth at the expense of the maternal tissue. Almost all 
of its secondary growth is centripetal ; the internal or foetal 
face of the placenta grows towards and projects into the 
blastodermic cavity. This is proved by the fact that the 
vegetative epiblast throughout gestation is continued into 
the placental cytoblast, not at the level of the internal sur- 
face of the placenta, but at the level of its external surface. 
This conclusion accords, he says, with those deduced by 
Duval for Carnivora, by Hubrecht for the shrew, and by 
Vernhout for the mole. 

Tn all these instances the nutritive material for the centri- 
petal growth of the ectoplacental trophoblast is furnished by 
the maternal trophospongia, which forms a cushion upon 
which the placenta rests. In the beaver this trophospongia 
has a twofold origin in a vascular dermatic proliferation and 
a lobular epithelial proliferation. The latter is now broken 
up into polygonal blocks by the capillary network, producing 
an areolated structure in section. Many of the epithelial 
areole have degenerated into syncytia in which numerous 
dark-stained nuclei are densely aggregated. Sometimes a 
syncytial mass is extensive and then there appears a curved 
band of nuclei in the midst of it. The epithelial areole, the 
syncytia, and the aggregations of deeply staining nuclei are 
very characteristic of the early trophospongia of the beaver’s 
placenta. To these features should be added the presence at 
certain points of brown granules chiefly surrounding large 


BLASTOCYST AND PLACENTA OF THE BEAVER. 223 


degenerating maternal nuclei. In the deeper zone of the 
maternal trophospongia the cell-islands are not yet syncytial ; 
some are entirely cellular, others are partly syncytial. 

There are interesting analogies between events in the pre- 
placental and euplacental periods, as under : 


Preplacental period. Euplacental period. 

1. Uterine glands partly normal, . Epithelial areole partly cellular, 
partly necrotic. partly syncytial. 

2. Centripetal dermatic prolifera- . Centrifugal allantoic prolifera- 
tion conveying maternal capil- tion conveying fcetal capil- 
laries. laries. 

3. Centrifugal epithelial prolifera- . Centripetal ectoplacental pro- 
tion into the dermatic tissue. liferation into the allantoic 

tissue. 


At the junction of trophoblast and trophospongia a 
symplasma is formed. Schoenfeld (1902) discusses the 
use of the terms syncytium and plasmodium which were 
introduced by Prof. Haeckel; and symplasma suggested by 
Graf Spee. A syncytium is immobile, a plasmodium is mobile. 
In connection with the phenomena of placentation, Schoenfeld 
applies syncytium to maternal formations, plasmodium to 
foetal formations. Only when the fcetal plasmodium comes 
into contact with the maternal syncytium does the latter 
undergo degeneration and become converted into a symplasma 
which is defined as a syncytium in retrogression. ‘The 
symplasma is brought about by the incorporation of maternal 
protoplasm and nuclei into the substance of the plasmodium. 
Until the degeneration of the maternal elements is complete, 
the peripheral zone of the plasmodium is a symplasma con- 
taining active foetal nuclei and passive maternal nuclei. This 
interpretation involves perhaps a slight extension of the 
original meaning of the term symplasma. In the dog, 
Schoenfeld described the fusion of decidual cells with the 
plasmodium, thus converting the latter, for the time being, 
into a mixed “ fceto-maternal plasmodium” ; but there can be 
no harm in calling it a symplasma.! Good illustrations 
of a typical symplasma in the sense here indicated were 

' To avoid possible ambiguity perhaps the term “symplasmodium ”’ 
might he preferred. 

VoL. 60, parT 2.—NEW SERIES. 16 


224 ARTHUR WILLEY. 


given by A. Maximow (1900, Taf. xxx, figs. 1 and 2) for 
the rabbit. 

At many points the junction of plasmodium and syncytium is 
rendered conspicuous by the presence of the crowded syncytial 
nuclei, which contrast in colour, size and shape with the 
smaller, paler, oval, and more evenly distributed plasmodial 
nuclei. The conjoint symplasma may be observed to surround 
superficial capillaries of the trophospongia together with 
adjoming decidual cells. At such places the plasmodial 
nuclei may be seen intruding into the syncytium, and no 
boundary can be drawn between the plasmodial protoplasm 
and the syncytial protoplasm. The inclusion of maternal 
capillaries seems to be effected by the tortuous growth of the 
allantoic tissue which pushes the trophoblast before it into 
the trophospongia, so that we find in the border zone between 
trophospongia and placental labyrinth islands of allantoic 
tissue surrounded by trophoblast, and outside that a mantle 
of symplasma (PI. 20, fig. 75). In the diagrammatic figure 
(Pl. 20, fig. 76), the allantoic villi, with their trophoblastic 
investment, are seen to be tipped by syncytial groups of 
nuclei, rendered in black. The characteristic trophospongial 
islands or areole show, in the sections, various grades of 
syncytiation at different levels, there being no regularity in 
the distribution of the syncytia throughout the trophospongia, 
except at the symplasma or zone of contact. 

The trophospongial islands, whether syncytial or cellular, are 
separated by capillaries with normal endothelium. At the 
base of the trophospongial cushion, between it and the massive 
dermatic proliferation, there is a basal sinus-like blood-space, 
to which large capillary vessels pass vertically through the 
dermatic tissue from the direction of the mesometrium. This 
arrangement is indicated in PI. 20, fig. 76. Near each pole 
of the placenta a large vessel is found receiving its affluents 
from the sanguimaternal lacune in the swollen tips of the 
trophoblastic trabecule ; from thence it descends into the 
mucosa. The two polar vessels possess a proliferated endo- 
thelium which they retain until they reach a point deep in the 


BLASTOCYST AND PLACENTA OF THE BEAVER. 225 


mucosa, where each of them is continued into a narrow vessel 
with normal endothelium which passes directly and abruptly 
into the proliferated walls (Pl. 20, fig. 77). There is an 
equally abrupt transition from the proliferated endothelium 
to the plasmodial pseudo-endothelium near the fcetal peri- 
phery of the placenta. At one pole the vessel penetrates the 
placenta near its right margin; at the other pole it enters 
near the left margin. 

The data at my disposal do not permit of a direct com- 
parison of the placental circulation of the beaver with other 
types which have been investigated. Such comparisons 
should be made at equivalent stages. The circulation in the 
early established placenta will necessarily differ in its details 
from that in the mature finished placenta. As for the latter, 
excellent figures have been given by Tafani of the injected 
placente of various mammals, less successful perhaps as 
regards the rabbit, but remarkably clear as regards rat, 
guinea-pig, and bat. The placenta of the guinea-pig possesses 
its own special features. That of the rat (Mus decumanus), 
as represented in Tafani’s tav. v, fig. 2, 1s more like that of 
the bat (Vespertilio murinus, tay. vi, fig. 2) than the 
guinea-pig (Tav. v, fig. 1). In Mus and in Vespertilio 
there is a central maternal artery penetrating through the 
middle of the discoplacenta to its foetal aspect, where it 
spreads out into the afferent sanguimaternal lacune. 


To left and right of the placental insertion the areolated 
trophospongia merges imperceptibly into a marginal zone of 
hollow crypts, the walls of which are partly necrotic. In the 
obplacental and periplacental regions of the gestation sac the 
dermatic cells constitute an epithelioid tissue comparable to 
Nolf’s epithelioid layer in the bat. In the mesometric region 
the dermatic tissue, though proliferating, still retains a 
primitive aspect, and does not present an epithelioid mosaic 
pattern in section. 

Median sections through the placenta show that the der- 
matic tissue in which the vessels are lodged projects like a 


226 ARTHUR WILLEY. 


cone into the trophospongial cushion (PI. 21, fig. 78). After 
leaving the basal trophospongial sinus, which now appears as 
an arched line between the dermatic axis and the areolated 
cortical substance, the vesse!s branch and enter into numerous 
anastomoses with each other, forming a rete mirabile 
within the dermatic cone continuous with the trophospongial 
sinus, which is itself retiform. With a simple lens the 
trophospongial crescent with its peripheral zone of syncytia, 
its interstitial meshwork, and its basal sinus, can be seen to 
perfection following the contour of the dermatic cone with its 
rete mirabile. : 

Another series, cut nearly longitudinally through the gesta- 
tion sac, showed the placenta in a slightly more advanced 
condition—4"50 mm. in length, 3°75 mm. high. In tangential 
sections, i. e. such as do not pass through the central tropho- 
spongial cone, the allantoic mesoblastic villi penetrate deeply 
and tortuously into the embryotrophic cap of areolated 
trophospongia, so that allantoic islands, surrounded by their 
trophoblastic investment, appear in section amongst the areole. 

Towards the centre of the placenta, the concentric strata of 
which it is composed stand out very clearly. Beginning at 
the foetal aspect, there is first a narrow zone of allantoic 
mesoblast containing the superficial allantoic vessels; then 
the swollen ends of the trabecule dilated with the sangui- 
maternal lacune; thirdly, the labyrinth of anastomosing 
trabeculze ; fourthly, the border zone of symplasma; fifthly, 
the trophospongial areolz ; sixthly, the dermatic cone. Most 
of these parts are to be found in figures already referred to, 
and again in Pl. 21, fig. 79. Since the polar vessels enter 
the placenta near its right and left margins, neither of them 
is seen in a median section. Noteworthy is the abrupt 
transition from the areolated trophospongia to the dermatic 
tissue. 

The position of the embryo in the blastocyst conforms to 
the rule which applies to those Vertebrates in which there is 
a great disparity between the animal and vegetative poles, 
i.e. when an omphalon or a yolk-sac is present. It has been 


BLASTOCYST AND PLACENTA OF THE BEAVER. Papa 


mentioned above, and the figures show, that the umbilical 
membrane or area vasculosa stretches across the foetal sac 
midway between placenta and obplacenta. The body of the 
embryo, surrounded by its amnion, lies sideways in the 
exocceelom between placenta and umbilical membrane, with 
its right side towards the placenta and its left side towards 
the omphalon (Pl. 20, fig. 73, and Pl. 21, fig. 79). In the 
Sauropsida, at the gill-slit or branchiotrematic stage, the 
embryo also comes to he with its left side towards the yolk- 
sac. In the rabbit, at the same stage, the forebody of the 
embryo by reason of the cervical flexure projects into the 
umbilical vesicle surrounded by the proamnion, or, more pre- 
cisely, the proamniotic omphalopleure. In the beaver there 
is no trace of a proamnion at this stage; and when it is 
remembered that in the preplacental blastocyst the mesoblast 
is continuous within the circuit of the sinus terminalis, 
except at the notochordal contact, and that there exists 
already an anterior as well as a posterior median extension of 
the exoccelom, the formation of a proamnion at any stage 
would seem to be excluded. 

In Pl. 21, fig. 79, we see the amniotic membrane stretching 
between the umbilical membrane and the placenta. In the 
same figure the opening of the omphalomesenteric duct into 
the omphalon and that of the allantoic canal into the flattened 
allantois are indicated, though they do not occur actually on 
the same section. The allantoic canal communicates with 
the hypoblastic cloaca, into which the Wolffian ducts now 
open, and forms the hypoblastic axis of a mesoblastic stalk, 
accompanied by two arteries anda vein. At the distal end 
of the stalk the canal] widens out as the allantoic sac which 
is imbedded in the thickness of the mesoblast at the feetal 
surface of the placenta, causing no protuberance. ‘This 
flattened sac has a longitudinal extension of about 0°75 mm. 
During part of its course from the cloaca to the sac, the 
Jumen of the allantoic canal is occluded by cellular prolifera- 
tion, so that it becomes solid; this condition has been 
observed in two series. 


228 ARTHUR WILLEY. 


In one series (VIIa in the Utrecht catalogue) the omphalo- 
mesenteric duct has a partially occluded lumen; but in 
another series (V 8) the lumen is open, and contains maternal 
blood-corpuscles, which it conveys from the omphalon to the 
mid-gut. The presence of red blood-corpuscles in the mid- 
gut could not easily be attributed to accident, and it appeared 
at first a mystery how they came to be there. The clue to 
the mystery was found in the behaviour of the obplacental 
trophoblast, which at this stage consists of flattened megalo- 
karyocytes closely applied to the uterine wall; here and 
there they are greatly distended with erythrocytes, and there 
is evidence of the transfusion of maternal corpuscles into the 
omphalon. They pass into this cavity across the obplacental 
trophoblast, and are in fact to be found scattered in the 
midst of the coagulum. 

I will conclude this chapter with some further details 
concerning the embryo in order to define the stage of develop- 
ment which it has now reached with greater precision. Its 
actual age cannot be estimated, and can only be roughly 
guessed at. If we assume the period of gestation to be 
one hundred days, then these embryos will certainly fall 
within the first twenty-five days, and probably within the 
first fifteen days; the preplacental blastocysts are likely to 
belong to the first ten days. The relative age is best 
reckoned according to the size of the gestation sac. 

The embryo is now in the gill-slit or branchiotrematic 
stage; the mouth is open, but there is no anus. In section 
the mouth-cavity, flattened dorso-ventrally, appears like a 
pair of gill-slits, but this is entirely deceptive, since the 
formation is that of the stomodeum with its pituitary 
diverticulum meeting the blind end of the infundibulum. 
The four pairs of true gill-pouches just fail to open to the 
exterior, being closed over externally by a narrow cellular 
bridge. The auditory sacs are closed; the optic stalk is 
hollow, the optic cup and choroid fissure are formed, and 
the lens invagination is still connected with the ectoderm. 
Suitable sections show a length of the spinal cord flanked by 


BLASTOCYST AND PLACENTA OF THE BEAVER. 229 


somites resembling the corresponding figure of a 12 mm. pig 
embryo in C. 8. Minot’s ‘ Laboratory Text-book of Embryology,’ 
p. 230, fig. 185. The buds of the fore-limbs are rather more 
advanced than those of the hind-limbs, which appear as 
broad crescentic thickenings. 


XIII. InveRMepiate StaGces. 


Two stages, intermediate between the establishment of the 
growing placenta and the final period towards the term of 
gestation when it ceases to grow, were obtained, but have not 
yet been worked out in section. Their external characters 
present two points of special interest concerning the early 
relations of the chorion leve or diplotrophoblast of the 
umbilico-placental areas, and the appearance of a spherical 
allantoic vesicle. At the stage described in the preceding 
chapter, the allantois occurred as a narrow tube communi- 
eating proximally with the cloacal region and expanding 
distally into a flattened sac immersed in the allantoic meso- 
blast at the surface of the placental labyrinth, not causing 
any additional protuberance into the exoccelom. 

The next older stage is represented by three gestation sacs, 
one in the right horn of the uterus, two in the left. These 
were despatched to me in a tin containing 10 per cent. 
formalin on March 23rd. The swellings were equal, some- 
what shrunken, 60 mm. in long diameter. Upon cutting 
through the antimesometric wall, 2°50 mm. thick, the 
umbilical membrane (area vasculosa) was seen to be 
expanded and in contact with the mucosa, though devoid of 
attachment. Along the courses of the vessels low ridges are 
developed, which press against the mucosa, so that the 
impressions of the larger vessels are distinctly visible upon 
the smooth inner surface. When first opened the mucosa 
appeared dark reddish aud spongy, with white giant-cells 
showing against the ground colour when viewed under a 
simple lens. 


230 ARTHUR WILLEY. 


When the feetal sac was cut into, the umbilical membrane 
was found to be adherent to the amnion, from which it could 
be peeled off. In the mesometric division of the fcetal 
sac the amniotic membrane bends inwards to its insertion 
upon the dome of the placenta, and here it becomes free from 
the umbilical membrane, so that a considerable exoccelomic 
cavity exists in this region. Thus the baggy amnion hangs 
loosely in the mesometric exoccelom, but elsewhere adheres 
by simple contact to the area vasculosa, obliterating the 
exoccelomic cavity. A linea alba marks its insertion upon 
the placenta. 

By slicing away the wall of the gestation sac, which 
becomes thicker (5 mm.) towards the mesometrium, and then 
removing a portion of the vascular membrane, the fcetus, 
38 mm. long, is exposed, loosely enveloped by the spacious 
amnion. On the mesometric side, the area vasculosa ends 
with a villous rim at the two connecting membranes which 
surround the aditus of the interutricular segments. In this 
region the mucosa presents a rugose or convoluted surface, 
and the folds are continued as opaque whitish buttresses 
from the mucosa to the connecting membranes, extending as 
far as the edge of a narrow clear zone below the villous rim 
of the umbilical membrane. At this stage, therefore, simple 
dissection suffices to demonstrate the dual composition of the 
definitive connecting membrane (PI. 21, fig. 80). 

On laying open the whole cavity of the exoccelom the 
polar areas of the foetal sac are exposed from the inside. 
They cover over the aditus of the interutricular segments. 
The dissection was made shortly after the preservation in 
formol, and the polar areas (umbilico-placental areas) appeared 
as clear, non-vascular membranes forming turgid protuber- 
ances into the exocceelom near each end of the placenta, each 
being surrounded by part of the sinus terminalis. The 
pyriform caudad protuberance was 9 mm. high, 10 mm. long ; 
the convex cephalad protuberance was about 5 mm. high 
by 11 mm. long (Pl. 21, fig. 81). In another specimen 
opened several months later the polar areas were flattened 


BLASTOCYST AND PLACENTA OF THE BEAVER. Dene 


out. In fig. 81 the caudad end of the fcetal sac is also the 
caudad end of the gestation sac; the head of the embryo is 
directed towards that segment of the uterus which leads to 
the Fallopian tube. The cephalad protuberance lies over the 
opening or aditus of that segment, while the caudad protu- 
berance arches over the posterior opening. The placenta 
measured 33 mm. in length, 21 mm. in width, and 21 mm. in 
height. Besides its remarkable form and bold eminence, the 
most noticeable feature was the absence of an allantoic 
vesicle. The position of the allantoic sac at the placental 
surface was indicated by a flat membrane without fluid, 
situated beyond the centre of distribution of the vessels. In 
a second specimen from the same tract there was a very small 
incipient allantoic vesicle with a transparent wall. In both 
of these instances the head of the foetus was directed towards 
the oviduct. 


The next intermediate stage is represented by four gestation 
sacs, one right and three left, sent to me in 10 per cent. 
formalin on April 6th, having an average length of 81 mm. 
Soon after the arrival of the material I opened all the sacs 
by cutting out an obplacental piece from each, and punctured 
each vascular foetal sac, collecting the clear fluid that squirted 
out. The amnion was again closely adherent to the umbilical 
membrane, so that in puncturing the latter the amnion was 
also pierced, and the clear fluid, which appeared faintly straw- 
coloured en masse, was the liquor amnii. In the later 
stages, when the foetus has nearly attained its limit of growth, 
the amnion loses its adhesion to the umbilical membrane, so 
that the exoccelom then becomes the main cavity of the feetal 
sac. At the present stage the amniotic cavity is the main 
cavity of the foetal sac, the exocceelom, even to the base of 
the placenta, being temporarily obliterated. 

The wall of the gestation sac has now become reduced to a 
mean thickness of about 2 mm. _ In the first specimen 
examined the head of the foetus was again directed towards 
the Fallopian tube. The foetus as it lies in its amnion has a 


Tae ARTHUR WILLEY. 


length of 75 mm., measured from the crown of the head to 
the base of the tail. The flattened tail is not bent under the 
body, but simply curves round against the posterior end of 
the foetal sac, extending a little beyond the hind feet; it 
measured 15 mm. in length. The placenta now has a length 
of 42 mm. 

The most conspicuous difference between this and the 
preceding stage is the presence of a spherical allantoic vesicle, 
about 16 mm. in diameter, with thin pellucid wall, projecting 
boldly into the amniotic cavity. It les just in front of the 
centre of distribution of the allantoic vessels into the placental 
labyrinth (PI. 21, fig. 82). 

The three gestation sacs in the left uterus held their foetus 
in the following positions: In the one nearest to the ovary 
the head of the foetus is directed towards the cervix uteri; 
in the middle sac the foetus lies in the same position, with 
the head towards the cervix; in the last of the three, i.e. 
the one nearest to the vagina, the head of the foetus is 
directed towards the oviduct, the tail towards the cervix. 


XIV. Tat Connecting MEMBRANE. 


At the early placental stage represented in PI. 20, fig. 76 
and fig. 79, the umbilical membrane is invaginated about 
halfway into the omphalon and stretches like a diaphragm 
across the blastocyst, separating the omphalon from the 
exocceelom. The latter is limited towards the aditus of the 
gestation sac by a sheet of somatopleure or diplotrophoblast, 
which constitutes one of the umbilico-placental areas described 
in my former paper (1912). At the poles of the feetal 
sac, beyond the range of the placenta, this membrane 
stretches free across the cavum uteri within the circuit 
of the sinus terminalis (Text-fig. 5). 

From the outer edge of the sinus terminalis two other 
membranes arise—the umbilical membrane and the ompha- 
lopleure. The short free zone of omphalopleure intervening 
between the sinus and the uterine wall is the primordial 


BLASTOCYST AND PLACENTA OF THE BEAVER. 233 


connecting membrane. It is clearly derived from the 
adcarinal membrane of the preplacental blastocyst; it is 
what is left of this membrane after the peripheral encroach- 
ment of the primordial area vasculosa has ceased. It 
continues to grow in later stages, and, by accession of 
material from the uterine mucosa, becomes converted into 
the definitive umbilico-uterine connecting membrane. The 
latter is therefore a composite membrane of dual origin, 


TEXxT-FIG. 5. 


Section through region of one of the aditus of the gestation 
sac at the early placental stage. 1. Omphalon. 2. Umbilical 
membrane. 3. Exocelom. 4. Sinus terminalis. 5. Con- 
necting omphalopleure (representing adcarinal membrane). 6. 
Diplotrophoblast. 7. Aditus. 


as was indicated in the preceding chapter. This fact is 
proved by its early history as well as by its histological 
structure, into the details of which I will forbear to enter 
at present, although illustrations have been prepared. The 
manner of junction of splanchnopleure, omphalopleure and 
somatopleure at the level of the sinus terminalis in the 
early placental stage is shown in Text-fig. 6. 

From the earliest to the latest days of placentation the 
Sinus terminalis is situated at the junction of the three 
membranes whose primary names have just been given. Their 


234 ARTHUR WILLEY. 


secondary designations are respectively—umbilical mem- 
brane, umbilico-uterine membrane, and umbilico-placental 
membrane. 

The omphalopleure, as a whole, constitutes the inferior 
(anti-mesometric) wall of the omphalon and of the entire 
blastocyst. It is present in all its integrity during the early 
placental period but disappears in later stages, with the 
exception of the persistent connecting membrane. ‘The 
origin of the latter is thus explained. It is a secondary 
product of the primary obplacental implantation of the 


TEXT-FIG. 6. 


Part of periplacental wall of blastocyst, at the early placental 
stage, to show the junction of membranes at the sinus 
terminalis. 1. Cavity of omphalon. 2. Exocelom. 3. 
Cavum uteri. 4. Mucosa. 5. Sinus terminalis. 6. 
Splanchnopleure. 7. Omphalopleure. 8. Somatopleure. 


blastocyst, and directly comparable with the adcarinal 
omphalopleure of the preplacental blastocyst. The umbilical 
membrane forms the major part of the outer wall of the mature 
foetal sac. In my former paper on the beaver (1912, p. 207) 
I called it the vascular chorion or endochorion of the rodent 
blastocyst, not being at that time aware of the circumstance 
that the more suitable name, umbilical membrane, had been 
applied to it for the rabbit by E. van Beneden and C, Julin in 
1884. 

The umbilico-placental membrane is in its origin a non- 
vascular somatopleure or diplotrophoblast, but in the mature 
foetal sac there are intrusive capillaries proceeding into it 


BLASTOCYS! AND PLACENTA OF THE BEAVER. 235 


from the area vasculosa. They form anastomosing loops, 
and bear a resemblance to the capillary loops in the chorion 
leve of the bat’s blastocyst as figured by van Beneden and 
Julin. 

The reason for the existence of the connecting membrane 
which supplements the placental implantation in the attach- 
ment of the foetal sac to the gestation sac, may perhaps be 
sought in the semi-aquatic habits of the beaver and in the 
fact that the female retains her activity, leaving the lodge 
and swimming under the ice to procure food from the sub- 
merged stock of winter-provender, throughout the period of 
gestation. Were it not for the additional support afforded by 
the connecting membrane, the narrow deciduous root of the 
massive placenta might easily be torn asunder. The con- 
necting membrane must relieve the placenta of much of the 
stress and strain to which it would otherwise be exposed. 

The nearest approach to the condition of having a con- 
necting membrane seems to be represented by a temporary 
formation, which was described in the early gestation sac of 
Sciurus by F. Muller in 1905. It appears that the peri- 
placental mucosa forms a ring-shaped thickening to which 
the thickened trophoblast adheres before the folding of the 
amnion. In the following stage this ring-shaped zone otf 
implantation, with the sinus terminalis close to it, con- 
tinues to extend, leaving the placental part of the cavum 
uteri still unoccupied, but with numerous crypts opening 
into it. 

In the young gestation sac of Sciurus, when the blasto- 
cyst is attached to the obplacental wall, the cavum uteri, 
as had been previously observed by Fleischmann, becomes 
constricted by a circular periplacental thickening into a 
smaller mesometric and a larger anti-mesometric portion. 
The zone of adhesion or omphalo-placental ring (‘ ompha- 
loiden placentatiering,”’ F. Muller, op. cit., p. 395) occurs at 
the lip of the mesometric cavum. In this ring an epithelial 
syncytium is formed, with which the trophoblast becomes 
intimately united. At length the syncytium is destroyed and 


236 ARTHUR WILLEY. 


then the plasmodial union is exchanged for simple adhesion. 
Later still even this is given up. This early periplacental 
implantation occurs at the level of the sinus terminalis, 
and is interpreted by Muller as a vestigial omphaloid placenta- 
tion such as had not been described in rodents before. The 
periplacental connection has nothing to do with the allantoic 
placenta which forms subsequently. It extends below and 
thus embraces the openings or aditus of the gestation sac. 
Its limited centripetal growth and its transitory endurance 
are its most remarkable features. 

There can hardly be a doubt that the temporary omphalo- 
placental adhesion of Sciurus is comparable with the per- 
manent umbilico-uterine connection of Castor. 


XV. SPECIAL CONSIDERATIONS. 


The two leading characteristics of the preplacental blasto- 
cyst of the beaver, the obplacental implantation with differen- 
tiation of erythrocytophagous and leucocytophagous megalo- 
karyocytes, and the placental keel, are of specific physiological 
importance to the growing embryo, but they also have a 
comparative value which can only be elucidated by a brief 
discussion. In order that the reader may be orientated with 
regard to the systematic position of the beaver, a few words 
of introduction are desirable. 

Although the beaver, in structure and habits, is unique 
amongst the Rodentia, displaying in high degree the qualities 
of intelligence and adaptability to local conditions, yet it would 
be wrong to suppose that it is farthest removed from a primi- 
tive organisation. On the contrary, in its monotrematous 
cloaca and pentadactyle limbs, to mention these two external 
features only, it retains the marks of a very ancient mam- 
malian type. In my former paper (1912) it was suggested 
that “the beaver occupies a position amongst Rodentia com- 
parable with that attained by man amongst primates.” It is 
well known that man has retained some primitive features in 
limbs, teeth and digestive tract, as compared with many other 


BLASTOCYST AND PLACENTA OF THE BEAVER. 237 


mammals. Just as Hubrecht’s discovery of the blastocyst of 
Marsius with its “ Haftstiel ” indicated an excessively remote 
origin for the primate ontogeny, so the blastocyst of the 
beaver with its exostyle may have an analogous bearing upon 
rodent ontogeny. 

It is generally agreed, under the support of such authori- 
ties as Cope, Winge, and Tullberg, that the beaver family is 
related to the squirrel family, both of these families being 
associated under J. F. Brandt’s suborder, Sciuromorpha (1855). 
Tullberg divides the Simplicidentata into two great tribes, 
Hystricognathi and Sciurognathi. The latter comprises two 
sub-tribes, Myomorphi and Sciuromorphi ; and the latter falls 
into three sections, Sciuroidei, Castoroidel, and Geomyoidei. 
Winge (quoted by Tullberg) held that rodents are to be derived 
from primitive Mammalia resembling the least specialised 
Insectivora, from which they diverged by cumulative merease 
of the gnawing habit. 

F. Muller (1905), adopting Haeckel’s (1895) generalisations 
maintained that of all Rodents the Sciuromorpha have 
diverged least from the ancestral type; and he added that the 
genus Sciurus in particular occupies the most central position 
and has preserved the most primitive form. On the other hand, 
Max Weber (1892) regarded the scales on the tail of the beaver 
as the remains of a primitive scaly covering of the body. 


The preplacental blastocyst of Lepus and Sciurus is a 
plano-convex blastocyst, that of Mus and Cavia is an 
inverted blastocyst, that of Castor is an everted blastocyst. 
The task before us is not to decide which of these three is the 
most primitive type of blastocyst, but to consider which of 
them offers the readiest comparison with the blastocyst of 
Tarsius. It may be premised that in one respect the 
euplacental blastocyst of the beaver is the most primitive 
known amongst existing Rodents by reason of the persist- 
ence of the umbilico-uterine connecting membrane which 
is a consequence of the periplacental implantation of the 
trophoblast. 


238 ARTHUR WILLEY. 


The keel extends from end to end of the elongated balloon- 
shaped blastocyst, along its superior or mesometric side, 
dipping into the deep placental groove. One of the funda- 
mental relations of this remarkable formation is so obvious 
that its importance in establishing the reality of the pheno- 
menon might escape attention—namely, the coincidence in 
the configuration of the wall of the blastocyst and that of the 
cavum uteri or cavity of the gestation sac. The keel may 
thus be accepted at once as a fact, and need not be regarded 
with suspicion as an artefact. 

There are four or five principal structures concerned in the 
constitution of the keel: the epiblastic keel, the mesoblastic 
keel, the hypoblastic keel and the embryomic keel, to which 
may be added the exoccelomic keel. The interpretation and 
comparison with other forms will hinge upon the massive 
mesoblastic keel which follows behind the primitive streak. 
It is convenient to anticipate conclusions to some extent by 
assuming that the keel represents an ancient or primitive 
mechanism, and that the exserted mesoblastic keel of the 
beaver is comparable with the “ Haftstiel” of Tarsius, 
monkeys and man. 


History of the “ Haftstiel.”—So far as I have been 
able to ascertain, the first use of the term “ Haftstiel” as 
applied to the mammalian blastocyst occurs in Selenka’s 
memoir on the opossum (1887, vide his text-fig. C, on p. 136 
op. cit.). In this case the main cavity of the blastocyst is 
the omphalon, into which the allantoic vesicle, surrounded by 
a narrow exoccelom, hangs freely. In the early Primate 
blastocyst the main cayity is the exoccelom, into which the 
reduced omphalon hangs freely. These conflicting relations 
depend upgn the varying degrees of development of the 
exoccelom, allantois and omphalon respectively. In the 
opossum the embryo with its allantoic vesicle and exoccelom 
is suspended from the chorion or wall of the foetal sac into 
the main cavity (omphalon) by a hollow exoccelomic stalk. 
In the Primates the embryo with its umbilical vesicle is sus- 


BLASTOCYST AND PLACENTA OF THE BEAVER. 239 


pended from the chorion into the main cavity (exoccelom) by 
a massive allantoic stalk. There is no massive allantoic 


mesoderm in the opossum and no hollow allantoic vesicle in 


ce P] 


man. Just as the term “chorion” is often used to denote 
the wall of the foetal sac, irrespective of the constitution of 
its membranes, so the term “ Haftstiel” denotes the mecha- 
nism by which the embryo is suspended within the cavity of 
the foetal sac. 

The real knowledge of the Primate “ Haftstiel”’ dates from 
His’s third memoir on the ‘Anatomie Menschlicher Hmbryonen ’ 
(1885). I have not seen the original memoir, but the subject 
is treated very fully by C. §. Minot in his contribution entitled 
“Uterus and Embryo” (1889). His made what Minot 
appraised as ‘‘the discovery of fundamental importance ”— 
that in the early human blastocyst the allantoic sac appears 
as a small endodermic tube lying in a posterior prolongation 
of the body which His called the “ Bauchstiel,” and that at 
this early stage the allanto-chorionic vessels already run to 
and branch out upon the chorion. Thus “the allantois is, 
from the first, continuous with the chorion” (Milnes Marshall, 
1893). 

F. Keibel has stated recently (1913) that the allantoic tube 
in the human blastocyst is an entirely vestigial structure ; 
he does not say explicitly of what it is a vestige, but the text 
imphes clearly that he considers it to be a vestige of the free 
allantoic vesicle of Sauropsida. The beaver may help us to 
another explanation, namely, that it is a vestige of that 
portion of the omphalon which descends into a keel-shaped 
“Haftstiel”” or exostyle. The latter term can be used in 
general as an equivalent rendering of ‘ Haftstiei,” its 
etymology being analogous to that of exoccelom, both terms 
referring to structures that lie outside the embryo. 

Selenka’s memoir on the “ Affen Ostindiens” appeared in 
1891. Inthis monograph it was shown that the characteristics 
of the monkey’s blastocyst are: (i) The early separation of 
the omphalon (‘ yolk-sac”’) from the chorion; it takes no part 
in the nutrition of the embryo and must be regarded as a 

voL. 60, parT 2.—NEW SERIES. Na 


24.0 ARTHUR WILLEY. 


vestigial structure ; it is scantily supplied with vessels and 
floats as a small stalked vesicle in the exoccelom, until by the 
expansion of the amnion it becomes pressed against the 
chorion and finally succumbs to resorption. 

(2) The spacious exoccelom acts as a reservoir of food-stuffs 
until the allantoic vessels take over the function of fcetal 
metabolism. 

(3) After the formation of the amnion (the method of 
which was not observed), the embryo retains its connection 
with the placental chorion by a solid cord of mesoblast into 
which a rudimentary allantoic tube penetrates. ‘lhis massive 
cord with its vestigial endodermic cavity is the allantoic stalk 
or “ Haftstiel,”’ the vehicle of the placental vessels. 

(4) The main cavity of the early blastocyst is the exoccelom. 
Subsequently the amniotic cavity dilates enormously and the 
amnion finally fuses with the chorion, so that the cavity of 
the foetal sac is then amniotic cavity. 

From the. above résumé it is obvious that Selenka’s 
‘‘ Haftstiel ” is identical with His’s ‘‘ Bauchstiel.” As an 
English equivalent the expression ‘‘ body-stalk” has been 
suggested by C. 8. Minot and adopted by J. W. Jenkinson ; but 
if the comparison with the beaver is accepted, the need for a 
more general term, e.g. exostyle, is indicated. 

We must now refer briefly to Hubrecht’s classical paper in 
Gegenbaur’s ‘ Festschrift’ (1896), entitled ‘ Die Keimblase von 
Tarsius.” The chief characteristics of the Tarsius blasto- 
cyst may be summarised : 

(1) Rauber’s layer, or the trophoblast over the formative 
epiblast, disappears in the very early stages, shortly after the 
delamination of the hypoblast. 

(2) The exoccelom has a remarkably precocious develop- 
ment and the hypoblastic sac occupies only a small portion of 
the spacious blastocyst. 

(3) The blastodisce is at first excentric. The ectoplacental 
proliferation of the trophoblast is situated some way behind 
the blastodisc, not diametrically opposite to it. 

(4) Between the ectoplacental proliferation and the posterior 


BLASTOCYST AND PLACENTA OF THE BEAVER. 241 


border of the blastodisc (embryonic shield), there extends a 
solid mesoblastic tract which at the same time forms part of 
the wall of the closed mesoblastic sac or exoccelom ; this solid 
cord is the “‘ Haftstiel.’ At this stage there is no mesoderm 
in the embryo and the region of the embryonic shield is still 
didermic. 

(5) The blastocyst is only attached to the uterine wall by 
the ectoplacental disc; otherwise it lies free in the uterine 
lumen. 

(6) A complete and close-meshed area vasculosa 
develops in the wall of the hypoblastic sac or umbilical 
vesicle, and is filled with blood-corpuscles long before the 
heart begins to beat; but vascular rudiments arise in the 
“ Haftstiel ” before the appearance of the omphaloidean net- 
work. 

(7) The mesoblast of the “ Haftstiel ” is at first contiguous 
with the adjacent trophoblast; but, pari passu with the 
formation of the amnion, it becomes separated from the 
trophoblast by insinuation of the exoccelom and so becomes 
converted into the primordial umbilical cord. 

(8) After the ‘“ Haftstiel”” has become vascularised and 
enlarged, a tubular outgrowth from the umbilical vesicle 
penetrates backwards into the connective tissue of the 
“Haftstiel.” This is the allantoic diverticulum of the 
umbilical vesicle. 

From this brief tabulation of Hubrecht’s discoveries re- 
garding the blastocyst of Tarsius, we gather that although 
the allantoic tube is a secondary outgrowth, it belongs to the 
omphalon. There is no question of its being a diverticulum 
of the hind-gut at its first origin. Again, the “ Haftstiel” 
is directly continuous with the primitive streak, and forms, 
at the beginning, part of the wall of the blastocyst behind the 
embryo. In 1889 Hubrecht, 4 propos of Hrinaceus, had 
defined the “ Haftstiel” as a caudal mesoblastic cord which 
grows backwards from the posterior end of the primitive 
streak in order to promote the early vascularisation of the 
chorion. To this definition Resink (1904) addedin italics: ‘‘ Der 


242 ARTHUR WILLEY. 


Haftstiel entsteht jedoch als die von Anfang an vorhandene 
Verbindung des entypierten Keimfeldes mit dem Chorion.” 

In the preplacental blastocyst of the beaver, the endo- 
dermic allantois does not appear as an outgrowth, but is 
actually a deep hypoblastic groove of the omphalon extending 
into the keel. It is, therefore, from the first an integral part 
of the omphalon. ‘lhe steps by which this hypoblastic 
groove becomes narrowed down to the allantoic canal with 
its distal flattened sac cannot be followed in the material at 
my disposal. It must take place simultaneously with the 
closure of the digestive tract of the embryo, which is 
similarly associated with the narrowing down of the omphalo- 
mesenteric duct. 

With reference to the above quotation from Resink, a pupil 
of Hubrecht, the important point in my estimation is his 
insistence upon the primary character of the “ Haftstiel,” 
which is not in its essence a secondary formation (cf. Milnes 
Marshall, 1893). The flattened “ Haftstiel ” of Tarsius may 
be directly compared with the keel-shaped exostyle of the 
beaver. Both of these structures are organs of fixation, by 
means of which the embryo attaches itself to the placental 
complex. The ectoplacental disc is derived from the 
proliferation of the exostylar trophoblast. The localisation 
of the “ Haftstiel” in Tarsius and the exostyle in the 
beaver affords strong confirmation of the view put forward 
by Minot in 1889, that the discoidal placenta is probably a 
primitive placental type. The chief characteristic of the 
blastocyst of Castoris the possession of a massive keel-shaped 
exostyle or “ Haftstiel” into which the endodermic allantois 
extends as a primary groove of the omphalon; and the 
exostyle itself is directly continuous with the proliferation 
of the primitive streak. In his useful, though doubtless 
ephemeral, speculations concerning the origin of the foetal 
annexes of Mammalia, Resink admitted that he was unable 
to explain the origin of the allantoic sac. Perhaps the 
beaver may offer a new point of view from which this 


problem may be envisaged. 


BLASTOCYST AND PLACENTA OF THE BEAVER. 243 


Within the limits of the order of the Rodentia, the con- 
ditions preceding the ectoplacental implantation of the 
beaver’s blastocyst are most readily compared with those 
obtaining in squirrels. According to Fleischmann (1893), in 
Spermophilus, where there is no preplacental keel, the 
gestation sac is none the less subdivided into an omphaloid 
cavum and a discoid placental cavum or “Scheibenhdhle.” 
The young blastocyst is attached in the omphaloid cavum, 
and is so orientated that the embryonic pole (blastodisc) lies 
over the narrow passage (called ‘‘ Schlossspalte”’) connect- 
ing the omphaloid cavum with the discoplacental cavum. 
The latter is comparable with the deep mesometric groove in 
the gestation sac of the beaver. 

The isolated position and high standing of the beaver 
amongst existing Rodents are prima facie evidence of an 
exceedingly remote origin. Other facts of organisation and 
paleontology are in harmony with such evidence. These 
circumstances lend peculiar significance to the character ot 
the preplacental blastocyst. 


We may now give sume attention to the behaviour of the 
inferior or obplacental hemisphere of the beaver’s blastocyst. 
Without greatly enlarging the scope of the illustrations 
accompanying this paper, it would be impossible to do 
justice to the wonderful scenes of substitution of trophoblast 
for uterine epithelium, which can be witnessed in almost 
every section. In the preplacental blastocyst of the hedge- 
hog, Hubrecht found a phagocytic trophoblast forming a 
complete trophosphere round the blastocyst. The ecto- 
placenta is a localised derivative of the trophoblast. Resink, 
however, applied the term “ectoplacenta”’ to the entire tropho- 
blast. On the present occasion I propose to confine my 
remarks to Rodentia. 

Obplacental ectodermal proliferations were described in 
the blastocyst of the rabbit by A. von Kolliker in 1882. 
According to Kolliker an event of fundamental importance 
for mammalian embryology was Rauber’s discovery in 1875 


24.4, ARTHUR WILLEY. 


‘of the so-called ‘‘ Deckschicht’’ outside the formative epi- 
blast of the rabbit’s blastocyst, showing that the traditional 
view till then maintained by Coste, Hensen, and Kolliker, that 
the wall of the monodermic blastocyst had the value of 
embryonic ectoderm, was wrong. Kolliker called Rauber’s 
“ Deckschicht ” the primitive ectoderm of the area embryo- 
nalis. Of next importance was Lieberkuhn’s demonstration 
in 1879 of the origin of the permanent or formative ectoderm 
of the rabbit out of the inner cell-mass which lies against 
the primitive ectoderm and splits into two layers—the 
permanent ectoderm and endoderm.  Kolliker’s primitive 
ectoderm of the rabbit’s blastocyst thus prepared the way for 
Hubrecht’s celebrated conception of the trophoblast of the 
mammalian blastocyst. In the extra-embryonic primitive 
ectoderm of the rabbit Kélliker discovered numerous ecto- 
dermal proliferations in the form of elongate, villiform eleva- 
tions due to local nuclear divisions. 

These papille were next described by E. van Beneden and 
C. Julin in 1884. On opening an eleven-day gestation sac of 
the rabbit under picrosulphuric acid, by a crucial incision 
through the antimesometric wall, a liquid escaped and a 
coagulum was produced. ‘This was the fluid from the blasto- 
dermic cavity, which becomes the umbilical or vitelline 
vesicle or ‘‘ yolk-sac.” The cavity had been opened because 
the obplacental wall of the blastocyst was intimately united 
to the uterine mucosa. The union between the wall of the 
umbilical vesicle and the mucosa is effected by means of the 
epiblastic buds, which, as the authors stated, Kolliker first 
noticed. These buds arise upon the whole surface of the 
inferior hemisphere of the blastocyst, commencing about the 
eighth day. The degeneration of the cells, whose prolifera- 
tion engenders the buds, begins at the ninth day, and leads 
to the rapid degeneration of the entire epiblastic membrane, 
which, at the fourteenth to fifteenth day, detaches in shreds 
from the mucosa. 

M. Duval (1890, pl. i, fig. 28) described and figured the 
bilaminar inferior hemisphere of the rabbit’s blastocyst at the 


BLASTOCYST AND PLACENTA OF THE BEAVER. 245 


tenth day, and observed the plasmodial swellings connected 
by intervening thin tracts, but he let them degenerate without 
having effected any useful purpose. 

R. Assheton (1895) also mentioned and illustrated (‘ Quart. 
Journ. Micr. Sci.,’ 37, Pl. 19, fig. 7) the obplacental papillee 
of the rabbit. He says the first attachment of the blastocyst 
takes place between the lower parts of the blastodermic 
vesicle and the periplacental and obplacental folds of the 
uterus. It is effected by means of epiblastic papillae about 
the eighth day. Assheton suggested that the papillae became 
wedged into the uterine epithelium by reason of hydrostatic 
pressure from within the vesicle. 

H. Schoenfeld (1902) made the first complete cytological 
study of the festoon-like adhesions of the preplacental blasto- 
cyst of the rabbit on the seventh and eighth days. At seven 
days the zona pellucida becomes broken on the antimeso- 
metric side of the blastocyst and multinucleated ectodermal 
thickenings have appeared. At 7 days 4 hours the ecto- 
dermal buds come into intimate adhesion with the uterine 
syncytium and push into it with pseudopodium-like processes 
so as to reach the maternal capillaries, thus fixing the blasto- 
cyst to the mucosa in a manner comparable to the rooting of 
a plant. From the seventh day of gestation until 8 days 
4 hours, the blastocyst remains attached to the mucosa on the 
obplacental side. At 8 days 6 hours numerous plurimitotic 
figures occur in the epiblast of the obplacenta. At 8 days 
22 hours Minot’s giant-cells appear in the obplacental mucosa. 
They are of fcetal origin, resulting from retrogressive 
changes of the obplacental multinucleated epiblastic buds, 
now embedded in the mucosa. At their origin they are much 
more voluminous than any elements belonging to the mucosa 
or submucosa. According to Maximow, they reach their 
maximum development, with a diameter of 100 yw, on the 
sixteenth to twentieth days. 

J. Rejsek (1904) describes the obplacental implantation of 
the very young blastocyst of the Souslik (Spermophilus 
citillus). It continues until the blastocyst acquires a 


246 ARTHUR WILLEY. 


diameter of 2 mm., when it is given up simultaneously with 
the incipient placentation, which coincides in time with the 
cervical flexure of theembryo. The obplacental implantation 
is effected by means of a single multinucleate trophoblastic 
plasmodial proliferation at the antimesometric pole, sending 
processes into the mucosa. The number of nuclei increases 
by amitosis from 22 in a blastocyst of 0°126 mm. to more 
than 500. ‘This obplacental implantation of the souslik’s 
blastocyst would seem to be a phenomenon sui generis, the 
initial plasmodial thickening corresponding to a single 
papilla of the rabbit’s blastocyst. 

F. Muller (1905) also observed the antimesometric attach- 
ment of the preplacental blastocyst of Sciurus (vide his 
pl. vii, fig. 8). 

For the characteristic features of the obplacental implanta- 
tion of the preplacental blastocyst of the beaver, I must 
refer the reader to the foregoing text. The obplacental 
trophoblast is still connected with the uterine wall at the 
early placental stage, so that the implantation endures for a 
much longer period than in the rabbit. Eventually, when 
the invagination of the umbilical membrane is completed and 
the obplacental omphalopleure has vanished, giant-cells 
remain in the mucosa, appearing as white specks peppered 
over the internal surface of the gestation sac, scattered or in 
groups, not evenly distributed. 

In the preplacental stage the trophoblast, at the areas of 
attachment, is not more than one layer in thickness and its 
cells remain distinct, though they may have several nuclei. 
It does not send processes to any depth into the mucosa, but 
asa tule it has a flat insertion upon the decidual surface, the 
maternal capillaries pressing towards the trophoblast rather 
than otherwise. Wherever the uterine epithelium is displaced 
by the obplacental trophoblast, it disappears without previously 
forming a syncytium. It is surprising to find the trophoblast 
planted like a pseudo-epithelium upon the decidual surface, 
and, at its borders, normal uterine epithelium. This probably 
means that the uterine epithelium becomes necrotic in 


BLASTOCYST AND PLACENTA OF THE BEAVER. 247 


advance of the trophoblastic attack and that the trophoblast 
does not kill normal epithelium. When the mesometric 
glands degenerate, they do not form syncytia in the beaver, 
but the glandular epithelium becomes necrotic. The syncytia 
in the trophospongia at the early placental stage arise froma 
special proliferation of uterine epithelium, which takes place 
by mitosis in the mesometric region at the preplacental stage, 
before the ectoplacental proliferation of the trophoblast has 
set in. 


XVI. BrsLioGRAPHY. 


[The subjoined references are supplementary to the list given by me 
in 1912, and the items are not repeated here. Those marked with an 
asterisk have not been seen. | 


*1, Beneden, Edouard van.—* Sur les Placentas discoides,” ‘Comptes 
rendus Soc. Biol.,’ November 9th, 1888. 

and Charles Julin.—* Recherches sur la formation des 
annexes fctales chez les Mammiféres (Lapin et Cheiropteéres).” 
‘Archives de Biol.,’ v, 1884, pp. 369-434. 

- Bischoff, Theodor von.—(1) ‘ Entwickelungsgeschichte des Kanin- 
cheneies, Braunschweig, 1842. (2) ‘ Entwickelungsgeschichte 
des Meerschweinchens,’ Giessen, 1850. (3) ‘Neue Beobach- 
tungen zur Entwickelungsgeschichte des Meerschweinchens,’ 
Miinchen, 1866. 

. Bonnet, Robert.— Lehrbuch der Entwickelungsgeschichte,’ 2° Aufi., 
Berlin (Paul Parey), 1912. 

. Chipman, Walter.—‘ Observations on the Placenta of the Rabbit. 
with Special Reference to the Presence of Glycogen, Fat and 
Tron,” ‘Studies from the Royal Victoria Hospital, Montreal, 
vol. i, No. 4, 1902, 261 pp., 8vo. 

Fernandez, Miguel. Beitriige zur Embryologie der Giirteltiere. 
I. Zur Keimblatterinversion und spezifischen Polyembryonie der 
Mulita (Tatusia hybrida Desm.),” ‘Morph. Jahrb., xxxix, 
1909, pp. 302-333. 

7. Fleischmann, A.—‘ Embryologische Untersuchungen, III,” ‘ Die 
Morphologie der Placenta bei Nagern und Raubtieren,’ Wies- 
baden, 1893. 

*8. His, W.—‘ Anatomie menschlicher Embryonen,’ i, 1 
ili, 1885. 


2. 


* 


ie) 


nN 


On 


& 


io 6) 
CO 
He 
pas 
— 
19 
19 6 
tho 


248 


ARTHUR WILLEY. 


9. Hubrecht, A. A. W.—(1) “The Placentation of Erinaceus euro- 


pus, with Remarks on the Phylogeny of the Placenta,” ‘Quart. 
Journ. Mier. Sci.,’ 30, December, 1889, pp. 283-404, London, 1890. 
(2) “The Placentation of the Shrew (Sorex vulgaris L.),” op. 
cit. 35, 1894, pp. 481-537. (3) “ Die Keimblase von Tarsius: 
Kin Hilfsmittel zur schirferen Definition gewisser Saugetier- 
ordnungen,’ ‘Gegenbaur’s Festschrift, pp. 148-178, Leipzig, 
1896. 


10. Jenkinson, J. W.—(1) “ Observations on the Histology and Physio- 


logy of the Placenta of the Mouse,” ‘ Tijdschrift d. Ned. Dierk. 
Vereen.” (2), vii, 1902, pp. 124-198. (2) ‘ Vertebrate Embryology,’ 
Oxford (Clarendon Press), 1913. 


*l1. Jhering, H. von.—(1) “ Uber die Fortpflanzung der Giirteltiere,” 


‘Sitzungsber. Kon. Preuss. Akad. Wiss.,’ Heft 47, 1885, p. 105. 
(2) “Uber Generationswechsel bei Saugetieren,’ ‘Arch. f. Anat. 
u. Physiol. Phys. Abth.,’ 1886, pp. 442-450. 


12. Julin, Charles.—See Beneden, E. van, and C. Julin. 


13. Keibel, Franz.—* Die Entwicklungsgeschichte der Wirbeltiere,” 


pp. 333-398, in “ Zellen- und Gewebelehre, Morphologie und 
Entwicklungsgeschichte, II,’ ‘Zoologischer Teil. Leipzig 
(Teubner), 1913. 


14, Kolliker, Albert von.—* Die Entwicklung der Keimblatter des 


Kaninchens,” ‘ Wurzburg Festschrift,’ Bd. i, Leipzig, 1882. 


15. Lieberkiihn, Nathanael. Ueber die Keimblitter der Siugethiere,”. 


Marburg, 1879. [Abstract in ‘Zoologischer Jahresbericht’ for 
1880, Leipzig, 1881}. 


16. Maximow, Alexander.—* Die ersten Entwicklungstadien der Kanin- 


chen Placenta,” ‘Arch. f. mikr. Anat.,’ lvi, 1900, pp. 699-740. 


17. Minot, Charles Sedgwick—‘“ Uterus and Embryo: I, Rabbit. 


IT, Man,” ‘ Journ. Morph.,’ ii, 1889, pp. 341-462. 


i8. Muller, F.—‘‘ De Wederzijdsche Verhouding tusschen ei en uterus 


bij de Knaagdieren meer in het bijzonder bij Sciurus vulgaris,” 
‘Tijdschrift d. Ned. Dierk. Vereen.’ (2), ix, 1905, pp. 329-585. 


19. Newman, H. H., and J. T. Patterson.—(1) “A Case of Normal 


Identical Quadruplets in the Nine-banded Armadillo, and its 
Bearing on the Problem of Identical Twins and of Sex 
Determination,” ‘ Biol. Bull.,’ xvii, 1909, pp. 181-187. (2) “ The 
Development of the Nine-banded Armadillo from the Primitive 
Streak Stage to Birth; with Especial Reference to the Question 
of Specific Polyembryony,” ‘ Journ. Morph.,’ xxi, 1910, pp. 359- 
423. 


BLASTOCYST AND PLACENTA OF THE BEAVER. 249 


20. Nolf, Pierre.—* Etude des modifications de la muqueuse utérine 
pendant la gestation chez le murin (Vespertilio murinus),”’ 
‘Archives de Biol.,’ xiv, 1896, pp. 561-693. 

21. Patterson, J. T.—See Newman and Patterson. 


*22. Rauber, August.—‘ Ueber die erste Entwicklung des Kaninchens,”’ 
‘Sitzungsber. Naturf. Gesell.,’ ii, pp. 103-109, Leipzig, 1875. 

‘28. Rejsek, J.—‘Anheftung (Implantation) des Siaugetiereies an die 
Uteruswand, insbesondere des Eies von Spermophilus 
citillus,” ‘Archiv f. mikr. Anat.,’ lxiii, 1904, pp. 259-273. 

24. Resink, A. J—‘‘ Die Stammentwickelung der embryonalen Organe,” 
‘Tijdschrift d. Ned. Dierk. Vereen.’ (2), viii, pp. 159-201, Leyden, 
1904. 

25. Schoenfeld, H.—‘ Contribution 4 l'étude de la Fixation de uf 
des Mammiféres dans la cavité utérine, et des premiers Stades de 
la Placentation.” [Rabbit and Dog.] ‘Archives de Biol.,’ xix, 
1903, pp. 701-830. 

26. Selenka, Emil.—(1) ‘Studien iiber Entwickelungsgeschichte der 
Thiere. Heft. 4, “Das Opossum (Didelphys virginiana),’ 
Wiesbaden, 1887. (2) Id., Heft 5, 1° Halfte, No. 4, “ Affen 
Ostindiens,’ Wiesbaden, 1891. 

27. Tafani, P. A—‘Sulle condizioni uteroplacentari della vita fetale,’ 
Firenze, 1886. 

28. Tullberg, Tycho.—‘ Ueber das System der Nagethiere,’ Upsala, 1899. 

29. Vernhout, J. H.—“ Uber die Placenta des Maulwurfs (Talpa 
europea L.),’ ‘Anat. Hefte, V,’ Heft 1, 1894, pp.3-49. 

30. Weber, Max.—* Beitrage zur Anatomie und Entwickelung des Genus 
Manis,” ‘Zoologische Ergebnisse,’ Bd. ii, pp. 1-117, Leyden, 1892. 
[See p. 15 and Taf. ii, fig. 17, for reference to beaver’s tail. | 

31. Willey, A.—‘‘ Foetal Membranes of the American Beaver (Castor 
canadensis),” ‘Zool. Jahrb., Suppl. XV, Bd. ii, pp. 191-218; 
‘Spengel’s Festschrift, Jena, 1912. [With analysis of the 
amniotic fluid by Professor R. F. Ruttan.] 


EXPLANATION OF PLATES 14 ro 21, 
Illustrating Mr. Arthur Willey’s paper on “ 'The Blastocyst 
and Placenta of the Beaver.” 


[Some of the drawings were reversed under the Edinger apparatus, 
and others were not reversed. It has not been thought necessary to 


250 ARTHUR WILLLEY. 


mention this in each case. Figures executed by Mr. John Prijs are 
accredited to him under the explanations. | 


PLATE 14. 

Fig. 1.—Section across middle of gestation sac with blastocyst in 
sitt. The darker shading round the cavum uteri indicates maternal 
proliferation. The upper circle surrounds the erypt-like mouth of a 
gland near which there was a trophoblastic implantation. The lower 
circle surrounds the keel which projects into the very deep placental 
groove. (J. Prijs, del.) 

Fig. 2.—Portion of pericarinal zone to show isolated intrusions of 
trophoblast. 1. Hypoblast. 2. Trophoblast. 3. Interstitial substance 
different from the coagulum within the omphalon. 4. Uterine epithe- 
hum. 5. Subepithelial capillary vessel full of red blood-corpuscles. 
The cytoplasm of the uterine epithelium stains darkly with orange G. ; 
that of the trophoblast lightly. 

Fig. 3.—Fragment showing phagocytic attack on the part of megalo- 
karyocytes. 1. Uterine epithelium undergoing destruction. 2. Mucosa. 
3. Capillary. 4. Megalokaryocytes. 5. Coagulum in cavum uteri. 
6. Portion of gland. (J. Prijs, del.) 

Fig. 4.—Portion of coronal region, showing complete substitution of 
trophoblast for uterine epithelium on each side of the mouth of a gland 
which retains its epithelium intact and contains groups of leucocytic 
granules in its cavity. The abruptness of the cellular changes is not 
exaggerated. Note the flat contact between trophoblast and denuded 
surface of mucosa. 1. Obplacental gland. 2. Obplacental trophoblast. 
3. Flattened trophoblast bridging mouth of gland. 4. Two trophoblast 
cells charged with maternal erythrocytes. 

Fig. 5.—Plan of regions of blastocyst and mucosa about the middle 
of the gestation sac. The hypoblast is indicated by a broken line 
(substage C). 1. Coronal region. 2. Pericoronal cavum. 3. Peri- 
omphaloid zone. 4. Pericarinal festoons. 5. Adcarinal omphalopleure. 
6. Keel. 7. Bottom of placental cavum at the mesometric side of 
gestation sac. 8. Cavity of omphalon. 9. Peripheral limit of mesoblast 
or sinus terminalis. 10. Obplacental glands. 11. Uterine epithe- 
lium. 12. Megalokaryocytes (trophoblast). 

Fig. 6 (substage A).—Keel in anterior region of blastocyst. 1. Peri- 
pheral trophoblast. 2. Hypoblastic groove. 3. Mesoblast. 4. Formative 
epiblast rounding edge of keel, showing scattered vesicular growing cells. 

Fig. 7—Same. Near front end of primitive streak. Formative epi- 
blast and mesoblast tinted. 1. Hypoblast doubled by unsplit mesoblast. 
2. Primitive streak. 


BLASTOCYST AND PLACENTA OF THE BEAVER. 745 | 


Fig. 8—Same. Wedge-shaped proliferation of formative epiblast in 
the primitive streak. 

Fig. 9—Same. A cleft appears in the centre of the primitive streak. 
- Fig. 10—Same. Embryonic area 50 » farther back. Cell outlines 
are visible in the formative epiblast, but not rendered in the drawing. 
Mitoses in the primitive streak. The asterisk (*) is placed at the limit 
of the formative epiblast on that side of the keel; there is a vesicular 
cell at that point. 

Fig. 11—Same. The peripheral limit of the carinal half of the 
formative epiblast is nearing the edge of the keel. An intermittent 
groove is seen upon the surface of the primitive streak. 

Fig. 12—Same. The peripheral limit of the carinal half of the 
formative epiblast has reached the edge of the keel. The mesoblast 
band of that side has shortened and become thicker. Another groove 
is seen at the primitive streak. 

Fig. 13.—Same. Plan of cavum uteri with contained blastocyst. 
Drawn freehand as seen under microscope. The other sections of this 
series were outlined under Edinger’s apparatus with slide reversed. 
The blastocyst is only attached at the pericoronal angles. The hypo- 
blast is indicated by a broken line. (1) Coronal uterine epithelium. 
(2) Coronal cavum with coagulum. (3) Leucocyte in coronal cavum. 
(4) Coronal chromatophile trophoblast. (5) Omphalon. (6) Embryonic 
shield with primitive streak and mesoblast. (7) Placental groove. 

Fig. 14.—Same. Massive tissue at posterior end of primitive streak. 
This is the commencement of the “ haftstiel”’ or “ exostyle ” formation. 

Fig. 15.—Same. The last of the mesoblast, a thin band on one side 
of the hypoblastic groove, behind the primitive streak. 


PLATE 15. 


Fig. 16.—Sub-stage B. Anterior region of blastocyst, lying free. 
(1) Omphalon. (2) Trophoblast. (3) Hypoblast. 

Fig. 17—Same. Farther back in the anterior region, the blastocyst 
still lying free. (1) Omphalon. (2) Peripheral trophoblast composed 
of megalokaryocytes. They are more numerous than in the figure. (3) 
Hypoblast passing over the cavity of the trophoblastic keel. (4) Keel 
composed of cubical trophoblast, its walls agglutinated distally. (5) 
Chromatophile megalokaryocytes. «. Peculiar plasmodial effect at the 
pericarinal region, overlapping the cubical trophoblast. Semi-dia- 
grammatic. 

Fig. 18.—Posterior half of embryonic region showing deep groove 
over the primitive streak. Formative epiblast tinted dark, mesoblast 


952 ARTHUR WILLEY. 


lighter. The half of the formative epiblast which surrounds the keel 
is longer and thinner than the portion which lies above the groove in 
the figure, i.e. it is more stretched out, the proportion being 550 pu to 
350 nu. The epithelial change at (*) consists in this, that the nuclei of the 
peripheral trophoblast lie at one level instead of at different levels, and 
they are usually stained paler than those of the formative epiblast. 
There is a potentiai linear celom indicated at the top of the massive 
carinal mesoblast on the left of the figure. 


Fig. 19. Same. Carinal region in the posterior part of the blasto- 
cyst showing relations of trophoblast (black), mesoblast (pale), and 
hypoblast. 


Fig. 20.—Substage C. The didermic keel in the posterior region of 
the blastocyst lying freely between the walls of the placental groove. 
Subepithelial maternal capillaries indicated in black. (J. Prijs, del.) 

Fig. 21—Same. The tridermic keel in the posterior region. (1) 
Posterior end of mesoblast. (2) Loose hypoblast. 

Fig. 22—Same. The keel with the posterior exocelom. The right 
side of the figure is the embryonic side of the keel. The slide was 
reversed under Edinger’s apparatus. (1) Solid proliferation marking 
the peripheral limit or sinus terminalis of the mesoblast. (2) 
Hypoblast (represented by broken line). (3) Posterior exoceelom. 


Fig. 23.—Same. Portion of obplacental trophoblast in the coronal 
region with adjoining part of the cavum uteri and uterine epi- 
thelium. (1) Uterine epithelium (nuclei omitted). (2) Coagulum in 
cavum uteri. (3) Trophoblast. (4) Leucocytes in mucosa crossing 
uterine epithelium, lodged in uterine coagulum, and entering the 
trophoblast. 

Fig. 24—Same. Plan of cavum uteri and blastocyst at the level 
of fig. 23. A band of slime extends from the edge of the keel nearly 
to the bottom of the placental groove. (1) Coronal epithelium. (2) 
Coronal cavum. (3) Coronal trophoblast. (4) Omphalon. (5) Keel. 
(6) Slime. 

Fig. 25——Same. Part of obplacental region showing flattened 
megalokaryocytes with ingested leucocytic nuclei appearing as dark- 
stained granules, sometimes surrounded by a vacuole. (1) Uterine 
epithelium. (2) Coagulum in cavum uteri. (3) Chromatophile 
trophoblast. 


PLATE 16. 


Fig. 26.—Same. Section through keel in region of massive exostyle. 
The cubical epiblast changes its character abruptly, so that the exo- 
style proper is not covered by a cubical epithelium. (1) Sinus 


BLASTOCYST AND PLACENTA OF THE BEAVER. Zoo 


terminalis on embryonic side. (2) Reduced exocelom. (3) Hypo- 
blastic groove. (4) Exostyle. (5) Uterine epithelium. 

Fig. 27.—Same. Section through exostyle in siti, showing detached 
‘shreds. The piece of uterine wall to the right of the figure is drawn 
nearer to the keel than it was in reality when seen under the 
“Zeichenapparat.” The position of the nuclei opposite the lower shred 
seems to indicate traction or stress. (1) Gland opening into placental 
cavum. (2) Space at apex of keel. (3) Placental cavum. (4) Uterine 
epithelium on wall of placental cavum. (5) Section of maternal capillary. 

Fig. 28.—Same. Section through middle of exostyle, which is here 
nearly solid throughout, with a linear interrupted cavity towards the 
apex. (1) Exocelom. (2) Hypoblastic groove. (3) Cavity of exostyle 
shut off at this level from (2). (4) Uterine epithelium. (5) Slime at 
apex of exostyle. 


Fig. 29.—Same. Anterior region of exostyle, showing central cavity 
continuous with the omphalon. (1) Omphalon. (2) Sinus terminalis. 
(8) Exoccelom on embryonic side of keel. (4) Cavity of exostyle. (5) 
Uterine epithelium. (6) Loose hypoblast. 


Fig. 30.—Same. The keel in the region of the primitive streak and 
formative epiblast. The asterisks mark the limits of the formative 
epiblast or embryonic shield. (1) Exoccelom. (2) Primitive streak. 
(3) Hypoblastic cavity. (4) Sinus terminalis; unsplit mesoblast 
extends a short way beyond it. (5) Anti-embryonie trophoblast. (6) 
Anti-embryonic unsplit mesoblast. N. B.—On the embryonic side the 
sinus terminalis lies at a higher level, out of the range of the 
drawing. 

Fig. 51—Same. The keel in front of the primitive streak. (1) Peri- 
pheral trophoblast. (2) Parts of the celom. (3) Proximal groove on 
embryonic face of keel. (4) Hypoblastic cavity. (5) Anti-embryonic 
part of the marginal proliferation of mesoblast or sinus terminalis. 
(6) Epiblastic cavity of keel. 

Fig. 32.—Same. The keel at the level of the anterior limit of meso- 
blast, in front of the celom. (1) Sinus terminalis. (2) Hypoblastic 
cavity. (3) Carinal trophoblast. 

Fig. 33.—Same. Proximal groove of the epiblast in front of the 
mesoblast, showing proliferation with mitoses. 


TEI ANIM DH Alz/e 
Fig. 34.—Same. Keel in front of the mesoblast. 
Fig. 35.—Substage D. Section through keel in the post-stylar region 
cutting the exostyle tragentially. The ccelom appears in two discon- 
tinuous cavities, owing to the intrusion of the massive mesoblast. 


254 ARTHUR WILLEY. 


(Freehand.) (1) Sinus terminalis. (2) Posterior exocelom. (3) 
Exostylar mesoblast. (4) Unsplit mesoblast. (5) Hypoblastic groove. 


Fig. 36.—Same. The keel in the region of the exostyle. Fifty-five 
sections of 10, intervene between this section and that drawn in 
fig. 35; in this interval the exostylar mesoblast has been increasing in 
amount. Note the abrupt passage of cubical trophoblast into the 
flattened epiblast of the exostyle. The exostyle does not exhibit such 
a luxuriant development in this series as in others. (1) Sinus 
terminalis. (2) Exocelom. (3) Exostyle. (4) Hypoblastic groove. 
(Freehand.) 

Fig. 37.—Same. Section through keel in region of primitive streak. 
There are twenty-six sections of 10 between this and fig. 36. (Free- 
hand.) (1) Exocelom. (2) Pericardial primordium bounded by cubical 
epithelium outside, columnar on the inner side. (5) Regular hypoblast 
co-extensive with the formative epiblast. (4) Primitive streak. (5) 
Distal limit of formative epiblast, with abrupt epithelial transition. 

Fig. 38—Same. Anterior region of embryonic shield. In the next 
section the two sections of the ccelom fuse over the carinal hypoblast, 
which then ends in a tangentially cut mass of cells. (Freehand.) (1) 
Amniotic folds. (2) Carinal embryonic hypoblast. (3) Pericardial 
primordium. (4) Omphaloidean hypoblast. (5) Exocelom. There are 
thirty-nine sections of 10, between this and fig. 37. For position of 
embryonic shield with reference to the keel-formation, compare Pl. 14, 
figs. 6 and 7. 

Fig. 39.—Same. Anterior end of pre-embryonic ceelom with epiblastic 
keel-extension. Freehand. (1) Exocelom. (2) Broken line indicating 
omphaloidean hypoblast. 

Fig. 40.—Substage E. Section through the keel at the level of the 
anterior region of the embryonic shield. (1) Sinus terminalis. (2) 
Linear occluded exocelom. (3) Formative epiblast. (4) Pericardial 
primordium. (5) Cubical embryonic hypoblast. 

Fig. 41.—Same. Keel at level of notochordal primordium (1). The 
hypoblastic epithelium changes its character on each side of the noto- 
chordal contact zone. The asterisk denotes the proximal limit of 
embryonic shield. 

Fig. 42.—Same. Keel at the anterior part of the primitive streak (1). 
(2) Epiblastic keel. 

Fig. 43.—Same. Keel about middle of primitive streak. 


Fig. 44.—Same. Post-embryonic keel at the level of the anterior end 
of the exostyle. (1) Exostyle or “ Haftstiel” tissue. For the relations 
of the surface membrane at such places of close union, compare PI. 16, 
fig. 29. 


BLASTOCYST AND PLACENTA OF THE BEAVER. 


bo 
Or 
Or 


Fig. 45. 

Fig. 46—Same. Posterior end of exostyle. The whole extent of 
mesoblastic bands is shown. (1) Sinus terminalis. (2) Exoccelom. 
(3) Hypoblastic groove. (4) Massive mesoblast. (5) Exostyle. (6) 
Trophoblastic keel. 


Same. Mid-region of exostyle. 


PLATE 18. 

Fig. 47—Same. Region of post-stylar celom. (1) Post-stylar exo- 
celom. (2) Hypoblastic groove. 

Fig. 48.—Same. Posterior termination of mesoblast and exoccelom. 
The hypoblast has been reconstructed in the drawing. (1) Sinus 
terminalis. (2) Exocelom. (3) Uterine epithelium. 

Fig. 49.—Same. Didermic keel behind the mesoblast. The flattened 
epiblast cells graduate into flattened megalokaryocytes (not shown). 

Fig. 50.—Same. Towards posterior pole of blastocyst. The hypo- 
blast withdraws from the keel. (1) The last of the carinal hypoblast. 

Fig. 51.—Substage F. Section through entire gestation sac with 
blastocyst in sitt. (J. Prijs, del. (1) Omphalon. (2) Omphaloid 
cavum; the keel is seen to the left of the index number. (3) Placental 
groove; to be compared with Pl. 14, fig. 1. (4) Placental tropho- 
spongia. (5) Periplacental wall; the fine tubes in the walls are partly 
glands, partly capillaries. (6) Obplacental wall. (7) Mesometrium. 

Fig. 52.—Same. Appearance of the keel in the anterior polar region 
of the blastocyst. About 90 sections of 10 precede this in the series. 
(1) Commencement of mesoblast on embryonic side. (2) Hypoblastic 
groove. (3) Trophoblastic cavity. (4) Trophoblastic keel. 

Fig. 53.—Same. Keel at level of anterior part of embryonic shield. 
In an earlier section the pericardium communicated at its upper end 
with the exocclom. (1) Sinus terminalis; the exocclom, real or 
virtual, extends up to the sinus terminalis on each side, but the solid 
mesoblast may go a short distance beyond it. (2) Amniotic fold. (3) 
Embryonic hypoblast. (4) Formative epiblast. (5) Pericardial pri- 
mordium. (6) Distal exoccelom. 

Fig. 54.—Same. The pericardial primordium is separating into two 
parts. (1) Proximal moiety of pericardium. (2) Distal moiety of same. 

Fig. 55—Same. The pericardium consists of right (proximal) and 
left (distal) halves. (1) Proximal half of pericardium. (2) Distal half 
of same. (3) Sinus terminalis. 


PLATE 19. 
Fig. 56.—Same. Posterior quarter of embryonic shield. (1) The 
two limbs of the pericardial primordium. (2) Notochordal plate. (3) 
VoL. 60, PART 2,—NEW SERIES. 18 


256 ARTHUR WILLEY. 


Medullary groove. (4) Sinus terminalis. (5) Unsegmented em- 
bryonic mesoblast. 

Fig. 57—Same. Near posterior end of embryonic shield. (1) Noto- 
chordal plate in front of the primitive streak. (2) Coagulum in hollow 
at foot of keel. 

Fig. 58.—Same. Through the primitive streak. (1) Primitive streak. 
(2) Dense coagulum in hollow at foot of keel. (38) Hypoblastic groove. 
(4) Exoccelom. 

Fig. 59.—Same. The keel with the massive exostyle. (1) Exocelom 
on one side only; the sinus terminalis of this side lies beyond the 
range of the drawing. (2) Axial hypoblast. (3) Distal trophoblastic 
cavity. (4) Epiblastic folds. (5) Sinus terminalis. (6) Proliferating 
mesoblast of exostyle; this is typical ‘* Haftstiel” tissue. 

Fig. 60.—Same. Transition from exostylar to post-stylar region. (1) 
Exocelom. (2) Hypoblastic groove. (3) Massive mesoblast. (4) 
Trophoblastic extension of keel. 

Fig. 61—Same. Anterior part of post-stylar region. (1) Sinus 
terminalis. (2) Hypoblastic groove with cubical epithelium. (3) 
Post-stylar eelom. (4) Trophoblastic keel, somewhat folded. 

Fig. 62.—Same. Middle of post-stylar region. (1) Post-stylar ccelom. 
(2) Very long and straight trophoblastic keel. (3) Mucus at edge of keel. 

Fig. 63.—Same. Near the posterior end of the post-stylar region. 
(1) Sinus terminalis. (2) Post-stylar celom. (3) Trophoblastic 
keel with agglutinated walls. 

Fig. 64—Same. The two portions of the sinus terminalis nearly 
meeting below the hypoblast. (1) Sinus terminalis. (2) Straggling 
mesoblast cells. 

Fig. 65.—Same. Posterior extremity of mesoblast; the sinus ter- 
minalis confluent below the hypoblast. 


PLATE 20. 


Fig. 66—Same. Posterior didermic region of blastocyst. The 
cubical carinal hypoblast passes gradually into the flattened adcarinal 
hypoblast. (1) Adcarinal omphalopleure. (2) Carinal omphalopleure. 

Fig. 67.—The lumen of a non-gravid segment of the uterus from a 
cornu which contained young embryos with established placenta. The 
pits are the openings of uterine glands. The dermatic cone on the 
mesometric side is free from glands. (J. Prijs, del.) This is the stage 
when the connective tissue is uniformly proliferated round the lumen, 
but this zone containing numerous subepithelial capillaries is not ren- 
dered in the drawing. 


BLASTOCYST AND PLACENTA OF THE BEAVER. 257 


Fig. 68.—Part of section through the posterior end ofja preplacental 
gestation sac to show the general aspect of the mesometric glands 
which have become atrophic. They reach the epithelial surface though 
devoid of a lumen. (J. Prijs, del.) 


Fig. 69.—Part of section through a preplacental gestation sac showing 
epithelial proliferations at the sides and bottom of the placental groove. 
Between the lobular growths capillaries are lodged. Capillaries are 
also seen ascending towards the surface. (Freehand.) 


Fig. 70.—Detail of epithelial proliferation at bottom of placental 
groove. A few mitoses are seen. 

Fig. 71.—Epithelial mesometric wings (tinted) with dilated capillaries 
charged with red corpuscles. Substage F. (1) Bottom of placental 
groove. (2) Dermatic cone. (3) Dilated capillary. (4) Epithelial 
proliferation. 

Fig. 72.—Ampulla of a periplacental gland failing to open into the 
placental groove. From substage B. 

Fig. 73.—Gestation sac opened from the antimesometric side showing 
embryoin sittt. (J. Prijs, del.) 

Fig. 74.—The same viewed from the side to show how the adjacent 
segments of the uterus are inserted into the mesometric half of the 
gestation sac at this stage. (J. Prijs, del.) 


Fig. 75.—Detail from the border-zone between trophospongia and 
placental labyrinth, showing chain of syncytial nuclei (black). (1) 
Trophoblast. (2) Maternal capillaries containing red corpuscles. (3) 
Fetal capillaries containing nucleated red corpuscles and surrounded 
by allantoic tissue. To right of figure the latter appears as an island 
surrounded by trophoblast. (4) Symplasma with many syncytial 
nuclei and a few intrusive plasmodial nuclei (seen to right of figure). 
(5) Trophospongia. One maternal capillary is included, the other is 
being included within the trophoblastic labyrinth. 


Fig. 76.—Semidiagrammatic section through blastocyst and placenta 
between one end and the centre of the latter. The hypoblast, indicated 
by broken line, becomes indistinct in the obplacental region. The 
height of the blastocyst, measured from the fcetal surface of the placenta 
to the obplacental wall, was 8mm. (1) Omphalon filled with dense 
coagulum, 7 mm. wide. (2) Obplacental wall of blastocyst, applied 
to obplacental wall of uterus. (3) Umbilical membrane (area vascu- 
losa). (4) Sinus terminalis. (5) Umbilico-placental membrane 
(diplotrophoblast). (6) Opening (aditus) of interutricular segment of 
uterus into gestation sac. (7) Section of embryo in its amnion in 
the posterior region. (8) Placenta showing diagrammatically radial 
allantoic villi (tinted) alternating with radial trophoblastic trabecule 


258 ARTHUR WILLEY. 


excavated by sanguimaternal lacune. Height of placenta, 3°5 mm. 
(9) Areolated trophospongia with capillary network and subjacent 
capillary vessels in the dermatic tissue. (10) Primordial connecting 
membrane. 


Fig. 77.—Section through placenta near one of its poles to show the 
polar vessel entering it excentrically. (J. Prijs, del.) (1) Polar vessel 
with proliferated endothelium. (2) Diplotrophoblast. (8) Areolated 
trophospongia. (4) Sinus terminalis. (5) Placental labyrinth. (6) 
Allantoic mesoblast at foetal aspect of placenta. 


PLATE 21. 


Fig. 78.—Section through middle of placenta, showing the ecto- 
placental labyrinth resting upon the areolated trophospongia, which is 
supported by the dermatic cone containing branching and anastomosing 
vessels. (J. Prijs, del.) 


Fig. 79.—Diagram of sagittal-section through gestation sac. The 
amniotic cavity extends between the umbilical membrane and the 
placenta. (1) Opening of omphalomesenteric duct into the omphalon. 
(2) Sinus terminalis; the index numbers are in the exocelom. (3) 
Amnion. (4) Allantoic canal opening into a flattened allantoic sac. 
(5) Umbilico-placental diplotrophoblast ; the index numbers are at the 
openings of the gestation sac. (6) Mesometrium. (7) Omphalo- 
mesenteric vein. (8) Omphalomesenteric artery. (9) Omphalopleuric 
primordium of connecting membrane. 

Fig. 80.—Fetus of 38 mm. lying in its baggy amnion, exposed by 
cutting away the wall of the gestation sac and removing part of the 
umbilical membrane. (1) Cut surface of wall of gestation sac, 5°25 mm. 
thick behind. (2) Cephalad connecting membrane. (3) Rugose roof of 
gestation sac. (4) Caudad connecting membrane. (5) Linea alba, 
marking the insertion of the amniotic membrane upon the placenta. 
(6) Cut edge of umbilical membrane. Note: At this stage the foetus 
already possesses two pairs of mammary pits. 

Fig. 81—Same. The fetus has been removed, and the amnion, 
detached from its adhesion to the umbilical membrane, has been cut 
short and drawn taut, the object being to exhibit the polar umbilico- 
placental protuberances. (1) Cut edge of amnion. (2) Outer surface 
of umbilical membrane; the inner surface is shaded. (3) Cut edge of 
ditto. (4) Linea alba or insertion of amnion upon placenta. (5) 
Cephalad protuberance. (6) Caudad protuberance. 


Fig. 82.—Later stage; fetal length 73 mm. View of placenta and 
allantoic vessels from the side after removal of the foetus. There are 
three layers of vessels at the surface of the placenta; superficial allan- 


BLASTOCYS! AND PLACENTA OF THE BEAVER. 259 


toic veins (black) ; allantoic arteries (clear) in the middle level; maternal 
vessels (tinted) at the lowest level. The main allantoic vascular trunks 
are two arteries (clear) and one vein (black). The amnion, together with 
the vascular area (umbilical membrane), has been cut open and reflected. 
The vessels of the area show through the amnion, but this is only 
indicated in the figure at the region of distribution of the omphalo- 
mesenteric vessels. (1) Cut edge of amnion together with umbilical 
membrane. (2) Inner surface of amnion. (3) Allantoic vesicle with 
vein in the thin wall. (4) Surface of placenta projected into the 
amnion. (5) Cut surface of umbilical cord. (6) Omphalomesenteric 
vein. (7) Omphalomesenteric artery. 


ACROSSOTA LIPOSCLERA. 261 


On Acrossota liposclera, a New Genus and 
Species of Alcyonarian with Simple Tentacles. 


By 


Gilbert C. Bourne, M.A., D.Sc., F.R.S., 


Linacre Professor of Comparative Anatomy and Fellow of Merton 
College, Oxford. 


With Plate 22. 


Tue Alcyonarian Colony to be described in the follow- 
ing pages was collected by Prof. A. Willey in the 
d’Entrecasteaux group of islands, British New Guinea, and 
sent to me, | am ashamed to say how many years ago, along 
with a collection of Actiniz and Zoanthee. Partly because 
of the difficulty and uncertainty of identifying spirit pre- 
served actinians, partly because I had turned my attention 
to another group of animals, I have neglected this collection 
until, a short time ago, 1 began to prepare such specimens 
as were suitable for cutting into sections with a view of 
identifying and describing them. Among the Zoanthee I 
found a colony of small translucent polyps apparently grow- 
ing in a bunch attached to the fronds of a species of 
Halimeda. Some of these polyps were partly expanded and 
exhibited a few simple digitiform tentacles, and this 
character, together with their general appearance, suggested 
that they were non-incrusted Zoanthids. On closer examina- 
tion I found that the polyps arose at intervals from a long, 
sparsely branched adherent stolon, and that the apparently 
single colony consisted of several such stolons so closely 
intertwined that the polyps arising from them were closely 


262 GILBERT C. BOURNE. 


bunched together, giving the deceptive appearance of a 
compact encrusting colony. And, as I shall describe in 
detail, I found that the tentacles, though simple finger- 
shaped structures, are always eight in number, and that the 
structure of the polyps, in almost all other respects, presents 
the well-known characteristic Aleyonarian features. 

The entire absence of lateral pinnee on the tentacles of the 
autozooids of an Alcyonarian has not, to my knowledge, been 
previously recorded, and as my colony presents some other 
interesting characters which differentiate it from any 
described species, I must make a new genus for its reception. 
It seems necessary also to make the genus the type of a new 
family, for the possession of simple digitiform tentacles is a 
unique feature among the Alcyonaria. But, as will appear 
from what follows, the general characters of the colony and 
polyps place the new family in the order Stolonifera, in 
close juxtaposition to the Cornularide and Clavularide. 

The specimen may be described as follows : 


Order—Stotonirera, Hickson. 


Family—Acrossotide, the tenacles simple, digitiform, with- 
out lateral pinne. 


Genus—Acrossota! n. gen. 


Zooids borne at intervals on a simple sparingly branched 
radiciform adherent stolon; no spicules or calcareous 
skeleton of any kind; tentacles digitiform, without lateral 
pinne. 


Acrossota liposclera n. sp. 


Stolon subcircular where free, flattened where adherent, 
its cavity formed by a single solenium, but traversed by 
_mesogloeal trabecule ; the stolon branched and the branches 


1 d, without ; cpooowrdc, fringed. 


ACROSSOTA LIPOSCLERA,. 263 


intertwined but not forming a network by anastomosis. 
Zooids subcylindrical, of various lengths, frequently giving 
off stolonar outgrowths from their proximal moieties, the 
distal moiety of each zooid invaginable within the proximal 
moiety. Tentacles digitiform, completely invaginable. 
Outer wall of stolon and zooids strengthened by a gelatinoid 
supporting tissue formed by the ectoderm, and in addition a 
thin parchment-like external cuticle, the latter forming the 
chief supporting tissue in the basal parts of the zooids, the 
gelatinoid tissue more largely developed in the subtentacular 
region and in the stolon; the walls everywhere relatively 
thin, and not differentiated to form calices into which the 
zooids can be withdrawn. Stolon and zooids translucent and 
colourless in spirit. 

Locality, d’Entrecasteaux Group, British New Guinea. 

The form and habit of the specimen collected by Prof. 
Willey are so irregular that they are difficult to define. The 
central part of the specimen consists of several relatively 
large zooids, measuring 5 mm. in length and 1°75-2 mm. in 
diameter and rather closely crowded together. From the 
bases of these zooids several stolons are given off which creep 
along the frond of a Halimeda and are closely attached to it 
at intervals. Where attached the stolons are flattened and 
tape-like, but in the greater part of their courses they are 
round and simply twined round their support. The stolons 
branch and the branches may subdivide several times in an 
irregular manner, but the branches do not anastomose with 
one another. The branches and their subdivisions are twined 
round one another and tangled together, and bear other 
zooids at irregular intervals. From place to place a branch 
of the main stolon or a stolonar outgrowth of one of the 
zooids projects for some distance from the support as a long, 
free, thin-walled tube, near the end of which a zooid is 
developed, and from this zooid other stolons are given off, 
which may in turn bear zooids of various sizes. These free 
tubular outgrowths are frequently constricted in places, and 
it is probable that they are eventually separated off from the 


264 GILBERT C. BOURNE. 


parental colony at the points of constriction, and so give rise 
to new colonies, for the specimen included several short 
lengths of stolon bearing two, three, four or more zooids, and 
quite independent of the maincolony. But these and the stolons 
of the main colony were all tangled together and with the 
Halimeda, and it required some care to disentangle them. 
Acrossota, therefore, is remarkable, though not unique, 
amoung Alcyonaria for producing independent colonies by a 
form of gemmation. The stolons usually end in blunt, slightly 
swollen extremities, but sometimes are fixed by one extremity, 
which is then closely flattened against the surface of support 
(fig. 1). Evidently each stolon is at first a simple tubular 
outgrowth of the proximal part of the wall or of the base of 
a zooid ; this outgrowth is lined by endoderm and forms what 
I have called a “solenium.”! In Cornularia? the stolons 
are simple solenia, but in Clavularia, Sarcodictyon and other 
members of the Stolonifera the stolons are compound and 
contain several solenia. In Acrossota the stolons are not as 
simple as those of Cornularia, nor so complicated as those of 
Clavularia. The simple cavity is largely occluded, especially 
in the older parts of the stolons, by the formation of a spongy- 
looking tissue within the originally simple cavity. This 
tissue, as is shown in fig. 3, is formed by numerous fine 
branching trabecule, which pass from wall to wall and 
anastomose with one another. The trabecule are composed 
of a core of mesogloea, covered by endoderm, and are more 
or less flattened in section. Some of the larger trabecule 
contain strings or islets of endoderm cells embedded in the 
mesogloea, and the branching of the trabecule is apparently 
brought about by the formation of cavities in these intrusive 
endoderm cells, but there is no evidence of the formation of 
definite solenial outgrowths of the endoderm penetrating into 
the mesoglcea and forming compound stolons. The trabeculez 


' G. C. Bourne, “ Anthozoa,” in Lankester’s ‘ Treatise of Zoology,’ 
Part II, 1900, p. 16. 

? G. von Koch, “ Die Aleyonacea des Golfes von Neapel,” ‘ Mitth. a. d. 
Zool. Stat. zu Neapel,’ ix, 1891, p. 645. 


ACROSSOTA LIPOSCLERA. 265 


become less numerous in the terminal and most recently 
formed portions of the stolons, and are absent or very few in 
number and imperfectly developed in the swollen extremities 
(fig. 3). They are also scantily developed in the long free 
outgrowths which give rise to separate colonies. On the 
other hand, in the flattened adherent portions of the stolons, 
the trabecule tend to fuse together to form lamine, dividing 
the lumen of the stolon into several canals, running, on the 
whole, parallel to the long axis of the stolon. 

It is assumed rather than demonstrated that the compound 
stolons of Clavulariide are formed by the union of the walls 
of a close-meshed reticulum of simple solenia such as is found 
in Cornularia. It is probable that many band-shaped or flat 
encrusting stolons are formed in this manner, and all 
encrusting stolons must be formed in part by the fusion of 
the walls of radial outgrowths containing solenia from the 
bases of the zooids, but it is equally probable that a large part 
of the inosculating channels seen in sections of such stolons 
have been formed as in Acrossota, by the ingrowth of 
trabecule from the walls of a primitively simple solenium. 

The general structure of the zooids, with the exception of 
the tentacles, conforms to the Aleyonarian type. 

The Actinopharynx (as van Beneden! has renamed the 
** Stomodeum ” of the Anthozoa) is long, and its upper third 
is lined by a ciliated epithelium, of which the character is 
shown in fig. 5. The epithelial cells are elongated and 
attenuated, their nuclei stain deeply and appear closely 
crowded together in sections. The cilia take their origin 
from minute deeply staining granules, which give the free 
border of the epithelium a striated appearance. In this region 
there isa distinct sulcus or siphonoglyphe in which the cilia 
are specially long. But in the lower two thirds of the actino- 
pharynx (fig. 6) the siphonoglyphe dies out; the epithelium 
is composed of less elongated prismatic ciliated cells, the 
nuclei are no longer crowded together and stand nearly on 


1 E. van Beneden, “ Die Anthozoen d. Plankton Expedition,” ‘ Ergeb- 
nisse d. Plankton Expedition des Humboldt-Stiftung,’ ii, 1897. 


266 GILBERT C. BOURNE. 


the same level; the cilia are short and of uniform length. 
In contracted specimens this region of the actinopharynx is 
thrown into longitudinal folds and its lumen diminished, but 
the lumen is always considerably larger than in the upper 
third. At the lower border of the actinopharynx the. pris- 
matic ciliated epithelium passes rather abruptly into the 
endoderm, but is clearly continued down the edges of the two 
“ dorsal” or asulcar mesenteries to form the grooved filament, 
which in section presents precisely the same characters as 
those figured by Hickson! for Alcyonium. In the grooved 
filaments the ciliated cells again change their character, 
becoming small and narrow, with small deeply staining nuclei. 
The remaining mesenteries have thickened edges covered by 
endoderm cells, but no distinct filament. As is invariably the 
case in Alcyonarian zooids, the two “ dorsal” asulcar mesen- 
teries bearing the grooved filaments extend much further 
down in the body of the zooid than the remaining six. 

‘he section represented in fig. 4 shows that in other respects 
the mesenteries present the usual Alcyonarian features. The 
mesogloeal thickenings forming the muscle-banners are feebly 
developed, and only distinguishable in the region of the 
actinopharynx, but there they can be seen to be borne on the 
“ventral” or sulear faces of all eight mesenteries.. The 
peripheral parts of the mesenteries are very thin. Only one 
of the Zooids that I used for cutting sections was sexually 
mature. In it testes, in the form of small spherical follicles, 
were borne on short pedicles near the free edges of all but 
tle dorsal asulear mesenteries. 

The tissues of the specimens I have examined were only 
moderately well preserved, and the cell elements so minute 
that I have not been able to work out histological details to 
my complete satisfaction. The structure of the body-wall in 
particular has proved very difficult to interpret, and the 
structure of the various regions into which the body of the 
zooid may be divided appears to differ in invaginated and 


1 §. J. Hickson, “The Anatomy of Alecyonium digitatum,” 
‘Quart. Journ. Mier. Sci.,’ 37, 1895, p. 343. 


ACROSSOTA LIPOSCLERA. 267 


extended specimens, from which fact I infer that the tissues 
are very elastic and extensile, capable of being drawn out 
into an exceedingly thin layer or contracted into thicker 
sheets, according to the state of contraction of the animal. 

From a study of longitudinal and transverse sections of 
both extended and retracted specimens it appears that the 
following regions may be recognised: (1) The tentacles and 
the oral disc. (2) The portion of the body of the zooid 
immediately below the tentacles. This portion is invaginated 
in retracted specimens, but in both extended and retracted 
examples the body-wall in this region is moderately thick, 
and the tissues, both ectodermic and endodermic, are fairly 
clearly differentiated. In extended specimens this region is 
approximately of the same length as the actinopharynx, but 
in retracted specimens the latter structure is pulled down into 
the lower part of the ccelenteron and les below the tube 
formed by the invaginated distal portion of the body-wall of 
the Zooid. (3) The middle and basal region of the Zooid, in 
which the body-wall is extremely thin. (4) The stolons, in 
which the mesoglcea is thickened, the endoderm fairly con- 
spicuous, but the ectoderm reduced or absorbed in the 
formation of a gelatinoid supporting tissue. 

The first two regions are clearly the seat of the chief phy- 
siological activities. The cavities of the tentacles, the walls 
of the actinopharynx and the central moieties of the mesen- 
teries where attached to the actinopharynx are clothed with 
a thick highly vacuolated endoderm containing a profusion 
of zooxanthellee (fig. 4). The mesoglea is a thin layer exhibit- 
ing a faint striation, but not including any cellular elements. 
The ectoderm of the tentacles is formed by a layer of small 
cells, cubical in the extended but more columnar in the 
invaginated tentacles. The ectodermal muscular fibres are 
but feebly developed in the tentacles, the circular layer of 
endodermic muscular fibres being rather better represented, 
but still feeble. Minute oval refracting bodies can be recog- 
nised in the ectoderm of the tentacles, and I interpret them 
as nematocysts, but I was unable to resolve their structure 


268 GILBERT C. BOURNE. 


with the highest powers of the microscope at my disposal. 
The ectodermic lining of the actinopharynx has been described 
already. The ectoderm of the body-wall below the tentacles 
is illustrated in fig. 7, The continuous layer of cubical cells 
of the tentacles and.oral disc here gives place to what I can 
only describe as a reticulum of vacuolated protoplasm con- 
taining nuclei, but in which the limits of the cell bodies are 
with difficulty or not at all recognisable. ‘he body-wall is 
raised into a number of thickened ridges, one of which is 
shown in fig. 7. In this figure the mesoglcea is seen as a 
thin but distinct structureless layer, produced on the inner 
side into small processes to which the muscular fibres of the 
vacuolated endodermic cells are attached. On the outer side 
of the mesoglcea is an irregular aggregation of what may be 
called ectoderm cells, though, as mentioned, the cell outlines 
are not distinguishable. Some of these abut on the meso- 
gloea, without, however, forming a continuous layer. Others, 
more peripherally situated, form a discontinuous lining to the 
thin darkly staining external cuticle. The intervening space 
is filled up by a homogeneous more or less vacuolated sub- 
stance which seems to be formed by the dissolution and 
conversion into supporting tissue of the ectodermic cell- 
fusion. The homogeneous supporting substance stains more 
faintly than the mesogloea, but is probably of the same or 
similar chemical composition, for in older and more differen- 
tiated parts of the body-wall the ectodermic cell-fusion 
becomes scantier, the homogeneous substance increases in 
amount, and eventually fuses with the mesogloea, and can only 
be distinguished from it by its staining less deeply. Usually, 
however, tracts of ectoderm looking like cell ingrowths are 
included in the homogeneous supporting tissue. 

Fig. 8 represents a portion of the extremely thin body- 
wall of the lower middle and basal part of the zooid. In com- 
parison with fig. 7, it is seen that the endoderm is reduced 
to a very thin non-vacuolated flattened epithelium. The 
mesogloea is a distinct, very thin structureless layer. In one 
part of the section it seems to consist of two layers, between 


ACROSSOTA LIPOSCLERA. 269 


which are a few flattened nuclei with remains of cytoplasm 
surrounding them. ‘The outer layer, however, is nothing 
more than the deeper part of the homogeneous supporting 
tissue formed by the ectodermic cell-fusion, and the nuclei 
between it and the mesogloea are ectodermic nuclei. Ex- 
ternally may be seen a few nuclei of ectodermic cells which 
have obviously been used up in the formation of local 
thickenings of the supporting tissue, and externally is the 
thin but tough deeply staining cuticle. 

In the stolons the endoderm again becomes thicker and 
vacuolated, but does not contain zooxanthellae. The external 
cuticle remains, but the nuclei of the ectodermic cell-fusion 
have almost entirely disappeared, and the ectoderm appears 
to have been nearly wholly used up in the formation of 
supporting tissue, which is now so closely applied to and 
fused with the mesogloea as to be scarcely distinguishable 
from the latter. One would say at first sight that the section 
shows only endoderm, mesoglea and cuticle, the ectoderm 
having disappeared altogether. In this region strings of 
endoderm-cells make their way into the mass formed by the 
fusion of mesoglcea and supporting tissue, and these in- 
growths extend into the trabecule, which are themselves 
formed as local out-growths of mesogloea covered by endo- 
derm. ‘l'his ectodermic gelatinoid supporting tissue of 
Acrossota is comparable in essential respects to the support- 
ing tissue of Stereosoma celebense as described and 
figured by Hickson.! But there is this difference: that in 
Stereosoma there is a distinct external layer of cubical 
ectoderm cells, but no cuticle, whereas in Acrossota the 
definite external ectodermic epithelium is absent, but there 
is a distinct and continuous cuticle. 

There are no spicules in Acrossota. J have searched for 
them in teased-up specimens, in sections, and in the residue 
left after boiling in caustic potash. A felt-work of filamentous 
algee enclosing numerous diatoms, sponge spicules, and here 


' §. J. Hickson, “ A Revision of the Aleyonaria Stolonifera,” ‘ Trans. 
Zool. Soe. Lond.,’ xiii, 1895. 


270 GILBER1 C. BOURNE. 


and there an alcyonarian spicule, covers the older parts of the 
stolons and the basal moieties of the older zooids. But all 
these spicules are adventitious and le external to, though 
often closely adherent to, the cuticle. 

It may be concluded, from the foregoing description, that 
Acrossota is an Alcyonarian exhibiting primitive Cornularian 
characters in the mode of growth and habit of the colony, in 
the simplicity of its stolons (which, however, are not so simple 
as those of Cornularia), and in the absence of calcareous 
spicules, the last character being shared by such clearly 
primitive forms as Cornularia and Protocaulon molle, 
and also by Stereosoma celebense, Clavulariareptans, 
Clavularia celebensis, and the variety of Clavularia 
australiensis described as variety B by Hickson! The 
absence of spicules may be regarded as evidence of primitive 
organisation, but not certain evidence, for the spicules 
present in variety A appear to have been lost in variety B 
of Clavularia australiensis. The absence or disappear- 
ance of spicules appears to be correlated in the Clavulariidz 
with the formation of a supporting tissue derived from vacuo- 
lated branched ectoderm cells, and Acrossota shares this 
feature with Stereosoma and Clavularia australiensis, 
var. B. The feature peculiar to Acrossota is the absence of 
tentacular pinne, and it is a moot point whether this may 
be regarded as a primitive character. Stereosoma has but 
few pinne, and those spaced at considerable intervals along 
the tentacles. Further reduction of such pinne might lead 
to their ultimate disappearance, and the tentacles would then 
be simple and digitiform. On the other hand, it may be 
argued that simple tentacles must have preceded pinnate 
tentacles in phylogeny, and we have in Acrossota an otherwise 
primitive Alcyonarian with simple tentacles. The evidence 
seems to point to this being a primitive character, but in the 
absence of any criterion which shall enable us to distinguish 
between what is primitive and what is secondarily acquired 
by degeneration in such structures as these, one cannot be 

Siprde 


ACROSSOTA LIPOSCLERA. 271 


dogmatic. ‘lhe one thing certain is that in the possession of 
simple tentacles Acrossota stands as an exception to a rule 
otherwise universal for Aleyonarians, and it was this that led 
me to make as exhaustive a study as I could of its structure. 


EXPLANATION OF PLATE 22, 


Illustrating Mr. G. C. Bourne’s paper ‘‘On Acrossota 
liposclera.” 


[All the figures refer to Acrossota liposclera nov. gen. et sp.| 


Fig. 1.—An independent stolon attached by one of its extremities to 
a fragment of Halimeda. The stolon bears four zooids, three of which 
are partially expanded and exhibit the simple tentacles. 


Fig. 2—Two zooids from a specimen stained in hemalum and mounted 
in oil of cloves. One of the zooids is expanded, and exhibits the eight 
simple tentacles; the other is retracted, and shows the invaginated 
tentacles. d.m. Dorsal or asulcar mesenteries. 


Fig. 3.—A retracted zooid and part of a stolon from a specimen stained 
in hemalum. The body-wall of the zooid and part of the wall of the 
stolon nearest the observer have been cut away. d.m. Dorsal mesen- 
teries. st. Stolon. ¢. The invaginated tentacles. tr. Trabecule in the 
cavity of the stolon. 

Fig. 4.—A transverse section passing through the upper part of the 
actinopharynx of a retracted zooid, showing the actinopharynx with its 
sulcus, the eight mesenteries with the muscle banners, the eight invagi- 
nated tentacles and the thin external body-wall. The endoderm clothing 
the actinopharynx, the central moieties of the mesenteries and the 
tentacles is thick, and contains numerous zooxanthelle. aph. Actino- 
pharynx. d.m. “ Dorsal” mesenteries. v.m. “ Ventral” mesenteries. 
t.t. The invaginated tentacles. 


Fig. 5.—A_ section through the upper third of the actinopharynx, 
highly magnified, showing the elongated ciliated cells, the sulcus or 
siphonoglyphe, and the vacuolated endoderm. z. Zooxanthelle. Other 
letters as in fig. 4. 


Fig. 6.—A section through the lower part of the actinopharynx, 
showing the disappearance of the suleus and the changed character of 
the ciliated epithelium. Lettering as in the preceding figures. 

voL. 60, parT 2.—NEW SERIES. 19 


pay Ps GILBERT C. BOURNE. 


Fig. 7.—A longitudinal section through part of the subtentacular 
region of the body-wall, magnified 750. cu. The external cuticle. ee. 
Ectodermic cell-fusion with nuclei. en. Endoderm. h.s. The homo- 
geneous supporting substance formed by the ectoderm. mg. Mesoglea. 

Fig. 8.—Part of a longitudinal section through the body-wall of the 
middle of the zooid, magnified 750. Lettering as in fig. 7. 


THE PROBOSCIDIAN SYSTEM IN NEMERTINES. 273 


The Proboscidian System in Nemertines. 


By 


Dr. Gerarda Wynhof, 
Utrecht. 


With 36 Text-figures. 


Tae proboscis, together with its sheath and the rhyn- 
chodeeum, are among the most characteristic features in the 
anatomy of Nemertines, so characteristic even that the system 
is not absent in any known species, neither of Anopla nor of 
Enopla. And wherever it is found, the whole system shows 
the same advanced development of construction; all three 
organs are completely developed, the proboscis being an intro- 
vertible tube, which is fastened to the body-wall at the 
anterior end of the rhynchoceel, and is connected posteriorly 
by a retractor muscle to the wall of the sheath. The rhyn- 
chodzeum is a kind of atrium to the apparatus, through which 
the proboscis is everted (l'ext-fig. 1). Both proboscis sheath 
and proboscis itself possess a muscular wall, and the ]Jumen of 
the sheath is lined by an endothelium. The cavity of both 
rhynchodeum and proboscis is lined by continuous epithelium 
that shows a differentiation of glandular elements in different 
parts of the system. The proboscis, therefore, consists of 
three layers—an epithelium, a muscular coat and an endo- 
thelial layer ; the wall of its sheath of two—an endothelium 
and a muscular coat. The whole system is imbedded in the 
body-parenchyma. It has for a long time been assumed 
that the proboscis had developed out of a retractible part 


274. DR. GERARDA WYNHOFP. 

cf the head independently of corresponding structures in 
other invertebrates. Hubrecht (16) shared this opinion, 
and tried to explain, by the histological facts shown by 
the “Challenger” material, that the sheath had developed out 
of muscular elements in loco, and was an independent 
structure. However, Salensky (25) had published another 
hypothesis, comparing the whole proboscidean system of 


TEXxT-FIG. 1. 


i 
1 \\ 
1 is 8 prob.w. 
i iH iE 
At HE 
i \ " rhynch. 
ie ‘ia 
I 
le 
al 


Schema of the proboscidian system in Nemerteans after Salensky 
(26). rhynch. Rhynchodeum or proboscis introvert cavity. 
rhynch. c. Rhyncho-celomic cavity. prob. Epithelium of the 
introverted proboscis. prob. w. Proboscis wall. rhynch. w. 
Wall of the proboscis cavity (rhynchocel). retr. Retractor 
muscle of the proboscis. Compare with Text-fig. 36. 


Nemerteans with the proboscis of the Rhabdocclida 
proboscida. He founded his theory on certain embryo- 
logical facts denied by Hubrecht (15), who tried to find 
support for his view in anatomical facts. Biirger (8) looked 
for a homolegue of the whole system in the pharynx and 
pharyngeal sac of the Polyclads, but neither Birger nor 


Hubrecht could gain the agreement of Salensky, who published 


THE PROBOSCIDIAN SYSTEM IN NEMERTINES. 275 


two articles in 1909 (27) and 1912 (28) in which he maintains 
his views of 1884 (25). As Salensky only gives embryological 
facts, we shall have to look for further judgment to the 
anatomical data. Wishing to draw my own conclusions as 
to the value of Salensky’s theory, I have compared whatever 
was known about rhynchodeum, proboscis and proboscis 
sheath in Nemerteans. 


TExtT-Fic. 2. 


Schema of the proboscis of Paleonemerteans. p.e. Proboscis 
epithelium. o.c.m. Outer circular musculature. i.1.m. Inner 
longitudinal musculature. 


The first fact that came to light was the great difference 
in structure between the proboscis of armed and unarmed 
Nemerteans. For the difference between the two does not 
consist only in the presence or absence of the so-called stylet ; 
if is shown also in the structure of the muscular coat, and in 
the differentiation of the tube into different parts. It is even 
possible in Hnopla that the armature has got lost, as seems 
to be the case with the genus Planktonemertes (10a). Mala- 


276 DR. GERARDA WYNHOFF. 


cobdella also does not possess an armed proboscis. The 
Anopla show a great variety of structure for such a plainly 
built organ. The tube always consists of two parts, the 
anterior part sometimes possessing a differentiated region 
near the insertion to the body-wall. That a difference of 
cpithelial glands exists between these parts is nearly certain ; 
the exact nature of this difference, however, is not sufficiently 
known. Very often the beginning of the hinder part is 


TEXT-FIG. 3. 


Section of the proboscis of Carinoma after Bergendal (6, text- 
fig. 55). prob.n. Proboscidian nerve. p. end. Proboscidian 
endothelium. o.¢.m. Outer circular musculature. 7.l.m. Inner 
longitudinal musculature. m. cr. Muscle crosses. 


marked by a constriction. The principal seat of the variety 
of structure, however, is the muscular wall of the proboscis. 
In all Paleeonemerteans the muscular sheath of the proboscis 
consists of two layers, a circular and a longitudinal, of which 
the circular layer is situated directly beneath the epithelium 
(Text-fig. 2). 
Exactly this type of proboscis is found in several species of 
Tubulanus (9), in Procarinina atavia (5), in Carinina (9), 
Hubrechtia (9) and Hubrechtella (3). he other genera 
possess a more complicated proboscis, though, along the 


THE PROBOSCIDIAN SYSTEM IN NEMERTINES. DAY 


greater part of each, the two muscular layers described above 
are developed. Moreover, Bergendal described in Carinoma 
(6) a proboscis in which the circular muscle-fibres show a 
tendency to divert into the longitudinal layer and build mus- 
cular crosses (Text-fig. 5). A circular layer outside the longi- 
tudinal fibres is, however, not present. As the tendency to 
form crosses is also known in the circular fibres of the body- 
wall, their presence in the proboscis of Carinoma cannot alter 
our opinion that only two muscular coats are present. 


The genera Cephalothrix (83), Cephalotrichella (83) and 


TEXT-FIG. 5. 


TEXT-FIG. 4. 


Text-FIG. 4.—Proboscidian section of Carinesta anglica, 
Wynhoff. p.e. Proboscis epithelium. 7. 1. m., prob. n. and 
p.end, as in Text-fig. 3. 

TEXtT-FIG. 5—Proboscidian section of Carinesta orientalis, 
Punnett. .e., prob.n., p.end. and 7.1. m. as before. par. 
Parenchyma. constr., Constrictor muscle. 


Procephalothrix (88), Callinera (2), Carinesta and Carino- 
mella (12) differ more from the above-mentioned type. In all 
of them the circular muscle-fibres are wanting in the portion 
directly behind the insertion of the proboscis (Text-fig. 4). In 
the family Cephalotrichidee, as well as in the three other genera, 
this part is followed by a conspicuous thickening of the 
circular layer, which acts as a constrictor (Text-fig. 5). At 
different places in this neighbourhood, in Cephalotrichidee just 
before (Text-fig. 6), in Callinera behind (Text-fig. 7), in Cari- 
nomella (T'ext-fig. 8) before and behind the constrictor muscle, 
the parenchyma is particularly well developed, causing a regres- 


278 DR. GERARDA WYNHOFF. 


sion of the muscular fibres. he longitudinal fibres, however, 
are present over the whole length of the proboscis (‘Text-figs. 
6 and 7), though often much reduced, as at the places men- 
tioned above. In Carinomella (‘Text-fig. 8), however, the 
whole circular musculature has degenerated in the anterior 
part of the proboscis, and instead of four bundles of muscle- 


TEXT-FIG. 6. TEXT-FIG. 7. TEXT-FIG. 8. 


p-e--: 
constr.- 


par 


TEXxT-FIG. 6.—Schema of the proboscis of Cephalothrix. pe., 
par., 0.¢c.m., i. 1.m., and constr. as in Text-figs. 3 and 5. 


TExt-FIG. 7.—Schema of the proboscis of Callinera. Letters as 
in Text-fig. 6. 


Text-FIG. 8.—Schema of the proboscis of Carinomella. Letters 
as in Text-fig. 6. 


fibres, as in Cephalotrichide and Callineride (Carinesta [Text- 
fig. 4] and Callinera), Coe (12) found only two longitudinal 
muscles in the cavity of the sheath, connecting the wall of the 
sheath with the anterior part of the proboscis (Text-fig. 9). 
The other parts of the proboscis, however, consist, in all these 
genera, of the same layers as in Tubulanus (Text-fig. 2); it 


THE PROBOSCIDIAN SYSTEM 1N NEMERTINES, 279 


seems, therefore, not too presumptuous to say that in all Paleeo- 
nemertines the muscular coat of the proboscis is composed of 
two layers, a circular and a longitudinal, the latter being 
situated beneath the endothelial surface. he deviations from 
this type, found in the families of Cephalotrichidie, Calli- 
neridee, and the genus Carinomella, have merely been brought 
about by the disappearance of the circular fibres in the region 
directly behind the insertion, accompanied in Carinomella by 
the egression of the longitudinal fibres. 


Uop-arsinices, 8). 


Section through Carinomella after Coe (12), fig. 54. 1. m. Longi- 
tudinal muscle. rhynch.m. Rhynchodeal musculature. c¢./.m., 
Central long musculature. 7. ¢. m. Inner circular musculature. 
bl. v. Blood-vessel. dig. Digestive tract. par. Parenchyma. 


In Heteronemertea the proboscis shows a much greater 
variety of structure. A certain number of species possess a 
proboscis exactly like the Paleonemertea ; the outer layer of 
the ejaculated duct consists of the glandular epithelium, next 
to it is the circular muscle-layer, limited at the other side by 
the longitudinal fibres; such are the proboscides in Micrella 
rufa, Punnett (22), Oxypolia beaumontiana Punnett 
(22), and in Huborlasia (9). The genera Lineus, Cerebratulus 
and Micrura, which are so closely connected, that they cannot 
even be distinguished from each other by anatomical features 


280 DR. GERARDA WYNHOFF. 


alone, show a marked resemblance in their proboscides too. 
In all three genera species are known with two muscular 
layers, others possessing a third muscular coat (Tex-figs. 10 
and 11). Representative of the Paleeo-type are Cerebratulus 
urticans (J. Mill.) (9), and C. greenlandicus Punnett 
(23), Lineus scandinaviensis Punnett (24), and L. 


Trext-Fic. 10. 


| 
Schema of the proboscis of Heteronemerteans (Cerebratulus 
marginatus). o.l.m. Outer longitudinal musculature. p. e. 


Proboscis epithelium. 0. ¢. m. Outer circular muscular layer. 
i.l.m. Inner longitudinal muscular layer. 


bilineatus McIntosh (9), Micrura varicolor Punnett 
(24), M. bergenicola Punnett (24), and M. atra Punnett 
(24). Variations on this scheme are found in Heteronemer- 
teans as in Paleeonemertines. Miss Thompson (80) described 
the interesting genus Zygeupolia. The circular muscle-layer 
fails in the region behind the insertion, exactly as in Cephalo- 
trichide and Callineride (Text-fig. 12). Other genera and 
species possess three muscular layers in the proboscis; to this 


THE PROBOSCIDIAN SYSTEM IN NEMERTINES. 281 


group belong the Cerebratulus-, Lineus-, and Micrura-species 
as Cerebratulus marginatus Ren. (9), and C. eisigi 
Biirger (9), Micrura fasciolata Ehrenberg (9), and 
Lineus versicolor Birger (9). Two of these three layers 
consist of longitudinal fibres and they are separated by the 
circular muscle-fibres (‘l'ext-fig. 10). 

The presence of this third layer in the proboscis of Hetero- 


TErxt-rie. 11. 


Weare 


° 
eee 
qe. 


* 
Cae. 


Section of the proboscis of the introverted Cerebratulus mar- 
ginatus after Birger (9, taf. 23, fig. 1). m.ecr. Muscle 
crosses. 0. ¢.m. Outer circular muscular layer. 0.1. m. Outer 
longitudinal muscular layer. 7.1. m. Inner ditto. n. 1. Nerve 
layer. p.l. Epithelium (outer surface) of the proboscis. 


nemerteans ts rather striking. The acquisition of a longitu- 
dinal muscle-layer between the epidermis and the circular 
muscle-layer is characteristic of the body-wall in this group 
of Nemertines. The new layer in the proboscis wall, there- 
fore, being also a longitudinal layer, can immediately be com- 
pared with the outer longitudinal layer of the body-wall, and, 
as its surroundings are exactly the same, it is laid down as 
part of the outer longitudinal muscle-layer. The circular 
coat of the proboscis in that case must be the outer circular 
muscle-layer of the body-wall, and the original longitudinal 


282 DR. GERARDA WYNHOFF. 
layer is part of the inner longitudinal layer of the body-wall. 
The Paleotype of proboscides consists of the epithelium, the 
outer circular and inner longitudinal muscle-layers, that also 
form part of the body-wall. The newly obtained layer of the 
Heteronemerteans has not yet been acquired by all pro- 
boscides ; those with the Paleotype have not got it at all, the 
genus Parapolia Coe (11), is in the act of acquiring it (Text-fig. 
13). In the hinder part this proboscis shows the Paleotype, 
in the anterior part the Heterotype. 

A pecuharity of the circular muscle-layer of Hetero- 
nemertine-proboscides is shown in fig. 11. Some circular 


Trext-Fic. 12. 


Section of the proboscis of Zygeupolia after C. B. Thompson (80). 
p. end. Proboscis endothelium. J. m. Longitudinal muscles. 
p.e. Proboscis epithelium. probn. Proboscis nerve. 


fibres are seen to divert from their original stratum and to 
form crosses by entering the inner longitudinal coat, as I shall 
henceforth call it, conformable to the corresponding layer of 
the body-wall. These fibres of the muscular crosses continue 
even outside the longitudinal layer, constituting something 
like a very thin and very incomplete circular layer just 
beneath the endothelium. As Zygeupolia has muscular crosses 
in its proboscis, the thin circular coat beneath the endothelium, 
described by Miss Thompson (80), might be comparable with 
these circular fibres. 

Still another type of proboscis is found in the genera Baseo- 
discus (9) and Joubinia (9), at least in two of the three 
species belonging to this genus. The muscular coat consists 
again of two layers, one with longitudinal and the other with 


THE PROBOSCIDIAN SYSTEM IN NEMERTINES. 283 


circular fibres. Instead of the circular fibres lying beneath 
the epithelium as in Paleonemertines, the longitudinal 
muscles are found there (Text-fig. 14). The explanation would 
have been difficult, for either the layers might have been the 
inner longitudinal coat with a better-developed new circular 
layer as described in Lineids and Zygeupolia, or the outer 


TExt-FiGc. 13. 


p-e. ---- 


o.l.m - 


Schema of the proboscis of Parapolia. p.e. Proboscis epithelium 
(outer surface). 0.1. m. Outer longitudinal muscular layer. 
o.c. m. Outer circular ditto. 7. 1. m. Inner longitudinal ditto. 


longitudinal and the outer circular muscle-layer, had not a 
form hke Joubinia rubens Coe (11) or Oxypolella (3) 
existed. Both species show a different arrangement of the 
muscular fibres in the two parts of the proboscis (‘l'ext-fig. 15). 
The second part possesses the three muscle-layers character- 
istic of the body-wall of Heteronemertines. In the anterior 
part, however, the inner longitudinal fibres are absent, the 
wall showing the Baseodiscus type. ‘The process of disappear- 


284 DR. GERARDA WYNHOFF. 


ance of these longitudinal fibres, actually seen in Joubinia 
rubens, culminates in their total absencein J. longirostris 
and blanca and in Baseodiscus.! Soif the Paleotype consti- 
tutes the first stage in its development, the three-layered 
muscular coat of Cerebratulus marginatus gives the 
second stage, Baseodiscus the third. 

Putting all genera of Heteronemertines whose proboscis 
structure is known together in one list, we may find the links 
between the three above described stages. 

Paleotype (Text-fig. 2): Micrella rufa Punnett, Oxy- 


TrxtT-Fic. 14. 


So 


Fo. 
ay 


Section through the introverted proboscis of Baseodiscus after 
Birger (9, taf. 23, fig. 2). p.e. Proboscis epithelium. 1. L. 
Nervous layer. ¢.m. Circular musculature. l,m. Longitudinal 
musculature. 


polia beaumontiana Punnett, Huborlasia, Cerebratulus 
urticans (J. Muiler) and other Cerebratulus species, Lineus 


1 Bergendal (4) describeda Valencinia longirostris in which the 
so-called circular layer shows an interlacing of longitudinal and cireular 
fibres something like the proboscis sheath of Drepanophoride. The 
principal difference between Bergendal and Birger (9), who gave this 
peculiarity of the proboscis of Valencinia longirostris in his figure, 
loc. cit., taf. 23, fig. 9, comes to this: that the circular fibres, according 
to the first-named author, are arranged in two planes, neither coinciding 
with the horizontal section of the organ, and that Birger had not 
observed this fact. This, however, seems not to be of any importance 
to our conclusions, since both writers agree as to the circular nature of 
this musculature. The interlacing of fibres is, as has often been stated 


THE PROBOSCIDIAN SYSTEM IN NEMERTINES., 285 


bilineatus McIntosh, and other Lineus species, Micrura 
atra Punnett and other species; variations on this type, Para- 
lineus (29) and Zygeupolia (Text-fig. 12). 
Transition to Heterotype: Parapolia (Text-fig. 13), 
Heterotype (Text-fig. 16): Cerebratulus marginatus 
Renier, and other species, Lineus versicolor Birger and 


TExt-F1IG. 15. TExt-FieG. 16. 


=| 
pet 

= 

= 
olm.=S 

= 
oc.m- = 

= 

= . 

=|, 

tI 

= 

I i 

=i 
2 El: 
ulm i - 


Text-Fic. 15.—Schema of the proboscis of Joubinia rubens. 
Letters as in Text-fig. 13. 


TExt-FIG. 16—Schema of the proboscis of Joubinia longi- 
rostris and Basiodiscus. Letters as in Text-fig. 15. 


related species, Micrura fasciolata Hhrenberg Langia; 
variety Joubinia longirostris Berg. 


by other investigators, the result of the dissolving of muscular crosses, 
and seems therefore not to be of any consequence to our explanation. 
More compromising is the fact that in Baseodiscus the nervous layer, 
which I have not taken into consideration in this article, but which gives 
only support to my views, does not lie in the expected place between 
longitudinal and muscular fibres, but beneath the epithelium, as in 
Paleonemertines. Is this a reminiscence of a more primitive state ? 


286 DR. GERARDA WYNHOFF. 


Transition to Baseodiscus type: Oxypolella and Jou- 
binia rubens (Text-fig. 15). 

Baseodiscus type (‘l'ext-figs. 14 and 16): Baseodiscus 
species, Joubinia longirostris and blanca Birger. 

Two genera of Heteronemerteans, Poliopsis and Valencinura 
have not been mentioned yet. As regards Poliopsis this is 
due to our ignorance about its proboscis; in all the specimens 
which I know, it has been thrown off before capture. As 


Trext-FIG. 17. 


Section through the introverted proboscis of Valencinura 
bahusiensis after Bergendal (4, taf. 1, fig. 15). p. end. 
Proboscis endothelium. p.c. Proboscisepithelium. n.l. Nerve 
layer. Other letters as in Text-fig. 13. 

regards Valencinura its proboscis has been described accu- 
rately by Bergendal (4). The anterior two parts of this organ 
possess a muscular coat very much like the anterior part of 
Parapolia’s proboscis. The third part, however, lacks all 
circular musculature. From the description given by 
Bergendal I cannot possibly make out whether the longitudinal 
fibres of this last part are the continuation of the outer or of 
the inner longitudinal layer. Two figures are given of the 
region in which the circular musculature is getting thinner 
before it disappears. 

The order in which different coats are arranged is: Epi- 


THE PROBOSCIDIAN SYSTEM IN NEMERTINES. 287 


thelium, outer longitudinal layer, nervous layer, circular 
muscles, inner longitudinal muscle layer and endothelium 
with endothelial circular fibres (Text-fig. 17). The next section 
that is figured, however, gives: Epithelium, circular fibres, 
nervous layer, longitudinal muscle layer and endothelial circu- 
lar fibres (Text-fig. 18). Probably, therefore, both outer cir- 
cular and outer longitudinal muscle-fibres have disappeared, 
or, as in Parapolia, the outer longitudinal muscle-fibres have 
not yet been developed in the hinder part of the proboscis, and 
the outer circular muscular layer has disappeared, as it does in 


TExtT-FIG, 18, 


Section through the proboscis of Valencinura bahusiensis 
after Bergendal (4, taf. 1, fig. 14). e.m. Outer longitudinal 
muscles. e. p. Proboscis epithelium. Other letters as in 
Text-fig. 17. 

the other Nemertines. The place of Valencinura in our list 
should then be next to Parapolia; the nervous layer in 
changing places with the circular muscle-fibres, however, 
makes one cautious as to this supposition. 

The facts taught by comparative anatomy, therefore, lead 
us to the conclusion that the proboscis of the Anopla is a 
structure of the body-wall in which both epithelium and 
muscular coat have taken part. We shall now have to con- 
sider whether all muscle-layers or only part of them helped to 
form the proboscis. 

The body-wall of Anopla is, as Miss Thompson (80) tried to 
make certain, composed of four different muscle-layers. Of 
these all Paleonemerteans, with the exception of Carinoma, 

VOL. 60, part 2.—NEW SERIES. 20 


288 DR. GERARDA WYNHOFF. 


only possess three, namely, the outer circular, the inner longi- 
tudinal and the inner circular muscle coat. In Heteronemer- 
teans, as in Carinoma, an outer longitudinal layer of fibres 
has been developed. In the proboscides of Palzeonemerteans 
we have found traces of no more than two of the three layers 
of the body-wall. Of the third or inner circular muscle-layer 
I have not been able to detect any traces in the proboscis of 
any Heteronemertean either, endothelial circular muscle- 
fibres being of anentirely different nature. Here embryology 
comes in to tell us where to look for the missing part. 


TrExt-FiG. 20. 


TExT-FIG. 19. cer Migs 
eer 
LM Sw 


eb 10eceee-—--——— 


Pye Ly 


NY 


TExT-FIG. 19.—Longitudinal section through an embryo of 
Prosorochmus after Salensky (28, taf. 3, fig. 27.4). p.mes. 


Proboscidian mesoblast. ent. Entoblast. prob. e. Proboscidian 
ectoblast. ep. Epiblast. 


TEXxtT-FIG. 20.—Schema of an embryo of Lineus obscurus 
after Hubrecht (15, fig. 101). rh.w. Rhynchocelomic wall. 
mes. Mesoderm. cer. Cerebral ganglia. jp. e. Proboscis epi- 
thelium. Other letters as in Text-fig. 19. 


A description of the development of the proboscis and its 
sheath is given by Salensky (25, 26, 27, 28), by Hubrecht (15), 
Birger (7), Lebedinsky (18, 19, 20),and Arnold (1). Of these 
authors, Hubrecht, Biirger and Arnold studied the embryology 
of Lineus ruber, Salensky (26,28) Pilidium gyrans (pro- 
bably the larva of Cerebratulus marginatus) and Pili- 
dium pyramidale. The Hoplonomertea subject to these 
investigations were Prosorochmus viviparus (Salensky, 


THE PROBOSCIDIAN SYSTEM IN NEMERTINES. 289 


25, 27), Prostoma vermiculum (Lebedinsky, 19, 20), 
and Drephanophorus spectabilis (Lebedinsky, 18, 20). 
All these investigations are essentially in agreement with each 
other as regards our subject, with the exception of Hubrecht’s. 
This author, like the others, described the origin of the pro- 
boscis as an invagination of the ectoderm lying between, or, 
according to other authors, being part of, the two ectoblastic 
discs which give rise to the epidermis of the adult worm. 
All authors agree also as to the fact that this ectoblastic 
invagination, never becoming separated from the epiblast, 


TEXT-FIG. 21. 


Section through the proboscidian system of an embryo of Lineus 
after Arnold (1, taf. i, fig. 18). retr. Retractor muscle. ect. 
Eetoblast. ent. Entoblast. Other letters as in Text-fig. 20. 


gets its own mesoblastic investment, which is ectodermic in its 
origin (Text-fig. 19). At a later stage the ectodermic invagi- 
nation of the proboscis has got two mesoblast layers separated 
from one another by a space. Hubrecht (15) considered that 
this second layer had grown out from the mesoblastic layer 
of the body-wall, cutting off part of the body-cavity, or, 
according to Hubrecht, part of the archoccel (T'ext-fig. 20). 
Later investigations, however, have shown that Hubrecht made 
amistake. Biirger (7) studied Hubrecht’s sections, and came 
to conclusions in perfect accordance with those of Salensky 
and all later investigators. The second mesoblastic layer takes 
its origin from the first one by delamination. Lebedinsky has 
described the presence of two “ urmesoblasts” at the hinder 


290 DR. GERARDA WYNHOFF. 


border of both layers. ‘Iwo mesoblastic sacs take in this way 
the ectoblastic proboscis between them; when these sacs 
reach each other behind the proboscis, the retractor muscle 
is formed at the place where the walls fuse (Text-fig. 21). The 
inner layer gives rise to the muscular layers of the proboscis- 
wall; the outer to the wall of the sheath. Originally, how- 
ever, these two structures belong together, their separation 
being brought about by a splitting of the mesoblast. 


TExT-FIG. 22. 


Section through Carinina grata after Biirger (9, pl. xi, fig. 3). 
rhynch. Rhynchocelomic cavity. rh.c.m. Rhynchoccelomic 
circular musculature. xhynch. end. Endothelium of the rhyn- 
chocel. dig. Epithelium of the digestive cavity. 7. ¢.m. Inner 
circular muscle layer. nephr. Nephridium. 


Here we have found the evidence we looked for on p. 4. 
The proboscis has revealed itself as a part of the body-wall, 
of which the innermost layer of the muscular coat, the mner 
circular musculature, is absent. Embryology teaches that 
the proboscis sheath belongs to the proboscis, as both take 
their origin from the same tissue. We must, therefore, look 
for the missing part in the rhynchoccelomic wall. If the 
splitting took place between longitudinal and inner circular 
muscle-layer, the sheath must consist of a circular layer alone ; 
if the longitudinal layer was the seat of these changes, 


THE PROBOSCIDIAN SYSTEM IN NEMERTINES. 291 


the wall of the rhynchoccelom should have an inner longi- 
tudinal and an outer circular muscular coat. The following 
facts anatomy discloses: 

Procarinina possesses a proboscis-sheath, consisting of 
nothing but a very thin layer of circular fibres (5); in 
Carinina nothing but circular fibres are found (Text-fig. 22), 
Tubulanus linearis has a rhynchocelomie wall, as Pro- 
carinina (9); in Tubulanus polymorphus inside this layer 


Text-rie. 23. 


‘a 
‘2 


y 


34, 


TEXT-FIG. 23.—Part of a section through Procephalothrix 
linearis. rh.c.m. Rhynchocelomic circular musculature. 
rh.l.m. Rhynchocelomie longitudinal musculature. — ¢.l. m. 
Central long. musculature. 7.1. m. Inner, and ec. 1. m. central 
longitudinal muscles. par. Parenchyma. ent. Entoblast. 


Trext-FIG. 24.—Section through Carinesta anglica. rhynch. 
caw. Rhyncocelie cavity. rh.c.m. Rhyncocelomic muscu- 
lature. 7z.c.m. Inner circular muscles. prob. Proboscis. 
bl. v. Blood-vessels. dig. tr. Digestive tract. 


one row of minute longitudinal muscle-fibres can be found. 
Cephalotrichide (Text-fig. 23) agree with Tubulanus poly- 
morphus, and so does Hubrechtia; but the Callineridz (2) 
do not possess any longitudinal muscle-fibres inside the 
circular layer of the proboscis-sheath (Text-fig. 24). 

The body parenchyma, in which the proboscidian system 
is imbedded, develops, as a rule, longitudinal fibres in the 


292 DR. GERARDA WYNHOFF. 


Paleonemerteans. A localisation of these fibres is known to 
develop in several cases as the longitudinal septum between 
sheath and intestine, but also, as I have tried to demonstrate 
(31), as a longitudinal muscle-layer around intestine and 
proboscis-sheath, or as longitudinal. muscle-bundles around 
the latter. In Procarinina these fibres have no connection at 
all with the sheath; neither have they in Carinina, in which 
a septum has developed. ‘hey seem to fail in Hubrechtia, 


Trxt-rre. 25, 


Section through Cephalotrichella after Birger (9, taf. 11, fig. 
14). rh.c. Rhynchocelomic cavity. long.m. Outer rhyncho- 
cclomic longitudinal musculature. J.n. Lateral nerves. es. n. 
(Esophageal nerves. m. Mouth. 


but are present in Carinina in the septum, as in some 
Cephalotrichide. In Callinera and some Tubulanide longi- 
tudinal fibres are present between the inner circular muscle 
coat of the body-wall and the rhynchoccelom, without build- 
ing a distinct layer, as is the case in Cephalotrichella 
signata (Text-fig. 25), Carinomella (‘l'ext-fig. 9), and Cari- 
nesta (Text-fig. 26). 

In all these cases they are regarded as being a new 
acquisition of the proboscis-sheath, developed out of the 
central longitudinal musculature. A strong support to this 
opinion is given in the behaviour of this musculature in the 


THE PROBOSCIDIAN SYSTEM IN NEMERTINES. 293 


family Cephalotrichidz (81). Even different species of one 
genus as the two so nearly related species, Procephalo- 
thrix filiformis and P. aliena, show the development of 
a longitudinal musculature in the first and an ordinary 
septum in the other (81, 33). The Tubulanide show the same 
differences in one genus. 


TEXT-FIG. 26. 


Section through Carinesta anglica. rhyne. Rhynchodeum 
rh. c.m. Circular muscles of same. 1. m, Longitudinal muscles 
of same. dig. Digestive canal. nephrid. Nephridium. 0. ¢. m. 
Outer circular muscular layer of body. 7. c. m. Inner ditto. 
i. l.m. Inner longitudinal ditto. bl. v. Blood-vessel. 1. n. 


Lateral nerve. 


Therefore, if a longitudinal muscle-layer is developed out- 
side the circular musculature of the proboscis-sheath, we 
must regard it as a secondary one, that has not been present 
ab initio, and not as an inherent part of this organ. 

In Heteronemertini the two layers are noted. The circular 
muscle-layer is always present, an inner longitudinal coat has 
been recorded for Euborlasia, several Lineus and Cerebratulus 


294. DR. GERARDA WYNHOFF. 


species, Joubinia longirostris, Micrura and Langia, 
Valencinura, Parapolia, Zygeupolia, ete. 

A secondary longitudinal muscle-layer does not seem to be 
frequent; as far as I know it is not present in any Hetero- 
nemertean. Therefore we may conclude that as a rule the 
proboscidian sheath of the Nemertea anopla consists of 
two layers, an inner longitudinal layer, which may be absent 


TEXxT-FieG. 27. 


rad 7 
rhb H 1 
: 1 vlan 
ch.c ty, - 


oo 
0 OR 


oO 
oo” 


° 


6 
pee 
L} 


o 


2 45 go8 


Section through Emplectonema gracile after Birger (9, pl. 
xv, fig. 27). rad. m. Dorso-ventral musculature. gl. Glands. 
Other letters as in other figures. 


(as in some Paleesonemerteans), and has sometimes the thick- 
ness of one layer of fibres, and an onter circular muscle-coat. 

If ever theory were in accordance with the facts it is m 
this case. We were led to suppose that the proboscis was an 
inverted part of the body-wall by anatomical facts. One 
layer, however, failed, and embryology taught us that we 
might find it in the proboscis sheath, the two structures 
belonging together and being one in ontogeny. The missing 
part is, as a matter of fact, found in the proboscis sheath, 
and this consists of the two layers we knew a priori it must 
consist of, if the supposition were right. We conclude, 


THE PROBOSCIDIAN SYSTEM IN NEMERTINES. 295 


therefore, that in Nemertea anopla the proboscis and its 
sheath have phylogenetically the same origin, both being 
part of an inverted portion of the body-wall; that the two 
structures became separated by a rent in the muscular coat, 
which took place in the inner longitudinal or between the two 
inner muscle-layers. The probosvis has, at least, three layers, 


TEXT-FIG. 28. 


Longitudinal section through the proboscis of Prosorochmus 
after Birger (9, pl. xxiii, fig. 14). st. Stylet. gl. Glands. 
constr. Constrictor. c.m. Circular muscle layer. Other letters 
as in previous figures. F 


if reduction does not take place: the epithelium, the outer 
circular and the inner longitudinal muscle-layer, an outer 
longitudinal layer not necessarily developing, if it is present 
in the body-wall. The proboscis sheath has two or one 
muscular coat, an inner longitudinal layer, if present, part of 
the inner layer of the body-wall, and an outer circular muscle 
layer, the inner circular muscle-coat of the body-wall. 


296 DR. GERARDA WYNHOFF, 


I have purposely not confused these facts and conclusions 
in Anopla with those in armed Nemerteans, for the develop- 
ment of the armature in Hoplonemertea has so specialised 
this organ, that we must expect great deviations from the 
original conditions. 

The genus Malacobdella also lives under such unnatural 
circumstances for a Nemertine that ail kinds of anomalies 
may be expected. 

Of the three original muscle-layers of the body-wall, the 


TEXT-FIG. 29. 


Section through the proboscis of Amphiporus pulcher after 
Birger (9, taf. 23, fig. 3). Letters as in other figures. 


inner circular fibres, as such, are never present; we have to 
look for them, as has been demonstrated by Miss Thompson (80), 
in the dorso-ventral musculature of the body (Text-fig. 27). 
The outer longitudinal muscle-layer of Heteronemertini has 
nowhere developed, and we can recognise the other longi- 
tudinal layer by the nervous system being imbedded in it. 
Therefore we shall have to look for three muscular layers in 
the proboscidian system: a circular and a longitudinal layer, 
the latter possibly giving notice of its nature by the seat 
of the nerves, both in the proboscis, and probably remnants 
of the same longitudinal layer, and a circular layer in the 
wall of the sheath. 

Now let us examine the facts. Malacobdella has a proboscis 


THE PROBOSCIDIAN SYSTEM IN NEMERTINES. 297 


as described above, an epithelium with a circular muscle- 
layer underneath, and the longitudinal layer divided into two 
parts by a nervous layer. The endothelium, which lines both 
proboscis and sheath, is in the latter structure followed by a 
circular muscle-layer, as the presence of longitudinal muscle- 
fibres between them seems to be doubtful (9). These facts 
are in perfect accordance with those in Anopla. 

In Hoplonemertea some diversity exists as to the structure 
of the sheath. In all species, however, the wall of the pro- 
boscis has developed in avery similar way. As a rule we can 


TExtT-FIG. 30. Text-Fia. 31. 


ulm. 


_ eric 


Text-Fic. 30.—Section through the hinder part of the proboscis 
of Amphiporus marmoratus after Birger (9, pl. xxiii, fig. 
18). p.n. Proboscis nerves. Other letters as in previous 
figures. 

Text-FIG. 31.—Section through the rhynchocelomie wall of Dre- 
panophorus after Birger (9, taf. 23, fig. 37). Letters as before. 


divide the proboscis into three parts (Text-fig. 28), the 
middle part being the seat of the characteristic structures. 
Of these parts the first region has a wall consisting of three 
muscle-layers, two of which are circular coats separated by 
the only longitudinal coat in which the nervous layer is present 
(Text-fig. 29). The longitudinal coat, therefore, must be 
regarded as the inner longitudinal layer of the body-wall, and 
naturally one looks upon the circular layer between epithelium 
and longitudinal fibres as the outer circular muscle-coat. The 
other circular layer is present throughout the whole length 


298 Dk. GERARDA WYNHOFF. 


of the proboscis, as is the longitudinal layer. The outer cir- 
cular muscle-layer, however, has absolutely disappeared in 
the hind part of the proboscis (Text-fig. 30), and is found in 
the middle part in the shape of two distinct sphincters (Text- 
fig. 28). One might feel inclined to regard the second cir- 
cular layer as the inuer circular coat of the body. ‘This, 
however, is not in accordance with the embryological facts, 
that taught us to regard the sheath as part of the body- 
wall. Moreover, this sheath has a musculature exactly like 
that of Anopla. In all Hoplonemertea two muscular layers are 
present—an inner longitudinal layer often feebly developed, 
and an outer circular layer (Text-fig. 29). There is no 
reason not to regard them as identical with the same layers 
in Anopla, and therefore the inner circular muscle-coat of 
the body-wall is represented by this circular layer of the 
sheath, the longitudinal fibres being part of the inner longi- 
tudinal musculature. The family Drepanohporide, however, 
has a differently constructed sheath, in which, as in the 
hinder part of the same structure in Carinoma, longitudinal 
and circular fibres have become interlaced (Text-fig. 31). 

We have already discussed a similar process in connection 
with the proboscis of Joubinia longirostris. As two 
kinds of fibres are present, and in all related genera these 
fibres are situated in the way stated above, we might regard 
their arrangement in Drepanophorus as a complication of 
that stage, which cannot disturb our view of the facts. 

So homologues of all layers of the body-wall are found in 
Hoplonemertea in exactly the same place as in Bdellonemertea 
and in Anopla; the second circular layer of the proboscis, 
however, cannot be homologised with any part of it, neither 
do we know any such layer in Anopla. However, two cases 
are known, in which a tendency seems to exist to form a new 
circular layer underneath the endothelium; Bergendal (6) 
described the presence of circular fibres at that position, 
caused by the existence of faint muscle crosses (Text-fig. 3). 
C. B. Thompson (80) showed the presence of a layer of 
endothelial muscle-fibres. Though both are probably of a 


THE PROBOSCIDIAN SYSTEM IN NEMERTINES. 299 


totally different nature, we may look for an analogy between 
the development of this second circular muscle-layer in the 
proboscis of Hoplonemertea and the above described 
structures. Moreover, the Hoplonemertea have such a highly 
developed proboscis with a stylet, reserve stylets, the presence 
of the whole middle part (Text-fig. 28) with its glands and 
ejaculatory duct, etc., that the obtaining of a new muscle- 
layer seems nothing compared with the development of the 


TEXT-FIG. 32. 


Soy, vhynch .cav, 


oo 2 
= 3 ? = ea a ea 
7 2 oy BEE d.t.comny,. 
2 : Pokaan ae ‘ 
S 5 : ee = S25 525000 S50Esemnoe 
. = 7 va . 5 a E a 
ae 2 . So Le Ge = =e 4 IT, rhynchy. 
‘ pit Oy aia . . dy Be 
x 2 « a= , : ‘ o : : 2, ras . ewe > 
2S ROA Ug eee ee oe oe eas 
a Nos ° - sos - - Seine 
| TT Poo Wooo ae _ & — ee a 
=a aa se ie 
= 2S oe eo eS eee 2 SS ee we 


Schema of a longitudinal section through Cerebratulus margi- 
natus after Birger (9, pl. xxi, fig. 1). d.n. comm. Dorsal nerve 
commissure. v.n.comm. Ventral nerve commissure. rhynch. 
Rhynchodemu. dig. Digestive tract. prob. Proboscis. rhynch. 
cav. Cavity of the rhynchoceel. 


armature. And if we regard this layer as an acquisition of 
the proboscis itself, the facts found in Hnopla are in perfect 
accordance with those in Anopla. 

After all, it seems to me that the structure of the proboscis 
in all Nemerteans together with the structure of the rhyncho- 
coelom can only lead to this conclusion, that they both took 
their origin from the body-wall, a separation being brought 
about in the musculature, which, as in embryology, caused the 
development of a proboscis and of a sheath. 

Proboscides have been described in many invertebrates ; 


300 DR. GERARDA WYNHOFF. 


they are known in Cestodes, in Turbellaria, in Hirudinea, in 
Acanthocephala, in Sipunculids and Priapulids, in Echiurids 
and in Kinorhyncha, and in Nemertines. There can be no 
doubt that these proboscides are not homologous with one 
another. In Kinorhyncha, Priapulids and Sipunculids the 
so-called proboscis consists of the anterior part of the body, 
bearing the mouth at its top. The body-cavity extends into 
this part, and muscles extend from this region of the body- 
wall to the wall of the trunk, causing the invagination of 
the anterior part as a prolongation of the pharynx. 


TErxtT-FIG. 33. 


tf / d TE. comm. 
fe = sro > 
[rit sanee 2982 S0Re= === 2S SSae Bae - ood 


Schema of a longitudinal section through Malacobdella grossa 
after Birger (9, pl. xvill, fig. 2). atr. Atrium. rhynch.w. 
Rhynchocelomic wall. d. n. comm. Dorsal nerve commissure. 
v.n.comm. Ventral ditto. dig. Digestive tract. 


In Hirudinez the so-called proboscis is quite another 
structure. The wall of the pharynx of Rhynchobdellida 
shows a protrusible part, which bears the opening of the 
intestine at its extremity when erected, in the manner of the 
pharynx of Turbellaria. Now, Birger (8) suggested that 
the pharynx of Turbellaria is homologous with the pro- 
boscidean system of Nemertea. Both structures, therefore, 
must be considered more closely in this connection. 

Biirger was brought to this conclusion by certain very 
characteristic features in Enopla. For, though in all Anopla 
the proboscis is extruded through a separate proboscis-pore 
situated in front of the brain (J'ext-fig. 52), and the mouth is 


THK PROBOSCIDIAN SYSTEM IN NEMERTINES. 301 


always found behind the cephalic ganglia, digestive and 
proboscidean systems having no connection whatever with 
each other, they are closely connected in all Enopla (Text- 
figs. 33 and 34), with the exception of Drepanophoride. 
The genus Malacobdella, in many features so widely different 
from all Hoplonomertea, shares this character, though the 
way in which digestive tract and proboscidian system are con- 
nected is quite exceptional in Malacobdella (Text-fig. 33). The 
mouth is situated at the anterior extremity of the head, being 


TExT-FIG. 34. 


oes. 


Schema of a longitudinal section through Nemertopsis 
peronea after Birger (9, pl. xv, fig. 1). rhynch. Rhynchodeum. 
rhynch. w. Rhynchoccelomic wall. prob. Proboscis muscles. 
oes. cesophagus. 


wide and slit-like as in many Heteronemerteans, and giving 
entrance to the large cesophagus which lies underneath the 
cerebral ganglia, and into which the intestine proper opens. 
The rhynchodeum, which, in fact, is so small that as such 
it does not seem to exist, opens into the dorsal wall of the 
cesophagus, the anterior part of the latter being called atrium 
because of this fact. 

In Hoplonomertea exactly the reverse of this condition is 
found (Text-fig. 34); here the proboscis-pore opens to the 
exterior, giving entrance to a well-developed rhynchodeum, 
in the hinder part of which the proboscis is inserted. Through 


_Pproe. rhynsh,.w. rhynch. 
ee 
ays 
Mi 
S 5 - LT 
He - >. 
aA. - re F H 
Eee ee =sssa0 ; ee TESS y 
ee ate = : 7 & 
SSeS —_— 2S ee J 
oan ae Soe aa ii 
; 2 cw 
ars) 


302 DR. GERARDA WYNHOFF. 


an opening in the ventral wall of the rhynchodzeum the ceso- 
phagus gets an open connection with the rhynchodzum. 
Biirger took the rhynchodwum to be the homologue of the 
atrium of Malacobdella, and declared the proboscidian system 
to be a kind of appendix of the digestive organs. As both 
cesophagus and rhynchodeeum originate by invagination of 
the ectoblast, lined by the mesoblastic layer which gives rise 
to the muscular coat of these organs, Biirger’s hypothesis 
seems to be not impossible. 

Biirger (8), who stands for the Turbellarian relationships of 
Nemertea, compared the intestine of both classes, and found 
a pharynx in the first and a proboscidian system in the other 
class without any homologues. He then suggested that they 
were themselves homologous, both being derived from the 
epiblast and the mesoblast underneath. ‘The pharyngeal sac 
of Turbellaria should be the homologue of the atrium s, 
rhynchodeum in Nemertea. The great difference between 
the two organs, however, consists in their different relation to 
the cesophagus, which passes right through the pharynx, and 
has no direct connection with the proboscis; for Birger 
suggested, as we have done, that a splitting must have taken 
place in the muscular coat of the pharynx, giving rise to a 
rhynchoceelom. This seems altogether very plausible, but 
how is the emancipation of the pharynx from the digestive 
tract to be explained? Biirger says: ‘‘ Denken wir den 
Pharynx nicht in den Vorderdarm gestiilpt, sondern tiber 
oder vor der Anlage des Vorderdarms in das Parenchym 
gvewachsen und dann einen Spalt im Mesoderm des so ver- 
schobenen Pharynx entstanden, dieses in zwei Blatter teilend, 
so bekommen wir den Pharynx in einer vor oder iiber dem 
Vorderdarm befindlichen Héhle mit mesodermaler Wandung 
zu liegen, welche dem Rhynchocélom homolog sein wiirde, 
der Pharynx selber aber verhielte sich vollstandig wie der 
Nemertinenriissel.” The way in which the cesophagus must 
attain a direct communication with the pharyngeal sac is 
rather absurd: ‘“ Es ergiebt sich ohne Weiteres, dass wir in 
dem Pharynx ein Zuviel haben, denn nur seine Tasche, die, 


THE PROBOSCIDIAN SYSTEM IN NEMERTINES. 303 


nachdem wir den Pharynx exstirpiert haben, direct 
mit dem Darm communiciert, entspricht dem Nemertinen 
vorderdarm.” Even if Biirger did not suppose this mode of 
arriving at the described situation to have really taken place, 
as I am sure he never did, I cannot see in what other way 
the pharynx could be emancipated from the digestive tract. 
It is quite impossible to picture the way in which these 
changes had to take place without any such lesion, and on 
this ground alone the theory of Biirger seems to lose vitality. 
But there is more. Birger gives three reasons to support his 
views. ‘The first one we have already given above. The 
second support is found in embryology: ‘‘ Der Nemertinen- 
riissel entsteht stets aus einer Hctodermeinstiilpung, die 
mit einem diese umgebenden Mesodermwulste verschmilzt. 
Die Anlage des Riissels erfolgt bei den Metanemertinen am 
selben Orte wie die des Vorderdarms und mit ihr gemein- 
schaftlich.” ‘he similarity of histological structure of the 
two parts is the third one. As to the latter, the histological 
resemblance of pharynx and proboscis consists only in the 
presence of an epithelium and the muscular layers of the 
body-wall, and therefore the same statement as before, that 
both are structures of the body-wall, must be repeated. 
Real histological resemblance there cannot possibly be, as 
the epithelium of the proboscis in Nemertines shows a great 
variety of elements in different parts and different species. 
That the muscular tissue of Nemertines and Turbellaria agree 
in structure is too well known to have any proving impor- 
tance in this case, for the proboscis musculature is built up 
in exactly the same way as the muscles of the body-wall. 
Biirger’s third reason, therefore, is not at all convincing. 
The embryological facts, which follow in the second place, 
have since been contradicted by Lebedinsky and Salensky. 
In the first place, Biirger compares the invagination of the 
proboscis-ectoderm with the extirpation of the pharynx- 
ectoderm; then the pharyngeal sac is the primarily planned 
organ, the pharynx developing afterwards at the bottom. 
In Nemertines the proboscis is the organ that appears first, 
VoL. 60, PART 2,—NEW SERIES, 21 


304 DR. GERARDA WYNHOFP. 


a later invagination of ectoderm giving rise to the rhyncho- 
dzum, the homologue of the pharyngeal sac. 

But even if one takes into consideration the plain embryo- 
logical facts in Nemertines, it seems difficult to come to the 
same conclusions as Biirger. For in what way do Nemerteans 
get a mouth? In Anopla the blastopore sometimes, after 
becoming closed, is transferred to the inside of the animal by 
the invagination of the cesophagus, In Knopla the facts are 


TEXT-FIG. 35. 


Longitudinal section through Prosorochmus after Salensky (27, 
fig. 8). @s.atr. Hsophageal atrium. rhynch. Rhynchodeum. 
vh.c. Rhynchocelomic cavity. gl. Glands. dig. Digestive 
canal. prob. Cavity of inverted proboscis. rhe. Rhyncocel. 


less simple. Drepanophorus, the genus in which cesophagus 
and rhynchodezum open separately, shows no connection at 
all between the two systems, not even in embryology ; for 
here the blastopore is closed, the narrow entodermic part 
giving rise to the blind gut by being removed forward. The 
primary ectodermic cesophagus invaginates near the probo- 
scidian system, but perfectly separate. The place where 
cesophagus and entoderm communicate in Drepanophorus 
has, therefore, nothing to do with the same spot in Hetero- 


THE PROBOSCIDIAN SYSTEM IN NEMERTINES. 305 


uemerteans. In all other Hoplonemertea the primary ceso- 
phagus originates in exactly the same way ; the mouth closes 
afterwards, and the primary cesophagus gets a new opening 
to the exterior through the rhynchodeum.  lLebedinsky 
described in Tetrastemma vermiculus how the rhyn- 
chodzeal epithelium grows towards the cesophagus, a bifurca- 
tion of the cavity being produced. Salensky (27), however, 
is of opinion that in Prosorochmus the communication is 
brought about by the cesophagus, which has produced through 
histological differentiation an atrium of its own (Text-fig. 55). 
The difference of opinion between these two authors does not 
matter in the least for our decision, for it is evident that the 
conditions in Drepanophorus are primitive, and that the facts 
on which Birger based his theory are not primitive at all, as 
the Drepanophorus stage is passed through in ontogeny. 
This fact, the difference between the communication in 
Malacobdella and Hoplonemertea, the impossibility of a 
pharynx developing into a proboscis, the difference of origin 
between the two organs, the differences in the connection 
between pharyngeal sac and pharynx at one side, proboscis 
and rhynchodeum at the other, the insufficiency of the grounds 
on which Biirger bases his theory, make me very sceptical as 
to its value. In fact, I cannot find any other point of resem- 
blance between the two organs than that they are derivatives 
of what the Germans call die Hautmuskelschlauch. 

There are, however, other organs in Invertebrates, even in 
Platyhelminthes, which have a similar origin, and we shall 
have to look for comparison between them. For example, the 
proboscis of Echiurids is a structure of the body-wall—at least 
it consists of an epidermis and muscles only ; moreover it is 
situated in front of the mouth, and is known to be very 
elastic. As, however, this proboscis is supposed to be the 
preoral lobe of the larva and cannot be retracted into the 
body, it does not give any special indication of affinity to 
the same structure in Nemerteans. In Acanthocephala (14) a 
proboscis is found, which can only partly be retracted into 
the body. The proximal part is surrounded by a sheath of 


306 DR. GERARDA WYNHOFF. 


muscles, the circular fibres of which act as protractors; the 
retractor muscle, which retracts proboscis and sheath, the 
two parts being continuous, is inserted at the hind end of the 
sheath, and extends through the body-cavity to the body-wall. 
Though some agreement in these structures in Hchinorhynchus 
and Nemertea cannot be denied, I do not suppose there is 
any relationship between them. The proboscidian system 
of Acanthocephali is entodermic in origin and only secon- 


TEXT-FIG. 36. 


prob. atr. 


Schema of the proboscis of Macrorhynchus croceus after 
von Graff (18, pl. x, fig. 12). prob. atr. Proboscidian atrium. 
prob. Cavity of the introverted proboscis. prob. w. Probos- 
cidian wall. vad. m. Radial musculature. rhynch. w. Rhyn- 
chocelomic wall (Muskelzapfen). Compare with Text-fig. 1. 


darily gets its epidermis. This fact alone seems to exclude 
all possibility of homology with the proboscis of Nemer- 
teans. 

Coming to the consideration of two other structures, both 
in Platyhelminthes, the proboscides of Tetrarhynchus and 
of Proboscida, the four proboscides of the characteristic 
genera of Cestodes superficially show a great resemblance 
to those of Nemerteans. The proboscis is an introvert, 


THE PROBOSCIDIAN SYSTEM IN NEMERTINES. 307 


being retracted by a muscle-bundle at the bottom of the 
tube just as in Nemertines. A muscular sheath is present, 
separated from the proboscis by a cavity filled with a liquid. 
Histologically, however, a’ great difference seems to exist. 
For instance, the proboscis itself has no muscles at all, if 
I understand Pintner (21) right; the wall of the sheath is 
differently constructed at its anterior and posterior part, 
a muscular sheath only existing in the hinder part, the 
anterior being a derivative of the ectoderm. Moreover the 
proboscides of these Cestods are paired, those of Nemertines 
being unpaired. 

Much more striking is the likeness to the proboscis of Tur- 
bellaria proboscida. Salensky (25 and 26) was the first 
to compare them, and though he already proclaimed them 
homologues in 1884, nobody seemed to agree with his opinion. 
Hubrecht (16) raged against it, and afterwards Biirger (8) 
gave a new theory. Salensky, however, was not convinced, 
and tried to give fresh support to his hypothesis in his 
articles of 1909 and 1912 (27, 28). 

The proboscis, which is found in certain genera of Rhabdo- 
ccelida, is localised at the anterior end of the head. Von 
Graff (13) described the organ in this way: “Der Probisciden 
riissel ist nichts weiter als eine bleibend gewordene Einstiil- 
pung des Vorderendes, wie wir sie voriibergehend bei Meso- 
stomum rostratum entstehen sehen.” It consists of two parts, 
the proboscis and a kind of atrium, through which the pro- 
boscis is everted (Text-fig. 36). The cavity of both is lined by 
_ the somewhat changed epiderm. The wall of the atrium has, 
like the body-wall, a muscular coat, which is the direct con- 
tinuation of the muscular layers of the body-wall. At the base of 
this sac or atrium the proboscis is inserted. Here the muscular 
coat is broken up into two layers; the inner layer continues 
along the epidermis, the other one forms the outer lining of 
the proboscis. A thick layer of radially placed muscle-fibres 
fills the space between the two parts of the muscular coat of the 
body-wall. In Nemertines (Text-fig. 1) we find the same 
arrangement: an atrium to the proboscis, being part of the 


308 DR. GERARDA WYNHOFF. 


body-wall, that has grown inward and is called rhynchodzum. 
The continuation of the epithelium is found in the epithelium 
of the proboscis. The muscular coat of the body-wall is split, 
and has given rise to the muscular wall of the proboscis and of 
the sheath. The cavity of the rhynchoccel is even traversed by 
a band of radially placed muscle-fibres, the retractor muscle. 
So Salensky (26) proposed the following homologies. 


Rhabdocéles probosciféres. Némertiens. 


‘1. Poche de la trompe. 1. Vestibule de la trompe. 
2. Epithélium de latrompe. 2. Epithélium de la trompe. 
3. Couche interne dela calotte 3. Couche musculaire de la 
musculaire (Muskelzapfen trompe. 
v. Graff). 
4, Couche externe dela calotte 4. Parois de la gaine de la 
musculaire. trompe. 
5. Muscles radiaires de la 5. Bride musculaire. 
calotte. 


When von Graff (18) described the structure of the pro- 
boscis of Macrorhynchus and other Rhabdoccelida the origin 
of the proboscis and sheath in Nemertines was not known. 
He could not find any homologue of the sheath in his genera, 
aud on this ground denied any relationship between the 
proboscides in these two classes of Platyhelminths. Another 
ground for denying it was found in the division of the pro- 
boscis of Nemertines into two parts, the posterior glandula 
part not being present in Turbellaria proboscida. This 
ground, however, seems not to be very important; the pro- — 
boscis of Rhabdoccelida represents rather a simple stage of 
organisation in comparison with the same structure in Nemer- 
tines. It does not seem to be necessary to go into such details 
of. structure, especially not since we know what a great 
variety of epithelial structure is to be found in different 
genera of Nemertea. 

Salensky looked for support for this theory to embryology. 
And certainly the fact that the muscular coats of rhyncho- 
ccelomic and proboscidian walls take origin out of the same 


THE PROBOSCIDIAN SYSTEM IN NEMERTINES. 309 


layer of mesoblastic elements, a splitting giving rise to the 
cavity of the sheath, is highly in favour of his views. It was 
exactly on this point that Hubrecht did not agree with him ; 
therefore Hubrecht, who, moreover, claimed the Annelida- 
relations of Nemertea, could not possibly follow Salensky. As, 
however, later investigations proved that Salensky was right, 
this reason for disagreeing with his theory gave way. But 
Hubrecht gave other reasons. In the ‘ “Challenger”’ Report’ 
he wrote, p. 104: ‘Salensky would probably not have made 
his startling hypothesis above alluded to, based on ontogene- 
tical observations of a scission in the proboscidian wall, by 
which (1) a muscular proboscidian sheath surrounding the 
proboscis becomes separated from, and independent of, the 
musculature of the proboscis itself, and (2) an isolated ccelome 
—the proboscidian cavity—is originated, if he had been as 
well acquainted with the comparative anatomy of the animals 
about which he writes as he is with certain details of their 
ontogeny.” The anatomical facts, that tell against Salensky, 
Hubrecht summarises in this way (loc. cit., p. 103): “There 
can hardly be any doubt, when we take into consideration all 
the morphological data at our disposal, that the muscles com- 
posing the proboscidian sheath gradually took their origin by 
the increase and modification of pre-existing muscular elements 
which belonged to the body-wall and to the body-parenchyma 
before the proboscis, modified from a tactile organ, as it 
appears to have primitively been, had yet become evolved, 
through the growth inwards of the anterior tip of the body, 
into an aggressive weapon with stylet or nematocysts, etc.” 
Since this was written many facts have been disclosed as to 
the anatomy of the proboscidian system, and we have seen 
that they lead us to the conclusion, that the sheath is 
part of the body-wall, and has not originated in loco out 
of muscular fibres in the parenchyma of the body. The 
central parenchyma does not possess any circular muscle- 
fibres. Wherever other than longitudinal fibres have been 
found in it they have been demonstrated to be part of one 
of the layers of the body-wall, as a rule the inner circular 


310 DR. GERARDA WYNHOFF., 


muscle-layer. Longitudinal muscle-fibres may be present in 
the central connective tissue, and we have already seen that 
in some cases they are arranged around the proboscis sheath 
so as to form a secondary longitudinal coat. This, however, 
is not the rule, not even in Anopla, where a central longitu- 
dinal musculature is developed in the majority of genera. 
Our conclusion is exactly the reverse of the above statement 
of Hubrecht; the facts described in the first part of this 
article seem not to allow the denial of our conclusion. Both 
on anatomical and embryological grounds we must declare the 
muscular coats and walls of the proboscidian system of Nemer- 
tines to belong together ab origine; the facts put forward by 
Hubrecht against the theory of Salensky have been refuted by 
later embryological investigations. I hope to have given a 
new support to Salensky’s ingenious theory by this compara- 
tive study of the anatomy of the proboscidian system. The 
fact, stated by Hubrecht, that “we find the shorter pro- 
boscides and the less significant proboscidian sheaths among 
the more primitive genera of Nemertea,” is, though mentioned 
by him as telling against the ‘'urbellarian relations, in perfect 
accordance with, if not in favour of, Salensky’s theory. 


LITERATURE. 


Arnold (1898).—* Zur Entwicklungsgeschichte des Lineus gesse- 
rensis O. F. Miller,’ ‘Trav. Soc. Imp. Nat. Pétersb.,’ vol. 
xxviii, livr. 4. 


P 


2. Bergendal (1900)—‘Callinera biirgeri Bergendal, en rapre- 
sentant for ett afvikande sligte bland Palaionemertinerna,” 
‘Lunds Univ. Arsskr.,’ Bd. xxxvi, Afd. 2, No. 5. 


3h (1902).—** Zur Kenntniss der nordischen Nemertinen ITI,” 
‘Bergens Mus. Aarbog.,’ No. 4. 

4. (1902).—** Valencinura bahusiensis Brgdl., ein Beitrag 
zur Anatomie und Systematik der Heteronemertinen,” ‘ Lunds 
Univ. Arsskr.,’ Bd. xxxviii, Afd. 2, No. 3. 

5. (1902).—* Eine der construirten Urnemertine entsprechende 


Paleonemertine aus dem Meere der schwedischen Westkiiste,” 
‘Zool. Anz.,’ Bd, xxv. 


THE PROBOSCIDIAN SYSTEM IN NEMERTINKES. 31] 


6. Bergendal (1903).—‘* Beobachtungen iiber den Bau von Carinoma 
Oudemans nebst Beitrage zur Systematik der Nemertinen,” 
‘Lunds Univ. irsskr.,’ Bd. xxxix, Afd. 2. 


7. Burger, O. (1894).—* Studien zu einer Revision der Entwicklungs- 
geschichte der Nemertinen,” ‘ Ber. Naturf. Ges. Freiburg i B., 


1894. 

8. (1895).—* Die Verwandtschaftsbeziehungen der Nemer- 
tinen,’ ‘ Verh. Deutschen Zool. Ges. Strassburg.’ 

9. (1895).—** Die Nemertinen,” Fauna u. Flora Neapel., Mon. 
22. 

10. (1897-1907).—** Die Nemertinen,” ‘ Bronn Klass. u. Ordn. 
Tierreich.’ 

10a (1900).—** Die Nemertinen,” ‘ Wiss. Erg. D. Tiefsee Ex- 


pedit. Valdivia,’ Bd. xvi, Lief. 2. 
1l. Coe, W. R. (1895).—* Descriptions of Three New Species of New 
England Paleonemerteans,” ‘ Trans. Conn. Acad.,’ vol. ix. 
(1905).—‘*‘ Nemerteans of the West and North-west Coasts 
of America,” ‘ Bull. Mus. Harv. Coll.,’ vol. xlvii. 

13. von Graff, L. (1882).—‘ Monographie der Turbellarien.’ I, ** Rhab- 
docelida,” Leipzig. 

14. Hamann, O. (1891).—‘* Monographie der Acanthocephalen,” ‘ Jen. 
Zeitschr. Naturw.,’ Bd. xxv. 

15. Hubrecht, A. A. W. (1885).—‘ Proeve eener ontwikkelingsgeschie- 
denis van Lineus obscurus, Utrecht, Leeflang. 

16. ——— (1887).—** Nemertea,” ‘ Rep. Sc. Res. Challenger Zool.,’ vol. 
abe 


17. Joubin, L. (1890).—‘* Recherches sur les Turbellariés des cétes de 
France (Némertes),” ‘ Arch. Zool. expér. gén.’ (2), Bd. viii. 

18. Lebedinsky (1896).—* Zur Entwicklungsgeschichte der Nemer- 
tinen,” ‘ Biol. Centralbl.,’ Bd. xvi. 

(1897).—Ibid., Bd. xvii. 

(1897).— * Beobachtungen iiber die Entwicklungsgeschichte 

der Nemertinen und Nachtrag.,” ‘Archiv Mikr. Anat. u. Entw. 

gesch.,’ Bd. xlix. 


20. — 


21. Pintner (1895).—“ Versuch einer morphologischen Erklirung des 
Tetrarhynchenriissels,” * Biol. Centralbl.,’ Bd. xvi. 
22. Punnett, R. C. (1901).—*On Two New British Nemerteans,” 
‘Quart. Journ. Mier. Sci.’ (N.s.), vol. 44. 
(1901).—**On Some Arctic Nemerteans,” ‘ Proc. Zool. Soe. 
Lond.,’ 1901, vol. ii, Part I. 
VOL. 60, pAkT 2.—NEW SERIES, 22 


23. 


351 


24 


2 DR. GERARDA WYNHOFF. 


. Punnett, R. C. (1903).—* On the Nemerteans of Norway,” ‘ Bergen 
Mus. Aarbog.,’ 1903, No. 2. 


25. Salensky, W. (1884).—‘* Recherches sur le développement du Mono- 


pora vivipara,” ‘Arch. de Biologie,’ v. 


26. (1886).—** Bau und Metamorphose des Pilidium,” ‘Z. W. Z., 
Bad. xliii. 

2 i (1909).—** Ueber die embryonale Entwicklung des Pro- 
sorochmus viviparus,” ‘ Bull. Acad. Imp. Sciences, Petersb., 
1909. 

28. (1912).—** Ueber die Morphogenese der Nemertinen. I. 


Entwicklungsgeschichte der Nemertine im Pilidium,” ‘Mém. 
Acad. Imp. Sciences, Pétersb.,’ vol. xxx, No. 10. 


29. Schutz, V. (1912)—‘*Paralineus elizabethe,” ‘Z. W. Z., 


30 


31. 


Bd. cio. t 
. Thompson, C. B. (1901).—‘“* Zygeupolia litoralis, a New Hetero- 
nemertean,” ‘Proc. Acad. Nat. Sc. Philad.,’ vol. liii, p. 2. 
Wynhoff, G. (1910).—** Die Gattung Cephalothrix und ihre Bedeu- 
tung fiir die Systematik der Nemertinen; I, Anatomischer Teil,” 
‘Zool. Jahrb.,’ Bd. xxx, Abt. f. Anat. 


32. (1912).—** Die Systematik der Nemertinen,” ‘Zool. Anz.,’ 
Bd. xl. 
23h (1913)—** Die Gattung Cephalothrix, ete.; II, Systema- 


tischer Teil,” * Zool. Jahrb., Bd. xxxiv, Abt. f. Syst. 


GAMETOGENESIS OF GRANTIA COMPRESSA. 


313 


Observations on the Gametogenesis of Grantia 


compressa. 


By 
Arthur Dendy, D.Sc., F.R.S., 


Professor of Zoology in the University of London (King’s College). 


With Plates 23 to 26. 


CoNtvTENTs. 


) GENERAL INTRODUCTION 
) MarerraLt AND METHODS OF INVESTIGATION 
a THE BREEDING SEASON AND LIFE-CYCLE 
(D) THE DISTRIBUTION OF THE GERM-CELLS IN THE See 
(E) THE RELATIONS BETWEEN THE DIFFERENT TISSUE- 
ELEMENTS 
(F) aa 
(a) Historical . : 
(b) Origin and Growth of the ftine: Casati 
(c) Multiplication of the Odgonia 
(d) Growth and Feeding of the Young Onestes in the 
Flagellate Chambers 2 
(e) Growth and Feeding of the Oacee in the Mesc glea . 
(f) Maturation of the Odcytes 
(g) The Nurse-cells and their Origin ; Pliers eens 
(h) Summary and General Remarks on Odgenesis . 
(G) SPERMATOGENESIS 
(a) Historical . : 
(b) Origin and Growth of the Berrerey cnauaeane 
(c) Formation of the Spermatocysts or Cover-cells 
(d) Development of the Sperm-morule from the Primary 
Spermatogonia : ; 
(e) The Formation of Speniioraa Comparison with 
Sycon, ete. 
(H) FERTILISATION OF THE Ovum: 
(k) List oF LITERATURE REFERRED TO 
(L) EXPLANATION OF PLATES : : : 
VOL. 60, PART 3.— NEW SERIES. 23 


314 ARTHUR DENDY. 


(A) GeneraL Intropuction. 


Tue problem of the gametogenesis of Sponges is one which 
has attracted a good deal of attention from time to time, but. 
it can hardly be said that our knowledge of the subject is as 
yet by any means in a satisfactory condition. As far back as 
1851 Huxley described the occurrence of supposed sperma- 
tozoa in Tethya, and in 1854 Carter did the same for 
Spongilla, but Haeckel is probably right in supposing that. 
both these authors were in reality describing flagellate collared 
cells, a mistake which, owing to the difficulty of detecting the 
collar in certain conditions, it’ is very easy to make. 
Haeckel, indeed, claims for Lieberkiihn the original discovery 
in 1856 of the sexual differentiation of sponges and of the 
ova and spermatozoa (in Spongilla). 

Haeckel points out the great difficulty in finding the 
spermatozoa of sponges, a difficulty which has been, and 
doubtless still is, experienced by many spongologists, and 
which arises chiefly, no doubt, from their extremely minute 
size, the fact that they occur scattered throughout the 
sponge and not segregated in definite gonads, and their 
lability to confusion with other tissue elements. The ova, on 
the other hand, in their later stages of growth, are readily 
recognisable by their large size, and, at certain periods, their 
conspicuous vesicular nuclei, although these again are not 
collected in definite ovaries, but occur scattered through the 
sponge-tissues. 

The spermatozoa of calcareous sponges were discovered by 
Haeckel himself in 1871, and at about the same time by Eimer. 
In his well-known monograph, “ Die Kalkschwimme ” (1872) 
the former author gives an account of all that was previously 
known of the germ-cells of sponges, and describes his own 
observations and conclusions with regard to the Calcarea. 
He finds reasons for deriving the spermatozoa from 
collared cells of the so-called entoderm (gastral layer), and 
gives several figures of sperm-morule lying amongst the 
ordinary collared cells in the gastral layer of several species. 


GAMETOGENESIS OF GRANTIA COMPRESSA. ey 


With regard to the ova he remarks that the question of the 
origin and of the original position of the egg-cells is the 
most difficult and darkest part in the histology of calcareous 
sponges, an opinion with which those who have studied the 
question will hardly feel inclined to quarrel. Haeckel himself 
at first regarded the ova as originating in the “ exoderm,”’ 
in which he, of course, includes everything but the “ ento- 
derm,” but his more mature conclusion is that they really 
originate in the “ entoderm.” He says (1872, p. 159): “ Hin- 
zelne Geisselzellen des Entoderms vergréssern sich, ziehen 
thren schwingenden Geisselfortsatz ein, umd entwickeln sich 
direct durch Aufblaihung des Kernes und bedeutende Volums- 
Zunahme des Protoplasma zu Hizellen.” 

Since Haeckel wrote, opinion as to the origin of the germ- 
cells in sponges has changed, and though very few writers 
have recorded any detailed observations, it is generally held, 
in accordance with the teachings of F. E. Schulze and others, 
that both ova and spermatozoa arise from amceboid wandering 
cells in the mesoglcea, which, of course, is part of Haeckel’s 
‘*‘exoderm.” No one, since Haeckel’s time, appears to have 
seen the sperm-morule lying amongst the collared cells in 
the walls of the flagellate chambers, and no one has traced 
the stages by which collared cells might be supposed to have 
been converted into spermatogonia. Haeckel himself seems 
to have been by no means satisfied that his own observations 
on the subject were conclusive, and Poléjaeff, writing ten 
years later (1882), evidently regarded them with grave 
suspicion. 

Nevertheless, I believe that Haeckel was essentially right 
in maintaining that odgonia and spermatogonia both arise 
from collared cells. It is, at any rate, quite certain that 
bodies resembling spermatogonia are frequently to be found 
in small morula-like clusters amongst the collared cells in the 
walls of the flagellate chambers of Grantia compressa, 
very much as figured by Haeckel for other Calcarea, except 
that I have never observed the tails of the spermatozoa in 
this situation, and, indeed, have only seen the sperm-morule 


316 ARTHUR DENDY. 


in properly prepared, stained sections. Between the original 
collared cell and the mature germ-cell of either sex amceboid 
stages intervene, which are frequently to be found in the 
mesogloea, at any rate in the case of the odgonia. It is an 
easy matter to trace the growth of the amceboid odgonia into 
mature ova, but the question is—Whence come the earliest 
amoeboid stages ? 

The modern answer to this question is that they are 
derived from primitive amceboid cells (“ archzocytes ”’), but 
I cannot help suspecting that this answer is merely an echo 
of Weismann’s well-known views as to the early segregation 
of the germ-cells and the continuity of the germ-plasm. It 
depends, so far as | am aware, not upon direct observation, 
but upon a process of reasoning by exclusion. The germ- 
cells are supposed to remain over after the somatic cells have 
been subtracted from the sum total of cells derived from the 
segmenting ovum. 

I believe that this particular theory of the origin of the germ- 
cells in sponges originated with Dr. Otto Maas (1894), who 
sums up his conclusions as follows (p. 35) : 

“© (1) Wir koénnen hier eine directe Abstammung der Keim- 
zellen der einen Generation vom Ei nachweisen, indem durch 
Subtraction aller somatischen specialisierten Hlemente schli- 
esslich eine Anzahl indifferent gebliebener Elemente iibrig ist, 
die Urgeschlechtszellen. 

“ (2) Der Hauptunterschied zwischen den somatischen und 
den Geschlechtszellen zeigt sich vom Anfang wie spater im 
Kern, und zwar in der Quantitiét und Anordnung des Chro- 
matins.” 

Anyone who has studied the gametogenesis of sponges knows 
how greatly the condition of the chromatin varies in different 
stages, and I do not believe it is possible to indicate any nuclear 
character by which primordial germ-cells can be definitely 
distinguished from somatic cells, at any rate inthe present state 
of our knowledge. The alleged distinction between the two 
groups of cells is a purely theoretical one. Indeed, the latest 
writer on the subject, Max Jérgensen (1910), has already come 
to the conclusion that the distinction which Maas endeavours 


GAMETOGENESIS OF GRANTIA COMPRESSA. Bays 


to draw between germ-cells and somatic cells cannot be 
maintained in sponges, and “ dass es sehr wohl zur Bildung von 
Geschlechtszellen aus somatischen Zellen kommen kann, wie 
man dies ja bei dem primitiven Charakter der Schwamme von 
vornherein erwarten wiirde” (p. 170). This author derives the 
odgonia of Sycon from ordinary ‘“‘mesoderm”’ cells rather than 
from primitive undifferentiated amcebocytes (archeocytes), 
though he thinks that the latter may also give rise to odgonia. 
He does not deal with the spermatogenesis. 

My own attention was first particularly directed towards 
the problem of the origin of the germ-cells in calcareous 
sponges under the following circumstances. Some years ago 
Prof. Herdman invited me to write a memoir on the common 
British species, Grantia compressa. In the course of 
this work it became evident that although the ova of this 
sponge were easily recognisable in the maternal tissues in 
various stages of growth, and also embryos in varicus stages 
of development, no one had ever been able to find the sperma- 
tozoa, though they were searched for repeatedly by Mr. Carter 
(1875, p. 25). For the sake of completeness it appeared 
highly desirable that this gap in our knowledge should be 
filled. 

In April, 1912, accordingly, I paid a visit to the laboratory of 
the Marine Biological Association at Plymouth, and although 
I was unable to find spermatozoa in the living specimens, of 
which I examined a large number, I preserved material which 
on subsequent investigation furnished the clue to the solution 
of the problem. I found that the sponge is hermaphrodite, 
producing both male and female germ-cells simultaneously, 
but that the spermatozoa are produced in comparatively small 
numbers, the minute sperm-morule being scattered here and 
there, enclosed in spermatocysts, between the collared cells of 
the chamber walls, and also occurring free in the flagellate 
chambers. The spermatogonia in these sperm-morule are 
extremely minute, andI am unable to say anything with regard 
to their mode of division. Apparently they are sometimes 
transferred as sperm-morule to the inhalant canals of the 


318 ARTHUR DENDY. 


Same or of another individual, where they break up into 
Spermatozoa, but it is probable that they may sometimes 
break up into spermatozoa before leaving the parent sponge. 
The evidence on these points is, however, curiously scanty, 
and I conclude that the spermatozoa are rarely, if ever, 
liberated in large numbers. 

With regard to the process of odgenesis, on the other hand, 
I found a number of very interesting stages, and with the 
exception of the second maturation spindle (described by Max 
Jorgensen in Sycon), am able to give a fairly complete 
account. My observations agree in many respects with those 
recorded by Jérgensen on the oogenesis of Sycon, but I am 
able to add a good deal to his account, and in some respects 
my interpretations are different. ‘Though I have repeatedly 
observed mitotic figures of various types I have always found 
the chromosomes very small and ill-defined, and I cannot 
pretend to give such precise descriptions of the mitotic 
phenomena as Jérgensen has done. It is very difficult to 
understand his account of these phenomena, or to harmonise 
it with currently accepted views, but I must leave to specialists 
in cytology the detailed criticism of his work in this respect. 
I may perhaps say, however, that many of his figures appear 
to me to be very diagrammatic. 

One of the most remarkable phenomena observed during 
my investigations is the nutrition of the growing odcyte by 
means of phagocytic nurse-cells. Very extensive phagocytosis 
has also been observed in the case of certain large amcebo- 
cytes, which seem to devour the young germ-cells in some 
cases in a wholesale fashion. 

Though well aware that I am not able to give by any 
means a complete account of the history of the germ-cells in 
Grantia compressa, I hope that the following pages may 
not only help to fill a conspicuous gap in our knowledge of 
this common British sponge, but also throw some new light 
on the difficult problem of gametogenesis in sponges generally. 

I must take this opportunity of thanking my friends at the 
Plymouth Laboratory, especially Dr. Allen and Mr. Orton, 


GAMETOGENESIS OF GRANTIA COMPRESSA. a19 


for their hospitality and assistance during my visit. I am 
also greatly indebted to the University of London for the use 
of their table at the Laboratory. 


(sp) MarertaL and Meruops or INVESTIGATION. 


My observations at the Plymouth Laboratory, made from 
April 10th to April 22nd, 1912, were mainly directed towards the 
discovery of the spermatozoaof Grantia compressa. ‘The 
sponge may be obtained in large numbers close to the Labo- 
ratory, at low water, and I also received a number of fresh 
specimens from Drake’s Island and Rum Bay. They varied 
greatly in size, from quite small to as much as 80 mm. in 
height by 18 mm. in breadth (a specimen from Rum Bay). 
A large number were microscopically examined in the hving 
condition, either by teasing or by means of hand sections, or 
by pipetting out the contents of the central gastral cavity, 
but my search for living spermatozoa was fruitless. 

I preserved a considerable number of specimens, however, 
for future examination, and the results recorded in the present 
paper are based almost entirely upon the study of these by 
means of paraffin sections. 

The material that turned out satisfactorily was fixed either 
in strong Flemming’s solution, or in a mixture of Fiemming, 
formol and sea-water. In the former case it was graded up, 
after washing, to 70 per cent. alcohol; in the latter it was 
preserved in formol sea-water. 

The sections were, for the most part, cut of a thickness 
of 5y and stained on the slide. I found that iron-brazilin 
gave excellent results, but iron-hematoxylin was also used. 

For staining in bulk borax-carmine or paracarmine was 
employed, the latter being sometimes followed on the slide 
by picro-indigo-carmine, but without much effect. 

I desire to express my thanks to my skilful laboratory 
assistant, Mr. Charles Biddolph, for the care which he has 
taken in preparing the sections. 

Although many specimens were preserved, only five have 


320 ARTHUR DENDY. 


actually been used for the purposes of the present investiga- 
tion, numbered in my notes 11, 21, 22, 25 and 24 respectively. 

It will save repetition if I give particulars concerning 
these specimens at once, and of the mode of treatment of the 
material. 

No. 11.—A very large specimen from Rum Bay, 80 mm. 
high by 18 mm. in breadth; brought in about mid-day (April 
15th) and examined the same afternoon (about 1 o’clock). 
When examined a vigorous stream was coming out from the 
main vent. A section of the living sponge from near the base 
showed large ova, but no embryos were seen. A section near 
the vent showed smaller germ-cells. The specimen was cut 
in half lengthwise, and half fixed in strong Flemming’s solu- 
tion (in the dark) for half an hour, then washed for an hour 
or more in tap-water and graded through 30 per cent. and 
50 per cent. to 70 per cent. alcohol. 

No. 21.—A rather small specimen, about 14 mm. high, 
collected about mid-day on April 18th, and examined the same 
afternoon. There was an active current issuing from the 
vent, bringing with it a quantity of fine yellowish-grey sedi- 
ment, which collected at the bottom of the glass dish. In 
hand section numerous large rounded cells were seen free in 
and projecting into the flagellate chambers, apparently 
actively moving, but probably only owing to the movements 
of the flagella of the collared cells. Fixed in strong Flem- 
ming’s solution and preserved in 70 per cent. alcohol. 

No. 22.—A moderate-sized specimen, about 25 mm. high 
by 18 mm. broad ; collected about mid-day on April 18th and 
examined the same aftrnoon. Fixed entire in Flemming and 
sea-water formol' for about a quarter of an hour, then washed 
in formol and sea-water and preserved in same. 

No. 23.—A moderate-sized specimen, collected on April 
18th and examined and fixed on the 19th, having been kept in 
the circulation of the aquarium overnight. Stream coming 
from osculum when examined. Found no amecebocytes or 


1 Take 10 c.c. commercial formaldehyde in 90 ¢c.c. sea-water and add 
20 c.c. strong Flemming’s solution. 


GAMETOGENESIS OF GRANTIA COMPRESSA. 321 


other cells in water pipetted from gastral cavity. On exa- 
mination of living section found very numerous rounded 
amoebocytes with coarse granules, and a few with pseudo- 
podia, in the chambers, some, if not most, attached to walls 
by short peduncles. Half of specimen fixed in strong Ilem- 
ming’s solution for about an hour, washed and graded up to 
70 per cent. alcohol. Another part fixed in absolute alcohol. 

No. 24.—A good-sized specimen from Rum Bay, examined 
and preserved on April 22nd, after having been kept in the 
aquarium circulation since April 15th. The collared cells were 
found to be still active and did not show the characteristic 
signs of suffocation. Bulk of specimen fixed in Flemming 
and sea-water formol. 


(c) THe Breeprna Season anvD Lire-Cycie. 


There is strong, indeed, I think conclusive, reason for 
believing that Grantia compressa is an annual sponge, 
growing rapidly during the winter and spring and breaking 
up and perishing in the autumn, after producing numerous 
embryos. I have already given particulars as to the sizes of 
some of the specimens met with at Plymouth in April, 1912, 
and my colleague, Mr. R. W. H. Row, tells me that about the 
middle of August, 1913, when he visited Plymouth, they 
were already breaking up, and it was difficult to obtain a good 
specimen of any considerable size—indeed, most of them had 
apparently already disintegrated. 

The breeding season at Plymouth would seem to begin in 
the first half of April; germ-cells are then being produced in 
enormous numbers, but comparatively few embryos are found. 
At least that was my experience in 1912. 

Previously, in 1911, I had observed mature embryos being 
shot out of the osculum of a specimen which I examined in 
the laboratory in the first week of June. I also find plenty of 
advanced embryos, along with germ-cells in various stages of 
growth, in a specimen collected for me by Mr. Row about the 
middle of August, 1913. The germ-cells (ova), however, are 
nothing like so abundant as in material taken in April. 


322 ARTHUR DENDY. 


It seems, therefore, that the breeding season lasts through- 
out practically the whole of the spring and summer. 

According to Mr. Orton, who kindly allows me to make use 
of information about to be published in the ‘Journal of the 
Marine Biological Association’ (1914), there are really two 
breeding seasons at Plymouth for Grantia compressa. In 
June embryos are discharged from large specimens (which sub- 
sequently disintegrate). These embryos develop into indi- 
viduals which, while still very small, produce numerous 
embryos in October. Mr. Orton has also obtained data sup- 
porting the view that the same specimen may breed twice 
during its life-history—once in late autumn and again in the 
following summer. 


(D) THe DisrrrpuTion or THE GERM-CELLS IN THE SPONGE. 


As a result of my observations I think I have been able 
to establish the fact that Grantia compressa, unlike 
certain non-calcareous sponges, such as Oscarella lobularis 
(Schulze, 1877), is hermaphrodite, producing male and female 
gametes simultaneously. In the case of Sycon raphanus, 
Poléjaeff, as far back as 1882, came to the same conclusion, 
but considered that that sponge afforded an example of in- 
complete sexual separation, some individuals being predomi- 
nantly male and others predominantly female. The former 
were extraordinarily rare, but produced an immense quantity 
of spermatozoa, as well as numerous ova. The latter pro- 
duced very few, or even (apparently) no spermatazoa, but a 
large number of eggs. 

It is quite possible that the same relations may exist in 
Grantia compressa, but if so I have never been fortunate 
enough to find the predominantly male individuals. All 
that I have examined appear to be predominantly female, 
the sperm-morulz occurring only in comparatively small 
numbers. 

Poléjaeff derives both the sperm-morule and the ova (in 
Sycon) from ordinary amceboid wandering cells, and figures 


GAMETOGENESIS OF GRANTIA COMPRESSA. 323 


them scattered in the mesogloea, apparently without any 
special arrangement. 

Goérich (1903) also accepts the usual views as to the origin 
of both male and female germ-cells from amcebocytes, but he 
states that in Sycon raphanus ova are produced in the 
lower two thirds and sperm-cells in the upper third of the 
sponge. He has not, however, followed the spermatogenesis 
beyond its earliest stages. The only generalisation that | 
can make about the distribution of the germ-cells in Grantia 
compressa is that the younger parts of the growing sponge, 
towards the osculum, only contain immature germ-cells, exactly 
as might be expected. 

It may be admitted that in Grantia compressa also the 
germ-cells, both male and female, can be traced back to 
amoeboid wandering cells, but, according to my own obser- 
vations, these amcebocytes can, in their turn, be traced back 
to collared cells of the gastral epithelium lining the flagellate 
chambers. Presumably any collared cell may become directly 
transformed into a primary odgonium or spermatogonium, 
losing its collar and flagellum and becoming ameeboid. The 
evidence for this statement will be presented in the next 
section. 

Having become ameeboid, the young germ-cells are free to 
wander about. The primary odgonia first migrate into the 
mesogloea from the gastral epithelium, and later on, when 
fully grown, they migrate back into the chambers, where 
they undergo repeated division and give rise to small odcytes. 
The young oécytes remain, feeding and growing, for some 
time in the chambers; then they migrate once more into the 
mesogloea, where they undergo enormous growth, followed by 
maturation and fertilisation. 

The migrations of the spermatogonia appear to be of a less 
extensive character. I have reason to believe that they may 
migrate into the mesogloea and there become provided with 
their cover-cells or spermatocysts, but they appear to spend 
most of their existence in, or attached to, the walls of the 
fla ellate chambers. 


324. ARTHUR DENDY. 


It is easy to observe, in hand-cut sections of living specimens 
in the early part of the breeding season, that the flagellate 
chambers contain large numbers of amcebocytes hanging, as 
it were, from their walls. These are, for the most part, 
germ-cells of both sexes in various stages of growth, although 
other amcebocytes may also occur in the chambers. 

There is thus no localisation of the germ-cells in Grantia 
compressa, nothing that can be spoken of as gonads, 
neither ovaries nor testes. Just before undergoing matura- 
tion, however, the relatively enormous ova withdraw their 
pseudopodia and round off, each one taking up a definite 
position behind the gastral epithelium of an adjacent chamber, 
and causing the layer of collared cells to bulge out into the 
chamber. Here fertilisation and the earlier stages of de- 
velopment take place, the embryo becoming surrounded by 
an endothelial capsule, derived from the mesoglcea, during the 
latter process. Finally the ciliated amphiblastula breaks 
through the layer of collared cells into the chamber cavity, 
and is discharged into the sea through the central gastral 
cavity and vent. 

The sperm-morule are also discharged into the flagellate 
chambers, and doubtless find their way out through the vent. 
I have found them, not only in the chambers, but also 
adhering to the outer surface of the sponge and in the 
inhalant canals, though, except in the chambers themselves, 
only in very small numbers. I have also found some evidence 
of their breaking up into spermatozoa in an inhalant canal. 

There can be little doubt that fertilisation is effected by 
spermatozoa which enter the sponge (perhaps as sperm- 
morulz) through the dermal pores with the inflowing stream 
of water, but whether these spermatozoa are derived from 
the same sponge as the eggs which they fertilise, or from 
another individual, would seem to be a matter of pure 
chance. 

In a paper on the ‘Anatomy of Grantia labyrinthica, 
etc.,”’ published in 1891, I expressed the opinion that the ova 
migrated through the walls of the inhalant canals and were 


“<> 
a 


re 


j 
; 


GAMETOGENESIS OF GRANTIA COMPRESSA. 325 


fertilised while suspended in the inflowing stream of water, 
subsequently migrating back to undergo their development 
in the mesoglea. I certainly did observe amcebocytes sus- 
pended in this position, but I now realise that they were far 
too small to be mature ova, and must unreservedly withdraw 
my interpretation of the observation. It is evident that 
maturation and fertilisation (in Grantia compressa) both 
take place after the ovum has taken up its definitive position 
in the mesoglcea behind the gastral epithelium. Possibly the 
spermatozoon has to penetrate a thin layer of dermal epithelium 
and mesogloea in order to reach the ovum, or perhaps the 
presence of the enormous ovum causes some rupture in the 
wall of an adjacent inhalant canal. It is impossible to say 
exactly what takes place. 

One more point may be mentioned in this section, and that 
is the tendency of particular stages of gametogenesis, or at 
any rate of odgenesis, to occur in large numbers in certain 
specimens, or in certain parts of the sponge, while more or 
less completely absent from others. Instances of this pheno- 
menon will be given in the followimg pages; it seems to 
indicate that the odgonia are produced in successive crops 
which go through their developmental stages synchronously. 


(e) THe RELATIONS BETWEEN THE DirrerENT TissuE ELEMENTS. 


There can be no doubt that the tissues of sponges are far 
less definite and less permanent than those of typical Metazoa. 
Without going back to the old view that the sponge is nothing 
more than a colony of Protozoa, which seems to be negatived 
by the degree of histological differentiation that they exhibit 
and by the facts of sponge embryology, we may safely say 
that the individual cells of which the sponge is composed 
often exhibit a remarkable power of changing their relative 
positions and also a high degree of polymorphism. Thus it 
will be remembered that Minchin (1898) has shown that the 
cells (scleroblasts) which secrete the triradiate spicule-systems 
in calcareous sponges migrate into the mesoglwa from the 


326 ARTHUR DENDY. 


dermal epithelium (“ectoderm”), and that the porocytes in 
Leucosolenia, when the sponge contracts, migrate through the 
gastral epithelium and fill up the central cavity (Minchin, 
1900). It is also well known that the epithelial cells of 
the so-called ectoderm are highly contractile and capable of 
great change of shape, and Maas has shown (1900) that in 
the developing Sycon the epithelial cells lining the central 
gastral cavity are derived by immigration from the dermal 
epithelium on the outer surface of the sponge. 

In the case of Grantia compressa it is hardly possible to 
speak of permanent tissues at all. According to my observa- 
tions any of the constituent cells cf the sponge may become 
amceboid and wander off to some new situation. This may 
very easily be demonstrated for the collared cells by examin- 
ing teased preparations of the living sponge in sea-water, 
when the collared cells can be seen putting out long, hyaline, 
finger-shaped pseudopodia in an extremely characteristic 
manner.! The fact that the collar may still be present along 
with the pseudopodia, as shown in fig. 1, affords unmistakable 
proof of the origin of these amceboid cells in teased prepara- 
tions. In stained sections we sometimes see something of 
the same kind, and fig. 2 represents a collared cell sinking 
into the mesogloea from between its fellows of the gastral 
epithelium. In this case pseudopodia and flagellum are seen 
to be present simultaneously but the collar is not visible. In 
fig. 3 an amcebocyte, probably derived from acollared cell, but 
apparently without collar, flagellum and pseudopodia, is seen 
lying in the mesoglcea behind the gastral epithelium. 

Appearances such as are represented in figs. 4,6 and 7 
also indicate very clearly that the cells of the so-called ecto- 
derm are not only contractile, but may become converted 
into amcebocytes and wander off into the mesogloea. In fig. 
4 is shown one of the epithelial cells lining the central gastral 
cavity in the contracted or “ flask-shaped ” condition (a), and 
in the subjacent mesogloea a typical amcebocyte (b). Fig. 6 
shows how such amcebocytes may be directly derived from 


1 Compare Carter (1875, p. 22) for a similar observation. 


GAMETOGENESIS OF GRANTIA COMPRESSA. o27 


epithelial cells which have migrated inwards, the identity of 
the two being clearly indicated by the presence of the 
numerous darkly stained granules that characterised the 
epithelial cells of the central gastral cavity, at any rate in 
this specimen. Fig. 7 shows a similar relation between the 
much less granular epithelial cells and amcebocytes around an 
inhalant canal. 

The mesoglcea of Grantia compressa contains, of course, 
a large number of amcebocytes, and there is no need to 
suppose that all of them are merely amceboid phases of either 
collared or pavement epithelial cells (compare figs. 58-61). 
Sometimes small amcebocytes may appear to form connective- 
tissue networks of stellate cells, but I doubt very much if they 
really do so, at any rate more than temporarily, and even in 
this condition they bear such a close resemblance to the 
epithelial cells lining the inhalant canals that I entirely fail 
to see how they can be distinguished cytologically. 

Under these circumstances it is, of course, quite impossible 
to say what cells of the adult sponge are derived from each 
of the cell-groups recognisable inthe larva. It also seems quite 
inadequate to say that the germ-cells are derived from 
wandering cells in the mesoglcea. 

The one constant and characteristic feature about sponge 
histology is, of course, the collared cell, and that is only 
constant in the sense that its typical form is that which 
possesses collar and flagellum. The sponge is, after all, not 
very much more highly advanced in organisation than a 
colony of choano-flagellate Protozoa. In such a colony we 
should certainly, I think, expect the germ-cells to be derived 
from collared cells, either directly or indirectly through an 
amceboid phase, and in a later section I hope to be able to 
show that this is how they actually originate in sponges. 

There is one feature about the collared-cells which appears 
to have attracted but little attention from sponge histologists 
but which deserves notice in this connection. I refer to the 
accumulation in them of what appear to be granules of reserve 
food-material. I have observed these as a very constant 


328 ARTHUR DENDY. 


feature in sections stained in a variety of ways, as polygonal 
bodies scattered more or less abundantly in the cytoplasm 
(figs. 2, 3, 8, etc.). They vary much in size and in the 
intensity with which they stain. In sections of material 
(spec. 23) fixed in absolute alcohol and stained with borax- 
carmine followed by picro-indigo carmine they are distinctly 
recognisable, and stain a pale, greyish colour. In sections of 
Flemming material without further staining they can easily 
be detected, though only very lightly stained. 

They are quite distinct in Flemming material stained with 
iron-brazilin, but, I think, more so after counter-staining with 
picro-indigo carmine, when they appear of a dark grey colour 
(spec. 11). They are hardly affected by nuclear stains and 
are evidently not chromidial in nature. The depth to which 
they stain exhibits a curious variation in some cases, as will 
be seen by reference to fig. 8, from a section of material (spec. 
21) fixed in strong Flemming and stained with paracarmine and 
picro-indigo carmine, where variation in this respect is visible 
even in one and the same cell. They are usually most 
abundant in the lower part of the cell (figs. 3, 8). 

These observations on their staining reactions appear to 
me to be quite in harmony with my view that the bodies in 
question are ‘‘reserve granules.” The question next arises, 
Are such granules characteristic of the collared cells, or do they 
occur also in the other tissue-elements? We have already 
had occasion to notice the presence of numerous granules 
in the epithelial cells lining the central gastral cavity and 
in the amcebocytes derived from these (fig. 6). ‘These granules, 
it will be observed, are of a different character from those now 
under consideration, being smaller and more highly refractive. 
My observations lead me to believe, however, that granules 
similar to those in the collared cells may also occur in the 
epithelial cells both of the gastral surface and of the inhalant 
canal system, though perhaps less abundantly. They 
certainly occur in many of the amcebocytes, and I had at first 
hoped that their presence might have served as a means of 
distinguishing those amcebocytes that originate from collared 


GAMBTOGENESIS OF GRANTIA COMPRESSA. 329 


cells from those that do not. I fear, however, that this hope 
must be largely abandoned, though the granules in question 
may perhaps serve as supplementary evidence of origin in 
certain cases. 

Another point to be noticed in connection with the collared 
cells concerns the condition of the nucleus. Usually in my 
sections this appears to be darkly and almost uniformly 
stained, but frequently it exhibits a distinctly reticulate 
character, with small, scattered, chromatin granules at the 
nodes of the reticulum, while intermediate conditions occur 
between the two extremes. As all conditions may occur close 
together in the same preparation it is difficult to account for 
the differences, but such variations have to be borne in mind 
in considering the origin of the germ-cells as indicated by 
nuclear characters. 

Miss Muriel Robertson and Prof. Minchin (1910) have 
given an account of the division of the collared cells in 
Cilathrina coriacea, and Miss Robertson (1911) has dealt 
with the corresponding phenomena in Grantia compressa 
and Sycon sp. Though figuring the collared cells in detail, 
neither of these: authors refer to the “reserve granules,” 
which I find to be such a constant feature in Grantia. This 
is perhaps due in part to the fact that their preparations 
were stained especially with a view to demonstrating the 
phenomena of mitosis. They figure much the same variations 
in the appearance of the nuclei of the collared cells as I have 
seen, but I cannot agree with their views as to the general 
occurrence of a single, relatively large karyosome (nucleolus). 
I have myself only occasionally seen such a body in these 
nuclei and do not attribute any special importance to it. 

Alike in Clathrina, Sycon and Grantia the division of the 
collared cells was found by Robertson and Minchin to take 
place longitudinally and to be accompanied by a typical 
mitosis. In Clathrina the number of chromosomes was found 
to be “ about sixteen,” but in Grantia and Sycon the chromo- 
somes are said to be “ not very distinct,” and the number is 
not given. Miss Robertson’s figures, however, especially 

voL. 60, PART 3.—NEW SERIES. 24 


330 ARTHUR DENDY. 


fig. 12 on Pl. 19, suggest eight or ten as the number rather 
than sixteen. I have myself hardly ever observed mitosis in 
the collared cells, but my fig. 9 represents a probable case, 
and it will be seen that so far as my very limited observations 
go they agree with those of Miss Robertson. 

I do not, of course, profess to give a complete account of 
the collared cells in this place; much more might be said 
about them, but the only thing needful here is to emphasise 
those points which have a direct bearing upon the problem of 
the origin and maturation of the germ-cells. 


(r) OOGENESIS. 
(a) Historical. 


By far the most complete account that has yet been given 
of the odgenesis in any sponge is contained in Dr. Max 
Jérgensen’s memoir (1910). As this account refers to a type 
(Sycon) closely related to Grantia compressa, I propose 
to give a brief review of Jérgensen’s results before passing on 
to describe my own. 

As already noticed, this author derives the primary o6gonia 
from so-called mesoderm cells, either resting stellate con- 
nective-tissue cells, or amcebocytes, between which he con- 
siders that no essential difference exists. The mesoderm cells 
multiply mitotically and become converted into odgonia of the 
first order. These increase in size and presently wander into 
the flagellate chambers, where they undergo mitosis and divide 
into odgonia of the second order. The latter divide again 
into oédcytes. The two o6dgonial divisions are said to be 
“atypical,” but quite similar to one another. During the 
mitosis eight chromosomes make their appearance “in Form 
von 'l'etraden”’ ; these are believed to be formed by fusion of 
several segments of a segmented spireme. Lach “ tetrad” is 
figured as dividing into two tetradiform daughter-chromo- 
somes on the spindle. 

Although the division of the oégonia frequently occurs in 


GAMETOGENESIS OF GRANTIA COMPRESSA, 331 


the flagellate chamber, the author evidently regards this as a 
more or less accidental circumstance, which takes place only 
if the 6ogonia have not got room enough to divide in the 
mesogloea. I think myself that the migration of the odgonia 
into the chambers before dividing must have a deeper signifi- 
cance than this. 

After their divisions are over the daughter-cells of the 
oégonia wander back into the mesoglcea as young oocytes. 
Here they undergo a short resting stage, and then they pass 
through the prophases of what must apparently be regarded 
as the mitosis belonging to the first maturation division. A 
long spireme is formed which arranges itself in a characteristic 
“bouquet ”’ form, and even shows a contraction phase similar 
to synapsis, though this may be due to the action of reagents ; 
at the same time “chromidia” appear in the cytoplasm, 
probably ejected from the nucleus. A large nucleolus is like- 
wise present, but this is also a characteristic feature of the 
odgonia. Some of these young odcytes now degenerate, 
possibly furnishing nutrient cells for the older odcytes. In 
others the spireme breaks up again into chromatin granules, 
and the odcyte continues its growth. 

In the later stages the oOcyte increases enormously in size 
and puts out long, branching pseudopodia. When the 
growth of the cytoplasm has reached its completion the 
nucleus enters upon the so-called “critical stage,” in which 
the nucleolus has completely disappeared and the nucleus is 
almost entirely devoid of chromatin. Apparently the chro- 
inatin has migrated through the nuclear membrane into the 
surrounding cytoplasm, where it is represented by “ chro- 
midia.” Chromatin-granules and nucleolus now reappear in 
the nucleus, which becomes very large and vesicular. Then 
a second diminution in the amount of chromatin takes place, 
this time apparently effected by solution or absorption within 
the nucleus itself and not by extrusion into the cytoplasm. 
The nucleus, with its diminished quantity of chromatin 
arranged once more in the form of tetrads, approaches the 
surface of the odcyte, which has in the meantime rounded 


Bo2 ARTHUR DENDY. 


itself off. The nuclear membrane now disappears, and the 
eight tetradiform chromosomes, now much diminished in 
size, arrange themselves on the equator of the first maturation 
spindle. Each chromosome divides into two daughter 
tetradiform chromosomes and the mitosis is completed in the 
ordinary way. Then the first polar body, containing eight of 
the daughter-chromosomes (‘tetrads’’), is twisted off from 
the surface of the odcyte in a very characteristic manner. A 
second maturation division now takes place and eight chro- 
matin elements finally remain in the fertilised egg, eight 
having passed out into the second polar body, but the forma- 
tion of the second polar body was not fully observed. 

It is difficult to understand completely the author’s views 
as to the manner in which the reduction of the chromosomes 
is carried out, and, indeed, although he discusses the problem 
at some length, he does not profess to have come to any 
certain conclusions: ‘ Leider ist mein Objekt zu klein und 
ungiinstig, um diese wichtige Frage sicher zu entscheiden.” 
At any rate he evidently considers that the somatic number 
of chromosomes is sixteen and the reduced number eight, for 
he finds sixteen in the segmentation nuclei of the embryo. 
The most curious thing appears to be that the number is 
already reduced to eight in the odgonial mitoses, but it is 
suggested that this may be a pseudo-reduction. Possibly it 
is connected with the tetrad formation which is supposed to 
take place at this stage. 

The author describes an interesting process of nutrition of 
the growing odcyte by ingestion of nutrient cells. I have 
observed a somewhat similar phenomenon myself in the case 
of Grantia and shall discuss it at length later on. 

As regards the fertilisation of the ovum and the subse- 
quent segmentation stages the most interesting feature 
appears to be the splitting up of the pronuclei and segmen- 
tation nuclei into karyomeres, a phenomenon which I have 
also observed in Grantia. 


GAMETOGENESIS OF GRANTIA COMPRESSA. 333 


(b) Origin and Growth of the Primary Oégonia. 


Before proceeding to describe the origin and growth of the 
primary odgonia as observed in Grantia compressa ib 
may be well to point out the great difficulties that arise with 
regard to the problem of seriation in the earlier stages of 
odgenesis. These difficulties are accentuated by the absence 
of definite localisation of the germ-cells and the consequent 
mingling of different stages in the mesoglcea or in the chambers, 
so that it is often impossible to be certain even to which 
generation a particular cell belongs. We can only fit the 
different observed stages together in what seems to be the 
most probable order on the sometimes scanty evidence avail- 
able. 

In the later stages of growth of the odcyte there is less diffi- 
culty, because the size of the cell not only shows that it is an 
odcyte, but indicates at the same time its place in the series. 

Inasmuch as amceboid cells of various sizes frequently 
immigrate into the chambers and come to he between the 
collared cells, the mere fact of the occurrence of a young 
germ-cell in such a position affords no conclusive evidence 
that it has been derived from a collared cell. The only way 
in which such an origin can be demonstrated, as it seems to 
me, is by finding cells which exhibit the characters of young 
germ-cells while still retaining the collar or flagellum, or 
both, of the collared cells. It must be admitted that it is not 
often that such intermediate forms are met with, and it is 
obvious that they can only be hoped for in very carefully pre- 
pared sections. I have, however, seen a few such cases, 
which seem to me to place my conclusions almost beyond 
doubt. Figs. 10, 11, and 12 are all taken from sections of 
specimen 21, stained with paracarmine and picro-indigo car- 
mine. In Fig. 10 two collared cells are represented side by 
side. In both the collar and flagellum are clearly visible, and 
in one the body of the cell is already considerably enlarged, 
apparently by the accumulation of reserve material. This 
enlarged collared cell I take to bea primordial germ-cell, but 


334 ARTHUR DENDY. 


whether it would have turned into an o6gonium or a spermato- 
gonium cannot be decided. The appearance of a vacuole 
around an unusually large granule of reserve material, however, 
suggests to my mind the latter.’ 

Figs. 11 and 12 represents a later stage, in which the body 
of the cell is greatly distended with reserve granules similar to 
those which occur in the ordinary collared cells, and the remains 
of collar and flagellum are, unless J am mistaken in my inter- 
pretation of the appearances, still visible. The nucleus has 
retreated to about the middle of the cell and a distinct 
nucleolus or karyosome has appeared. The whole cell pro- 
jects conspicuously beyond its neighbours into the cavity 
of the flagellate chamber, and there is a clear indication 
of pseudopodium formation represented in fig. 12, the 
pseudopodium being formed by a drawing out of the proximal 
end of the cell between the adjacent collared cells. (An 
apparent pseudopodium shown in fig. 11 may belong to 
another cell not seen in thesection.) In fig. 11 it will be seen 
that the young odgonium les actually next to the exhalant 
aperture of the chamber, and the one represented in fig. 12 
also lies close to an exhalant aperture, but I have not sufficient 
evidence to show whether or not there in any constancy in 
this position. 

The primary odgonia appear to leave the layer of collared 
cells very soon after their origin and migrate into the 
mesogloea (fig. 13). The collar and flagellum completely dis- 
appear and pseudopodia are put out (fig. 14). The nucleus 
at first appears uniformly stained except for the large 
nucleolus, but presently small granules of chromatin appear 
scattered between the nucleolus and the nuclear membrane, 
though the nucleolus appears to be surrounded by a narrow 
ring free from granules (fig. 15). The reserve granules are 
still abundant and easily recognisable in the cytoplasm. 

The primary o6gonium now appears to round itself off more 
or less completely before entering upon mitosis, as shown in 


‘ Compare the account of the origin of the primary spermatogonia, 
later on. 


GAMETOGENESIS OF GRANTIA COMPRESSA. 335 


fig. 16. In specimen 22 the mesoglcea between the chambers 
is crowded with o6gonia in this condition, and their numbers 
suggest that it may be a resting state. It will be seen that 
the cytoplasm is only slightly granular, but may still contain 
reserve material in the form of polygonal granules. The 
nucieus is faintly reticulate, and there is a very large and con- 
spicuous nucleolus and a thin nuclear membrane. 


(c) Multiplication of the Odgonia. 

Having reached the stage represented in fig. 16 the odgonia 
begin to prepare for division, which is effected by mitosis. 
During the progress of this mitosis they migrate through the 
layer of collared cells into the flagellate chambers (fig. 21), 
where the cell-division actually takes place. Although the 
earlier stages of the mitosis are found in the mesoglea (figs. 
17, 18), I have never seen the actual division taking place 
except in the chambers themselves, and cannot therefore 
agree with Jorgensen that the odgonia only migrate into the 
chambers when they are short of room in the mesoglea. This 
view is not in harmony with the fact that the fertilised eggs, 
which are many times larger than the odgonia, manage to 
find room for their development in the mesogloea by pushing 
out the gastral epithelium without rupturing it. I therefore 
think that there must be some special reason for the migra- 
tion of the odgonia into the chambers. I would suggest that 
it may enable the young odcytes to find abundant nutriment 
in the chambers in the first instance, and subsequently to dis- 
tribute themselves more readily throughout the sponge by 
creeping along inside the walls of the chambers and re-enter- 
ing the mesoglea at various points. 

Jorgensen is of opinion that there are two generations of 
oogonia in Sycon, and the large number of odcytes present 
seems to indicate that there must be at least two odgonial 
divisions in Grantia. Further evidence of this is to be found 
in the small size of the youngest odcytes as compared with 
the daugchter-cells formed by division of the primary odgonia 
(compare figs. 29 and 26). 


336 ARTHUR DENDY. 


Before proceeding to describe the mitosis of the primary 
odgonium it 1s necessary to say a few words with regard to 
the character of the chromatin substance in the nucleus. The 
growing odgonium contains, as we have seen, a large spherical 
nucleolus or karyosome (fig. 14, etc.), which stains very darkly, 
and in its later stages at any rate minute granules of chro- 
matin may also appear in the nucleoplasm (fig. 15). During 
the prophases of mitosis all the darkly staining granules of 
chromatin disappear from the nucleus, while the nucleolus 
may be cast forth from the odgonium altogether (figs. 17, 
18). There remains behind a quantity of more lightly stain- 
ing chromatin which forms the spireme, and, subsequently, the 
chromosomes. It is, I think, impossible to avoid the conclusion 
that there are here two totally different kinds of chromatin, for 
which we may accept the terms “trophochromatin ” and 
‘“jdiochromatin ” respectively, the former being concerned in 
the metabolism and growth of the cell and the latter in the 
processes of reproduction. The trophochromatin is represented 
chiefly by the nucleolus. Small granules of chromatin are 
possibly cast out into the cytoplasm as chromidia, for there is 
some evidence that chromidia may be found at this stage, but 
they are nothing like so conspicuous as they are in the 
oécyte. In the latter the chromidia are, of course, supposed 
to be concerned in yolk-formation, which has hardly com- 
menced in the oégonium. The large nucleolus appears to be 
bodily cast forth from the odgonium during mitosis (fig. 18). 
It can, therefore, hardly be supposed to be directly concerned 
in yolk-formation at this stage. It may possibly be of a 
different nature from the chromidia, and represent a mass 
of waste products accumulated in the nucleus during the 
growth and metabolism of the odgonium. In the young 
odcyte, however, as we shall see shortly, the nucleolus 
definitely gives rise to chromidia (“yolk nucleus ”), while in 
the maturing odcyte it appears to undergo degeneration and 
absorption in the cytoplasm. 

Jorgensen gives an interesting discussion on the nature and 
behaviour of the chromatin in the odcytes of Sycon, to which 


GAMETOGENESIS OF GRANTIA COMPRESSA. 387 


I must refer the reader, but he appears to have paid very 
little attention to this question so far as the odgonia are con- 
cerned, and does not seem to have observed the bodily ejection 
of the nucleolus. 

We may also say a few words here with regard to the 
formation of the chromosomes. Jérgensen figures both 
spireme and chromosomes as being stained perfectly black— 
as black, in fact, as the chromidia. His material, like mine, 
was fixed in Flemming’s solution, and he used iron-hzema- 
toxylin for staining (controlled by borax carmine and 
saffranin preparations). My own preparations were for the 
most part stained with iron-brazilin, but I have also used 
iron-hematoxylin. There is, of course, a good deal of 
variation in the results obtained by either of these methods, 
but my experience is that, as a rule, the spireme thread and 
chromosomes stain comparatively lightly as compared with 
the nucleolus and chromidia. I have never seen the chromo- 
somes so sharply defined as Jorgensen figures them, and I 
have seen nothing of the so-called tetrad formation which he 
describes during the odgonial mitoses and in the maturing 
oocyte. The chromosomes have always appeared to me much 
more like those figured by Prof. Minchin and Miss Robertson 
for the dividing collared cells—small subspherical or irregular 
bodies, so crowded together and ill-defined that it is impos- 
sible to count them accurately. The number characteristic 
of the odgonia appears to be about eight, as will be seen by 
reference to figs. 19-23, and this is possibly the somatic 
number (compare fig. 9). 

After these preliminary observations the actual division of 
the primary odgonia may be described very briefly. The 
prophases of the mitosis occur while the odgonium is still 
lying in the mesoglcea outside the chamber which it is about 
to enter, and while it is in a more or less amceboid condition 
(figs. 17, 18). Whether or not there is any casting out of 
chromidia into the cytoplasm I am not certain, but as a few 
small, densely staining bodies resembling chromidia sometimes 
appear in the cytoplasm in later stages of this mitosis, it 


338 ARTHUR DENDY. 


seems not unlikely that such may occasionally be the case. 
Apart from the nucleolus, however, there is very little 
chromatin in the nucleus to be cast out. 

The large vesicular nucleus approaches the surface of the 
odgonium until it is bounded on the outside by only a very 
thin layer of cytoplasm. In the meantime a spireme thread 
makes its appearance, and the nucleolus also approaches the 
surface (fig. 17). The nuclear membrane disappears, and 
the nuclear sap merges into the cytoplasm. A very curious 
phenomenon now takes place, the nucleolus being expelled, 
not only from the nucleus, but from the odgonium. Fig. 18 
shows it in the process of extrusion, surrounded by a drop of 
nuclear sap. ‘The pear-shaped form of the nucleolus at the 
moment of extrusion suggests that it must be a very soft, 
perhaps a semi-fluid, body in life. I have only seen this 
phenomenon exhibited very rarely in what can be considered 
as at all a conclusive manner. I have frequently seen the 
nucleolus apparently cast out of the odcyte, but minute 
inspection shows that this is (? always) an artificial result 
brought about in the act of cutting the sections. During the 
process of fixation, etc., the nucleolus appears to become very 
hard, and the knife then tears it bodily out of the odcyte. 
In the case represented in fig. 18, however, I think there can 
be no question of the normality of the process of extrusion ; 
indeed, that the nucleolus must be extruded at this stage 
seems to be indicated by its complete absence in later stages 
of the mitosis. 

It is extremely difficult to determine whether a particular 
spireme stage under observation belongs to an odgonial or to 
au odcyte mitosis. Jam inclined to think, however, that in 
the former case there are few or no chromidia, while in the 
latter the chromidia are fairly strongly developed (cf. figs. 
44,45). I must admit again, however, that the sorting out 
of these stages is to a large extent arbitrary. 

The spireme thread now breaks up into chromosomes, a 
typical spindle is formed with a minute centrosome at each 
pole, and the chromosomes arrange thems:2lves in the usual 


GAMETOGENESIS OF GRANTIA COMPRESSA. 339 


“equatorial plate” (figs. 19, 20). The chromosomes now pre- 
sumably divide, though I can hardly claim to have seen the 
actual division (cf., however, fig. 21), and the two groups of 
daughter-chromosomes migrate towards the two centrosomes 
(figs. 22, 28). The spindle disappears, and we are left with 
two closely aggregated groups of chromosomes as the founda- 
tions of the two daughter-nuclei (fig. 24). Constriction of the 
cytoplasm between these two daughter-nuclei now follows 
(fig. 25), and finally the entire odgonium becomes divided 
into two daughter-cells (fig. 26). 

Apparently not until the prophases have been passed 
through does the odgonium migrate through the layer of 
collared cells into the adjacent flagellate chamber, as shown 
in fig.21. Here it rounds itself off, usually into an oval form 
(figs. 19, 20), before completing the mitosis. At this stage 
the cytoplasm exhibits a fairly uniformly and densely granular 
character, shown especially well in fig. 22. A few small 
densely staining granules, resembling chromidia, are some- 
times visible in it (fig. 19), and may even persist in the 
daughter-odgonia after completion of the division (fig. 26). 

It seems almost certain that a second odgonial division 
takes place in the flagellate chambers very shortly after the 
first one. Fig 27 represents a stage which I interpret as an 
odgonium of the second generation with reconstituted nucleus, 
and fig. 28 represents what I take to be such a secondary 
oégonium in mitosis. Without the intervention of such a 
stage it would be difficult to explain the origin of the next 
series of stages (figs. 29-39), which occur very abundantly 
in the chambers, and which I interpret as young oocytes. 
With regard to this period of the odgenesis my conclusions 
differ widely from those of Jorgensen, who makes his odgonia 
of the second order larger than those of the first order, and 
his young odcytes larger still, which would be very difficult 
to understand in view of the repeated division, and the 
apparent absence, so far as his account goes, of any process 
of nutrition. 


340 ARTHUR DENDY. 


(d) Growth and Feeding of the Young Odcytes in 
the Flagellate Chambers. 


The growth of the odcytes may.be divided into two very 
distinct periods, during the first of which they are found in 
the flagellate chambers, while during the second they lie in 
the mesogloea between the chambers. 

As already indicated, it is by no means an easy matter to 
sort out all the very numerous amceboid cells that occur in 
the flagellate chambers, some representing stages in oégenesis 
and others stages in spermatogenesis, into their proper cate- 
gories. Amongst them, however, may be distinguished a type 
which occurs very abundantly and exhibits certain peculiarities 
by which it is more or less readily recognised. The cells in 
question are small and distinctly amceboid, of irregular form, 
and usually attached to the wall of the chamber by pseudo- 
podia. They have a rather small nucleus, with a relatively 
large nucleolus surrounded by a narrow clear space, and then 
by a broad ring of minute granules extending to the nuclear 
membrane. The cytoplasm usually exhibits more or less 
numerous inclusions which may be surrounded by vacuoles. 
Some of these inclusions, which I interpret as chromidia or 
yolk-nuclei, stain nearly black, others may stain much more 
lightly, and look like food-particles undergoing digestion. A 
typical series of these curious cells is shown in figs. 29-39. 
I interpret them as young odcytes engaged in feeding opera- 
tions. 

Figs. 29 and 30 show what appears to be the youngest 
stage of this series, a stage which may be derived from the 
mitotic division of a secondary odgonium such as is repre- 
sented in fig. 27 or 28. The cytoplasm at this stage is seen 
to be uniformly and rather coarsely granular and contains no 
chromidia or other inclusions of any kind. The structure of 
the nucleus even at this early period, with its large nucleolus 
and broad zone of chromatin granules distinctly separated 
from it, appears to me to be essentially that of an immature 
female gamete. Presently the characteristic inclusions make 


—————— = 


ee 


GAMETOGENESIS OF GRAN'ITTA COMPRESSA. 341 


their appearance in the cytoplasm, which otherwise may come 
to exhibit a more homogeneous appearance (figs. 31-39). 
Some of these inclusions are comparatively lightly staining 
bodies enclosed in vacuoles (figs. 31, 56, 37), and in one case 
(fig. 89) a nucleated cell was distinctly recognised amongst 
other bodies. I therefore believe that the lghter coloured 
inclusions are food-particles undergoing digestion, captured 
by the young odcytes from the stream of water that flows 
through the chambers. 

The nature of the intensely black stained bodies that 
appear in the cytoplasm is easily interpreted. These 
resemble the nucleolus in appearance, but may be much 
larger (figs. 33-35). They may or may not be surrounded 
by distinct vacuoles. On the other hand, they may take 
the form of small granules or groups of granules (figs. 
35, 36). Fig. 38 gives the clue to the manner in which 
they are formed, for here the nucleolus is seen actually 
discharging part of its own substance into the cytoplasm 
through the nuclear membrane. I think there can be no 
doubt that in these cells very active metabolism, accom- 
panied by the formation of chromidia, or “ yolk-nuclei,” is 
going on, and that the bodies in question are of this nature, 
and are concerned in the elaboration of yolk-granules in the 
cytoplasm. 

Apparently feeding now ceases and the remains of food- 
particles disappear from the cytoplasm. The odcyte next 


migrates through the chamber wall into the mesogloa (fig. 
40). 


(e) GrowrH AND FrepinG oF THE OocyrEs IN THE MesocLaa. 

The formation of chromidia, or “ yolk-nuclei,” by extrusion 
of matter from the nucleolus, which was already commenced 
within the chambers (fig. 88), may now be continued very 
freely. Specimen 23 contains an immense number otf odcytes, 
lying in the mesogloea between the chambers, in which this 
chromidium-formation is going on (figs. 41, 42, 43), and often 
giving rise to very fantastic appearances. ‘The nucleolus is 


342 ARTHUR DENDY. 


evidently in a liquid or semi-liquid state and appears to be 
squeezed out into the cytoplasm in threads or drops, Just as 
an artist’s colours may be squeezed out of their tubes. During 
this process the nucleus itself may become distinctly pear- 
shaped (fig. 43). The drops squeezed out into the cytoplasm 
are at first enclosed in distinct vacuoles, apparently derived 
from the nuclear sap. In these vacuoles they disintegrate 
into granules (fig. 43), which probably became scattered 
through the cytoplasm. The nuclear membrane may become 
very indistinct during the process, and there are indications 
that the nucleus is passing into the spireme stage to be des- 
cribed next. It is doubtful whether the nucleolus is ever 
completely eliminated from the nucleus at this stage. It 
seems to me more probable that some of it always remains 
behind as the foundation of the huge nucleolus which forms 
such a conspicuous feature in later stages of the odcyte. The 
irregularity in shape of the odcyte during this process of 
chromidium formation indicates that it is still amoeboid, and 
it seems possible that the extrusion of the chromidia may be 
due to strong contraction of the cytoplasm, though it must be 
admitted that the mechanism of the process is very obscure. 

The whole of the series of stages representing the feeding 
and growth of the young odcytes in the chambers and the re- 
markable process of chromidium-formation just described 
appears to have been unobserved by Jérgensen. It is true that 
that observer worked upon Sycon, but it is unlikely that two 
such closely related types as Sycon and Grantia should differ 
in this respect, especially when they agree so closely as regards 
other features of the odgenesis. JOrgensen figures the young 
odcyte, supposed to be directly derived from the last o6gonial 
division, as being very similar to the stage represented in my 
fig. 47a (compare his fig. 32). This stage may very easily be 
derived, however, through the intermediate stages represented 
in figs. 41-43, from the last of the feeding stages observed 
in the chambers (fig. 39). 

It also seems very probable that shortly after leaving the 
flagellate chambers and undergoing the process of chromidium- 


GAMETOGENESIS OF GRANTIA COMPRESSA. 343 


formation just described, the young odcyte exhibits the pro- 
phases of a mitosis which really belongs to the first maturation 
division. J6rgensen puts this incomplete mitosis, represented 
by a well-marked spireme stage, immediately after the stage 
which, in Sycon, probably corresponds to my fig. 47a. I 
prefer to place it immediately before this stage, where it seems 
to me to fit in better. Figs. 44 and 45 show two of the pro- 
phases in question. ‘I'he former is evidently a leptotene phase 
and the latter a pachytene. The latter always shows the 
characteristic contraction of the spireme described by Jérgen- 
sen, which may possibly represent a synapsis or be due simply 
to the action of reagents. Both these figures show well- 
developed chromidia in the cytoplasm, which, as I have already 
said, inclines me to include them in the odcyte series rather 
than in the odgonial series, though I do not consider that the 
evidence is by any means conclusive. It is obvious from 
fig. 45 that the odcyte may still be highly amceboid during 
this phase, exhibiting a very irregular outline. 

Jérgensen considers that some of the odcytes at this stage 
undergo degeneration and may have something to do with 
forming the nutrient cells for the older odcytes, but I 
have obtained no good evidence of such degeneration and 
my observations on the feeding of the older odcytes do not 
support this view. The spireme thread now disappears 
(fig. 46) and the odcyte rounds itself off, takes up a definite 
position in the mesoglcea behind the layer of collared cells 
(fig. 47), and enters upon its main period of growth. 

Fig. 47a represents a condition of the odcyte which is very 
commonly met with. Jorgensen figures a similar condition 
in Sycon (fig. 52), and speaks of it as a resting condition, but, 
as I have already pointed out, he regards it as the direct 
product of the division of an odgonium of the second 
generation. I am not quite sure, however, that Jorgensen’s 
fig. 32 really represents the same stage as my fig. 47a, for he 
does not show, or if so only very faintly, the chromidia in the 
cytoplasm, which appear to me to be very characteristic of 
this stage. It is by the abundance of these chromidia, 


/ 


344 ARTHUR DENDY. 


indeed, that this stage is chiefly distinguishable from the 
primary odgonium just before mitosis, as represented in fig. 16, 
taken in conjunction with the fact that the latter occurs 
especially in association with the odgonial mitoses going on in 
the chambers, while the former occurs especially associated 
with the later stages of odcyte growth. 

In Grantia the young odcyte, passing out of this “ resting 
condition,” simply flattens out somewhat on one side (fig. 47b), 
and puts out long branching pseudopodia by means of which 
it attaches itself to the mesoglceal surface of the layer of 
collared cells, as shown in fig. 48. Whether these psendo- 
podia are merely “ anchoring ”’ pseudopodia, or whether they 
also serve to extract nutriment for the growing odcyte 
from the collared ceils with which they are in contact, must 
for the present remain an open question. The appearances 
represented in fig. 48, however, suggest to my mind the 
latter. We are reminded of the manner in which the super- 
ficial cells of the embryo of Stelospongus attach themselves to, 
and evidently draw nutriment from, the large capsule-cells 
by which they are surrounded, as described by me many 
years ago (Dendy, 1888). It will be seen that the collared 
cells contain abundant “reserve-granules,’ and _ similar 
granules appear in the odcytes, while the chromidia have 
entirely disappeared from the cytoplasm. The nucleus is 
distinctly reticulate, with numerous darkly stained chromatin 
granules and a large nucleolus, the latter surrounded by a 
clear space (possibly due to shrinkage ?). 

The odcyte continues to increase in size and the nucleus 
grows more rapidly than the cytoplasm. Presently we reach 
a stage which is very characteristic and very frequently met 
with, and which I propose to call the “ contraction stage of 
the odcyte.” This is represented in fig. 49. It will be seen 
that the entire cell has rounded itself off again and the 
pseudopodia have contracted into blunt knobs. The nucleus 
is very large and pretty uniformly granular, all or nearly all 
the darkly staining chromatin having evidently been expelled 
into the cytoplasm in the form of chromidia (yolk-nucleus), 


GAMETOGENESIS OF GRANTIA COMPRESSA. 345 


which at this period are apparently not derived, at any 
rate directly, from the nucleolus. ‘I'he latter is very large, 
spherical, and fairly darkly staining. ‘This stage, again, does 
not appear to have been observed by Jérgensen in the case 
of Sycon. 

It is at about this period that the feeding of the odcyte by 
means of nurse-cells begins, a process which continues right 
on until the odcyte has reached its maximum size or nearly 
so, and which will be dealt with separately in a later 
section (see, in the meantime, figs. 50, 51, 52). 

During this process the odcyte increases enormously in 
size, and long, root-like pseudopodia are put out again (fig. 54) 
in a plane parallel to the layer of collared cells against which 
the odcyte lies. The nucleus assumes the form of an enor- 
mous, thin-walled vesicle, containing a reticulum of lightly 
staining, flocculent material, which looks very much like a 
precipitate or coagulation, and includes small, darkly staining 
chromatin granules scattered through it. Most of the darkly 
staining chromatin, however, appears to be expelled into the 
cytoplasm in the form of chromidia, which may be extremely 
numerous (fig. 53). The nucleolus is very large,and frequently 
exhibits a differentiation into more and less darkly staining 
spheres, which may or may not be concentric (figs. 52, 53, 
54). Occasionally the nucleolus appears to be cast out into 
the cytoplasm, but, as already stated, I have come to the 
conclusion that this is an artificial result, due to the action of 
the knife in cutting. Sometimes the cytoplasm around the 
nucleus exhibits a faint radial arrangement of its granules, 
as shown in figs. 52 and 54. ‘This is a feature upon which 
Jérgensen has laid some stress,in the case of Sycon. 


({) Maturation of the Odécytes. 


The odcyte now withdraws all its pseudopodia and rounds 
itself off into a compact ellipsoid body about 0:045 mm. in 
maximum diameter, preparatory to maturation. The huge 
vesicular nucleus disappears and its contents are apparently 

VoL. 60, PART 3.—NEW SERIES. 25 


346 ARTHUR DENDY. 


diffused throughout the cytoplasm, which is dense, and, with 
the exception of certain inclusions to be mentioned imme- 
diately, uniformly and rather coarsely granular. Atno stage 
have I seen any vitelline membrane. ‘The cytoplasm contains 
a large number of very small chromidia, mostly aggregated 
in irregular groups or clouds (fig. 55), which are probably, 
in part at any rate, derived directly from the small chromatin 
granules that remained in the nucleus at the close of its 
career. The immense nucleolus in two cases (out of the very 
few met with) was clearly discernible in the cytoplasm, 
where it appeared to be undergoing absorption (fig. 55, no.). 
In one case a group of small bodies that might be chromo- 
somes (fig. 55, chr. ?) was observed in the cytoplasm at some 
distance from the nucleolus; but the nature of these bodies 
is really uncertain, as also is the nature of a protrusion of the 
surface of the odcyte opposite to the supposed chromosome- 
group. It is possible that we have here the beginning of the 
formation of the first polar body, but I think that that is 
extremely doubtful. 

Ata slightly later stage, however, the chromosomes appear 
unmistakably in connection with the first maturation spindle, 
which is represented in fig. 56. It is obvious that this agrees 
closely with the first maturation spindle as described by 
Jorgensen (cf. his fig. 59), but unfortunately I have only 
been able to find one really good maturation spindle in my 
preparations, and I am therefore unable to give any details 
with regard to the process. I am not even sure of the number 
of chromosomes, but there appear to be eight or ten in each 
daughter-group represented in fig. 56. Jérgensen makes the 
number eight in each group, but represents each chromosome 
as a sort of tetrad. From my own observations I can only 
say that the chromosomes are small irregular bodies, and it 
appears to me that, whenever chromosomes are concerned, 
Jérgensen’s figures must be somewhat diagrammatic. 

The formation of the first polar body in Grantia is repre- 
sented in fig. 57, and it is evident that it takes place very 
much as described by Jérgensen for Sycon. The polar body 


GAMETOGENESIS OF GRANTIA COMPRESSA. 347 


itself is seen to be of large size, and the remains of the spindle 
are seen as a very distinct cord, slightly thickened in the 
middle, connecting the group of chromosomes in the polar 
body with the group that remains behind in the odcyte. 

I have searched in vain for the second maturation spindle 
and second polar body, but, considering the rarity with which 
the first occurs in my specimens, their apparent absence has 
no significance. Kven Jorgensen, however, was not able 
fully to observe the formation of the second polar body, 
though he tells us that apparently it also contains eight 
chromosomes, while eight (dyads ?) remain in the mature 
ovum. 


(g) The Nurse-Cells and their Origin; Phago- 
cytosis. 


A process of feeding on the part of the growing oécyte at 
the expense of certain nutrient cells has been described in 
the case of Sycon by both Gorich (1903) and Jorgensen (1910). 
Gérich says that the oocytes (“ Hizelle”’) ingest entire cells, 
which are probably themselves egg-cells of smaller size. The 
process is represented as taking place very much as in the 
case of an Amoeba ingesting food-particles, except that no 
vacuole is formed around the ingested nutrient cell, whose 
protoplasm appears to mingle directly with that of the oocyte. 
Hach nutrient cell appears to be asmall nucleated amcebocyte, 
with a relatively large, darkly staining particle (? chromi- 
dium) in the cytoplasm. Numerous very similar bodies, 
evidently chromidia, are figured in the cytoplasm of the 
feeding odcyte, which appears to be at about the stage figured 
in my figs. 50 and 51 (compare Gorich’s figs. 1-6). 

Jorgensen tells us that the ingestion of nutrient cells by 
the odcyte takes place towards the end of the growth period, 
and goes on right up to the formation of the maturation 
spindle. ‘The ingested cells are believed to be mostly odgonia 
and the ingestion is supposed to be due to chemotaxis. The 
ingested cells are said to be taken into a preformed gullet in 


348 ARTHUR DIENDY. 


the odcyte, and not, as described by Gérich, by means of 
pseudopodia. Jdérgensen has, however, observed ingestion 
by means of pseudopodia in the case of Sycon setosum. He 
also finds the presence of a compact chromidium in the cyto- 
plasm to be characteristic of the nutrient cell. This chromidium 
may be taken into the gullet of the feeding odcyte without the 
nutrient cell, but usually the entire nutrient cell is taken in. 
The chromidia received by the odcyte in this way are distin- 
guished by their size from those cast out from the nucleus, 
but both undergo like degeneration in the cytoplasm. 

I have never seen anything resembling the formation of a 
gullet (“Schlund’’) in Grantia, and I think that Gdrich’s 
account of the taking in of the nutrient cells in Sycon seems 
the more probable of the two. 

In Grantia the process is complicated by the intervention 
of what I propose to term “ nurse-cells,’’ which capture the 
nutrient cells and bring them to the odcyte. Possibly the 
supposed taking in by the odcyte of the chromidium only 
from the nutrient cell really indicates something of the same 
kind for Sycon. 

I have observed the feeding of the odcyte by nurse-cells in 
specimens 1] and 23 and I have seen many instances of it, and 
I do not think there can be any doubt either as to the 
observations themselves or as to their interpretation. The 
peculiar arrangement shown in figs. 50 and 51 is frequently 
met with in my sections. I interpret it as indicating that a 
nurse-cell (n. c.) has captured a smaller cell which I regard as 
a nutrient cell or food-cell (f. c.), and is passing it into the 
cytoplasm of the odcyte. I shall discuss the origin of the 
nurse-cell directly ; in the meantime I may point out that it 
exhibits certain fairly well-marked characters of its own. 
Its cytoplasm is thin-looking and stains very lightly, and the 
nucleus is of moderate size, with a well-developed nuclear 
membrane and reticulum and a rather small nucleolus. The 
cytoplasm usually contains conspicuous inclusions which are 
evidently the remains of ingested and partially disintegrated 
cells, but nothing that can be identified with the “ chromidium” 


GAMETOGENESIS OF GRANTIA COMPRESSA. 349 


of the nutrient cells described by Gérich and Jorgensen in 
Sycon ; indeed, the nurse-cell does not appear to be itself a 
nutrient cell in the sense of these authors. The real nutrient 
cell (f. c.) is, however, a very conspicuous object, lying in the 
middle between the nurse-cell and the odcyte. It appears 
to be always an oval or spherical cell with a reticulate 
nucleus and rather dense, finely granular cytoplasm, and it 
appears to be passed on to the odcyte by the nurse-cell before 
it has undergone any disintegration, for its outlines are 
perfectly definite. Judging from its small size and the absence 
of chromidia in the cytoplasm I am inclined to think that the 
nutrient cell is probably a rounded-off collared cell, but it is 
impossible to be certain on this point, though I shall bring 
forward evidence presently to show that the nurse-cells do 
capture collared cells. 

The process of feeding by nurse-cells does not appear to 
begin until the odcyte has attained a considerable size, and it 
may be fed in this way both in the contracted and in the 
expanded condition, 1. e. when the pseudopodia are reduced 
to short knobs and when they are fully extended. Figs. 50 
and 51 represent the process of feeding by the nurse-cells 
as usually observed. A somewhat different condition is 
represented in fig. 52. The nurse-cell is here crushed in 
between the odcyte, which is now much larger, and the layer 
of collared cells against which it lies. The nutrient cell has 
passed completely into the cytoplasm of the odcyte, where it 
is surrounded by a small vacuole in which it appears to be 
undergoing disintegration, for instead of a distinct nucleus 
it exhibits two masses of chromatin—one at the middle and one 
at the side. Both nurse-cell and nutrient cell appear to be 
much smaller than in the cases previously described. The 
whole arrangement somewhat resembles the taking in of a 
chromidium from a autrient cell as described by Jorgensen, 
but it is probably simply a more advanced stage of the feeding 
process shown in the preceding figures. 

We come now to the question of the origin of the nurse- 
cells. I think there can be little doubt that they are derived 


350 ARTHUR DENDY. 


from small amcebocytes which occur scattered in the mesoglea, 
such as is represented in fig. 58. It will be seen that we 
have here a cell of about the same size as the nurse-cell shown 
in fig. 51, and, except for the absence of cytoplasmic inclusions, 
closely resembling it. The cytoplasm is faintly staining and 
thin-looking, the nucleus is reticulate, with a small nucleolus 
and a well-developed nuclear membrane, and even the some- 
what angular outline of the entire cell appears to be more or 
less characteristic, and is frequently met with again in the 
nurse-cells while in the act of feeding the odcytes (compare 
figs.50,51). Whence these amcebocytes come it is impossible 
to say. They do not look like metamorphosed collared cells 
or epithelial cells, and they are very likely derived direct 
from embryonic amcebocytes. They probably do not all 
become nurse-cells, for some, which appear to be of the same 
nature, grow to a large size and may put out long filiform 
pseudopodia (figs. 59-61) In this extended condition I have 
found them both inside the flagellate chambers (fig. 60) and in 
the mesogloea between them (fig. 61), while in a more or less 
rounded-off condition they are sometimes to be seen passing 
through the layer of collared cells, especially in the neighbour- 
hood of the exhalant apertures of the chambers (fig. 59). 

Such amceboid cells sometimes develop into very active 
phagocytes, apparently entirely on their own account. I 
have observed this in specimens 11, 22 and 24. Specimen 24 
in particular is crowded with large and small phagocytes, 
which have evidently been feeding voraciously, and apparently 
chiefly on young germ-cells. 

Fig. 62 represents a phagocyte from the mesogloea, which 
has ingested a single relatively large cell, too large, I think, 
to be a spermatogonium, and so large that the cytoplasm of 
the phagocyte is only able to stretch itself around its prey in 
the form of a thin envelope. In fig. 63 an actively amceboid 
phagocyte, with filiform pseudopodia, is apparently in the act 
of ingesting a collared cell. Fig. 64 represents a phagocyte 
rounded off in the mesoglcea, with two partially disintegrated 
cells in its cytoplasm. This one looks as if it mght have 


GAMETOGENESIS OF GRANTIA COMPRESSA. sor 


become a nurse-cell, though rather large. Fig. 65 represents 
a very actively amoeboid phagocyte, while fig. 66 shows one 
squeezing itself through the layer of collared cells. Fig. 67 
shows one with root-like pseudopodia, half in and half out of 
a chamber; this one has collected and ingested no less than 
six cells, which I judge from their size and appearance to be 
spermatogonia. Fig. 68 represents yet another hanging on to 
the inner surface of the layer of collared cells. 

It appears, then, that while some of these amcebocytes 
exercise their phagocytic propensities in favour of the odcytes, 
and become nurse-cells, others feed voraciously on their own 
account, even entering the flagellate chambers and appa- 
rently collecting the numerous young germ-cells that are found 
therein. It seems not improbable that this excessive phago- 
cytosis may be regarded as a perverted instinct, for the phago- 
‘cytes appear to take in food far beyond their own possible 
requirements, and yet I have never seen a large phagocyte 
feeding an odcyte. It is perhaps worth noting that speci- 
men 24, in which most of the cases of phagocytosis by large 
amcebocytes were observed, had been kept in the circulation 
for a week before being killed and preserved. The abnormal 
conditions may have stimulated the amcebocytes to an 
abnormal activity in phagocytosis. 

‘he possibility also occurs to one that some of the large 
phagocytes are parasitic Amoebze which do not belong to the 
sponge at all. Mr. Orton (1913) has recently described 
Amoebee from the gastral cavity of Sycon which certainly 
seem to be quite independent organisms, though I at first 
thought otherwise (Dendy, 1913), and it seems by no 
means improbable that such Amcebee may enter the chambers 
and feed upon the young germ-cells, or even force their way 
into the mesoglea. The chief argument against this view 
appears to be the impossibility of distinguishing between the 
larger and smaller phagocytes, and the apparent identity of 
the latter with the nurse-cells, which must certainly be 
regarded as belonging to the sponge itself. 


352 ARTHUR DENDY. 


(h) Summary and General Remarks on Odgenesis. 


It will be seen from the foregoing account of my observa- 
tions on the odgenesis of Grantia, that in their main features 
they agree with what has already been recorded, especially by 
Jorgensen, for the closely related genus Sycon. In some 
important respects, however, and especially as regards the 
derivation of the odgonia from collared cells, my observations 
differ from those of Jérgensen, and, while I have not been 
able to obtain anything like such precise results as he claims 
with regard to the mitotic phenomena, I have been able to 
describe a great deal that has either escaped his notice or does 
not occur in Sycon, concerning, for example, the feeding of the 
young odcytes in the chambers and the subsequent formation 
of chromidia by extrusion of nucleolar matter into the cyto- 
plasm, the very remarkable feeding of the odcytes by means 
of nurse-cells, and the process of phagocytosis in general. 

The process of odgenesis in Grantia may be briefly 
summarised as follows : 

The primary odgonia are directly derived from collared 
cells, which accumulate reserve material, enlarge, with- 
draw their collars and flagella, become amceboid and wander 
into the mesoglcea, re-entering the chambers before dividing 
mitotically into the odgonia of the second generation. Prior 
to this division the nucleolus appears to be bodily cast out of 
the odgonium. 

The odgonia of the second generation divide again while 
in the chambers, and probably almost immediately, into small 
oocytes. 

The small odcytes become amoeboid, and, while still within 
the chambers, attached by pseudopodia to the layer of 
collared cells, take in food-particles and form conspicuous 
chromidia in their cytoplasm. Having increased considerably 
in size they leave the chambers and enter the mesoglea. 

Here they continue to undergo a process of extensive 
chromidium-formation by extrusion of chromatin from the 
nucleolus into the cytoplasm. 


GAMETOGENESIS OF GRANTIA COMPRESSA. 353 


They also probably undergo about this time the prophases 
of the first maturation division. 

They now send out anchoring pseudopodia by which they 
became attached to the layer of collared cells and probably 
draw nutriment from them. 

Presently they undergo a remarkable contraction, the 
pseudopodia being almost completely withdrawn, and about 
this time they begin to be fed by special nurse-cells, which 
collect smaller cells and pass them into the growing oocyte. 

The oécyte increases greatly in size, and long root-like 
pseudopodia are again put out in a plane parallel to the layer 
of collared cells against which it lies. 

The nucleus becomes very large and vesicular, with a huge 
spherical nucleolus, and chromidia are abundantly formed in 
the cytoplasm, though apparently no longer directly derived 
from the nucleolus, but formed probably by extrusion of 
granules of chromatin from the nucleus. ‘he chromidia 
(yolk-nucleus), whatever their source, are probably concerned 
in the elaboration of the deutoplasm, with which the cyto- 
plasm becomes uniformly and densely charged, though 
definite individual yolk-granules can hardly be recognised. 

When the odcyte has reached its full size it withdraws its 
pseudopodia and rounds itself off, the nuclear membrane 
disappears and the contents of the nucleus disperse themselves 
through the cytoplasm. The nucleolus remains recognisable 
for some time after this event, but gradually becomes 
absorbed. 

Chromosomes have not been recognisable since the odgonial 
mitoses, when they appeared in the equatorial plate as a group 
of eight or ten minute, irregularly rounded bodies, each of 
which presumably divided into two. They now appear again 
on the first maturation spindle, but only one really good 
spindle was found, and that already in the anaphase. 

The first polar body is formed in apparently a typical 
manner, exactly as described by Jorgensen for Sycon, and is of 
large size. 

The second maturation spindle and second polar body, 


304 ARTHUR DENDY. 


described by Jérgensen for Sycon, were not met with, though 
they probably occur. 

No evidence was obtained of a reducing division, and, indeed, 
the number of chromosomes could never be accurately counted. 

In spite of the fact that we have no reliable information 
with regard to the phenomena of meiosis in sponges, we 
cannot fail to be struck with the close general agreement of 
the process of odgenesis with the same process as observed in 
higher animals. ‘The multiplication of oédgonia; the formation 
of chromidia or yolk-nucleus by the odcyte; the early inception 
of the first maturation division (if this be confirmed), inter- 
rupted by the long period of growth; the co-operation of other 
cells in the process of nutrition ; the character of the nucleus 
and the formation of the polar bodies; are all features which the 
sponges share with higher groups, and one can hardly avoid 
asking the question, Does this close similarity in odgenesis 
point to a nearer relationship of the sponges with the Enterozoa 
than is usually admitted in this country ? The question is well 
worthy of consideration, but we can hardly hope for a final 
solution of it in the present state of our knowledge. We can 
hardly abandon the choano-flagellate ancestry of the sponges 
without much stronger evidence than we possess, and we 
certainly are not justified in attributing a choano-flagellate 
ancestry to other groups of Metazoa. May we, then, suppose 
that all the essential processes of odgenesis already existed in 
pre-choano-flagellate Protozoon ancestors common to sponges 
and Enterozoa? Such a supposition would certainly be in 
harmony with the view now generally held that the germ-cells 
of the higher animals are really equivalent to so many 
Protozoa, for if this be so then the phenomena of odgenesis 
must be such as we might reasonably expect to find in 
Protozoa. Recent advances in protozoology, I think, show that 
such an expectation is likely to be fulfilled, for we already 
know that the gametes themselves may attain as high a 
degree of differentiation (into ovum and spermatozoon) in 
Protozoa as in Metazoa. We also know that the female 
gamete may accumulate yolk (e. g. in Coccidium), and that 


GAMETOGENESIS OF GRANTIA COMPRESSA. Bis) 


something that may reasonably be interpreted as a forma- 
tion of polar bodies may take place (e. g. Paramcecium). 
I venture to predict that a good deal of light will, in the near 
tuture, be thrown upon the complex phenomena exhibited in 
the life-history of many Protozoa, by comparison of the events 
that take place in the gametogenesis of higher animals. Some 
more satisfactory and uniform system of terminology will, 
however, have to be evolved before much progress can be 
made in this direction. We shall have to know, for example, 
exactly what we mean by “‘chromidia.” Prof. Minchin (1912) 
tells us that “in a great many Sarcodina, especially in those 
belonging to the orders Amcebxa and Foraminifera, chromidia 
may be present in the gamete-forming individuals as a 
permanent constituent of the body-structure. In such cases 
the chromidia represent, wholly or in part, the generative 
chromatin, and give rise by formation of secondary nuclei to 
the nuclei of the gametes.” In the present paper I have, 
following Jérgensen, used the term “ chromidia” for all the 
chromatin which occurs in the cytoplasm. This, I think, is 
probably ali extruded from the nucleus (and nucleolus), and is 
almost certainly concerned in yolk-formation, and therefore 
‘trophochromatin ” and not “ idiochromatin.” The difficulty 
of distinguishing between these two kinds of “chromatin” 
forms perhaps the chief obstacle in the way of further 
progress in the direction indicated. 


(G) SPERMATOGENESIS. 


(a) Historical. 

I have already referred to Haeckel’s discovery of the 
Sperm-morulee of calcareous sponges in the gastral epithelium 
of the flagellate chambers, and to his opinion that they arise 
by division of collared cells. These observations never met 
with general acceptance, and are usually peuerded as having 
been superseded by Poléjaeff’s well-known work, ‘ Uber das 
Sperma und die Spermatogenese bei Sycandra raphanus 


Haeckel’ (1882). 


356 ARTHUR DENDY. 


Poléjaeff lays stress upon the discrepancies between the 
account of the spermatozoa given by Haeckel and that given 
by Eimer. According to Eimer, the spermatozoa, if not 
isolated, occur scattered through the tissues, united in millions 
in oval balls; according to Haeckel, they lie between the 
collared cells with their tails projecting freely into the cavity 
of the flagellate chamber, and it is never possible to find them 
in considerable quantities. According to Himer, again, they 
are to be distinguished from the collared cells by the character 
of the movements of the flagella; while, according to Haeckel, 
these movements show no essential differences in the two 
cases. I think it almost certain myself that, although Haeckel 
saw the sperm-morule in the situation he describes, he did 
not see the tails of the spermatozoa, for the sperm-morule 
represent a comparatively early stage of spermatogenesis at 
which no tails have yet appeared. Haeckel probably mistook 
for spermatozoon tails some of the flagella of the collared 
cells amongst which the sperm-morule lie. The discrepancy 
as to numbers and position is not a very serious matter. 
Poléjaeff himself has shown how enormously the number of 
Spermatozoa produced differs in different individuals, and I 
find myself that in Grantia, although the sperm-morule are 
generally to be observed in the walls of the flagellate cham- 
bers, or lying free in the chamber-cavities, some of the early 
stages of spermatogenesis occur in the mesoglea, while later 
stages are found in the inhalant canals (possibly of different 
individuals), and these might well appear in sections to be 
lying in the tissues. 

Poléjaeff (1882) found that in Sycon raphanus, although 
the sponge is hermaphrodite, the vast majority of individuals 
are predominantly female, and only very occasionally a pre- 
dominantly male specimen is forthcoming. In the latter, 
however, the sperm-balls (‘‘ Spermaklumpen”) were so nume- 
rous that their whole development could be traced in a single 
section. He derives these sperm-balls from ordinary amcebo- 
cytes (“Wanderzellen”’) in the mesogloea (mesoderm), similar 
to those which, in his opinion, give rise to the ova. These 


GAMETOGENESIS OF GRANTIA COMPRESSA. 357 


cells have a diameter of 0°008-0:02 mm., and their bright, 
vesicular nuclei are distinguished by their relatively large 
size and their highly refractive nucleolus. Such a cell is 
represented in Poléjaeff’s fig.3a as a spherical body with an 
excentrically placed nucleus. ‘The nucleus now divides into 
two somewhat unequal parts, which take up their positions at 
opposite poles of the cell, which becomes differentiated into 
two corresponding parts, a cover-cell and a sperm mother-cell 
(“ Ursamenzelle”’). The cover-cell does not divide again, but 
the nucleus of the sperm mother-cell divides repeatedly, and 
finally gives rise to a large number of very minute, granule- 
like spermatozoon-heads, enclosed within a capsule formed 
by the cover-cell. Hach of these heads presently becomes 
provided with a cytoplasmic tail. During this process there 
is no increase in volume of the sperm-ball, and Poléjaeff 
remarks upon the extraordinarily small size of the sper- 
matozoa. He also points out that the spermatogenesis of 
Sycon as described by him diifers in several respects from 
that described for non-caleareous sponges by Schulze, 
Keller and others. Thus the division of the nucleus of the 
“Ursamenzelle” is not immediately followed by division of 
the cytoplasm, so that there arises a multinucleate mother- 
cell and not a true sperm-morula, but he remarks that this is 
not a matter of any very great importance. More significant, 
perhaps, is the absence of the endothelial capsule formed 
around the sperm-ball by the mesoglcea cells in some of the 
non-calearea. This, I think, is a matter of some importance 
in connection with the problem of how the spermatozoa are 
transferred from one sponge to another. This problem Polé- 
jaeff does not attempt to solve, and, indeed, we are left in 
doubt, after studying his paper, as to whether or not he 
considers that self-fertilisation takes place in Sycon. His 
fig. | shows immense quantities of what appear to be sper- 
matozoon heads in the inhalant canals, the chambers and the 
exhalant canals, as well as sperm-balls in the mesoglea, but 
no attempt is made to decide the question whether or not all 
this mass of sperm has been derived from the same sponge. 


358 ARTHUR DENDY. 


I think the investigation of this question would probably 
have gone a long way towards reconciling the discrepancies 
between the observations of Haeckel and those of Eimer, and 
have shown that the cover-cells retain their wandering pro- 
pensities for some time, and carry the spermatgonia from the 
mesogloea into the collared-cell layer, whence they are dis- 
charged into the water-stream and carried out of the sponge 
altogether, possibly to find their way back again, either into 
the same or into another sponge, through the inhalant canal- 
system. My own observations clearly indicate that this is the 
course of events, though there appear to be noteworthy 
differences in details of behaviour between the two genera 
Sycon and Grantia. 

During the thirty-two years that have passed since the 
publication of Poléjaeff’s memoir, the only contribution that 
has been made to the very difficult problem of the spermato- 
genesis in calcareous sponges is that contained in the paper 
by Wilhelm Gorich—* Zur Kenntnis der Spermatogenese bei 
den Poriferen und Coélenteraten nebst Bemerkungen tber die 
Oogenese der erstern”’ (1903). This author again deals with 
the process as exhibited in Sycon raphanus, and although 
he describes only a few of the earlier stages in this sponge, 
his results are in one respect strikingly at variance with those 
of Poléjaeff. He agrees with the latter in deriving the 
spermatogonia from mesogloal amoebocytes, which round 
themselves off at an early stage of their growth as compared 
with the odgonia. He also describes the formation of a cover- 
cell, but in a totally different manner from that described by 
Poléjaeff. The spermatogonium and the cover-cell, though 
both derived from amcebocytes of the mesogloea, differ from 
one another in certain particulars and do not arise by division 
of a common mother-cell. hey only come into relation 
with one another secondarily, the cover-cell spreading itself 
around the spermatogonium, and finally ingesting it, in a way 
which is evidently very similar to the process of phagocytosis 
already described by me for Grantia. The result, however, 
is a Spermatogonium enclosed in a mother-cell very much as 


GAMETOGENESIS OF GRANTIA COMPRESSA. 359 


described by Poléjaeff. The spermatogonium is represented 
as dividing mitotically into two and then incompletely into 
four parts, beyond which its history was not followed. 

This is all that is known of the spermatogenesis in 
Calcarea. Amongst other sponges the most frequently and 
most fully investigated form is Spongilla, and it is perhaps 
worth while to say a few words about what is known in this 
case, which seems to be typical of the non-Calcarea, before 
proceeding to describe my own observations on Grantia. 

As far back as 1888, Fiedler published his memoir, “ Uber 
Hi- und Spermabildung bei Spongilla fluviatilis.” He 
describes the formation of cover-cell and sperm mother-cell 
(““Spermatocyte”’) by division of a common mother-cell 
(“Spermatogonium”) exactly as described by Poléjaeff for 
Sycon. The cover-cell, however (of which more than one may 
be found), not infrequently disappears before the contained 
“spermatocytes ” have completed their development, and the 
mass of sperm, which may have been derived from several 
“spermatogonia,” becomes enclosed in a secondary follicle 
formed from ordinary mesogleeal cells (“‘ Parenchymzellen”). 
The original “spermatocyte” divides repeatedly by mitosis 
into daughter-cells which are completely separated from one 
another. The smallness of the objects, however, makes the 
examination of the process very difficult and the details of 
mitosis are not very satisfactorily given. The last generation 
of spermatocytes, the spermatids, develop directly into the 
spermatozoa. A compact chromatin-ball is formed by con- 
traction of the nucleus, as previously described by Schulze 
for Halisarca (= Oscarella) (1877) and Aplysilla (1878), and 
the enveloping cytoplasm is drawn outinto aslender tail. In 
the fully formed spermatozoon the chromatin-ball forms a 
minute spherical head. 

Gorich (1903) added some interesting particulars as to 
Spongilla, especially with regard to the structure of the fully 
formed spermatozoon, in the paper which I[ have already 
quoted. He finds that the number of cover-cells taking part in 
the formation of the capsule or spermatocyst varies from one 


360 ARTHUR DENDY. 


to about six, and maintains that these cover-cells are derived 
from mesoglceal cells distinct from the spermatogonium as in 
the case of Sycon. He brings forward very little evidence, 
however, in support of this view. He describes the often 
repeated mitotic division of the sperm-cells within the 
spermatocyst very much as it was described by Fiedler. In 
the fully developed spermatozoa, however, he finds a far more 
complex structure than had been observed by any of his 
predecessors, for in addition to the spherical head and long 
slender tail he describes and figures middle piece, apical body 
and centrosomes, thus bringing the structure closely into 
line with that of the spermatozoon in Enterozoa, as ex- 
emplified by Aurelia, which he describes and figures in the 
same paper. 

With regard to the explanation of the close resemblance 
thus established between the spermatogenesis of sponges and 
that of the Enterozoa, and its bearing upon the relationship 
of the two groups, I may refer to what I have already said in 
my summary on the odgenesis. 


(b) Origin and Growth of the Primary Spermato- 
gonia. 

In returning to Haeckel’s view that the spermatozoa are 
formed by division of collared cells, I must admit that it is 
extremely difficult to bring forward convincing evidence that 
this is really the case. Haeckel, of course, was of opinion 
that the collared cells become divided up into spermatozoa 
in siti, and he says nothing of the existence of the sper- 
matocyst or cover-cell described by Poléjaeff and Gorich, 
while both the latter hold that the spermatozoa develop from 
amcebocytes of the mesoglea. I believe that the view of 
each of these authors expresses part of the truth, and I hope 
that my own observations may serve to account for, and to a 
large extent to reconcile, the discrepancies between them. 
The manner in which I have interpreted these observations 
and arranged the different stages cannot even yet, however, 
be regarded as more than tentative. 


GAMETOGENESIS OF GRANTIA COMPRESSA. 561 


The first stage in spermatogenesis, as in that of odgenesis, 
appears to be the enlargement of individual collared cells in 
the lining epithelium of the chambers (figs. 69,70). At the 
same time the cytoplasm acquires a peculiar curdled appear- 
ance (if I may use this expression for want of a better), which 
looks as if it might be due to the running together of the 
reserve granules. Irregular inclusions of large size may thus 
be formed, around which vacuoles frequently make their 
appearance. By these appearances it is possible to distin- 
guish between what I believe to be the primary spermato- 
gonia and the primary odgonia respectively, for in the latter 
(figs. 11, 12) it will be remembered that the reserve granules 
remain separate and do not run together in irregular masses. 

The nucleus now becomes very distinctly reticulate as com- 
pared with that of neighbouring collared cells (figs. 71, 72), 
collar and flagellum are withdrawn, and the cell puts out 
pseudopodia and becomes ameeboid (figs. 72-74). In this con- 
dition the primary spermatogonia are to be found hanging 
into the chamber from the layer of collared cells by means of 
their pseudopodia. Definite inclusions disappear from the 
cytoplasm. 

The cell now rounds itself off and assumes a very charac- 
teristic appearance (figs. 75, 76). It is readily distinguished 
from the oogonia by its smaller size and by the character of 
the nucleus. It is also quite different in character from the 
young odcytes, which are of about the same size (figs. 29, 30), 
especially as regards the nucleus, which is coarsely reticulate 
and without a really well-defined nucleolus, while that of the 
young oocyte has a very conspicuous nucleolus surrounded by 
a clear zone and then by a zone of fine granules. 

I have found the primary spermatogonia in their rounded- 
off condition in the mesogloea as well as in the flagellate 
chambers, so that it seems probable that while still in the 
amoeboid state they may migrate through the chamber-walls 
as the o6gonia and odcytes so frequently do. 

It is interesting to observe that the primary spermatogonia 
exhibit a good deal of variation in size, as is shown in fig. 

VoL. 60, PART 3.—NEW SERIES. 26 


362 ARTHUR DENDY. 


75. In one case also (fig. 75a) I have observed mitosis in the 
free spermatogonium, but I have no evidence that the sper- 
matogonium ever actually divides until it has been provided 
with a cover-cell. 


(c) Formation of the Spermatocysts or Cover- 
cells. 


It will be remembered that according to Poléjaeff the 
original amcebocyte in Sycon divides into two parts, one of 
which forms the cover-cell and the other the primary sperma- 
togonium, while Gérich claims that the cover-cell is formed 
by an independent amcebocyte which approaches and enve- 
lopes the spermatogonium in the mesogloea. My own obser- 
vations strongly support the latter view, and I regard the 
process of envelopment of the spermatogonium by the cover- 
cell as a special case of phagocytosis. Indeed, as already 
pointed out, spermatogonia are frequently ingested by the 
phagocytes, and it is difficult, if not impossible, to distinguish 
an ordinary case of phagocytosis in which only a single 
spermatogonium has been ingested, from a case of cover-cell 
formation (cf. fig. 62, in which the ingested cell, however, 
seems too large to be a spermatogonium). Figs. 77, 80 and 
81 represent spermatocysts, with enclosed primary spermato- 
gonia, lying in the mesoglea. In these cases the cover-cell 
appears to resemble closely a small phagocyte such as gives 
rise to the nurse-cells (cf. fig. 58). 

The majority of the spermatocysts, however, are found 
lying in the walls of the flagellate chambers between the 
collared cells, with the enclosed spermatogonium either still 
undivided, as shown in figs. 78 and 79, or in process of 
division, as shown in fig. 82, or, much more frequently, 
divided up into a sperm-morula (figs. 84, 85). When in 
this position the spermatocyst certainly looks very much as 
if it were derived in siti from a collared cell. Sometimes, 
it is true, the nucleus is distinguishable from that of adjacent 
collared cells by its reticulate character (figs. 82, 85), but in 
other cases (figs. 78, 84) no such distinction can be made out. 


GAMETOGENESIS OF GRANTIA COMPRESSA. 363 


I have already pointed out, however, that the nuclei of the 
collared cells themselves vary very greatly in appearance, 
and that reticulate and umiformly dark-stained nuclei may 
occur in adjacent cells. Sometimes the cytoplasm of the 
cover-cell may even contain reserve granules like those found 
in the collared cells (fig. 85). 

The history of the spermatogonia themselves, however, and 
the phagocytosis observed in the mesogloea, seem to me to 
indicate very clearly that the spermatocysts so often seen 
lying between the collared cells have reached their position 
by migration, and there is no sufficient reason for concluding 
that the cover-cells are ever derived from collared cells. 

As the spermatocyst lies in the collared cell layer its 
nucleus is situate, usually, at any rate, towards the lumen 
of the chamber, just as are the nuclei of the collared cells 
themselves, and the spermatogonium or sperm-morula is 
enclosed in its basal portion (figs. 78, 82, 84). 

As the sperm-morula develops the cyst formed by the 
cover-cell becomes extremely thin (fig. 87) and finally 
ruptures, discharging the sperm-morula into the cavity 
of the chamber (fig. 85). Thus the liberation of the 
sperm-morula from the cover-cell takes place much earlier 
than in the case of Sycon, where, according to Poléjaeff, 
spermatozoa are formed while still within the cysts. 


(d) Development of the Sperm-morule from the 
Primary Spermatogonia. 


The first division of the primary spermatogonium at least 
appears to take place mitotically, as has already been 
described by Gorich for Sycon. At any rate one sometimes 
finds spermatogonia in which the nucleus has disappeared, 
and what appear to be small, scattered chromosomes, eight or 
ten in number, are scattered through the cytoplasm as shown 
in figs. 81 and 82, while fig. 80 represents what may be a 
spireme stage. 

I have only once observed what appears to be the two- 


564 ARTHUR DENDY. 


celled stage of the sperm-morula (fig. 85), and I regard this 
as a somewhat doubtful case; it shows, however, two distinct 
spherical bodies, which I take to be secondary spermatogonia, 
enclosed in what is presumably a cover-cell (cov.) lying in the 
layer of collared cells. 

The four-celled stage I have never been able to find, in 
spite of prolonged searching. The eight-celled stage, however, 
T have seen several times (figs. 84, 85, 86), and it appears 
that the sperm-morula may be liberated from the  sper- 
matocyst as early as this (fig. 86). 

A stage in which the morula consists of sixteen cells, or 
thereabouts, is the most frequent in my preparations. This 
stage sometimes occurs still enclosed in the cover-cell (fig. 87), 
but more frequently lying free in the cavity of the flagellate 
chamber (figs. 88-92). A remarkable feature of the eight- 
celled and sixteen-celled stages is the extraordinary distinct- 
ness with which the daughter-spermatogonia are defined, but 
the exact appearance evidently depends somewhat upon the 
method of preparation. Very often they appear as little 
heaps of highly refractive spherical balls, like small shot, 
stained black or nearly so (figs. 83, 88). At other times they 
are much more lightly stained, and one or more minute, more 
darkly stained granules appear within them (figs. 86, 87, 89, 
90). How the division takes place I cannot say, but I have 
seen no sign of mitosis after the first division of the primary 
spermatogonium. Whether or not the spherical bodies 
represent entire cells, or nuclei only, I am also unable to say 
positively, but I conclude from their general appearance that 
the former is the case. It will be noticed that there is con- 
siderable difference in size between the spermatogonia in 
different sperm-morule of apparently the same stage of 
development (cf. figs. 87, 88), but this is only what might be 
expected from the differences in size of the primary spermato- 
gonia already mentioned. Although usually spherical, the 
daughter-spermatogonia sometimes appear to be polygonal 
from mutual pressure (figs. 89, 91). 

A noteworthy feature is the appearance between them in 


GAMETOGENESIS OF GRANTIA COMPRESSA. 565 


the sperm-morula of a substance that looks like residual 
protoplasm (figs. 86-92). As development proceeds (at the 
sixteen-celled stage), this material seems to swell up so as to 
separate the spermatogonia more or less from one another. 


(e) The Formation of Spermatozoa; Comparison 
with Sycon, etc. 


As to how the spermatozoa are developed from the sperm- 
morule in Grantia I have no definite information to offer. 
My material does not suffice to settle this question, but I 
have some reason to believe that the spermatogonia of the 
sixteen-celled stage undergo further repeated subdivision. I 
believe that this usually takes place after the sperm-morule 
have been transferred by the water currents to the inhalant 
canals of another individual. I have occasionally found, both 
in the inhalant canals and flagellate chambers, small masses 
of darkly stained granules (fig. 93) which appear to be 
identical with the masses of granules which Poléjaeff showed 
to be spermatozoon heads in Sycon. If they be of this nature, 
as I think highly probable, their minute size certainly serves 
to indicate a further breaking up of the spermatogonia of the 
sixteen-celled stage. I have also a small amount of direct 
evidence of such breaking up of the sperm-morule in the 
imhalant canals, but it is not conclusive enough to bring 
forward definitely. 

Poléjaeff tells us that in Sycon the spermatozoa are formed 
by repeated subdivision of the spermatogonia within the cover- 
cell, during which process the spermatocyst does not increase 
in size. The final products of these divisions are represented 
as being extremely minute, and each becomes provided with 
a slender flagellum. 

I happen to have in my possession one of Poléjaeff’s own 
preparations of Sycon, sent by him to Mr. Carter in 1883. 
This preparation shows “sperm-balls”’ (sperm-morule) in 
the inhalant canals, and I suppose them to have come from 
another individual. The sperm-balls vary considerably in 


366 ARTHUR DENDY. 


size (fig. 94) and each is still enclosed in the cover-cell, whose 
nucleus is very distinctly visible. In the interior a number 
of ill-defined, faintly stained bodies are present, which may be 
daughter-spermatogonia, or, as Poléjaeff supposes, merely the 
nuclei of these. They are nothing like so definite as the 
spermatogonia which I find in the sperm-morule of Grantia, 
but this may be partly because they belong to a later stage 
of development, and partly because of differences in the mode 
of preparation. According to the information given in his 
memoir Poléjaeff’s preparations were made with osmic acid 
(0°01-0°05 per cent.) material stained with alum carmine, 
and this probably applies to the preparation in my possession. 

I give in fig. 94 drawings of two sperm-balls from this 
preparation. It will be seen that they are of just about the 
same size as the spermatocysts with enclosed primary sperma- 
togonia in Grantia (cf. figs. 77, 78, 81), and, allowing for the 
longer retention of the cover-cell and the further subdivision 
of the spermatogonia after the sixteen-celled stage, I think 
there is no serious discrepancy. 1 am unable to make a direct 
comparison between the earlier stages in the two genera as I 
have found none of these in Poléjaeff’s preparations. 

That the sperm-morulz found by me in Grantia are identical 
with the structures already referred to as described by 
Haeckel in the lining epithelium of the flagellate chambers of 
various calcareous sponges, appears to me to admit of very 
little doubt, although of course I cannot agree with him in 
the details of his account. 

It has occurred to me as just possible that someone may 
suggest that these bodies are not sperm-morulz at all, but of 
quite a different nature; that they may be Sporozoa living 
for a period as intracellular parasites in the collared cells and 
possibly also in the amcebocytes of the sponge. Apart from 
the fact, however, that sporozoan parasites have never yet 
been observed in sponges, I think the developmental history 
of the bodies in question is sufficient to negative this view. 
No infection of the sponge-cells by small forms that might 
be young Sporozoa has ever been observed, and the cell which 


GAMETOGENESIS OF GRANTIA COMPRESSA. 367 


is surrounded by the cover-cell, and which I believe to be the 
primary spermatogonium, has nothing about it to suggest a 
sporozoon. 

I can only regret that I am unable to give a more satisfac- 
tory account of the spermatogenesis, but I hope that what I 
have said will arouse more interest in this extremely difficult 
problem, and enable future workers at any rate easily to find 
the structures in question, and perhaps fill up the many gaps 
which I have left. 


(4) FERTILISATION OF THE Ovum. 


I have not been fortunate enough to observe the actual 
entrance of the spermatozoon into the mature ovum. According 
to Jorgensen (in Sycon), the head swells up and gives rise to 
the male pronucleus, in which a nucleolus very soon makes its 
appearance. ‘I'he female pronucleus has in the meantime 
developed from the group of chromosomes that remained 
behind in the ovum after the formation of the second polar 
body. These unite together into a compact chromatin ball, 
around which a vacuole, enclosed in a nuclear membrane, 
makes its appearance. A nucleolusis very early differentiated 
in the midst of this mass, the remainder of which is broken up 
into chromatin granules, which become scattered over the 
nuclear reticulum. 

Sometimes a portion of one of the pronuclei (either male or 
female) is represented by a small separate “ karyomere,” so 
that there appear to be three nuclei in the fertilised ege. 
The number of nucleoli present in the pronuclei varies. 

So far as they go my own observations on Grantia are 
entirely in harmony with these results. I have been able to 
study four ova with well-developed male and female pronuclei. 
In only one of these cases were the two pronuclei both single, 
and in this case the number of nucleoli was either 4 + 1 or 
3 + 2,one of them having been displaced in the cutting. In 
the other three cases a karyomere was present in addition to 
the principal pronuclei, and the numbers and distribution of 


368 ARTHUR DENDY. 


the nucleoli were 1 + 14+ 1,14 14 2,and1+414 2 re- 
spectively. 

Fig. 95 shows one of these cases. The position of the 
degenerating polar body (p. b.) shows that the large pro- 
nucleus with the single large nucleolus is evidently the 
principal female pronucleus, and the small karyomere, with 
single nucleolus, doubtless belongs to it. The male pro- 
nucleus, on the right and below, has two nucleoli. 

The formation of karyomeres is a very curious and striking 
phenomenon, for further details as to which I may refer the 
reader to Jérgensen’s paper. That author points out that 
karyomere formation also takes place in Sycon in the process 
of segmentation of the fertilised ovum, and I| find the same 
to be true in the case of Grantia. 

.It is perhaps worth while pointing out that the nucleoli in 
the male and female pronuclei do not stain nearly so deeply as 
the nucleoli of the young odcytes. The same is also true of 
the nucleoli in the older odcytes. The difference may perhaps 
be correlated with the fact that in the young odcytes the 
nucleolus is actively engaged in the formation of chromidia or 
“ yolk-nuclei,’” while probably it is not directly engaged in 
this process in the older odcytes, and certainly not in the 
fertilised ovum, where yolk-formation has ceased. 


(x) List or LirERATURE REFERRED TO. 


1854. Carter, H. J.—‘* Zoosperms in Spongilla,” ‘Ann. and Mag. Nat. 
Hist.,’ vol. xiv. 

“ Notes Introductory to the Study and Classification of 
the Spongida,” ‘ Ann. and Mag. Nat. Hist.,’ vol. xvi. 

1888. Dendy, A—‘ On the Anatomy and Histology of Stelospongus 


flabelliformis, Carter; with Notes on the Development,” 
‘Quart. Journ. Micr. Sci.,’ vol. 29, N.s. 


1891. “On the Anatomy of Grantia labyrinthica, Carter, 
and the so-called Family Teichonide,” ‘Quart. Journ. Micr. 
Sci.,’ vol. 32, N.s. 

1913. 


“A meebocytes in Caleareous Sponges,” ‘ Nature,’ December 
4th and December 25th, 1915. 


1910. 


1856. 


1894. 


1900. 


1898. 


1900. 


1912. 


OS. 


1914. 


1882. 


1911. 


1910. 


GAMETOGENESIS OF GRANYTIA COMPRESSA. 369 


2. Kimer, Th.—‘‘ Nesselzellen und Samen bei Seeschwiimmen,” 


‘Arch. Mikr. Anat.,’ vol. vii. 


. Fiedler, K.—*‘ Uber Ei- und Spermabildung bei Spongilla 


fluviatilis,” ‘ Zeit. wiss. Zool.,’ vol. xlvii. 


. Gorich, W.—* Zur Kenntnis der Spermatogenese bei den Pori- 


feren und Colenteraten nebst Bemerkungen iiber die Odgenese 
der Ersteren,”. ‘ Zeit. wiss. Zool.,’ Bd. lxxvi, Heft iv. 


. Haeckel, E.—‘ Uber die sexuelle Fortpflanzung und das natiiwliche 


System der Schwamme,” ‘ Jenaische Zeitschrift fiir Medicin und 
Naturwissenschaft,’ Bd. vi. 


‘ Die Kalkschwimme.’ 


. Huxley, T. H.—* Zoological Notes and Observations made on 


Board H.M.S. ‘ Rattlesnake’; 2, ‘On Tethya,” ‘Ann. and Mag. 
Nat. Hist.,’ vol. vii. 

Jorgensen, M.—* Beitrage zur Kenntnis der Hibildung, Reifung, 
Befruchtung und Furchung bei Schwammen (Syconen),’’ ‘ Archiv. 
fiir Zellforschung,’ Bd. iv. 

Lieberkthn, M.—‘ Beitrage zur Entwicklungsgeschichte der 
Spongillen,” ‘ Muller’s Archiv.,’ 1856. 

Maas, O—‘Uber die erste Differenzierung von Generations- 
und Somazellen bei den Spongien,” ‘ Verhandlungen der 
deutschen Zoologischen Gesellschaft, Gottingen,’ 1893. 


“Die Weiterentwicklung der Syconen nach der Meta- 
morphose,” ‘ Zeit. wiss. Zool.,’ Bd. Ixvi. 


Minchin, E. A.—‘‘ On the Origin and Growth of the Triradiate 
and Quadriradiate Spicules in the Family Clathrinide,’ ‘ Quart. 
Journ. Micr. Sci.,’ vol. 40, n.s. 


“ Sponges,” ‘ Lankester’s Treatise on Zoology.’ 
‘An Introduction to the Study of the Protozoa.’ 


Orton, J. H.—‘‘ On a Habitat of a Marine Ameceba,” ‘ Nature,’ 
November 27th, 1913. 


* Preliminary Note on a Contribution to an Evaluation 
of the Sea,” ‘Journal of the Marine Biological Association,’ 1914. 


Poléjaeff, N.—‘* Uber das Sperma und die Spermatogenese bei 
Sycandra raphanus Haeckel,” ‘Sitzb. der k. Akad. der 
Wissensch. Wien,’ Bd. Ixxxvi. 

Robertson, M.—‘ The Division of the Collar-cells of Calcarea 
Heterocela,” ‘ Quart. Journ. Mier. Sci.,’ vol. 57. 

- and Minchin, EK. A.—* The Division of the Collar-cells of 
Clathrina coriacea,”’ ‘Quart. Journ. Micr. Sci.,’ vol. 55. 


370 ARTHUR DENDY. 


1877. Schulze, F. E.—* Untersuchungen iiber den Bau und die Entwick- 
lung der Spongien. Die Gattung Halisarca,” ‘Zeit. fiir wiss. 
Zool.,’ Bd. xxviii. 

“Untersuchungen iiber den Bau und die Entwicklung 


der Spongien. Die Familie der Aplysinide,” ‘Zeit. fiir wiss. 
Zool., Bd. xxx. 


(Lt) EXPLANATION OF PLATES 23 10 26, 


Illustrating Prof. Arthur Dendy’s paper, ‘‘ Observations on 
the Gametogenesis of Grantia compressa. 


[All figures are magnified about 1650 diameters, and, with the excep- 
tion of fig. 94, refer to Grantia compressa. | 


PLATE 23. 


Fig. 1.—Collared cell putting out pseudopodia, from teased prepara- 
tion of living sponge. col. Collar. (Unstained.) 

Fig. 2.—Collared cell (a), with flagellum still present, putting out 
pseudopodia and retreating into the mesoglea from the gastral epithe- 
lium. b., 6. Collared cells still in position in the gastral epithelium. 
(Specimen 24. Borax carmine.) 

Fig. 3—Ameebocyte (a) lying behind the gastral epithelium and prob- 
ably derived from a collared cell. b., b. Collared cells. (Specimen 24. 
Borax carmine.) 

Fig. 4.—Portion of section at right angles to the gastral surface, 
showing granular epithelial cell (a) in the “ flask-shaped ” condition at 
the surface, and amcebocyte (b) in the mesoglea. sp. Spicules. (Speci- 
men 24. Iron brazilin.) 

Fig. 5.—Ameebocyte in the mesoglcea just beneath the gastral cortex. 
(Specimen 24. Iron brazilin.) 

Fig. 6.—Portion of section at right angles to the gastral cortex, show- 
ing parts of two granular epithelial cells (a) in position on the surface, 
and a group of amcebocytes (b) lying in the mesoglea and evidently 
derived by immigration from the epithelial layer. p.g. A mass of 
pigment-granules apparently discharged from an amcebocyte. (Speci- 
men 24. Iron brazilin.) 

Fig. 7.—Portion of section through inhalant canal (7. ¢.) and adjacent 
mesoglea, showing transition from epithelial to ameeboid cells. (Speci- 
men 24. Iron brazilin.) 


GAMETOGENESIS OF GRANTIA COMPRESSA. 371 


Fig. 8.—Corresponding portions of two consecutive sections taken 
tangentially through the layer of collared cells, (a) through the nuclei, 
(b) through the cytoplasm below the nuclei; showing variations in the 
intensity of staining of the reserve granules. a, 8, y, 6, are identical 
cells in the two sections. (Specimen 21. Paracarmine and _picro- 
indigo carmine.) 

Fig. 9.—A collared cell (?) in mitosis, showing the two groups of 
daughter-chromosomes, seen end on. (Specimen 24. Borax carmine.) 


Fig. 10.—Two adjacent collared cells, one in the normal condition and 
the other beginning to enlarge to form a primary germ-cell. (Specimen 
21. Paracarmine and picro-indigo carmine.) 

Fig. 11.—Part of section through the exhalant opening (e. 0.) of a 
flagellate chamber, showing two normal collared cells and the con- 
version of another into a primary odgonium (p.o.). g.¢. gastral surface. 
(Specimen 21. Paracarmine and picro-indigo carmine.) 

Fig. 12.—Part of a similar section showing another primary odgonium 
(p. 0.) with remains of collar and flagellum and a pseudopodial process 
wedged in between the adjacent collared cells. (Specimen 2]. Para- 
carmine and picro-indigo carmine.) 

Fig. 13.—Primary odgonium (p. 0.) after complete loss of collar and 
flagellum, migrating into the mesoglea from the layer of collared 
cells (c. c.). (Specimen 24. Borax carmine.) 

Fig. 14.—Actively ameboid primary odgonium in the mesoglea. 
(Specimen 24. Borax carmine.) 

Fig. 15—Primary odgonium in which small granules of chromatin 
are beginning to appear in the nucleus around the nucleolus. (Speci- 
men 24. Borax carmine.) 

Fig. 16.—Primary odgonium rounding itself off and entering upon 
resting stage before entering flagellate chamber. (Specimen 22. Iron 
brazilin.) 

Figs. 17, 18.—Two primary oégonia in prophase of mitosis (spireme) ; 
the one on the right (fig. 18) showing the expulsion of the nucleolus. 
ec. c. Collared cells. (Specimen 22. Iron brazilin.) 


Fig. 19.—Primary odgonium in mitosis in flagellate chamber (late 
prophase), showing equatorial plate, spindle and centrosomes, with 
chromidia in cytoplasm. (Specimen 11. Iron brazilin.) 


Fig. 20.—Oogonium in mitosis in chamber. Same stage as last, but 
of smaller size. (Specimen 22. Iron brazilin.) 


Fig. 21.—Primary oé6gonium in mitosis (metaphase) forcing its way 
between the collared cells into a flagellate chamber. (Specimen 21. 
Paracarmine and picro-indigo carmine.) 


Ore ARTHUR DENDY. 


Fig. 22.—-Primary o6gonium in mitosis in chamber (anaphase), showing 
separation of the groups of daughter-chromosomes on the spindle. 
(Specimen 11. Tron brazilin.) 


Fig. 23—Primary ojgonium in mitosis in chamber (late anaphase). 
(Specimen 21. Paracarmine.) 


Fig. 24.—Odégonium in mitosis in chamber (late anaphase). (Speci- 
men 22. Iron-brazilin.) 


Fig.25.—Primary odgonium in mitosis in chamber (telophase), showing 
cell-division ; chromidia still visible in cytoplasm. (Specimen 22. Iron 
hematoxylin.) 


Fig. 26—Two daughter-cells (secondary oégonia) resulting from 
division of primary o6gonium, in chamber; nuclei not yet completely 
reconstituted ; chromidia still visible in cytoplasm. (Specimen 22. 
Iron hematoxylin.) 


Fig. 27.—Secondary odgonium lying in flagellate chamber, with 
reconstituted nucleus. c.c. Collared cells. (Specimen11. Iron-hema- 
toxylin and picro-indigo carmine.) 

Fig, 28.—Secondary odgonium in chamber, in mitosis. (Specimen 22. 


Ivon brazilin.) 


Fig. 29.—Young oécyte, attached to wall of flagellate chamber by 
pseudopodium. c.c. Collared cells. (Specimen 22. Ivon brazilin.) 


Fig. 30.—Young odcyte in chamber. (Specimen 22. Iron brazilin.) 
Fig. 31—Young odcyte in chamber, with one inclusion in the cyto- 


plasm. (Specimen 22. Iron brazilin.) 


Fig. 82.—Young odcyte in chamber, with cytoplasmic inclusions. 
(Specimen 24. Iron brazilin.) 


Figs. 33-35.—Young odcytes in chambers, with very darkly stained 
chromidia or ‘“ yolk-nuclei.” (Specimen 22. Iron brazilin.) 


Fig. 36.—Young odcyte in chamber, with food-particles in vacuoles, 
and chromidia. (Specimen 22. Tron hematoxylin.) 


Fig. 37.—Young odcyte in chamber, with food-particle in vacuole. 
(Specimen 22. Iron hematoxylin.) 


Fig. 38.—Young odcyte in chamber, showing formation of chromidia 
(yolk-nuclei) by expulsion of chromatin from nucleolus. (Specimen 23. 
Iron brazilin.) 


Fig.39.—Young oocyte in chamber, attached to wall at a,a, containing 
chromidia or “ yolk-nuclei ” and remains of ingested cells; nw., nucleus 
of oocyte; n.7.¢., nucleus of ingested cell. (Specimen 23. Iron brazilin.) 


GAMETOGENESIS OF GRANTIA COMPRESSA. Sie 


PLATE 24. 


Fig. 40.— Young odcyte leaving a flagellate chamber after feeding, and 
entering the mesoglea. c.c. Collared cells. (Specimen 11. Tron-haema- 
toxylin and picro-indigo carmine.) 

Figs. 41-43.—Formation of chromidia or “ yolk-nuclei”’ of the young 
oocyte (in the mesoglea) by extrusion of chromatin from the nucleolus 
nto the cytoplasm. (Specimen 23. Iron brazilin.) 

Fig. 44.—Young odcyte in prophase of mitosis. Leptotene stage. 
(Specimen 21. Paracarmine and picro-indigo carmine.) 

Fig. 45.—Young odcyte in prophase of mitosis. Pachytene stage, 
showing contraction of the skein. (Specimen 22. Iron hematoxylin.) 


Fig. 46.—Young odcyte immediately after the prophase of mitosis, 
the spireme broken up again. (Specimen 22. Iron hematoxylin.) 


Fig. 47.—a. Odcyte in “resting condition” behind wall of chamber. 
b. A similar odcyte flattening itself out against the layer of collared 
cells (c.c.). (Specimen 11. Iron hematoxylin and picro-indigo carmine. ) 


Fig. 48—Two odcytes in the mesoglea, attached by “ anchoring 
pseudopodia”’ to the layer of collared cells (c.c.). (Specimen 21. Para- 
carmine and picro-indigo carmine.) 

Fig. 49.—Odcyte in contraction stage, with pseudopodia reduced to 
blunt knobs. (Specimen 11. [ron hematoxylin.) 

Figs. 50, 51.—Odcytes being fed by nurse-cells (n.¢.), which have 
captured food-cells (f. c.) and are passing them into the odcyte. (Speci- 
men 11. Iron brazilin.) 

Fig. 52.—Odcyte of nearly full size, lying behind the layer of collared 
cells (¢.c.). ps. Base of pseudopodium which has been cut off. xn. c. 
Nurse-cell, with food-cell lying in the cytoplasm of the odcyte beneath 
it. (Specimen 11. Ivon brazilin.) 

Fig. 53.—Odcyte of nearly full size, showing very strong development 
of chromidia (“yolk-nuclei”) in cytoplasm. ps. Base of pseudo- 
podium that has been cut off; the other pseudopodia are altogether 
missed by the section. (Specimen 11. Iron brazilin.) 


Fig. 54.—Odcyte of approximately full size, just before withdrawal 
of the pseudopodia and rounding off. (Specimen 11. Iron brazilin.) 


PLATE 25. 


Fig. 55.—Odcyte after withdrawal of pseudopodia and disruption of 
the nucleus. xo. Nucleolus undergoing absorption in the cytoplasm 
(more distinct in next section). chr. (?) Group of chromosomes (°). 
(Specimen 23. Iron brazilin.) 


old ARTHUR DENDY. 


Fig. 56.—Tangential section of maturing oécyte with first matura- 
tion spindle in anaphase. (Specimen 11. Iron brazilin.) 

Fig. 57.—Tangential section of maturing odcyte, showing formation 
of first polar body (p.b.). (Specimen 11. Iron brazilin.) 

Fig. 58.—Small ameebocyte of the phagocyte (nurse-cell) type from 
the mesoglea between the chambers. (Specimen 11. Iron brazilin.) 

Fig. 59.—Large amcebocyte of the phagocyte type; forcing its way 
through the layer of collared cells (col.) near the exhalant opening of a 
chamber. (Specimen 21. Paracarmine.) 

Fig. 60.—Large ameebocyte of the phagocyte type, found in flagellate 
chamber. (Specimen 24. Borax carmine.) 

Fig. 61.—Large amebocyte of the phagocyte type, from the mesoglea 
between the chambers, stretched out against the layer of collared cells 
(c.c.), some only of which are shown. (Specimen 22. Tron hematoxylin.) 

Fig. 62.—Small phagocyte, with single large ingested cell; from 
mesoglea between chambers. (Specimen 24 Borax carmine.) 

Fig. 63.—Phagocyte (ph.) engaged in ingesting the end collared cell 
of a row, i. e. the one next to the exhalant opening of a chamber. . ¢. 
Collared cells. ps. Pseudopodium cut off from another amebocyte. 
(Specimen 24. Iron brazilin.) 

Fig. 64.—Resting phagocyte (nurse-cell?) in mesoglea just outside 
layer of collared cells (c. c.), with two ingested cells in process of diges- 
tion. n.p. Nucleus of phagocyte. (Specimen 11. Iron hematoxylin.) 

Fig. 65.—Active phagocyte in mesoglea outside chamber, attached 
by pseudopodia to the outer surfaces of the collared cells (c.c.), ¢.¢. 
ingested cells. n.p. Nucleus of phagocyte. (Specimen 24. Tron 
brazilin.) 

Fig. 66.—Phagocyte forcing its way through layer of collared cells. 
Lettering as before. (Specimen 24. Borax carmine.) 

Fig. 67.—Large phagocyte with six ingested cells, forcing its way 
through the layer of collared cells. Lettering as before. (Specimen 
24. Borax carmine.) 

Fig. 68.—Large phagocyte with four ingested cells, lying in flagellate 
chamber and attached by pseudopodia to the layer of collared cells. 
Lettering as before. (Specimen 24. Borax carmine and picro-indigo 
carmine.) 


PLATE 26. 


Fig. 69.—Section through parts of two flagellate chambers, adjacent 
to the exhalant openings, showing one of the collared cells enlarging to 
form a primary spermatogonium (p.s.). (Specimen 21. Paracarmine 
and picro-indigo carmine.) 


GAMETOGENESIS OF GRANTIA COMPRESSA. 375 


Fig. 70.—Another example of a collared cell enlarging to form a 
primary spermatogonium. (Specimen 21. Paracarmine and_picro- 
indigo carmine.) 

Fig. 71.—Section through lining epithelium of flagellate chamber 
showing one of the collared cells enlarging to form a primary spermato- 
gonium (p.s.). (Specimen 21. Paracarmine and picro-indigo carmine. ) 


Fig. 72.—Metamorphosis of enlarged collared cell into a primary 
spermatogonium (p.s.). ¢.¢. Ordinary collared cell. (Specimen 24. 
Borax carmine.) 

Figs. 73, 74.—Primary spermatogonia (p.s.) in the amceboid con- 
dition, attached by pseudopodia to the layer of collared cells (c. c). 
(Specimen 24. Borax carmine.) (Fig. 74 rather doubtful; it has a 
distinct but not very darkly staining nucleolus with a clear ring around 
it, and is rather large for a spermatogonium.) 


Fig. 75.—Group of primary spermatogonia rounded off; from flagellate 
chamber. Fig. 75a.—Another example from an adjacent chamber, in 
mitosis. (Specimen 24. Borax carmine.) 


Fig. 76.—Primary spermatogonium rounded off; from mesoglcea out- 
side chamber. (Specimen 24. Borax carmine.) 


Fig. 77.—Primary spermatogonium (p.s.) enclosed by cover-cell (cov.), 
behind layer of collared cells (c.c). (Specimen 24. Borax carmine.) 


Fig. 78.—Cover-cell (cov.) and primary spermatog6nium (p.s.) lying in 
layer of collared cells (c.c.). (Specimen 24. Borax carmine.) 


Fig. 79.—Cover-cell (cov.) and primary spermatogonium (p.s.) lying in 
layer of collared cells (c.¢.). (Specimen 24. Iron brazilin.) 


Figs. 80, 81—Primary spermatogonia undergoing mitosis within 
cover-cells ; from mesoglea. (Specimen 24. Iron brazilin.) 


Fig. 82.—Cover-cell (cov.) containing primary spermatogonium (p. s.) 
in mitosis, lying in layer of collared cells (c.¢.). (Specimen 24. Borax 
carmine.) 

Fig. 83.—Section through wall of chamber, showing cover-cell (cov.) 
containing nucleus and two secondary spermatogonia; also 16-celled 
sperm-morula (sp. m.) that has escaped from its cover-cell, the remnant 
of which (?) is shown at x. c.c. Collared cells. ep. Epithelial cell of 
inhalant canal. (Specimen 24. Iron brazilin.) 


Fig. 84.—Section through wall of flagellate chamber and inhalant 
canal, showing cover-cell (cov.) with sperm-morula of 8 cells (sp.m.), lying 
in layer of collared cells (c.c.). phag. Phagocyte with included cells 
lying in mesoglea. ep. Epithelial cells of inhalant canal. (Specimen 
24, Borax carmine.) 


« 


376 ARTHUR DENDY. 


Fig. 85.—Sperm-morula of probably 8 cells (sp. m.) enclosed in cover- 
cell (cov.), with adjacent collared cell (c.c.), in wall of flagellate chamber; 
seen en face. (Specimen 24. Borax carmine.) 


Fig. 86.—Hight-celled sperm-morula lying free in cavity of chamber, 
without cover-cell. (Specimen 24. Iron brazilin.) 

Fig. 87.—Sperm-morula of probably 16 cells still enclosed in cover- 
cell, in lining epithelium of chamber. n.cov. Nucleus of cover-cell. 
(Specimen 24. Borax carmine.) 

Fig. 88.—Sperm-morula of probably 16 cells, lying free in chamber 
and partially broken up. (Specimen 24. Iron brazilin.) 

Fig. 89.—Sperm-morula of probably 16 cells, after escape from cover- 
cell; seen in optical section. (Specimen 24. Iron brazilin.) 


Fig. 90.—Sperm-morula of probably 16 cells, lying free in chamber 
after escape from cover-cell. (Specimen 24. Iron brazilin.) 


Fig. 91—Sperm-morula of probably 16 cells, lying free in chamber 
after escape fron cover-cell, showing large amount of residual proto- 
plasm. (Specimen 24. Borax carmine.) 


Fig. 92.—Sperm-morula of probably 16 cells, lying free in chamber 
after escape from cover-cell, with swollen residual protoplasm. (Speci- 
men 24. Borax carmine.) 


Fig. 93.—Group of spermatozoon heads (?) from flagellate chamber. 
(Specimen 21. Paracarmine.) 


Fig. 94.—Two “sperm-balls” from inhalant canal of Sycon sp. 
n.cov. Nucleus of cover-cell. (From one of Dr. Poléjaeff’s prepara- 
tions.) 


Fig. 95.—Ovum during fertilisation, with male pronucleus containing 
two nucleoli, and female pronucleus represented by two karyomeres 
with one nucleolus each. p.b. Remains of polar body. x... x Line 


of contact of ovum with layer of collared cells. (Specimen 23. Iron 
brazilin.) 


THE CHROMOSOME COMPLEX OF CULEX PIPIENS. 3877 


The Chromosome Complex of Culex Pipiens. 


By 
Monica Taylor, S.N.D., B.Sc. 


With Plates 27 and 28, and 3 Text-figures. 


ContENTs. 

PAGE 
I. Introduction and Methods . : : OG 
II. The Reproductive Organs . ; : . 380 
III. Spermatogenesis . : 2 Hoo 
IV. Oogenesis 3 ; : . 389 
V. Somatic mitosis. : ; : . 389 
VI. Discussion : : Cee . 392 
VII. Summary : : : : . 394 

INTRODUCTION. 


Miss Srrvens, in a paper entitled “The Chromosomes in 
the Germ-Cells of Culex,” gave as one of her conclusions the 
following : 

“ Parasynapsis (parasyndesis!) occurs in each cell generation 
of the germ-cells, the homologous maternal and paternal 
chromosomes being paired in telophase and remaining so 
until the metaphase of the next mitosis.” 

Because of the importance of this discovery, especially in 
the support it lends to the recent theory of parasyndesis, and 
because of its bearing on the theory of the individuality of 
the chromosomes, Dr. Agar, in the summer of 1912, at Tay- 


1] have adopted the word “syndesis” for the conjugation of the 
chromosomes, and “ synizesis” for their clumping together. 


voL. 60, paRT 3.—NEW SERIES. 27 


378 MONICA TAYLOR. 


vallich, Loch Sween, collected and preserved material in order 
to investigate the germ-cells of Culex pipiens. This he 
very kindly gave to me, and in the spring and summer of 
1913 a supply of egg-rafts, larvee, and pupe from Milngavie 
and Skelmorlie has been used in conjunction with the original 
stock. 

The results obtained by a study of the Tayvallich material 
showed clearly that a much more extensive investigation than 
was originally intended would be necessary in order completely 
to elucidate the problems that incidentally presented them- 
selves. Hence many of the egg-rafts and larvee were placed in 
artificial ponds in order that greater control might be exercised 
over the material to be fixed, and single specimens were isolated 
so that their exact age could be determined, and their periods 
of ecdysis watched. 

The life-history of Culex pipiens is to be found set forth 
in innumerable text-books of Natural History, and much has 
been written about mosquitoes in connection with malaria, 
but in no case have I been able to find any adequate account 
of the behaviour and fate of the imagines which hatch out 
and live in captivity, nor of the time that elapses between 
the emergence of the imago and the deposition of eggs. 
Large numbers of pupz, developed, some under natural, others 
under artificial, conditions, from eggs obtained in May or 
August of 1913, were placed in small ponds which were 
covered over with large cages made of mosquito-netting so 
that the resulting imagines could be observed. From many 
hundreds of these captive-reared creatures I have not 
succeeded in obtaining any egg-rafts, although some of the 
imagines have lived for four months. Nor could they be 
induced to suck blood, which, according to the account of 
some naturalists, is necessary, even in non-tropical forms of 
gnats, for the development of the eggs. 

A comparison of the spermathecz of adults which have 
always been captive with those of adults taken in the open 
shows that the former never contain spermatozoa, although 
the latter do. Hence it would appear that captivity is not 


THE CHROMOSOME COMPLEX OF CULEX PIPIENS. 379 


favourable to fertilisation. The completion of this study by 
the detailed investigation of the maturation and fertilisation 
of the egg-cell will have to be postponed for an indefinite 
time until a developmental series of imagines caught in the 
open has been secured, or until the technique of artificial 
rearing has been mastered. 

Another interesting feature in connection with the larve 
is the want of uniformity in the periods of metamorphosis. 
Although the main body of any collection of larve will 
complete their development in the usual time, there are always 
some laggards who double or even treble the usual periods. 
‘emperature and food-supply do not wholly account for this 
retarded development. 

The fixatives employed have been Benda’s fluid, acetic 
bichromate, Gilson’s mercuro-nitric, Flemming, and Gilson- 
Petrunkewitch, the two latter being most successful for the 
cytology proper; the two former were useful for interpreting 
cytoplasmic details. 

Thionin, iron-hematoxylin (prolonged staining), as well as 
Mayer’s cochineal, Ehrlich’s hematoxylin and safranin were 
the stains employed. Many slides were first studied in 
thionin, and then the cover-slip was removed, the thionin 
washed out, the sections re-stained in iron-hematoxylin, and 
comparisons made between the results of the two stains. For 
certain stages after prolonged treatment with iron-hematoxylin 
much extraction was necessary, for others little, so that the 
same slide had often to be studied under various degrees of 
extraction. 

Although aceto-carmine preparations of the whole gonad 
are very useful for mapping out quickly the main facts of 
spermatogenesis, and although this stain has the advantage, 
as Miss Stevens has pointed out, and as my experience has 
confirmed, of increasing the size of the cellular elements and 
of thus rendering them easier of examination, they are not 
permanent, not so good for finer details, and not useful for 
somatic mitosis. Hence the figures given in this paper have 
been taken from sections (thickness ranging from 4 to 12 4) 
and not from aceto-carmine preparations. 


380 MONICA TAYLOR. 


Miss Stevens worked on Culex pungens,' a form which 
is very nearly allied to C. pipiens. It is probable that this 
close relationship between C. pipiens and C. pungens 
accounts for the great similarity which exists between her 
figures of primary and secondary spermatocyte anaphases and 
telophases and those that occur in C. pipiens. 


Trext-FiG. 1. 


amm 


Outline of testis of pupa of Culex pipiens to show level at 
which the various stages in spermatogenesis commonly occur. 
1-4. Synizesis stage. 5-9. Preparation for spermatocyte 1. 
10. Spermatocyte 1. 11,12. Spermatocyte 2. 13. Spermatids. 
14, etc. Spermatozoa in different stages of differentiation. 


Tur REPRODUCTIVE ORGANS. 


The post-embryonic development of Culex has been worked 
out by Hurst (1). The testes are paired, cylindrical in shape, 
and possess no receptacule seminales, the ripe spermatozoa 
being stored in the spermiducts. 

The ovaries (Text-figs. 2 and 3) are paired and cylindrical, 
each consisting of large numbers of ovarian tubes which all 
open into a single duct. The two ducts, one from each side, 
join to form a common oviduct into which three spermathece 
open. 

1 In a note appended to her paper, however, she expresses a fear that 
two species were used for her research. 


THE CHROMOSOME COMPLEX OF CULEX PIPIENS. 381 


The gonad proper is contained in the third segment from 
the hind end of the larva or pupa, which also is true for 
C. pungens (Stevens (2)). 


TEXT-FIG. 2. 


A 
2) BSS Po ea, 
AT Oe 


> ~ = 

; ' : @ : cco 

tt. ef. ot ear, < 
ae Ace ‘2mm 
- x o. t Ter OP pal Wo a 


B 
A. Section through ovary of Culex pipiens. c.o. Common 


oviduct. e. f. Egg-follicle. o. Oviduct of right ovary. 0. t. 
Ovarian tube. ¢.t. Tracheal tube. 3B. Diagram of ovarian 


tube with egg-follicles. 


TEXT-FIG. 3. 


‘Os ™m™M 


Section through ovary of Culex pipiens. e.f. Egg-follicle. 
e. f.e. Egg-follicle epithelium. o. Oviduct. 0. ¢ i 
tube. 7. c. Reproductive cells. ¢. ¢. Tracheal tube. 


SPERMATOGENESIS. 


A figure of a pupal testis drawn from a reconstruction is 
given in T'ext-fig. 1. It will be seen that. the testis is divided 
up into a number of small compartments by walls roughly at 
right angles to the long axis of the organ. These cysts 
contain cells in different stages of development, the contents 


3882 MONICA TAYLOR. 


of one cyst being presumably of the same age. In this par- 
ticular testis the posterior cysts are full of ripe spermatozoa ; 
higher up, the cysts contain spermatids; next to these are 
cysts with immature spermatids, while higher up the com- 
partments are full of interkinetic nuclei. Higher up still are 
cells, which, by their more anterior position in the gonad, and 
by their larger size, are presumably spermatocyte I, while at 
the head of the gonad are nuclei in synizesis. Thus the 
topographical relations of the cysts afford a means of identi- 
fying the different stages in the spermatogenesis. 

After consulting many sections of larve and pup, and 
carefully studying the topographical relations of the cysts, it 
has been possible to distinguish the usual spermatogenesis 
stages, and this having been done, the recognition of the 
different stages that occur in the development of each cell 
generation was not a difficult matter. Although the contents 
of one cyst are presumably of the same age, it is often possible 
to find in it extremes of any stage. ‘Thus in a cyst charac- 
terised by telophases of the first spermatocyte division there 
may be a few anaphases, and some metaphases, and possibly 
a few prophases. Working on this principle a cyst full of 
early prophases will sometimes contain cells undergoing pre- 
paration for prophases, and thus by linking up the informa- 
tion gleaned by a study of these individual variations a full 
series of stages can be obtained. Stress has been laid on the 
topographical relations of the cysts for reasons that will be 
apparent later. 

A comparison of Text-fig. 1 with fig.5 in Stevens’ paper (2) 
will show that in the case of Culex pungens any one gonad 
contains a greater range of spermatogenesis stages than is 
the case with C. pipiens. Only rarely in the latter case 
are spermatogonial divisions found in the pupe and old larve, 
these divisions taking place in younger larve. The usual 
distribution of stages in old larvee and pupe of C. pipiens 
is from synizesis to spermatozoa. 

Resting nuclei in the testes of young larve resemble the 
nuclei of the connective tissue (PI. 27, figs. 1, 2. and 3). In 


THE CHROMOSOME COMPLEX OF CULEX PIPIENS. 9383 


the densely staining nuclear sap the chromatin is more or less 
peripherally arranged except fora central mass (PI. 27, fig. 1). 
Resting nuclei are not found in old larve and pupe—the 
synizesis stage apparently replacing the resting stage—nor are 
they very common even in the young, as there is so much 
active growth. In Pl. 27, fig. 3, a cluster of resting nuclei is 
shown froma section of a larva stainedinthionin. The highly 
staining capacity of the cytoplasm at this stage renders the 
nuclei less conspicuous. Better differentiation is obtained in 
iron-hematoxylin (PI. 27, fig. 2). In the more posterior parts 
of the testis of young larve the cytoplasm of one cell is more 
or less clearly marked off from that of its neighbours, but at 
the head of the gonad the cytoplasm, densely provided with 
metabolic products, forms a kind of syncytium in which large 
numbers of small nuclei are indiscriminately scattered. It is 
important to notice that there are no large dark-staining 
bodies in the cytoplasm, which at this stage is uniform, these 
dark bodies being confined to older larva and pupe. 

Synizesis.—The telophases of the last spermatogonial 
divisions are particularly interesting, because they show how 
the synizesis nucleus has been formed, and, as the synizesis 
nucleus marks the commencement of preparation for the first 
meiotic divisions, the whole history of the nuclear changes 
from the last spermatogonial division to the formation of the 
spermatid can be traced. 

A figure of one such telophase is given on Pl. 27, fig. 4. 
The two daughter chromatin masses are still connected by the 
remains of the spindle, and round about these masses is a 
clear nuclear sap. The nuclear membrane is still present, 
this persistence of the membrane being characteristic of 
Culex mitosis as it is of that of many other insects. Hach 
potential daughter-nucleus is possessed of a fairly thin rim 
of cytoplasm in which are embedded dark-staining bodies. 
The final separation of these two constituents would result in 
the formation of two nuclei, each like that drawn in Pl. 27, 
fig. 5—i. e. the synizesis nucleus. In this nucleus a volu- 
minous and unstainable nuclear sap surrounds a “ coagulum” 


384 MONICA TAYLOR. 


consisting of chromatin and of a substance which is apparently 
derived from the spindle apparatus. 

This “coagulum”’ is frequently eccentric (Pl. 27, fig. 6). 
The cytoplasm of the synizesis nucleus, as was foreshadowed 
in the telophase of the mother-nucleus, is confined to a 
narrow rim. Certain dark-staining bodies differentiated by 
prolonged staining in iron-hematoxylin are to be found 
closely apposed to the nuclear membrane, the rest of the 
cytoplasm being more sparse. These dark-staining bodies 
are most readily discovered in material fixed in Benda, aceto- 
bichromate, and Flemming, and are only characteristic of the 
cytoplasm of the later stages of spermatogenesis, that of earlier 
stages being much more uniform, as has already been explained. 
Pl. 27, fig. 7, shows a cell somewhat older than that given in 
fig. 5, in which the dark mass of the synizesis nucleus has 
increased in size, the chromatin being now arranged more or 
less peripherally around a plasmosome, which stains very 
palely in the thionin sections, the chromatin being difficultly 
stainable except after prolonged treatment with iron-hema- 
toxylin. In this latter stain the chromatin is seen to form a 
fairly dense crown of closely matted fine threads around 
a central space in which les the plasmosome. 

The densely matted masses of chromatin around the plasmo- 
some now become disentangled to a great extent and occupy 
more space (Pl. 27, figs. 8 and 9). The underlying nuclear 
sap does not stain so deeply, so that the chromatin is more 
conspicuous. ‘The plasmosome stains more deeply in thionin 
than it did in the stage represented by fig. 7. Fig. 8 re- 
presents a section through, and fig. 9 an uncut nucleus of this 
stage. In the latter figure the chromatin threads are shorter 
and thicker than in fig. 8, in which the nucleus is slightly 
younger. The cytoplasm of the cell is increasing and is 
filled with small granules. It has a high staining capacity. 

Synizesis has now broken up; the chromatin threads, which 
have become thicker, lie against the nuclear membrane (PI. 
27, figs. 10 and 11). In only a few cases is it possible to 
count them, as they are so long and convoluted. Frequently 


THE CHROMOSOMH COMPLEX OF CULEX PIPIENS. 385 


the apices of the loops thicken and stain deeply. The 
plasmosome is a highly staining and conspicuous structure. 

Very often (Pl. 27, fig. 12) one chromosome thickens up 
before the others, which are still long and zig-zag, and not 
easily counted, or one part of the chromosome becomes locally 
thickened. This stage occurs in cysts along with nuclei 
containing fully formed chromosomes in late prophase. The 
threads frequently show a double character. 


Late Prophase (PI. 27, figs. 13 and 14). 


The chromosomes have now condensed sufficiently to make 
the investigation of their number easy. In every case it 1s 
possible to count three chromosomes, which are sometimes 
double, as is often the case in meiotic prophases. ‘The round 
nucleus is surrounded by a much greater quantity of cytoplasm 
than formerly. The chromosomes present great variety of 
shape. They are rod-shaped, club-shaped, dumb-bell-shaped, 
while crosses and rings are of frequent occurrence, these latter 
being formed by the partial separation of the chromosomes 
preparatory to metaphase. This preparation for metaphase 
is very often evident in the twisted character of the chromo- 
somes, three pairs of twisted chromosomes frequently appear- 
ing (compare Stevens’ fig. 14). 


Metaphase (PI. 27, figs. 15-18). 


The cell itself becomes spindle-shaped. This characteristic 
change of shape affords help in distinguishing meiotic from 
spermatogonial divisions, the metaphases and anaphases of the 
latter taking place in round-shaped cells, the shape of the cell 
being in no wise affected by the formation of the spindle. 
The spindle shape is well shown in figs. 15-18. 

A typical anaphase is illustrated by Pip 24, se. 19), Ewe 
of the chromosomes have each almost separated into daughter 
halves, the two daughter halves being merely united end to 
end. The third chromosome is not completely shown in the 
figure, part of it having been cut away. 


386 MONICA TAYLOR. 


Telophase (Pl. 27, fig. 20). 


The cell now becomes greatly elongated and the chromo- 
somes massed together. Spaces become apparent in the 
daughter chromatin masses, as shown in the left-hand mass 
in fig. 20. In the next stage these clear spaces have increased 
to such an extent that the central mass of the nucleus is free 
from chromatin, the latter occupying a peripheral position 
against the nuclear membrane (Pl. 27, fig. 21). It is inter- 
esting to compare what takes place here with what happened 
in the formation of the synizesis nucleus. In the latter case 
(Pl. 27, fig. 4) the vacuoles appeared round the daughter-mass 
of chromatin, converting this into the synizesis nucleus. In 
this case, however, the vacuoles appearing in the daughter 
mass, the chromatin is squeezed against the nuclear membrane 
(Plo 27; fig. 21). 

The cytoplasm in the cells, shown in figs. 12-20, is very 
conspicuous by its bulk, and in specimens fixed in Gilson- 
Petrunkewitch and Flemming (under certain conditions) it 
is homogeneous and deeply stainable in thionin. In some 
cells, notably those fixed in Benda, the cytoplasm is more 
like that of those cells represented in figs. 49-52, where it 
appears to be sharply differentiated into a more or less fluid 
substance and a few darkly stainable bodies. 


Second Meiotic Division. 


As has already been explained, no reticular resting stage 
follows the telophase of the first meiotic division. Vacuoles 
appearing in the daughter chromatin masses push the chro- 
matin against the nuclear membrane rendering it almost 
invisible (Pl. 27, fig. 21). Next a plasmosome, most readily 
demonstrated in sections stained in Ehrlich, appears, and the 
chromatin thickens somewhat (fig. 22). Later (fig. 23) a 
“ clock-face”’ stage similar to that which occurs in the develop- 
ment of the primary spermatocyte results. The chromosomes 
then thicken (fig. 24), though they are still in contact with 


THE CHROMOSOME COMPLEX OF CULEX PIPIENS. 387 


the nuclear membrane. Finally (fig. 25) three chromosomes 
appear in the prophase. Crosses and rings are again present, 
formed by the precocious longitudinal split, together with a 
divergence of the daughter halves. Often the chromosomes 
are so thick and close together that it is not easy to count 
them (figs. 26 and 27). 

In metaphase the spermatocyte II cells are again spindle- 
shaped (figs. 28 and 29), but they are roughly only two thirds 
the linear dimension of the primary spermatocytes, and, as 
already explained, their position in the gonad, irrespective 
of their size, would render their identification an easy matter. 
Fig. 30 shows an anaphase of this division. 


Formation of Spermatids (figs. 31 and 32). 


In the round daughter-nuclei which result from the second 
meiotic divisions the chromatin is again to be found against 
the nuclear membrane (fig. 31). The nucleus now becomes 
elongated, the chromatin, being still peripheral, gradually 
diminishes in size, and finally assumes a rod shape. 

Before going on to discuss the spermatogonial divisions, it 
may be well, in view of the difficulty of recognising the 
spermatogonial cells, to give some account of the nuclei in the 
undifferentiated gonads of young larve. 

In the following account whenever the sex is stated it must 
be remembered that it is only tentatively given. ‘The presence 
of cyst walls in the gonad has been the criterion for tentatively 
assigning the sex in the case of the male larve. 

The ovary is very richly provided with tracheal tubes, and 
at a very early stage it is possible to identify these tubes. 
The somatic cells forming the walls of the ovarian tubes can 
also be distinguished from the germ cells at a very early 
stage. By means of these criteria it is often possible to 
identify as an ovary the gonad of a very young larva. How- 
ever, in many cases it is not possible to say on which side lies 
the balance of probability. 

The general facts of mitosis in very young larve are illus- 


388 MONICA TAYLOR. 


trated by Pl. 27, figs. 33-40. Froma study of these figures it 
will be seen that the number of chromosomes in undifferen- 
tiated gonads, as well as in those of young male and female 
larvee, is 3. : 

Synizesis nuclei occur in very early stages of both male and 
female (figs. 33 and 40). The two telophases illustrated in 
figs. 37 and 38 do not suggest that the daughter-nuclei will 
pass immediately into the synizesis nuclei (cf. fig. 4). They, 
therefore, belong to early generations, spermatogonial. 

The effects of aceto-bichromate fixation are illustrated in 
figs. 59 and 40, where various metabolic constituents can be 
discovered in the cytoplasm (cf. figs. 53, 56, and 88). 


Spermatogonial Mitosis in Young Larve. 


Mitotic divisions in the fully differentiated, though imma- 
ture, gonad of fairly young larve will now be discussed. In 
such larve the hinder ends of the gonad contain the more 
advanced stages of spermatogenesis, the degree of differentia- 
tion of spermatozoa, of course, depending on the age of the 
larvee—but the anterior parts of the testis present a great 
difference in appearance from that of old larve and pupe. 
Conspicuous in such gonads is the absence of synizesis stages. 
The cellular elements are much smaller—the cytoplasm form- 
ing a syncytium, in which the small nuclei are embedded. 
This, the “ multiplication” stage of spermatogenesis, seems 
to be confined to young larvee. 

A series of division stages taken from this “multiplication” 
zone is given, the number of chromosomes being three (PI. 27, 
figs. 41-51, and figs. 2 and 3). 

The main point which emerges from a study of these nuclei 
is that the diploid number of chromosomes cannot be demon- 
strated in Culex pipiens. One cannot, therefore, use the 
number of chromosomes to distinguish spermatogonial from 
spermatocyte divisions. Still, a careful study of all the facts 
seems to show that the occurrence of a synizesis stage marks 
off the earlier divisions from those of spermatocyte I. The 


THE CHROMOSOME COMPLEX OF CULEX PIPIENS. 389 


divisions that occur in young larve are spermatogonial 
divisions, although they possess the haploid number. With 
regard to those synizesis stages in probable male larva, it 
must be remembered that some cells in quite young speci- 
mens differentiate very quickly. Hence the fact that synizesis 
stages do occur in such young creatures does not argue against 
the statement that the occurrence of a synizesis stage marks 
off spermatogonial nuclei from spermatocyte I. 


OoGENESIS. 


As already stated in the introduction, a full history of the 
facts of oogenesis can only be given when the necessary 
material has been collected. However, for present purposes 
it is only necessary to figure a few stages in order to show 
that the number of chromosomes in the germ-cells is three. 

Three chromosomes in prophase are shown in PI. 27, fig. 52, 
taken from a young larva before the “rosettes”? are well 
differentiated. The cell in which they occur clearly belongs 
to an early generation. 

A prophase drawn from a section of an imago, ten days old, 
is shown in Pl. 27, fig. 53. The cell is from a young egg- 
follicle, and as it is stained in thionin after fixation in Gilson- 
Petrunkewitch, the cytoplasm is very densely stained. 

A metaphase showing six chromosomes is given in Pl. 27, 
fig. 54. 

The drawing in PI. 27, fig. 55 is taken from an old larva. 
It shows a prophase where the number of chromosomes is 
three. The cytoplasm is abundant. * 

Pl. 27, fig. 56, taken from an egg-follicle of a captive-reared 
imago, illustrates the oldest stage obtained. The nucleus, 
with its deep-staining plasmosome, and its fine, well-dis- 
tributed reticulum of chromatin, forms a striking contrast to 
the deep-staining, voluminous cytoplasm that surrounds it. 


Somatic Mrrosts. 


Although there is much indirect evidence to show that the 
divisions described as spermatogonial are really divisions of 


390 MONICA TAYLOR. 


the earlier generations, in spite of the haploid and not diploid 
number of chromosomes, yet, in view of the possibility that 
they might be precocious spermatocyte I divisions, it has 
been thought advisable to work out the somatic mitosis of 
C. pipiens very thoroughly. 


Mitosis in the Somatic Tissues of the Ovary 
(Text-figs. 2 and 3). 


The somatic cells of the ovarian tubes of the ovary can 
be distinguished from the reproductive cells in very young 
larvee by their minute size. The developing ovarian tubes 
look like rosettes in sections cut from older larve, the large 
central cells of the rosette being reproductive cells, the peri- 
pheral cells, much smaller in size, being somatic. Later on 
groups of from four to eight of these reproductive cells 
become enclosed in an epithelium—the egg-follicle epithelium 
—also somatic in character. The number of these egg-follicles 
increases with the age of the creature, the ovary having 
assumed its definitive arrangement in late pupal life. The 
exceedingly thin walls of the ovarian tubes, the nuclei of 
which are very small, are greatly distended by the large egg- 
follicles, which are made up of one egg-cell surrounded by a 
varying number of nurse-cells. 

The ultimate fate of egg- and nurse-cells has still to be 
worked out. The somatic tissue of the ovary—whether the 
epithelium of the tubes, or the egg-follicle epithelium—is a 
prolific source of mitotic figures. In all cases the number of 
chromosomes is three. 

A cell belonging to an egg-follicle in telophase is shown in 
Pl. 27, fig.57, and near it is shown a reproductive cell (fig. 58), 
which also brings out the difference in size between the 
somatic and reproductive cell. 

Pl. 27, figs. 59 and 60, also show three chromosomes in 
typically somatic anaphase. 

Innumerable figures of prophases could be drawn from the 
preparations, it being characteristic of Culex pipiens that 


THE CHROMOSOME COMPLEX OF CULEX PIPIENS. 391 


metaphases and anaphases are comparatively difficult to find, 
prophases being much more abundant. 

In the case of all the cells of the ovary the number of the 
chromosomes is three. 


Somatic Mitosis (General). 


Somatic mitoses, apart from those in the ovarian tissue, are 
by no means easily found. They do not seem to be confined 
to any particular period of larval or pupal life, or to take 
place at any fixed hour of the day or night. 

The process, as observed in nerve-cells, is illustrated by 
Pl. 27, figs. 61-65, and Pl. 28, figs. 66-68. 

It will be seen from these figures that there is a great 
resemblance between the reproductive cells, at certain stages 
of development, and the somatic cells in the nerve ganglia. 

Similar likenesses between the body-wall cells and those of 
the gonad could also be demonstrated. Thus a very well- 
marked synizesis stage is characteristic, not only of the later 
stages of spermatogenesis and of the reproductive cells of the 
ovary, but also of the cells of the nervous system, and the 
gradual breaking up of the synizesis resembles, in the main, 
that which takes place in the gonad. Hence it would appear 
that synizesis has not the significance in the spermatogenesis 
of Culex that it is commonly believed to have in other 
creatures, since this phenomenon is not confined to repro- 
ductive cells. 

An investigation of the tracheal tube-cells shows that the 
number of chromosomes is three. Very frequently the 
daughter halves of the split chromosome can be seen in late 
prophase (PI. 28, figs. 69-71). 

Evidence as to the number of chromosomes derived from a 
study of the body-wall cells (illustrated by PI. 28, figs. 72-79) 
confirms the results already obtained elsewhere, while in the 
undifferentiated somatic cells in larvee just hatched (PI. 28, 
figs. 80 and 81) the number is again three. 

The alimentary canal wall splits into two layers during 


392 MONICA TAYLOR. 


pupal life, the inner layer undergoing disintegration (Hurst 
(1)). Some of the mitoses that occur in connection with this 
process are illustrated in Pl. 28, figs. 82-89. The number of 
chromosomes is three, and in some cases the precocious ten- 
dency of the chromosomes to divide in prophase for metaphase 
is again evident. 

Mitosis in Malpighian tubule cells is figured in Pl. 28, figs. 
90-95, while fig. 96 shows an equatorial plate from a muscle- 
forming cell. In all cases the somatic number of chromosomes 


is three. 


Discussion. 


In C. pipiens, as has been shown, there appear to be two 
maturation divisions, though no reduction of chromosomes can 
be demonstrated. This is contrary to general experience, for, 
as is well known, when, in any organism, the ripe germ-cell 
has the same number of chromosomes as the somatic tissues, 
one of the meiotic divisions is commonly omitted. It must 
be remembered, however, that while two divisions undoubtedly 
follow the synizesis stage in C. pipiens, the fact that they 
follow synizesis is the only one which has led to their separa- 
tion off from the apparently similar earlier divisions, and to 
_their being described as meiotic rather than spermatogonial. 
However, as nuclei closely resembling synizetic nuclei can be 
found in the somatic tissues of this creature, it is, therefore, 
possible that the synizesis nuclei in the testis have no real 
value in diagnosing the beginning of the meiotic phase, and 
that they merely represent a stage of inactivity. The rapid 
divisions in the “ multiplication ”’ zone result in the formation 
of large numbers of nuclei which, while they are awaiting 
differentiation into spermatozoa, remain as synizesis’ nuclei. 
When the time comes for this differentiation the synizesis 
nucleus begins to be active, and the stages in this awakening 
are analogous to the formation of spermatocyte I cells. 

From the foregoing account, it will be seen that the para- 
syndesis for each cell generation which Miss Stevens described 
for C. pungens cannot be demonstrated for C. pipiens. 


THE CHROMOSOME COMPLEX OF CULEX PIPIENS. 3893 


No figures indicating the presence of six chromosomes are to 
be found which cannot readily be interpreted as three chromo- 
somes precociously split for metaphase. For example, con- 
ditions like those Miss Stevens gives in her figs. 1 and 2 
(oogonia showing three pairs of chromosomes on equatorial 
plate) and in figs. 8, 9, and 10 (spermatogonial cells showing 
all three pairs in late prophase), in thelight of other evidence, 
must, when they occur in C. pipiens, be described as three 
chromosomes already divided for metaphase. The conditions 
of C. pungens, however, offer a suggestion as to how the 
haploid number of chromosomes in C. pipiens may have been 
derived. The permanent fusion of the paternal and maternal 
members of the’ pair, i. e. the conversion of parasyndesis into 
actual fusion, would result in the formation of three out of six 
chromosomes. 

Miss Stevens states that in C. pungens the intimate re- 
lationship of the two conjugants persists from one cell genera- 
tion to the next, the pairing taking place in telophase, and 
persisting until the metaphase of the next mitosis: From this 
it would seem that the conjugating chromosomes are only 
“unfused ” in metaphase. In the case of C. pipiens the 
pairsare fused throughout the whole mitosis, hence the haploid 
number. 

On the other hand, it is quite possible to give a different 
interpretation of Miss Stevens’ figures of parasyndesis from 
the one she offers. It is significant to note that she gives no 
figures in support of her statement that each of the six chromo- 
somes (i.e. each member of the three pairs of conjugating 
chromosomes) found in the oogonial and spermatogonial 
generations divides longitudinally. She merely states the fact 
that they do so. Unless this division can be demonstrated, it 
would seem as though the so-called conjugating chromosomes 
were merely the daughter-halves of a precociously split 
chromosome, as is the case in C. pipiens. 

An alternative suggestion, therefore, as to the chromosome 
complex of Culex pipiens and pungens is, that the somatic 
number is the same as that of the mature gamete, being three 


voL. 60, PART 3.—NEW SERIES. 28 


394 MONICA TAYLOR. 


an each case. This alternative would seem to involve the 
non-participation of one of the gametic nuclei in development. 

Whether this is the case, or whether the homologous 
chromosomes are temporally united in each cell-generation of 
C. pungens, and permanently so in C. pipiens, can only be 
settled by an examination of the process of fertilisation, which 
I hope to undertake in the near future. 

I am greatly indebted to Dr. Agar for much valuable 
‘criticism ; to many friends who have assisted me in collecting 
material; to Mr. P. Jamieson for cutting the more important 
sections ; and to Professor Graham Kerr for his sympathetic 
encouragement. 


SUMMARY. 


(1) The somatic number of chromosomes is three, both-in the 
male and female of Culex pipiens. 

(2) The number of chromosomes in the spermatogonia as 
well as in the primary and secondary spermatocytes and 
‘spermatids is three. 

(3) The spermatogonial cells are not characterised by a 
‘synizesis stage, which latter stage marks off the spermatogonial 
from the spermatocyte I stage. 

(4) The nuclear membrane persists throughout mitosis. 

(5) The synizesis stage represents an inactive phase of the 
nucleus in spermatogenesis. 

(6) A synizesis stage occurs in somatic nuclei. 


LITERATURE. 
1. 1890. Hurst, C. Herbert.—‘‘ Post-Embryonic Development of Gnat,” 
‘Trans. Liverpool Biol. Soc.,’ vol. iv. 


2. 1910. Stevens, N. M.—‘“ The Chromosomes in the Géaee cells of 
Culex,” ‘ Journ, Exper. Zool.,’ viii. 


THE CHROMOSOME COMPLEX OF CULEX PIPTENS. 3995 


EXPLANATION OF PLATES 27 anp 28, 


Illustrating Miss Monica Taylor’s paper on “ The Chromosome 
Complex of Culex pipiens.” 


[ All figures (except 3 and 6) were drawn with the Abbé camera under 
Leitz ;, oil-immersion objective and Zeiss compensating ocular 12, 
giving a magnification of 3300 diameters. This has been reduced 4, 
i.e. to a magnification of 2200. Figs. 3 and 6 are drawn to a magnifi- 


cation of 1900, giving a final magnification of about 1270. | 


PIA 27. 


Fig. 1.—Resting nucleus in connective-tissue. 

Fig. 2.—Resting nucleus from gonad of young ¢ larva. 

Fig. 3.—Group of nuclei drawn to same scale as fig. 6 from young 
& larva; thionin stain. 

Fig. 4.—Telophase of spermatogonial cell from head of testis of pupa. 


Figs. 5-32 are taken from pupal testes. 


Fig. 5.—A synizesis nucleus from head of testis of pupa. 

Fig. 6.—A group of synizesis nuclei from a pupa fixed in Benda, and 
‘deeply stained. 

Fig. 7—A nucleus showing early stage in the breaking up of 
synizesis. 

Fig. 8.—F rom a section 4 » thick, showing later stage in the breaking 
up of synizesis. 

Fig. 9.—From a section 8 » thick, showing nucleus slightly older than 
in fig. 8. 

Fig. 10.—View of nucleus from one pole, showing the chromatin close 
to the nuclear membrane. 

Fig. 11.—Optical section through a nucleus which at this stage is 
very characteristically like a “ clock-face.” 

Fig. 12.—Nucleus showing that the thickening up of the chromo- 
somes is not always uniform. 

Fig. 13.—Advanced prophase. 

Fig. 14.—Advanced prophase. 

Fig. 15.—Very early metaphase, showing the spindle-shape of the cell 
before the mitotic spindle has been organised. 


396 MONICA TAYLOR. 


Fig. 16.—A more elongated spindle-shaped cell. Early metaphase. 
g. 17.—Metaphase, with indistinguishable chromosomes. 
Fig. 18.—Metaphase in which one of the chromosomes has divided. 
Fig. 19.—Later anaphase. Only one of the three dividing chromo- 
somes is completely shown in the figure. 
Fig. 20.—Telophase. 
Fig. 21.—Interkinetiec nucleus. 
Fig. 22.—Interkinetic nucleus stained in Ehrlich, showing a plasmo- 


Fig. 25.—* Clock-face” stage for spermatocyte II nucleus. 

Fig. 24.—A nucleus of spermatocyte II, in which chromosomes are 
becoming visible. 

Fig. 25.—A nucleus of spermatocyte II, showing three chromosomes. 

Fig. 26.—Advanced prophase. 

Fig. 27.—Later prophase; the shape of the cell is just beginning to 
change. 

Fig. 28.—Early metaphase. 

Fig. 29.—Metaphase ; three chromosomes on equatorial plate. 

Fig. 30.—Anaphase. 
. 31.—Newly formed spermatid. 


oe 
5 
Fig. 52.—Immature spermatid. 


Figs. 33-40 are drawn from gonads of very young larve. 


Fig. 35.—A synizesis ina ¢ (?) larva. 


Fig. 34.—Prophase from a young ¢ (?) larva. 
Fig. 35.—Metaphase from a young ¢ (?) larva. 
Fig. 36.—Metaphase from gonad of larva. 


° de 


Fig. 37.—Anaphase from gonad of larva. 

Fig. 38.—Anaphase from ¢ (?) larva! 

Fig. 39.—Acetic bichromate and thionin preparation. 9 (?) larva. 

Fig. 40.—Acetic bichromate and iron-hematoxylin preparation. 
2 (?) larva. 


Figs. 41, 42.—Prophases, spermatogonial; the number of chromo- 
somes is three. 

Fig. 45.—Equatorial plate, showing three chromosomes; a young 
¢ larva, spermatogonial. 

Fig. 44.—Metaphase, spermatogonial, three chromosomes. 

Figs. 45, 46.—Late spermatogonial anaphases from young ¢ larva. 


THE CHROMOSOME COMPLEX OF CULEX PIPIENS. 397 


Figs. 47—51 are from head of testis of pupa. 

Fig. 47.—Prophase, spermatogonial, from head of gonad of pupa 
(cf. Stevens [fig. 10 (2) |). There are two long chromosomes which have 
split longitudinally, and one smaller one not yet divided into daughter 
halves. 

Fig. 48 shows the threads of the spindle just beginning to form 
inside the nuclear membrane, the chromosomes being too massed 
together to be counted. 

Fig. 49.—Metaphase; an equatorial plate where the three chromosomes 
are just separating into their daughter-constituents. 

Fig. 50.—Later metaphase, six chromosomes being easily counted. 

Fig. 51—EHarly anaphase. Note the round-shaped nucleus, and the 
somatic-like character of the V-shaped daughter-chromosomes. 


Figs. 52-56 are taken from ovaries. 


Fig. 52.—Prophase from young @ larva. 
Fig. 53.—Prophase from 2? imago. 


4.—Metaphase from ? larva. 
Fig. ? 
Fig. ¢ 


5.—Prophase from old 9 larva. 


= 
go 
(re HSne Sn.e Lng (ne 


6.—From 2 imago, two months old. 


Figs. 57, 58.—Ege-follicle. Epithelium cell with nucleus in late 
anaphase (57), and nurse or egg cell drawn to same scale (58). (2 
imago section.) 


Figs. 59, 60.—Ege-follicle epithelium cells in late anaphase, showing 
three chromosomes. (From a section of a 2? Imago.) 


Figs. 61-65 from nerve-ganglia. 

Fig. 61—Synizesis nuclei deeply stained in Ehrlich from old 
larva (ef. figs. 5, 6, and 40). 

Fig. 62.—Nerve-cell showing six blocks of chromatin in the “coa- 
gulum.” Old @ larva. 

Figs. 63 and 64.—Nerve-cells showing the breaking up of synizesis. 
Old 9 larva. 

Fig. 65.—EKarly prophase from ¢ pupa (cf. fig. 10). 


PLATE 28. 
Fig. 66.—Early prophase of a nerve-cell from a larva just hatched. 
Compare PI. 27, fig. 11. 


Fig. 67.—Prophase of a nucleus of a nerve-cell from a larva just 
hatched, Three chromosomes showing precocious splitting. 


398 MONICA TAYLOR. 


Fig. 68.—Prophase of a nucleus of anerve-cell from a young d larva 
Fig. 69.—Three chromosomes in the nucleus of a tracheal tube-cell 
from a young ¢ larva. Compare this with figs. 1, 9, and 10 of ‘The 
Germ Cells of Culex,’ Stevens (2). 
Fig. 70.—Tracheal tube-cell from young  (?) larva. 
Fig. 71.—Ditto. 
Figs. 72-79 are drawn from body-wall cells. 

Fig. 72.—Early prophase of a nucleus from a ? imago. 

Fig. 73.—Three chromosomes in prophase from a $ larva. 

Fig. 74.—One short and two long chromosomes. Prophase of nucleus- 
from a young ¢ larva. Note indication of two daughter-halves in one 
of the chromosomes. 

Fig. 75.—Three chromosomes in advanced prophase from ? larva. 

Fig. 76.—Equatorial plate. Side view showing three chromosomes ; 
one of them is cross-shaped. 

Fig. 77.—Anaphase from a @ larva. 

Figs. 78 and 79.—Telophases from @ larve. 

Figs. 80 and 81.—Undifferentiated cells of a larva just hatched, 
showing three chromosomes in prophase and late metaphase. 


Figs. 82-89 are drawn from cells in alimentary canal wall. 

Figs. 82, 83, 84.—Prophases from cells in alimentary canal wall of 
9 larve. 

Figs. 85, 86.—Prophases from cells in alimentary canal wall of a 9 
imago, ten days old. In fig. 85 one of the chromosomes has divided 
longitudinally. 

Fig. 87.—Prophase from alimentary canal wall of 9 larva. 

Fig. 88.—Metaphase from cell of same larva. 

Fig. 89.—An equatorial plate showing three chromosomes from wall 
of intestine. 


Figs. 90-95.—Malpighian tube-cells. 


Fig. 90.—From young d larva. 

Fig. 91.—From 2 larva. 

Fig. 92.—From young ¢ larva. 

Fig. 93.—From young ¢ larva. 

Fig. 94.—From larva. 

Fig. 95—From @ larva. Note cross-shaped chromosome. 

Fig. 96.—Equatorial plate showing three chromosomes; two of them 


ringed, from muscle-cell of young larva. 


STUDIES ON AVIAN HH MOPROTOZOA, 399 


Studies on Avian Hemoprotozoa: No. III.— 
Observations on the Development of Try- 
panosoma noctue (of the Little Owl) in 
Culex pipiens; with Remarks on the Other 
Parasites occurring.’ 


By 


H. M. Woodcock, D.Sc.Lond., 
Assistant to the University Professor of Protozoology. 


With Plates 29—31 and 1 Text-figure. 


CONTENTS. 
PAGE 
PREFACE P 5 ate) 
GENERAL hewamrn OF THE Senna, WorRK . . 400 
DESCRIPTION OF THE PARASITES: 
(i) Trypanosoma noctue . : . 406 
Comparison with the cultural dev elonrent of T. frin- 
gillinarum . . 409 
(ii) The ookinetes of Eialthontdi aan odie and Leuco- 
cytozoon ziemanni . : : . 421 
(iii) The “resting Flagellates ” . ; . 423 
GENERAL CONCLUSION REGARDING THE | Communes OR 
OTHERWISE OF THE DIFFERENT PARASITES . . 425 
PREFACE. 


I propose in the present paper to conclude the account of 
the work on the parasites of the Little Owl (Athene noctua) 
which I carried out at Rovigno during the spring and early 


‘For No. II of these Studies, “On the Trypanosome of the Little 
Owl,” etc., vide ‘Quart. Journ. Mier. Sci.,’ vol. 57, 1911, p. 141. 


400 H. M. WOODCOCK. 


summer of 1909. I have delayed publishing up to the 
present my observations on the developmental phases of the 
Trypanosome in the mosquito, as I hoped to be able, before 
now, to obtain the corresponding phases, and, indeed, the 
transmission back again to the bird, of Trypanosoma 
fringillinarum, here in England, to complete the work. 
This latter research, however, progresses, unfortunately, very 
slowly, so that I think it useful to publish my earlier observa- 
tions without waiting longer, more especially as the only 
other worker who has written anything of late on this subject, 
namely, Mayer (2), in his account of the parasites of another 
owl (Syrnium aluco), has upheld the view that the Flagel- 
lates occurring in mosquitoes which have fed on an owl are 
developed from Halteridia present in the blood of the bird. 
In my “ Notes on Sporozoa, No. IV ” (14), dealing with the 
nuclear structure of Halteridium and Leucocytozoon, I 
think I have shown clearly that, considered from the stand- 
point of these parasites, everything is against such a connec- 
tion of either with a Trypanosome. And the evidence which 
I have obtained from the Trypanosome side of the question is 
equally negative, and does not bear out Mayer’s contention 
in the slightest degree. 


GENERAL ACCOUNT OF THE EXPERIMENTAL WORK. 


The time at my disposal for experimental work with Culex 
was only very short—about the last three weeks of June. 
Earlier than this the bred-out females would not take blood, 
and at the beginning of July I was unfortunately obliged to 
leave Rovigno. Five owls were used, to which it will be 
convenient to refer by their numbers, viz., 15, 16, 19, 22, and 
23. The first two were quite free from any infection; No. 22 
had only Leucocytozoon, while Nos. 19 and 23 both had 
Halteridium noctue, Leucocytozoon ziemanni, and 
Trypanosoma noctue. As regards Owl 19, the Trypano- 
some was for some reason or other excessively scanty, and 
was not, I believe, present at all in the general circulation 


STUDIES ON AVIAN HAMOPROTOZOA. 401 


(cf. Study No. II). The Halteridial infection of Owl 19 was 
quite typical, the parasites, which, of course, were in the 
usual form of male and female gametocytes, being fairly 
numerous, and many of them full-grown and ripe for liberation 
from the host-cells. On the other hand, the Halteridial 
infection of No. 23 was not typical. In this owl nearly every 
red blood-cell was infected, usually with several (four or 
more) small forms, which in many cases, as they had increased 
in size, had united together into a kind of common plasmodium. 
This condition has been already described by me elsewhere 
(13). I observed hardly any full-grown, normal gametocytes 
in the blood of this bird. Owls 22 and 23 only arrived on 
June 12th. 

Karly in June, finding that my bred-out females would not 
yet bite, I left Owl 19 one evening ina cage into which I had 
put half-a-dozen or so “wild” females, caught in the same 
little outhouse, used as a ‘ dark-room,” where Schaudinn 
himself had caught many in time past. Before midnight two 
of these had fed. I may note here that the favourite place 
for feeding of the mosquitoes was the fleshy pad just above 
the bird’s nostrils. One of the two females was examined 
atter about thirty-six hours had elapsed, this being the period 
when, according to Schaudinn, the ookinetes become trans- 
formed into Trypanosomes. Its stomach contained a number 
of fully-developed ookinetes, i. e. which had lost all the 
pigment. he majority had the characteristic curved form, 
but did not appear at all active. A few, however, showed a 
certain amount of activity, which consisted in tending to 
straighten out, and again recurve, the body, either slowly or 
now and again spasmodically. I never observed any marked 
forward progression of the ookinetes. 

Besides the ookinetes certain other bodies were found to 
occur, very scantily, in the preparation. These elements 
were more elongated and spindle-shaped, somewhat resembling 
an Indian club in form, one extremity tapering finely, the 
other being rounded off. Further, some of them were very 
slightly curved or crescentic. Apart from the difference in 


402 H. M. WOODCOCK. 


shape, the general appearance of these bodies, observed 
living, was not at all unlike that of the ookinetes. They 
were quite non-motile. While I was in the act of observing 
one, and wondering whether it was a later stage in the 
development of an ookinete, it gave one or two very slight, 
jerky movements, and before I had fully assured myself that 
these really represented active, voluntary motion on the part. 
of the parasite, to my great surprise it had developed a 
flagellum. I thought I had just an indication of the tapering 
end of the body beginning to lengthen, but more than this I 
did not see. One moment the parasite had no flagellum, an 
instant later it had a fully-developed free flagellum, about 
three-quarters as long as its own body. ‘The process must. 
have been exactly comparable, in short, to that which I have 
since found to occur in Leptomonas (“Crithidia”) 
fasciculata (vide 15). I hurriedly brought a colleague, 
Dr. Reichenow, then also working at Rovigno, to look, and 
he, in turn, saw the process repeated in the case of another 
individual. ‘hese were the only two instances in which we 
saw the development of the flagellum, and we only found 
one other flagellated form in the preparation. We both 
carefully examined several of the other (non-motile) elements, 
and satisfied ourselves that they had no sign whatever of 
aflagellum. IfI may be permitted the personal reminiscence, 
I well remember how, in the excitement of the moment, 
we were both of us firmly persuaded that we had seen the 
most important stage in the transformation from an ookinete 
into a ‘l'rypanosome, as it had been described by Schaudinn. 
This confidence did not endure, however, for many days. I 
continued to watch certain ookinetes during the afternoon, 
and felt very disappointed that I could not see any indication 
whatever of a typical ookinete passing into one of the fusiform 
bodies. After keeping the preparation under observation for 
about three hours I removed the coverslip and made smears, 
which were fixed and stained. 

Most unfortunately, Owl 19 was taken ill and died on the 
following day, and for some days I had no owl infected with 


STUDIES ON AVIAN HZ MOPROUTOZOA. 403. 


Trypanosomes. During this interval I fed several caught 
mosquitoes on an uninfected bird (either 15 or 16), and 
these were examined at periods of from thirty-six to fifty- 
eight hours after being fed. In two cases, elongated, fusi- 
form elements, perfectly similar to those above-mentioned, 
were found; but none of them was seen to become active or 
develop a flagellum. In another female, examined about 
fifty-four hours after being fed, numerous active Flagellates 
were observed in the stomach; these differed considerably in 
appearance from the “resting Flagellates,”’ but, on the other 
hand, agreed closely with the characteristic developmental 
forms described below, and I have not the least doubt that 
they also belonged to the life-cycle of some Avian Trypano- 
some. It is important to note that no ookinetes. were seen 1n 
any of these mosquitoes. These observations showed not 
only that the fusiform resting Flagellates might probably have 
another origin than from the ookinetes of Halteridium, but 
also that it was essential to use only bred-out females, in 
order to follow the course of development taken by the blood- 
parasites in the Culex; hence, from this time onwards only 
such individuals were used. 

On June 12th Owl 23 arrived, and during the night of the 
15th—16th I found Trypanosomes in the peripheral circulation ; 
the parasites were of the characteristic fusiform, rather stout 
type (Pl. 31, fig. a), described by Minchin and myself (I. c.). 
The Trypanosomes were not at all infrequent—for Avian 
Trypanosomes, it must be remembered—in the peripheral blood 
at this time,! and were found also on other occasions. In both 
of the two first mosquitoes which were examined after being 
fed on this bird, after intervals of about thirty-four and forty 
hours respectively, numerous active Flagellates were found in 
the stomach. Digestion was proceeding normally and was 
about half accomplished, or rather more. (I may state here 
that females which had taken blood were always kept at a tem- 


' In any living preparation, consisting of a small drop of blood spread 
out into a thin layer under a coverslip seven-eighths of an inch square, 
there would be one or two Trypanosomes. 


404, H. M. WOODCOCK. 


perature of from 25°-27° C., at which temperature digestion 
took three to four days to be completed.) The develop- 
ment of the Trypanosomes appeared to be at about the same 
stage in both the females, and in both similar phases were 
observed. 

Ookinetes of Halteridium were also found, but in both 
mosquitoes they were few in number—quite scarce, in fact, 
when compared with the number present in the first female 
examined, which had fed on Owl 19 (cf. above). This was 
readily to be understood, bearing in mind the different con- 
dition of the Halteridial infection in the two birds; in spite 
of the very strong infection of Owl 23, there were nothing like 
so many full-grown, ripe gametocytes asin Owl 19. A living 
preparation made from one of these two stomachs was kept 
under observation for some time. Three ookinetes in different 
fields, all of which had lost their pigment grains, were noted 
at intervals during two hours, but none of them showed any 
change in form or the slightest indication of any development 
into one of the fusiform bodies or into a flagellate condition. 
The preparation was again looked at two hours later, with 
the same result in the case of two of the ookinetes; the third 
could not be found. By this time most of the Flagellates 
seemed to be dead—at all events, only two or three could be 
observed, and these were very sluggish. 

Altogether, twenty-six female Culex pipiens which had 
fed on Owl 23 were examined, after intervals varying from 
about twelve to eighty hours. Flagellates were found in 
twelve of these, i.e. in about 46 per cent.; sometimes they 
were numerous, in other cases only few were seen. On the 
other hand, out of thirty-two females fed on one of the other 
owls (Nos. 15, 16, 22), none of which was infected with 
Trypanosomes, the stomachs of which were carefully examined 
after different intervals (thirty-six to fifty-four hours), in not 
a single case were the parasites found! T'wo of these birds 
were quite free from any Heemoprotozoan infection ; Owl 22, 
however, had a fairly strong infection of Leucocytozoon 
ziemanni. Seven of the thirty-two mosquitoes examined 


STUDIES ON AVIAN HU®MOPROTOZOA. 4.05 


had fed on this latter bird,! and in four of them a few of the 
large ookinetes of Leucocytozoon were found. In form 
and appearance these were very similar to the ookinetes of 
Halteridium, but they were considerably larger. Those | 
observed were quite motionless, and did not change at all in 
shape. 

The fifty-eight mosquitoes examined were barely half the 
total number (about 120) which fed on blood. The mortality 
amongst these newly bred-out imagines was high, and it 
appeared to make little difference whether they took blood or 
the sugar-water, banana and prune juice with which they 
were supplied. More than a quarter of those which fed on 
blood died during the course of digestion. During the short 
time at my disposal I had, therefore, only very limited 
material for transmission-experiments. During the last fort- 
night between fifteen and twenty females which had fed on 
Owl 23 and successfully completed digestion, were given the 
opportunity of feeding on an uninfected bird (either 15 or 
16). To my very great disappointment, however, not one of 
these could be induced to feed again. Some of them drank a 
little water, or partook of the food-supply which was placed 
in the same cage for a few hours during the day-time for the 
males to feed upon; nevertheless, many of them gradually 
died off during this period. I was loath to sacrifice any more 
for examination, hoping to the very morning of my departure 
from Rovigno that one or more would bite again and give me 
the chance of seeing whether one of the uninfected birds 
would become infected. Unfortunately, the endeavour did 
not succeed. Owls 15 and 16 accompanied me back to 
England and lived for many months—much longer than a 
single bird did at Rovigno—but neither of them ever showed 
any parasites at all. 

While restricting myself to bred-out females for the experi- 
mental work, incidentally I examined a few more caught 
“ wild” mosquitoes. None of them was infected with active 
Flagellates (i. e. Trypanosome developmental forms), but again 

1 This owl died on June 22nd, so that I had it only ten days. 


406 H. M. WOODCOOK. 


two or three contained the peculiar, spindle-like resting forms, 
above noted; in no further instance, however, did I see one 
develop a flagellum. I have since wished that I had been 
able to devote more time to the study of this parasite, and to 
ascertain, for example, whether it occurred in male mosquitoes 
also. But I was intent on proving the origin of the Flagellates 
which developed in the blood-fed females, and in transmitting 
them back to the birds, if possible, and this took every minute 
of my time, as I had no assistance whatever and had every- 
thing to do myself. 


DESCRIPTION OF THE PARASITES, AS THEY WERE FOUND IN THE 
MosqQtitTors. 


(i). Trypanosoma noctuae. 


I pass on now to describe the developmental phases of the 
Avian parasites in the female Culex pipiens, and begin 
with those of Trypanosome noctunae. Considering the 
Flagellates, first of all, as they were seen in life, in mosquitoes 
examined about thirty-six hours after being fed on an infected 
owl, the most striking form, which at once held my attention 
on first seeing it, was a very long, extremely slender type, 
which progressed rapidly, by means of its flagellum and the 
undulating membrane along the anterior part of the body. 
The membrane appeared to be very narrow along the middle 
region of the body, and on account of this fact and the active 
movements of these individuals, I could not determine exactly 
where it ended, or whether this type was trypanomonad 
(crithidial), or trypaniform. In fixed and stained prepa- 
rations, however, it is seen to be distinctly and invariably 
trypaniform (figs. 13-21). Nevertheless, these attenuated 
trypanosomes were entirely different .in appearance from an 
ordinary, slender, elongated blood-form of Trypanosome, e. g. 
a piscine type such as T. granulosum. While the anterior 
part of the body, where the membrane was conspicuous, was 
‘sinuous and fiexible, the hinder part, as a rule, nearly half 


STUDIES ON AVIAN HAMOPROTOZOA., 407 


the entire length or more, was held quite stiffly and did not 
appear to be actively flexible at all. Frequently it was 
practically straight, but in some individuals:it was quite 
curved round, like a crook (cf. figs. 19-21); this posterior 
part would retain this shape, unaltered, even while the 
parasite was moving rapidly forwards. From a comparison 
of stained preparations it is evident that this portion of the 
body represents a prolonged extension of the cytoplasm 
behind (posterior to) the kinetonucleus. In view of my 
observations on the living parasites, I regard this cytoplasmic 
“‘tail ” as differing from the anterior half of the body in 
lacking anything of the nature of myonemes, and conse- 
quently any ability to bend or twist of itself. I consider 
that it is only, as it were, passively flexible, and that any 
curving or bending is produced mechanically, as the result of 
contact with the blood-cells and other elements among which 
the parasite happens to be working its way. What function 
this remarkable development serves, I was not able to 
ascertain. 

The above type of individual constitutes a fair proportion 
of the total number of flagellates present, even at this some- 
what early stage of the development. ‘The other types of 
parasite seen were for the most part relatively short trypano- 
monad (crithidial) forms and individuals representing every 
possible transition between such and the extremely attenuated 
forms. As is shown by the fixed and stained preparations 
made of stomachs at about this period, the majority of these 
intermediate forms are really trypaniform, i.e., the kineto- 
nucleus is on the aflagellar side of (posterior to) the tropho- 
nucleus. The shorter, more typically trypanomonad indi- 
viduals resembled the commonly occurring crithidial forms 
which develop in cultures of an Avian trypanosome, such as 
I have described in the case of T.fringillinarum. All 
these forms had what is usually distinguished as the crithidial 
type of movement. Progression was in a slightly zig-zag 
manner, the flagellum and the anterior part of the body 
(corresponding to the position of the undulating membrane) 


408 H. M. WOODCOCK. 


vibrating actively. The movement of these forms was not | 
nearly as rapid as that of the long, slender individuals. 
Lastly, other forms noted were short and pyriform, and 
moved jerkily, not displacing themselves to any extent; 
these were infrequent. Notwithstanding the great increase 
in number of the parasites which must have taken place 
since the blood entered the stomach, I found scarcely any 
individuals actually dividing. In one or two cases I saw try- 
panomonad individuals with two flagella, and in one instance 
a pair of such forms still connected together; in another 
instance which I noted, division was markedly unequal, a 
short, pyriform individual becoming separated off from a 
distinctly larger, broader, club-shaped form. 

At a later period of the development the long, attenuated 
forms predominate more and more, until in females examined 
fifty-five hours or subsequently after being fed, the stomach 
contained apparently only such forms ; and this is borne out by 
the study of fixed and stained preparations made of stomachs 
of this period or later. 

All my observations on the parasites relate to the stomach- 
phases of the life-cycle which occur between thirty-two and 
about seventy-six hours after the mosquito has fed. I 
examined four females from twelve to eighteen hours after 
feeding on Owl 23, but I could not find any Trypanosomes at 
this early period. It would be necessary to examine many 
individuals to find the earliest changes in the parasites, 
because, even if the development is proceeding all right, the 
Trypanosomes have not yet had time to multiply and give rise 
to any considerable number of parasites; and it must be 
remembered that only very few Trypanosomes are taken up 
by the mosquito from the blood to start with. Neither did I 
find any phases of the parasites in the intestine ; but this did 
not surprise me, because nearly .all the mosquitoes were 
examined, at any rate, some time before the stomach was 
quite empty. While digestion is still going on, the stomach 
is most certainly the principal, if not the only situation in 
which the Trypanosomes occur. If I had been able to 


STUDIES ON AVIAN HAMOPROTOZOA. 4.09 


examine a sufficient number of females after digestion had 
been completed, or after they had had another meal of blood, 
the parasites might have been found in the intestine ; I shall 
discuss this point later on. My observations are, | am aware, 
only incomplete ; but they were the fullest I was able to make 
in the circumstances, and bearing in mind the chief objects 
on which I concentrated my attention during the short time 
at my disposal, namely, to determine whether the 'Trypano- 
some-phases in the mosquito were derived from the Halteridial 
parasites or not, and to bring about, if possible, the trans- 
mission of the Trypanosomes back again to the owl. 

The Trypanosomes in Permanent Preparations; 
Comparison with the Developmental Stages found 
in Cultures.—lI think, however, by comparing the various 
forms which I did obtain in the mosquito, with the develop- 
ment which I have found to take place in cultures, in the case 
of another Avian Trypanosome, T. fringillinarum, that a 
general idea can be arrived at of the main course of the natural 
development in the stomach of the insectan host; as already 
indicated, the chief gap and element of uncertainty relates to 
the earliest changes undergone by the parasites. 

I have no hesitation in making use of the cultural develop- 
ment for this purpose, because practically all the different 
forms occur in both cases. The reason it is helpful is because, 
in the culture, the development continues over a much Jonger 
period owing to the fact that there is no absorption of the 
medium, such as occurs in the insect’s stomach, and therefore 
the intermediate forms are met with abundantly and stages 
in division are frequent, the latter being of much assistance 
in determining the sequence of the developmental changes. 
In the Culex, on the other hand, the course of digestion is 
comparatively rapid and completed in three to three and a- 
half days, by which time the stomach is empty of blood. It 
is undoubtedly in relation with this fact that we find the early 
and intermediate stages in the development of the parasites 
passed through very quickly, which leads on to the production 
of the ultimate stages found in the stomach. ‘This develop- 

vou. 60, PART 3.—NEW SERIES. 29 


410 H. M. WOODCOCK. 


ment proceeds along two lines, the result being the formation 
of two extreme types. The great difference in the relative 
frequency of certain forms, on the one hand in the culture, 
and on the other in the mosquito’s stomach, is entirely in 
accordance with the different conditions prevailing in the 
medium in the two cases. 

The earliest developmental forms obtained are trypano- 
monad individuals, such as are drawn in figs. 1-3, or b and ¢ 
of the scheme. ‘These are medium-sized, fusiform parasites, 
with a distinct membrane and the two nuclei close together, 
the kinetonucleus being usually just posterior to the middle 
of the trophonucleus. Both nuclei are either about the 
middle of the body or else slightly in the hinder part. This 
type of individual corresponds exactly with a particular 
crithidial form which occurs commonly in cultures, such as 
is drawn in figs. fj’ of the parallel series of stages from 
cultures, which I have reproduced ! for ready comparison with 
the developmental forms in the mosquito. In the mosquito, 
however, even in my earliest preparations (about thirty-two 
hours), the number of those trypanomonad individuals is 
relatively very small. This is in marked contrast to what is 
the case in the cultures, where the trypanomonad type is by 
far the predominating one. In the cultures many of these 
forms are distinctly larger than those with which I compare 
the mosquito forms just described (cf. figs. b’—-d’), and have 
the undulating membrane often better developed. In a small 
proportion of them, moreover, the kinetonucleus is just in 
front of the trophonucleus, instead of being alongside, a 
condition which I have not observed in this phase in the 
mosquito. 

As regards the immediate origin of these trypanomonad 
forms in the stomach, I think it is most probable that they 
arise by the division of rather larger, but otherwise similar 
forms. As is seen very clearly from the full series of the 
crithidial forms figured in my memoir (Il. c., Pls. 27, 29), the 

1 All these figures of cultural forms are taken from my Memoir in 
the ‘ Quart. Journ. Micr. Sci.,’ vol. 55, 1910, pl. 27, 29, or 30. 


STUDIES ON AVIAN HAMOPROTOZOA. 411 


division of the larger (earlier) individuals is at first by practi- 
cally equal binary fission, the only inequality being that the 
daughter-flagellum may be shorter than the parent one (cf. 
figs. e’—g’ of the scheme, Pl]. 31). Unfortunately I have not 
found any typical trypanomonad individuals in the act of the 
dividing in my preparations, but, as already mentioned, I 
observed an instance in life in which the parasite was just 
completing division, and in a condition practically identical 
with that seen in fig. 9’. 

However, I have very little doubt that the division of these 
trypanomonad individuals with the nuclei about the middle of 
the body is almost, if not quite, finished by the time of my 
earliest preparations. Because there are very few forms still 
left in this phase of the development; most of the parasites 
occurring are individuals which have lost the trypanomonad 
condition already and become trypaniform, and are either at 
some stage in the development of the elongated, attenuated 
trypanosome-phase, or have, in fact, attained the latter. I 
have been able to find, in stomachs of from thirty-two to 
thirty-six hours, a regular series of transitional forms in this 
connection (cf. PI. 29, figs.4-10). The elongated trypaniform 
condition is reached by a progressive modification of the form 
of the above trypanomonad individuals. I think this change 
is most probably unaccompanied by further division, at any 
rate, to any extent; otherwise, I ought to have found some 
indication of the process, as the parasites are fairly numerous, 
and all degrees in the gradual change of the type of form are 
to be met with. 

The earliest change is the passage of the kinetonucleus, 
and, of course, of the blepharoplast and attached flagellum, 
definitely behind the trophonucieus, the latter retaining its 
position (figs. 4 and 5). Next, the general cytoplasm in the 
posterior half begins to increase considerably in length ; in 
this way the parasite becomes longer, but it does not increase 
inuch in breadth. On the contrary, as the length increases, 
the body becomes ultimately not only relatively, but actually 
narrower and more slender. The change is most probably the 


412 H. M. WOODCOCK. 


result of two factors combined, namely, growth to some 
extent, and a thinning-out of the general cytoplasm. The 
latter factor is evident from the change in the form of the 
trophonucleus. This body, in the early stage, has the shape 
of a short and broad oval (figs. 4-6); as the trypaniform 
condition develops it becomes elongated, and appears as a 
narrow oval, compressed laterally (figs. 7-10). Meanwhile the 
kinetonucleus has passed well back behind the trophonucleus, 
but it never approaches anywhere near the posterior end of 
the body, the extension of the cytoplasm backwards, as the 
“tail”? develops, being so great (figs. 8-12). The final 
position of the kinetonucleus is usually about the middle of 
the body, or slightly in the posterior half. Nevertheless, the 
attached flagellum and the membrane are relatively long ; 
the membrane is always narrow and rarely at all wavy. 
Although, as the parasite assumes its final attenuated con- 
dition, the anterior part of the body also becomes thinned out 
and tapering, it does not appear to increase in length to any- 
thing like the extent that the posterior half does; so that the 
trophonucleus lies, as a rule, distinctly in the anterior part of 
the body, at times comparatively near to the anterior end 
(cf. figs. 13-20). 

As the final stage in the development of this form is nearly 
attained, a characteristic change usually occurs in the appear- 
ance of the trophonucleus. This organella, which has already 
become narrowed and compressed, becomes broken up into 
many oblong or rather rod-like blocks, arranged with their 
length transversely to the long axis of the body of the parasite, 
thus giving the whole nucleus a remarkable ladder-like 
appearance (figs. 13-17). Apparently this change repre- 
sents really the break-up of the original large karyosome, 
which is not seen, asa rule, in the nucleus when stained by 
Giemsa, into several small karyosomes arranged in a row. 
This is shown by the comparison of wet-fixed preparations 
stained by iron-hzmatoxylin, in which slender elongated 
trypaniform individuals, approaching the final form, have, 
instead of one large karyosome (as in Text-fig. 1a), four or 


STUDIES ON AVIAN HH MOPROTOZOA. 413 


more small karyosomatic bodies (Text-fig. 10). Probably 
each rod or block in the Giemsa-stained parasites corresponds 
to a karyosomatic grain of the greatly extended tropho- 
nucleus. Whether this change in the nucleus is absolutely 
necessary before the final form can be said to be completely 


TExt-FIG. 1. 


x 1600. Stages of Trypanosoma noctue in the mosquito, 
showing different conditions of the trophonucleus (from wet- 
fixed film, stained by iron-hematoxylin). A. Intermediate 
trypaniform phase. The karyosome of the trophonucleus is 
apparently budding off a small daughter karyasome. B. Ap- 
proximately final (propagative) form. There are five distinct 
karyosomes forming the elongated trophonucleus. 


formed, I do not feel certain ; occasionally individuals, which 
in other respects seem to have reached the final stage, are 
found where the nucleus is not ladder-like (cf. figs. 19-21). 
In my preparations of stomachs of fifty-five hours or later, 
scarcely any intermediate changes are present, practically all 
the individuals having attained the attenuated trypaniform 


4.14. H. M. WOODCOCK. 


condition. These are certainly very striking in appearance ; 
their thread-like form can be judged from the fact that their 
length is generally from 52 to 564, while they are rarely 
more than ly broad at the widest part, and often scarcely 
that! Superficially, they almost deserve the name of spiro- 
chetiform Trypanosomes, were that not so very misleading. 
The half of the body in front of the kinetonucleus—the 
actively motile part—is frequently spirally twisted or coiled to 
some extent; while the long, passive cytoplasmic “ tail” 
behind the kinetonucleus may be either practically straight 
(figs. 13, 14, 18 and 20) or bent up on itself in a curve 
(figs. 15-17, 19 and 21). 

The above-described trypaniform type also occurs in the 
cultural development. I have observed individuals both in the 
intermediate condition, and in what undoubtedly corresponds 
to the ultimate form attained in the mosquito’s stomach (cf. 
fies. k’-m’). It is important to note that the proportion of 
such forms to the trypanomonad individuals in the cultures is 
just the reverse from what is the case in the mosquito after 
thirty-two hours. This provides a very interesting com- 
mentary on the course of the cultural development as compared 
with that of the natural one. Whereas, in the latter, the 
enviromental conditions favour and lead on to the produc- 
tion of the attentuated trypdniform type, by the modifica- 
tion of the ordinary trypanomonad individuals, in the cultures 
there is not the same stimulus (the approaching completion of 
digestion) to the production of this form, and in consequence 
it is only developed, as it were, in isolated instances; the 
multiplicative (“crithidial”) form, on the other hand, per- 
sists and increases in numbers, far bevond what occurs 
under the natural conditions. The intermediate forms in the 
cultures resemble closely the corresponding individuals from 
the insect’s stomach (cf. figs. e-g). The few final forms 
which I have found (fig. m’) also agree quitefobviously in their 
chief features; there is the same long, cytoplasmic “ tail,” 
and even the ladder-like nucleus may be developed. The 
principal difference is that the individuals in the cultures are 


STUDIES ON AVIAN HA MOPROTOZOA. 415 


d 


much broader relatively than the natural “ spirocheetiform ’ 
individuals ; their body has apparently more bulk. I think 
this is simply an indication that these forms have grown 
more in the cultural medium than they do in the stomach ; 
and division having most probably ceased, the result is the 
marked increase in ‘‘ girth.” 

To consider next the other line of development followed by 
the parasites in the mosquito, this proceeds also from the 
original type of trypanomonad individual, by a modification in 
form and the mode of division. Here, again, I think the 
course of events can be understood from the cultural develop- 
ment. In certain of the ordinary trypanomonad individuals, 
the nuclei show a tendency to be in the hinder part of the 
body. This may be in the first place due to a shght oblique- 
ness in the direction in which the preceding nuclear division 
has occurred. At any rate, in such individuals the direction 
of the nuclear division is usually oblique, and we find not only 
the persistence of the nuclei in the hinder part of the body, 
but a distinct tendency already for the division of the body 
to be slightly unequal (cf. fig. 7’). These individuals with the 
nuclei definitely in the posterior half of the body pass, or 
grow into the characteristic club-shaped forms,' the general 
body-protoplasm tending to be concentrated in the region of 
the nuclei (fig. 0’). Now, in these forms, the division is 
always markedly unequal. It is important, I think, to note 
this particular point, that where the nuclei are situated about 
the middle of the body, the division is practically equal (as in 
the ordinary trypanomonad individuals) ; where, on the other 
hand, the nuclei are distinctly in the hinder part of a club- 
shaped individual, the division of the cytoplasm is unequal. 
This is seen in figs. p’ and q’ of my scheme and also in several 
other figures in my memoir (12). I consider this mode of 
division is primarily due to the drawn-back position of the 
nuclei, and consequently of the blepharoplasts, and to the 
fact that the new daughter-flagellum is, at any rate, largely 


’ I formerly distinguished these by the cumbrous term of “ accentu- 
ated trypanomonad ” individuals. 


416 H. M. WOODCOCK. 


formed by free, independent growth. Individuals which are 
manifestly the product of a similar division are seen in figs. 
7 and ¢’. 

Now, what I have found in the mosquito agrees entirely 
with the above description. One of the earliest stages along 
this line of development is the characteristic club-shaped 
trypanomonad form of fig. 22, ork. This individual corre- 
sponds very closely, it will be seen, to that of fig. o’ or p’; it 
is almost ready for division. As remarked in my account of 
the living observations, I saw one instance of such a form in 
the act of unequal division. The smaller of the daughter- 
individuals about to result wasa short, pear-shaped form, with 
scarcely any membrane; the larger one wasa rather club- 
shaped individual, long and tapering, and probably with a 
long, narrow membrane (cf. fig. q’). Unfortunately, I have 
not obtained any examples of individuals just dividing in my 
permanent preparations; in view, however, both of the living 
observation and of the fact that numerous individuals. 
belonging to both the distinct forms which are produced by 
such division occur in my preparations (cf. figs. 24, 26, and 
again, figs. 23, 25), I cannot doubt that such unequal division 
takes place as a normal and typical phase of the development 
in the mosquito. F 

By further division of the larger daughter-individual 
(possessing the long membrane), smaller forms are produced 
(tigs. 30-32 or n and 0) ; these can at first be recognised as 
corresponding to one or the other of the two types, but in the 
smaller individuals the distinction between the daughter- 
forms tends to be less marked. Here and there I have found 
a stout pyriform individual, one of the early-developed pear- 
shaped daughter-forms, just about to divide, after further 
growth (fig. 27) ; and also a quite small parasite, one of the 
ultimate stages, I should say, undergoing division (fig. 34). 
There can be no doubt that this line of development leads 
ultimately to the production of small, pear-shaped or oval 
parasites, with the nuclei close together and situated about 
the middle of the body, or nearer the posterior end, and with 


STUDIES ON AVIAN HASMOPROTOZOA. 417 


the flagellum drawn back but with practically no membrane 
(figs. 29, 32 and 33). Here we have, unmistakably, the 
haptomonad phase of the T'rypanosome, as I have proposed 
(16). to term the so-called ‘‘ gregariniform” phase, which 
serves for attachment (and coincident multiplication). If 
these figures, for instance, are compared with certain of the 
text-figures in my description (15) of Leptomonas fas- 
ciculata, as I found this parasite in Culex pipiens, it 
will be perfectly clear that they represent the corresponding 
phase of the Flagellates in both cases. This agreement, I 
may incidentally mention, affords an excellent example in 
favour of the point I have urged, that from the haptomonad 
phase alone (having regard only to the morphology) it cannot 
be said with certainty whether a particular Insectan parasite 
represents a Leptomonad, a Crithidia or a Trypanosome. 


From the above account it is clear that the early develop- 
ment of ''rypanosome noctue in the mosquito culminates 
in the production of two distinct and extreme types. Whether 
the above description includes all the modifications of form 
which occur in the life-cycle in the Insectan host, I am not 
able to say. I think, however, that it is not difficult to inter- 
pret the significance of the end-stages, and if the view I 
favour is correct, any further stages in the life-history repre- 
sent in the main a repetition, or re-development of the above 
sequence of forms, consequent on the persistence of the para- 
sites in the mosquito and their response to fresh meals of 
blood. 

The haptomonad forms most probably—as, indeed, is 
implied by thus designating them—become attached to some 
part of the wall of the alimentary canal, lose practically all 
the flagellum and enter upon the resting phase; ‘ resting,” 
that is to say, as regards locomotion, but not in regard to 
nutrition or multiplication. I think there is no reason to 
doubt that the chief situation favoured by these hapto- 
monads for attachment, as the digestion becomes finished, is 
the anterior end of the stomach and especially the invagi- 


418 H. M. WOODCOCK. 


nated epithelium of the proventriculus, as was so graphically 
described by Schaudinn (10) and illustrated by him in Text- 
fig. 14. Just because the parasites were in this situation, I 
feel sure that in regard to this point Schaudinn’s account 
relates to actual developmental phases of T. noctuz and not 
to a purely Insectan form (such as, possibly, Leptomonas 
fasciculata). In the case of the latter, on the other hand, 
the haptomonad forms are restricted to the intestine apparently. 
Whether the haptomonad phase of T. noctuz also invades 
the intestine to any extent, 1 am unable to say; considering 
that these forms tend to mass themselves at the anterior 
end of the stomach, it is quite likely that they do not— 
unless they are also able to form cysts for passage to the 
exterior. 

Returning now to the first lime of development, what is the 
further destiny of the remarkable thread-like individuals? I 
am decidedly of the opinion that they represent the propaga- 
tive phase of the Trypanosome in the mosquito, the form, 
that is, in which the parasite is finally transmitted back again 
by inoculation to the owl. Unfortunately, I have no proof of 
this, but certain considerations point strongly to this being 
the right explanation of the significance of the above very 
characteristic type. In the’ first place, a brief comparison 
with the known course of the life-history in piscine T'rypano- 
somes is most suggestive. It has been recognised for some 
time that a given species of fish Trypanosome shows—at all 
events, in many cases—very considerable polymorphism during 
that part of the cycle undergone in the blood of the Verte- 
brate host; thus certain individuals of a species may be quite 
small, slender forms (representing a young stage), others 
large, massive forms, with all intermediate grades between 
ef. T. granulosum, T. perce, Minchin [3]). Now I 
have shown that the same pronounced polymorphism also 
obtains in Avian Trypanosomes (vide [12] and, with Minchin 
[5)). Whereas earlier writers frequently described a small 
form from a particular host as one species, and a large, massive 
form from the same host as another species, the true meaning 


STTDIES ON AVIAN HAMOPROTOZOA. 419 


of these different types is that they are different stages in the 
life-history of one species. ‘Thus the large, massive, “ blue ” 
Trypanosomes of the little owl, originally described under 
the name T. ziemanni, and connected by Schaudinn with 
Leucocytozoon ziemanni, are really only the large forms 
of T. noctuze and not a separate species at all. And a 
corresponding great variation in form and size is found in 
T. fringillinarum, in the chaffinch. Hence there is a 
strong family resemblance between piscine and avian Try- 
panosomes, as regards the types met with in the Vertebrate 
host ; and it seems quite clear that both belong to the same 
group or division of ''rypanosomes, and stand somewhat apart 
from Mammalian forms, for instance. Turning now to the life- 
cycle of piscine forms in leeches, which has been fully worked 
out by Miss Robertson in two or three cases (e.g. T. raix 
in Pontobdella [7 and 8], T. danilewskyi in Hemi- 
clepsis [9], it is found that, after a varying period of multi- 
plication by the parasites in the trypanomonad (crithidial) 
phase, as the digestion approaches completion—the time 
occupied varying considerably according to the conditions—a 
trypaniform type is developed, which becomes very elongated 
and slender. ‘his passes forwards from the crop into the 
proboscis-sheath, and is the propagative, inoculative form, 
which transmits the infection to a fresh fish. 

This type of form is essentially similar to that above 
described in Culex, the significance of which I believe to be 
also the same; in the leech it does not apparently attain the 
same degree of tenuity and the remarkable thread-like 
appearance which it does in the mosquito, but it may at times 
show the same peculiar ladder-like nucleus. It may at first 
be thought that from thirty-four hours onwards is too soon 
for the final, propagative form in the mosquito to be already 
present. But this rapid development appears quite explicable 
when the habits of the mosquito are considered. Unlike most 
of the Invertebrate hosts (transmissive agents) of Trypano- 
somes, mosquitoes do not feed solely on blood. On the con- 
trary, it is generally recognised that, so far as the species of 


420 H. M. WOODCOCK. 


temperate climates are concerned, meals of blood are not by 
any means the rule, and in many cases only taken in con- 
nection with the development of the eggs. In the case of 
most females blood appears to be necessary for this purpose ; 
it certainly is so in Culex pipiens, and as I have recently 
shown (l. c.), two meals of blood will suffice to bring about 
the growth and oviposition of the fertilised ova. After laying 
one, or perhaps two, batches of eggs, most of the females 
either die or go into hibernation. Hence, it is perfectly clear 
that unless the propagative phase of the Trypanosome is 
developed in time for inoculation at the second (or at most 
the third) meal of blood, the chances of the parasite passing 
back to the bird are very uncertain—an entirely different. 
state of affairs from what is the case among tsetses or 
leeches. For the above reasons, therefore, I think there is no 
difficulty in assuming that the thread-like ‘‘ spirochetiform ” 
individuals which I have described represent the inoculative 
type, or in understanding their early development in the 
mosquito. 

While I consider it is quite likely that these forms are 
inoculated again into the blood at the second time of feeding, 
I do not overlook the possibility that they (or some of them) 
pass first into some other organ of the mosquito (such as the 
salivary glands, or cesophageal diverticula), and there await 
a third meal; but this has still to be ascertained. I do think, 
however, Schaudinn was mistaken in stating that the Try- 
panosomes cannot be transmitted back again to the owl until 
the fourth meal inclusive. I should say, if the parasites had 
to wait until the mosquito took a fourth meal of blood, they 
would rarely, if ever, have the opportunity of getting back 
again at all, unless, perhaps, they were in a female which was 
going to hibernate. It is obvious from Schaudinn’s account 
that he was following chiefly the multiplication of the parasites 
in the haptomonad condition (vide his description of their 
attachment in vast numbers to the walls of the alimentary 
canal). Schaudinn does not appear to have seen the real 
propagative phase at ail! He nowhere describes the 


STUDIES ON AVIAN H#MOPROTOZOA. 421 


attenuated forms with the characteristic cytoplasmic “ tail” 
(ef. on the other hand, Mayer’s account, referred to below). 
Schaudinn’s slender, “ spirocheetiform” individuals, which 
he describes under ‘ Spirocheta” (Trypanosoma 
ziemanni), agree fairly well with the intermediate try- 
paniform stages in the development of the final type (such as, 
e.g. figs. 7 and 8); but I must add, nothing has astonished 
me more than the lack of close correspondence, in the main, 
between the different forms of the Trypanosome occurring in 
the mosquito, as Schaudinn figured them, and as I have found 
them. 


Il. The Ookinetes of Halteridium noctue and 
Leucocytozoon ziemanni, 


A few words next concerning the ookinetes of Halteri- 
dium and Leucocytozoon, as they occur in my permanent 
preparations. As regards their general appearance, I have 
little to add to the description previously given by Mayer 
(l.c.). The ookinetes of both forms are fundamentally 
similar in type, as was of course to be expected, considering 
that two essentially similar parasites are concerned. In both 
cases the body is very frequently more or less coiled up in the 
form of a C. Practically the only difference between the two 
is that the ookinetes of Leucocytozoon are much larger, 
especially in regard to length, than are those of Halteri- 
dium (c.f. figs. 35-40 and 41 and 42). 

In my preparations of the Culex which fed on Owl 19, 
made from tlurty-three to thirty-six hours after feeding, the 
ookinetes of Halteridium are very numerous. In prepara- 
tions made from mosquitoes which fed on Owl 25, ookinetes are 
also usually to be found (up to thirty-six hours), but they are 
very scanty ; this difference is due to the different conditions 
of the infection in the two owls respectively (see above, pp. 400— 
401). Speaking generally, all the ookinetes observed are in 
the same phase of development, and with one or two exceptions 
have lost all the pigment. In none of them is anything like 


422 H. M. WOODCOCK. 


a kinetonucleus, let alone a developing flagellum, recognisable. 
In a certain number a small chromatinic area, or clump of 
chromatinic grains, is present, in addition to the nucleus (figs. 
35 and 36); but these stain quite similarly to the nucleus, 
and in no case have the characteristic appearance of a kineto- 
nucleus, as seen when stained by Giemsa (contrast the “ resting 
flagellates ” referred to below). Unfortunately I am unable 
to say whether the nucleus possesses a central karyosome or 
not, but at any rate there is certainly no excentric or extra- 
nuclear karyosome, such as occurs at certal periods both in 
Halteridium and Leucocytozoon when in the blood,! and 
which often simulates a*€inetonucleus in so marked a manner 
that it was formerly mistaken for one, until I showed clearly 
(14) what its true significance was. Frequently the ookinete 
shows one or two vacuoles in the cytoplasm; I have not been 
able to find any ookinetes in my preparations of mosquitoes 
which were made later than fifty-four hours after feeding. 

I have only come across very few ookinetes of Leuco- 
cytozoon on my smears; I was generally able to find one 
or two during the living examination of the stomachs of 
mosquitoes which had fed on an owl infected with this 
parasite, but unfortunately. in several cases, owing to their 
scantiness, | have not succeeded in obtaining any on the 
permanent slides made. ‘Those I have found are all practi- 
cally similar in form and size (figs. 41 and 42). The nucleus 
is always situated fairly near to the more rounded end of the 
body, as it is also in Halteridium; it varies in size to some 


1 Reichenow, in a note on Leucocytozoon ziemanni in his 
account of the Hzemogregarines in ‘Handb. d. Path. Protozoen’ (6) 
states that he was unable to find a karyosome in the male gametocytes. 
As I showed in my “ Notes on Sporozoa,”’ published about the same 
time, there is certainly such an organella present, excentric or even 
extranuclear as in the female forms. It is unmistakable in prepara- 
tions stained by iron hematoxylin, but it is rarely shown in Giemsa 
smears. Reichenow does not say whether the figure he gives is from a 
Giemsa smear or not, but from its general appearance (e. g. the hyper- 
trophied host-cell nucleus stains quite differently after iron-hema- 
toxylin), I should say it was. 


STUDIES ON AVIAN HAZMOPROTOZOA. 423 


extent. Now and again separate chromatinic granules occur 
close to it (cf. fig. 42). In these ookinetes, also, one or two 
vacuoles are often present. None of the ookinetes of Leuco- 
cytozoon, any more than those of Halteridium, are in any 
later stage of development. I have not seen the least indica- 
tion of the remarkable skein-like formation so graphically 
described by Schaudinn, accompanied by nuclear multiplica- 
tion and the eventual development of a number of very small 
Trypanosomes! I must say that I doubt very much now the 
correctness of all this. 


III. The “Resting Flagellates.” 


As was described in the account of the living observations 
(see above, pp. 401-402), in asmall number of female Culex 
pipiens caught wild, which were examined at an early stage 
of the experimental work before I restricted myself to the use of 
bred-out mosquitoes, certain characteristic and rather peculiar 
motionless bodies were sometimes present. I have mentioned 
how on one occasion such resting forms were actually seeu 
to develop into active flagellates, and the temporary illusion 
fostered by the observation. But not only were they seen in 
the first “ wild female,” which fed on Owl 19, well infected 
with Halteridia, they also occurred now and again in females 
which had fed on an uninfected bird, or which were examined 
before being allowed to feed on a bird at all! Hence, what- 
ever their origin, there is no reason for associating them with 
Halteridium. For some cause or other, with the above- 
noted exception, these resting forms were never observed to 
become active, nor were active flagellates corresponding to 
them ever found. 

In my permanent preparations, practically all these para- 
sites occur in the resting condition (figs. 43-45): in one or 
two instances, however, there is an unmistakable wavy 
border in the anterior part of the body, accompanied by a 
drawn-out, tapering anterior end (figs. 46-47). Iam strongly 
of the opinion that the flagellum is actually developed in both 


424, H. M. WOODCOCK. 


these cases, but it is difficult to be quite certain because for 
some reason or other, it has not stained red in the usual 
manner after Giemsa. Especially in the individual drawn in 
fig. 47, however, a definite line is evident along one edge of 
the anterior, tapering part, which is continued free for a short 
distance, which in all probability represents a flagellum. 
The general appearance of these forms, moreover, closely 
resembles that of the particular individual referred to above, 
just the instant before a well-marked, free flagellum became 
apparent. I think no one can have any doubt that these are 
really Flagellates, because they are certainly binucleate 
forms; in all cases, a definite kinetonucleus is present, 
usually immediately in front of the trophonucleus, which 
has the characteristic staining reaction to Giemsa. It may 
be pointed out that there is no question of this element being 
a karyosome, because, the karyosome of the nucleus can at 
times be seen, appearing, as is always the case in Giemsa- 
stained smears, as a clearer area, with a distinct centriole in 
the middle (cf. figs.43 and 46). The cytoplasm always stains 
more darkly and much more bluish-purple in tint than any of 
the developmental phases of Trypanosoma noctue, in 
which, by the way, aflagellate phases appear to be entirely 
lacking. In all my experience of crithidial forms, whether 
as developmental phases of Trypanosomes, or “ Crithidial” 
parasites, I have never observed any in which the cytoplasm 
stains in this peculiar dark manner, or in which the flagellum 
is so faint. 

This parasite is certainly a crithidial form; this is evident 
from the contiguity of the two nuclei about the middle of the 
body, as well as from the distinct undulating membrane, 
where an individual is in, or about to assume, the active 
condition. In one respect it appears to be unique, 1.e., in 
being non-flagellate, and moreover, without any trace of a 
rhizoplast, and quite motionless, when still possessing the 
elongated, fusiform shape, which is always associated in 
other Flagellates of this kind, with the active, flagellated con- 
dition; in all other cases of which I am aware, when the 


STUDIES ON AVIAN HAMOPROTOZOA. 425 


parasites are “resting” and non-flagellate they are in the 
short, more or less oval, haptomonad phase. ‘l‘his form does 
not ‘agree with any of the “ Crithidiz” hitherto described 
from mosquitoes ; until more is known about it I am inclined 
to regard it as a distinct parasite. At present I have 
obtained the impression that this may bea purely Insectan 
form, a parasite more particularly, perhaps, of the larval 
Culex. Somehow, it does not look like a form accustomed 
to a blood-medium ; its appearance and behaviour are so 
different from all the other crithidial forms which I have had 
occasion to study. 


GENERAL CONCLUSION REGARDING THE QUESTION OF A CONNEC- 
TION BETWEEN THE ‘l’RYPANOSOME AND THE H#MOSPORIDIAN 
PARASITES OF THE LirrLeE Owt. 


For the last time, I hope, that it will be necessary, I return 
to this subject, more particularly in order to point out the 
bearing upon it of the observations recorded above on the 
developmental phases of these parasites in the mosquito. It 
must be apparent, indeed, that these observations support 
and further strengthen the conclusion at which I had already 
arrived (14), that there is no connection whatever between 
these different types of parasite; and I have found nothing 
that in any way corroborates Mayer’s account (2), in which 
he has upheld the opposite view. 

In the first place, with regard to the mosquito which fed 
on Owl 19. As was to be expected from the typical ripe 
Halteridial infection of this bird, numerous fully-formed 
ookinetes were found without difficulty in the stomach when 
dissected. But not one of these ever showed any sign of 
passing into a flagellate condition; nor were any active 
Flagellates observed of the different types which I subse- 
quently found frequently in mosquitoes fed on an owl known 
to have Trypanosomes. This fact is very significant when it 
is remembered that no Trypanosomes were ever seen in the 
blood of Owl 19, either in life or in searching smears, and if 

VoL. 60, PART 3,—NEW SERIES. 30 


426 H. M. WOODCOCK. 


they were present, must have been so rare at the time as to 
be negligible. 

As was discussed above, the matter was complicated just 
at the outset by the use of a few “ wild ” females in which a 
“resting Flagellate” occurred, which on one or two occasions 
developed into the active condition. In my own opinion, how- 
ever, it is perfectly clear that these resting Flagellates have 
nothing to do with the ookinetes. Although there is un- 
doubtedly a general resemblance in appearance between these 
two bodies when observed in life, there are several important 
reasons for concluding that this is only a coincidence. In 
stained preparations the two types of element appear funda- 
mentally distinct, for the resting Flagellates show without 
exception the binucleate condition; the ookinetes, on the 
other hand, never do. In not a single instance, whether of 
Halteridium or of Leucocytozoon, have I been able to 
find an ookinete which possesses the binucleate condition, 
the first essential for it to be regarded as connected with a 
Heemoflagellate. Again, the staining reaction of the general 
cytoplasm in the two cases is entirely different ; and though, 
knowing what I do of the dangers, no less than the advantages, 
associated with the use of the Giemsa stain, I should be the 
last to lay stress upon casual staining differences in different 
cases ; nevertheless, where such a difference is constant and 
uniform, weight may be laid upon it. Further, these resting 
Flagellates occurred also in “wild” females fed on owls 
totally uninfected with any of the parasites under discussion, 
and in which no ookinetes, of course, were found. Finally, 
these resting Flagellates have certainly nothing whatever to do 
with the developmental phases of Trypanosoma noctue. 

Now, in the bred-out mosquitoes which fed on Owl 28, 
ookinetes were always scanty; particular individuals were 
carefully watched on different occasions, but in these also no 
change or development of any kind was ever seen. Never- 
theless, in about 46 per cent. of these mosquitoes 
examined, active Flagellates were observed in 
different stages of development; and it is to be re- 


STUDIES ON AVIAN HA MOPROTOZOA. 427 


membered that Owl 23 was the only birdin which Try pano- 
somes were found in the peripheral circulation, 
and actually at this period. If the Trypanosomes had 
indeed developed from ookinetes, it is difficult to understand 
why they should occur frequently when the ookinetes were 
scanty (and sometimes not actually noticed), and yet not at all 
in the case where the ookinetes were numerous (the female 
which fed on Owl 19, with the normal Halteridial infection). 
All my observations point clearly to these active Flagellates 
being the developmental forms of T. noctuz in the mosquito, 
and derived directly from the stumpy, or stout fusiform 
Trypanosome, present in the blood in the summer (as pre- 
viously described (5). I have never once seen the slightest 
indication of these developmental Trypanosome-stages in any 
mosquito fed on a bird which did not contain this stumpy, 
transmissive phase of the Trypanosome in the peripheral 
circulation—whether it contained Halteridium, or Leu- 
cocytozoon, or neither. 

A few remarks, in concluding this discussion, about Mayer’s 
account (l.c.). The most important statement of this worker 
is that, in hanging drops of blood from an owl infected with 
Halteridium, in which careful examination failed to reveal 
any Trypanosomes after four days or so Flagellates were found 
to be present; the inference is, of course, that these had 
developed from the Halteridium-ookinetes. I can only 
say that I feel absolutely convinced that Mayer was mistaken 
in supposing no 'l'rypanosome-individuals to be present in the 
drop at the beginning of the experiment. I know well from 
experience how easily one of these forms can be overlooked, 
especially if the blood-corpuscles are in a fairly thick layer. 
Another possible explanation would be that certain very 
minute or ultramicroscopic phases of the Trypanosome were 
present, which later gave rise to flagellates. But this idea is 
quite unnecessary when I have shown clearly that the same 
particular flagellate developmental stages occur regularly, 
both in cultures and in mosquitoes, as a result of inoculation 
with a definite Avian 'l'rypanosome form. For my own part, 


428 H. M. WOODCOCK. 


I say candidly that I can see no sufficient evidence up to the 
present for believing in “ cryptotrypanosomiasis 7! or in 
“infective granules,” etc. In all the life-cycles of Trypano- 
somes which are now known in the invertebrate host, there is no 
suggestion of such a thing, and such an explanation is not, I 
consider, required. In birds, and even more so in cattle and 
sheep, Trypanosome individuals may be so excessively scanty in 
the circulation that the greatest difficulty is entailed in 
finding them ; but let a single individual succeed in passing 
into a culture or into its right invertebrate host, and it will 
multiply so rapidly that before many days have passed the 
forms to which it gives rise are readily found. I may say 
also that I am no longer inclined to think that any small 
form of these Avian Trypanosomes occurs in the red blood- 
corpuscles, which might perhaps be mistaken for a young 
stage of Halteridium. Since very considerable doubt has 
been thrown upon the occurrence of such astagein Try pano- 
soma cruzi, it is becoming more and more probable that 
Trypanosomes do not get into the red cells at all; at all 
events I am disinclined now to postulate the occurrence of 
such a phase in the Avian Trypanosomes which I have 
studied until we have an authentic instance of it in some other 
case. As in the case of fish-Trypanosomes, with which, as I 
have shown above, Avian Trypanosomes have much in common, 
I firmly believe that unless the transmissive trypanosome- 
form of an Avian T'rypanosome passes into culture or the 
invertebrate host, no development of Flagellates will occur. 
Mayer describes and figures further certain “‘ large forms ” 
which developed in mosquitoes, which he regards as the de- 
velopmental stages of Leucocytozoon, in contradistinction 
to Halteridinm. But these forms are obviously the same 
as certain of those which I have described and figured as the 
developmental stages of Trypanosoma noctue; his fig. 56, 
for instance, represents an individual nearly arrived at the 
final attenuated form. The stages which Mayer associates 


1 By this term I understand some definite phase of the parasite, 
hitherto unrecognised and possibly ultramicroscopic. 


STUDIES ON AVIAN H#EMOPROTOZOA. 429 


with Halteridium, equally with those which he associates 
with Leucocytozoon, constitute part of a regular series, and 
belong to a definite life-cycle, as I have clearly shown. 
Correspondingly, it will be remembered, Minchin and I 
showed also that the small forms of the Trypanosome in the 
blood, associated by Schaudinn with Halteridium noctue, 
form part of a regular series with the large individuals, 
regarded by him as belonging to Leucocytozoon ziemanni, 
and altogether represent only one species—T ry panosoma 
noctue; there is no species I’. ziemanni. If, therefore, 
one still held to the idea that the small Trypanosomes are 
connected with Halteridium and the larger ones with 
Leucocytozoon, one would be led to the impossible position 
that Halteridium and Leucocytozoon are different 
phases of one and the same thing. 

I do not suppose that Mayer any longer considers that 
an ontogenetic connection exists between any of these 
different parasites, especially as in addition to the above- 
inentioned paper, I have also published my account of the 
cytology, in which I have shown that, after all, there is no 
nuclear dimorphism in Halteridium and Leucocytozoon, 
and that these are not related to the Binucleata. But as Mayer’s 
account is the only one published of late years on the develop- 
imental stages of these parasites of the owl in mosquitoes, I 
have been obliged to point out where it is erroneous in the 
light of my own observations on the whole subject. 

The conclusion of the whole matter is, that the three 
parasites of the little owl—Trypanosoma noctue, Halte- 
ridium noctuze and Leucocytozoon ziemanni—are 
entirely distinct and separate types; and the same is un- 
doubtedly true for other species of these parasites in other 
birds. So far as T. noctue is concerned I have been 
able to outline above the main course of its development 
in the mosquito (Culex pipiens), though there are, un- 
fortunately, gaps still to be filled up. The development 
of the Halteridium and the Leucocytozoon in the mosquito 
remains to be ascertained—supposing, that is to say, that 


430 H. M. WOODCOCK. 


there is any development beyond the ookinete stage. Other 
workers (e. g. the Sergents (11), Aragao (1) and Mayer (loc. 
cit.) have obtained the development of one or both of the 
parasites up to the same stage, but never any farther. As 
regards Halteridium, Aragao, who has succeeded in trans- 
mitting H. columbe from a Hippoboscid fly (Lynchia) 
back again to the pigeon, is doubtful whether there is really 
any further development in the Insectan host; it is possible, 
moreover, that the so-called schizogony in the lung represents 
the delayed sporogony of the ookinetes, as has been hinted at 
by Minchin, in his text-book on the Protozoa (4). If that be 
so, then I see no reason why the ookinetes of H. noctue 
should not behave similarly ; in which case, Culex pipiens 
may prove, after all, to be a true host—transmissive agent— 
of this parasite also. Atany rate, for all that one can yet say 
to the contrary, the ookinetes may be inoculated back again 
into an owl, at the same time as the final propagative forms 
of the Trypanosome. It is, perhaps, not without significance 
that in every owl which was infected with the one parasite, 
we found the other also to be present (cf. Minchin and Wood- 
cock, loc. cit.). The .development and transmission of 
Leucocytozoon are still more a matter of uncertainty. As 
far as Schaudinn’s observations are concerned, if such a 
remarkable nuclear multiplication and skein-development 
does occur, it is very strange that neither Mayer norI myself 
have seen any signs of it. One thing is, I think, practically 
certain; if the ookinete does produce a number of small 
elements by rapid division, these will not prove to be Try- 
panosomes or other binucleate Flagellates ! 


Tue Lister INSTITUTE; 
May 12th, 1914. 


10. 


19 le 


12. 


13. 


14. 


STUDIES ON AVIAN HAMOPROTOZOA, 43] 


REFERENCES TO LITERATURE. 


. Aragao, H. de B.—* Ueber den Entwickelungsgang und die Ueber- 


tragung von Hemoproteus columbe,” ‘Arch. Protistenk.,’ 
xii, 1908, p. 154, pls. 11-13. 


. Mayer, M.—* Ueber ein Halteridium und Leucocytozoon des 


Waldkauzes und deren Weiterentwickelung in Stechmiicken,” 
op. cit., xxi, 1911, p. 232, pls. 22 and 23. 


. Minchin, E. A.—* Observations on the Flagellates Parasitic in the 


Blood of Fresh-water Fishes,” ‘ Proc. Zool. Soc.,’ 1909, i, p. 2, pls. 
1-5. 


‘An Introduction to the Study of the Protozoa.’ London: 
Ed. Arnold, 1912. 


and Woodcock, H. M.— Observations on the Trypanosome 
of the Little Owl (Athene noctua), ete.,” ‘Quart. Journ. Mier. 
Sci., 57, 1911, p. 141, pls. 20 and 21. 


. Reichenow, Ed.—“ Die Hemogregarinen,” in ‘Handb. d. path. 


Protozoen,’ edited by Prowazek, 8. v., vol. ii, Lief. 1, p. 602, 2 pls. 


. Robertson, M.—*‘ Studies on a Trypanosome found in the Alimen- 


tary Canal of Pontobdella muricata,” ‘Proc. physic. Soc. 

Edin.,’ 17, 1907, p. 83, pls. 4-7. 

* Further Notes on a Trypanosome found in Pontobdella 

muricata,” ‘Quart. Journ. Mier. Sci.,’ 54, 1909, p. 119, pl. 9. 

“Transmission of Flagellates living in the Blood of certain 
Fresh-water Fishes,” ‘Trans. Roy. Soc., B. 202, 1911, p. 29, pls. 
1 and 2. 

Schaudinn, F.—*‘ Generations- und Wirthswechsel bei Try pano- 
soma und Spirocheta,”’ ‘Arb. kais. Gesundhtsa.,’ 20, 1904, p. 
087, text-figs. 

Sergent, E. and E.—* Etudes sur les Hématozoaires des Oiseaux,” 
‘Ann. Inst. Pasteur,’ 21, 1907, p. 251, pls. 6 and 7. 

Woodcock, H. M.—“Studies on Avian Hemoprotozoa: I. On 
certain Parasites of the Chaffinch (Fringilla cwlebs) and the 
Redpoll (Linota rufescens),” ‘Quart. Journ. Micr. Sci.,’ 55, 
1910, 641, pls. 27-31. 

“On an Unusual Condition observed in Halteridium,” 

‘Zool, Anz.,’ 38, 1911, p. 465, text-figs. 

‘ Notes on Sporozoa, Nos. II-IV,” ‘ Quart. Journ. Mier. Sci.,’ 

58, 1912, p. 171, pls. 9 and 10. 


432 H. M. WOODCOCK. 


15. Woodcock, H. Mi—* On ‘ Crithidia’ fasciculata in Hibernating 
Mosquitoes (Culex pipiens), ete.,” ‘Zool. Anz.,’ 43, 1914, p. 370, 
text-figs. 


“ Further Remarks on the Flagellate Parasites of Culex, 
etc.,” ‘Op. cit., 44, 1914, p. 26. 


EXPLANATION OF PLATES 29-31. 


Illustrating Dr. H.M. Woodcock’s paper on “Studies on Avian 
Heemoprotozoa : No. II].—Observations on the Develop- 
ment of Trypanosoma noctue (of the Little Owl) in 


Culex pipiens; with Remarks on the Other Parasites 
occurring.” 


[All the figures on Pls. 29 and 30 are magnified 2000 times linear; 
those of the scheme, PI. 31, are x 2000 ( nearly). Iam indebted to Miss 
Rhodes for kindly drawing and colouring most of the figures of the 
Halteridium ookinetes. | 


PLATE 29. 


All the figures relate to the development of T. noctuz in the 
mosquito (Culex pipiens). 

Figs. 1-3.—Typical trypanomonad forms. 

Figs. 4-6.—Harliest stages in the development of the trypaniform 
type. ; 

Figs. 7-10.—Intermediate trypaniform stages. 

Figs. 11, 12.—Approximation to the final inoculative form. Fig. 11 
shows an intermediate stage in the change in the nuclear condition ; 
three fairly large chromatinic masses (karyosomes) are present. 

Figs. 13-17.—Typical attenuated thread-like forms, with ladder-like 
nucleus. These are considered to be the inoculative type. 

Figs. 18-21.—Individuals which agree in character with the final 
forms, except for the fact that the nucleus has not become ladder-like. 

Figs. 22, 23.—Club-shaped trypanomonad forms. Early stages in the 
second line of development. (See text.) . 

Figs. 24, 26.—Pear-shaped forms, with the kinetonucleus tending to 
be in front of the trophonucleus and with hardly any membrane. These 
individuals result from the smaller member of an unequal division of 


STUDIES ON AVIAN HA MOPROTOZOA. 433 


such a form as that of fig. 22. They represent the beginning of the 
haptomonad phase. 


Fig. 25.—Individual resulting from the larger half of the same or a 
similar unequal division. It tends to be slightly club-shaped, and has 
the nuclei well back, with a long, narrow membrane. 


Fig. 27.—Pyriform individual commencing division. 
Figs. 28, 29.—Other pyriform haptomonad forms. 
Figs. 30, 31.—Small individuals which still show to a slight extent 


the characters of the differing products of an unequal division (probably 
of such a form as that of fig. 25). 

Figs. 32, 33.—More haptomonad forms, or more strictly, perhaps, 
forms which are going to become attached. 


Fig. 34.—One of the smallest forms found undergoing division. 


PLATE 50. 


Figs. 35-40.—Ookinetes of Halteridium noctuex. In fig. 38 part 
of the skin or pellicle of the ruptured erythrocyte is seen still attached 
to the ookinete. This occurs now and again. 


Figs. 41, 42—Ookinetes of Leucocytozoon ziemanni. 

Figs. 43-47.—“ Resting Flagellates” found in “wild” mosquitoes. The 
individuals of figs. 46 and 47 have almost certainly developed a flagellum, 
marginal, if not free. 


PLATE 31. 


Scheme comparing the development of T. noctuz in the mosquito 
with the developmental stages of T. fringillinarum as found in 
cultures. (For explanation, see text.) 


x 
ct 
s 


STUDIES IN THE EXPERIMENTAL ANALYSIS OF SEX. 439 


Studies in the Experimental Analysis of Sex. 
Part 11—On Stylops and Stylopisation. 


By 
Geofirey Smith, M.A., 
Fellow of New College, Oxford. 


And 
A. H. Hamm, 
Assistant in the Hope Department, Oxford. 


With Plates 32-35. 


It has long been known that the presence of Strepsipterous 
parasites on solitary bees and wasps may exert a remarkable 
influence on their hosts, and the paper by Professor J. Pérez 
in 1886 (1), gave a full account of the effect of Stylops 
melittee on the bees Andrena, which must always be 
referred to as a classical contribution to the study of parasitic 
castration. Although many authors have subsequently 
written on these parasites, the observations of Monsieur 
Pérez remain unsurpassed for fulness of detail and interest, 
and we are indebted to the veteran entomologist for 
additional information which he has kindly put at our 
disposal in answer to inquiries. The re-examination of the 
subject, which is undertaken here, seemed desirable, not so 
much for the purpose of bringing new facts to light, but for 
confirming reported facts and collating them with the newly 
acquired results on parasitic castration brought about by 
Sacculina, since the effect of Stylops on its hymenopterous 
hosts seemed to be parallel to that of Sacculina on crabs, 


436 GEOFFREY SMITH AND A. H. HAMM. 


and yet to differ from it in many important respects. The 
Strepsipterous parasites differ from Sacculina in that they 
have the sexes separate, whereas Sacculinais hermaphrodite, 
so that we can test whether the sex of the parasite has any 
influence on the effect excited. Moreover, it is clear that in 
certain cases the female host is induced to assume certain 
male characters as the result of parasitic castration by 
Stylops, an effect which is never observed in the case of 
Sacculina, and which requires careful examination. We 
have been able to study a large number of stylopised bees of 
various species, but the most fruitful material consisted of a 
large colony of heavily stylopised Andrena nigrownea, 
established in a grassy bank close to the University Museum, 
which was kept under observation for several years and 
afforded us a rich collection. The following observations on 
the structure and history of the parasite are based on the 
bees taken from this colony. 


1. Norges oN THE PARASITE. 


- Stylops melittez, like all the Strepsiptera, has the sexes 
separate; the adult malJe’ being an extremely active-winged 
insect, showing a possible relationship to the Coleoptera 
(Pl. 33, fig. 8), while the adult female is a degenerate grub- 
like creature which remains permanently inside the body of 
the bee in which it. is parasitic. The male, before hatching 
out as.a winged insect, also develops inside the abdomen of 
the bee, undergoing its larval stages and pupation in this 
situation. ‘Ihe male pupa, in fact, closely resembles the 
adult female parasite, and protrudes a little cap between 
the segments of the bee’s abdomen to the exterior, which 
closely resembles the head of the female parasite which is 
similarly protruded. When the adult-winged male emerges 
from its pupa and from the bee, it pushes off the protruded 
cap of the pupa and leaves the old empty pupal case inside 
the abdomen of the bee where it can often be recognised as 
a hollow cavity communicating with the exterior. Since the 
male Stylops emerge from their pupa soon after the bees 


STUDIES. IN THE EXPERIMENTAL ANALYSIS OF SEX. 437 


come out of their burrows for the first time in spring, the 
empty pupal case of the male Stylops is much more 
frequently met with in the bee than the pupa with its cap on 
containing the male itself. In the case of the female 
Stylops, it is quite different, since she remains permanently 
in the bee, and may be found all through the spring and early 
summer with her body distended with developing eggs or 
larvee. 

The appearance of a bee’s abdomen with three female 
Stylops in it is shown on PI. 32, fig. 1. ‘The heads of the 
female Stylops are seen protruding between the segments 
of the abdomen, while the rest of their bodies are hidden 
inside the abdomen of the bee. If we dissect out the whole 
of the female parasite from the bee (Pl. 32, fig. 2), we see 
that its body consists of two chief parts, a hard yellow 
chitinous portion, the cephalo-thorax, which is protruded to 
the exterior, and a soft white segmented portion or abdomen, 
which is buried in the bee’s abdomen, occupying a very large 
proportion of the hemocoel or body cavity of the bee. 

There was for a long time some doubt as to whether the 
hard protruded chitinous part was really the head or the tail 
end of the Stylops, but it is now quite certain that it is 
really the head end, as it is possible to prove by means of a 
median sagittal section, such as is shown on PI. 382, fig. 3, that 
the brain (gn. 1) and the subcesophageal ganglion (gn. 2) are 
situated there. We can also establish from the position of 
the gangha and the ventral nerve cord (n), that the surface 
of the cephalo-thorax exposed to view when the parasite is 
im sit on the bee is the ventral surface. 

Looking at this ventral surface of the cephalo-thorax we 
can see (PI. 32, fig. 2), certain tubercles and slits which are 
of importance. A median tubercle near the extreme anterior 
end is placed just behind a minute pore which represents the 
mouth (m). On each side of the mouth are a pair of lateral 
tubercles which probably represent the atrophied man- 
dibles (md). Behind these tubercles is a bow-shaped slit (0), 
the opening of the brood passage, by which the larve are 


438 GEOFFREY SMITH AND A. H. HAMM. 


finally liberated. This slit at maturity communicates with a 
superficially-placed passage running all along the ventral 
surface of the body. The lateral limits of this passage are 
indicated by the ridges (7) showninthe figure. The relations 
of this brood passage are shown in the sections on Pl. 32, 
figs. 3 and 4 (b). It will be seen that the passage is simply 
a space between the chitinous integument and the body wall, 
the epithelium of which (ep) is peculiarly modified by the 
cells being produced into spiny processes. The function of 
this curiously modified epithelium of the brood-passage has 
never been suggested, but possibly its rough surface is 
convenient for the young larve to travel over in their journey 
to the exit at o. 

The eggs complete their development in the body of the 
female, which at maturity consists of a mere sack contain- 
ing them, all the other organs being reduced to vestiges. An 
idea of the appearance of a fully ripe female with the body 
full of active larvee and embryos is given by the photograph 
on Pl. 33, fig. 6. The fully-developed larvee reach the 
brood passage from the body cavity of the parent by means 
of five peculiar trumpet-like invaginations which lead into the 
brood passage and acquire openings into the body of the 
parent at the moment that the larve are ready to escape. 
The five trumpets are shown attached to the epithelium of 
the brood passage, the rest of the body having been re- 
moved, in the photograph on Pl. 33, fig. 7, and a trumpet 
with its opening into the brood passage is shown in the 
sections on Pl. 32, figs. 3 and 4. 

We may make some mention of the other organs of the 
body of the female parasite. The skin, except where the 
epithelium of the brood passage is specially modified, is ex- 
ceedingly thin, and the nourishment of the body must take 
place by absorption through this thin skin. There are no 
special cells for seizing on or elaborating a special kind of 
food either in the skin or elsewhere, so that we may suppose 
that the hemolymph of the bee affords a ready-made medium 
which supplies the parasite with all that is requisite. This is in 


STUDIES IN THE EXPERIMENTAL ANALYSIS OF SEX. 439 


perhaps significant contrast to Sacculina, where a special 
and highly-developed system of roots ramifies through the body 
of the host, and is engaged in seizing on a special constituent 
of the food, viz. fat. 

The alimentary canal (Pl. 32, figs. 3 and 4, g), of the 
Stylops is clearly recognisable, but in a quite degenerate and 
useless state. There is a minute mouth opening (m), and an 
equally minute anus at the hind extremity, but the lumen of 
the gut through the body is obliterated. The whole apparatus 
is obviously functionless. There is a peculiar mass of cells 
with dark staining nuclei and eosinophilous cytoplasm (PI. 32, 
fig. 3, c) situated in the ventral part of the cephalo-thorax ; 
the function of these cells is unknown, but possibly they are 
of the nature of supporting-cells analogous to cartilage, to 
stiffen the exposed region of the body. Dorsally to the gut 
the remains of the dorsal blood-vessel or heart (Pl. 32, figs. 
3 and 4) can be recognised. The nervous system consists of 
three ganghonic masses (Pl. 32, fig. 3, gn 1, 2, and 3), the 
dorsally-situated ganglion, (gn 1), being connected by means 
of a thin commissure round the cesophagus with a large 
ganglion in the thoracic part of the cephalo-thorax (gn 2). 
From this ganglion a thin, nervous strand (n), representing 
the ventral nerve cord, passes toa third ganglion (gv 3) in the 
abdomen, and from this a thin filament passes away, but no 
other ganglia could be found in the adult female. 

Of great importance for the economy of the parasite is the 
tracheal system. ‘There are two main tracheal trunks which 
open on conspicuous tubercles (Pl. 32, fig. 2, tr.) on each side 
of the cephalo-thorax. These two trachez (see Pl. 32 and 
33, figs. 4 and 7, tr) pass right through the body, giving off 
numerous branches, which ramify among the developing eggs 
and supply them with oxyyen from the outside, quite inde- 
pendently of the bee. We may note again a contrast between 
Stylops and Sacculina here, since the developing Saccu- 
lina roots have to obtain their oxygen from the blood of the 
crab. 

The above account of the structure of the female parasite 


4.4.0 GEOFFREY SMITH AND A. H. HAMM. 


is sufficient to give an idea of its mode of life and nourish- 
ment; for further details the reader may refer to the papers 
of vy. Siebold (2), Pierce (11) Nassonow (10) and Brues (8). 
There are certain points relative to the life-history of the 
parasite which remain obscure but suggest features of great 
interest. It is known that the larve, which are called 
Triungulins and have the form shown in ventral view on 
Pl. 32, fig. 5, leave the body of the parent by means of the 
opening of the brood passage, and: find their way on to 
flowers visited by the bee. They then clamber on to another 
bee which visits the flower, and clinging on to its hairs are 
carried back to the burrow, where they ultimately enter the 
cells and infect the next generation. At exactly what stage 
they enter the young bee larve is not known, but it is 
presumably at an early stage of development. 

The really obscure part of the life-history concerns the 
mode of development of the eggs, and the question, how the 
female parasite is fertilised by the male, if, indeed, fertilisa- 
tion ever takes place. 

In our account of the structure of the adult female parasite 
it was shown that the portion of the body protruded to the 
exterior is the head, and not the tail, as certain authors have 
supposed. Apart from the anomaly of fertilisation taking 
place through the head end of an insect, we have been unable 
to find any opening or organ such as a spermatheca for the 
entrance or reception of the spermatozoa of the male, either 
on the protruded cephalo-thorax or on the rest of the body 
inside the bee. It has been suggested by some authors (16) 
that fertilisation may take place through the opening of the 
brood passage, but this appears to us improbable, as there is 
no means of entrance from the brood passage into the body of 
the parasite, the trumpet-like invaginations being completely 
closed until shortly before the larve are ready to emerge. 
It has also been suggested that fertilisation might take place 
through the mouth and alimentary canal, but this rather 
extravagant suggestion is not supported by the actual state 
of the alimentary canal, which appears in sections as a con- 


STUDIES IN THE EXPERIMENTAL ANALYSIS OF SEX. 441 


tinuous narrow tube passing through the body without any 
outlet into the body cavity where the eggs are contained. 

It has already been suggested that the eggs of Strepsiptera 
may develop parthenogenetically in certain cases. Thus Brues 
(8), describing the oogenesis in Acrochismus wheeleri 
(Pierce), a Xenid parasite on the wasps Polistes, contends 
that the eggs develop parthenogenetically, the second polar 
body re-entering the egg and fusing with the egg nucleus. 
Since Brues, however, did not follow polar-body formation his 
evidence is incomplete. 

Dr. R. C. L. Perkins (5) has expressed the view that the 
parasites of the bee Halictus must be parthenogenetic, at 
least in certain cases, as when the eggs begin to develop it is 
impossible that a male could have fertilised the female. Dr. 
Perkins also has kindly informed us that out of 500-1000 
specimens of this parasite seen by him only one or two were 
males, so that evidently the males are on the point of dis- 
appearance. ‘Taking these facts in conjunction with the 
practical impossibility, as it seems to us, of the spermatozoa 
in Stylops ever entering the body of the female at the time 
the eggs begin to develop, we are led to the conclusion that 
development is always parthenogenetic in the Stylopide. If 
this is correct, it follows that the active winged males are 
useless for the propagation of the species, a conclusion which 
few would accept without misgiving. 

Various observers have attempted to observe the act of 
copulation in the Strepsiptera, but mostly with no or very 
equivocal success. Pierce (11), in his valuable revision of the 
Strepsiptera, remarks—‘‘ That the female must be fertilised 
can hardly be doubted, and yet the nature of the act and the. 
fact itself has been but slightly proven.’ Observations on 
the behaviour of the male by Saunders and Crawford (11) 
show that he is actively attracted by the female Stylops, or, 
at any rate, by the bee on which a female Stylops is situated, 
and that he runs about on the body of the bee, evincing the 
greatest excitement. F. Muir (17) has made similar observa- 
tions, but is uncertain how copulation takes place. 

vou. 60, PART 3.—NEW SERIES. 31 


4.49 GEOFFREY SMITH AND A. H. HAMM, 


The following account, compiled by one of us (A. H.) from 
notes on the habits of the male, may be given: 

At the end of April and beginning of May, 1912, the male 
Stylops was not uncommon in the vicinity of the colony of 
Andrena nigroznea, being seen on the wing at mid-day in 
sunshiny weather. The singular flight of the male Stylops has 
often been seen and commented upon by many observers. When 
once recognised they can never be forgotten, the peculiar 
fight and milky-white wings at once distinguishing them 
from all other insects. None were observed actually flying 
over the burrows of the bee, and nearly all, when first seen, 
were some 10 or 15 feet from the ground, sufficiently high to 
prevent some of them being captured. All the specimens 
caught were boxed alive for further observation or experi- 
ment. On three occasions a male Stylops was, immediately 
after capture, introduced into a large glass-bottomed box 
containing a freshly caught bee, infected with one or more 
female Stylops. In each case the behaviour of the male was 
identically the same. The male Stylops, directly it was intro- 
duced into the box, fluttered on to the bee, and quickly ran 
over its body to where the head of the female Stylops was 
everted between the bee’s abdominal segments. At this time 
the male is rapidly vibrating its wings and protruding its 
last two or three apical segments, which are long and tapering 
like an ovipositor. The insect is thus quite unlike a dried 
specimen in which these segments are invariably telescoped 
into the body. Actual pairing did not occur on any of the 
three occasions. The bee, which had been resting quietly in 
the box, became extremely restless as soon as the male Stylops 
flew on to it, and kept repeatedly climbing to the top of the 
box and then suddenly dropping, as if in the endeavour to 
rid itself of its unwelcome rider. After about ten or fifteen 
minutes of ceaseless running to and fro over the bee, the 
male Stylops voluntarily quitted the Andrena, but still 
continued to run and vibrate its wings for about two hours 
longer, after which time it dropped apparently exhausted, 
and died shortly afterwards. The other males which had 


STUDIES IN THE EXPERIMENTAL ANALYSIS OF SEX, 443 


been boxed also continued to flutter and to vibrate their 
wings ceaselessly, until, in about two hours, all movement 
came to an end, and they were apparently dead. It will be 
seen from this account that the male Stylops, besides retain- 
ing its structure and activity unimpaired, also possesses the 
instinct for attaching itself to the bee, but in no case has 
actual copulation been observed. 

The copulatory apparatus of the male Stylops (fig. 8, a 
and B) consists of a hollow chitinous penis, shaped rather like 
a pick-axe, which can be everted on a hinge, but when with- 
drawn is covered by a grooved sheath. Although the penis 
is a slender organ it has a sharp point, and might be used 
for hypodermic injection of spermatozoa into the body of 
the female. It is, however, very difficult to see how the 
- Injection of spermatozoa into the body of the female would 
result in fertilisation, because the eggs never quit the ovary, 
and are always completely surrounded by follicular epi- 
thelium, which would prevent the access of spermatozoa 
casually injected into the body-cavity. ‘lhe male does not 
show any ‘trace of degeneracy in its internal reproductive 
organs, the vesicule seminales being crowded with active 
spermatozoa. 

In several females the eggs have been found in an early 
stage of development, the features of which strongly confirm 
our suspicion that development is parthenogenetic. In these 
cases all the developing eggs are at approximately the same 
stage of development, exhibiting two, or, in some cases, more 
segmentation nuclei (Pl. 32, fig. 4 a,b/.), while at the periphery 
of the egg a mitotic spindle is observed (p. 1), which invari- 
ably exhibits a single large chromosome and three, or four 
smaller ones, often in process of division. Hach egg is com- 
pletely invested by the follicular epithelium (/). 

Now, it is quite clear that the mitotic spindle inust represent 
the first polar body in process of division. There is, however, 
no trace of a second polar body, which there certainly ought 
to be if a second polar body was given off and fertilisation 
effected in the usual way. 

VoL. 60, PART 3.—NEW SERIES. B1§ 


444, GEOFFREY SMITH AND A. H. HAMM. 


It is possible to explain the appearances in these eggs on 
three suppositions: (1) That no second polar body is formed 
and that the female pronucleus develops parthenogenetically, 
(2) that a second polar body is formed but fuses with the 
female pronucleus, which then develops parthogenetically, or 
(3) that a second polar body is formed but disappears and 
leaves no trace, although the first polar body is still only im 
the metaphase of its mitosis, and that fertilisation is effected 
by a spermatozoon. 

It must be admitted that the third supposition is exceed- 
ingly improbable from what we know of the development of 
any other egg, the entire disappearance of the second polar 
body before the completion of the mitosis of the first polar 
body being altogether unknown. 

During the present year (1914) the burrows of the colony - 
of A. nigrownea were carefully watched at the beginning 
of the season, and the first bee carrying a female Stylops 
was captured, and the parasite preserved and its eggs 
examined by serial sections. These eggs were found to be 
already in a fairly late segmentation stage, which is strong 
presumptive evidence that development had already begun 
before the bee had left its burrow, and before the Stylops 
would have had a chance of being fertilised. 

It will therefore be seen that the cumulative evidences in 
favour of the parthenogenetic development of the eggs of 
Stylops are exceedingly strong, consisting in the following 
main heads: (1) There is no opening or apparatus in the 
female adapted for conveying the spermatozoa to the eggs; 
(2) the eggs remain throughout their development encased in 
the follicular epithelium of the ovary, so that access to them 
by spermatozoa which had entered the body cavity is very 
difficult to imagine; (3) parthenogenesis must occur as a 
normal rule in the parasites of Halictus; (4) the known 
stages in the polar-body formation of Stylops are inconsistent 
with the view that fertilisation by a spermatozoon has been 
effected; (5) actual copulation by the male has never been 
adequately observed. 


STUDIES IN THE EXPERIMENTAL ANALYSIS OF SEX. 445 


We may finally note that in a large number of colonies of 
infected Andrena it would appear that the male parasite is 
very much scarcer than the female, and in certain cases may 
have almost entirely died out. This rule is, however, not 
invariable, and in the case of Xenos, Wheeler (12) records 
an actual preponderance of males over females. 

The most difficult thing to explain, on the assumption that 
the males are now useless, is their marked instinct for 
clambering on the body of the bee when infected by a female 
Stylops. This surely indicates that at some time pairing 
took place in this situation. In explanation of this it must 
be remembered that under present conditions the female 
Stylops never develops beyond what is really a larval or rather 
pupal stage, and that at some previous period in the history 
of the parasite it is certain that the female developed further 
and probably issued from the bee as a fully-formed and pos- 
sibly winged imago. It may well have been that when this 
was the case the males waited for the emergence of the 
females and paired with them directly they issued from the 
bee, and that they still retain the instinct of attaching them- 
selves to the infected bees and waiting for the appearance of 
the female imago, an appearance which now is never realised. 
For an explanation we are forced to fall back upon some such 
hypothesis as this, since no transitional forms are known to 
exist in nature which might show the intermediate steps by 
which the endo-parasitic habit and arrested development of 
the female parasite have been acquired. 

If we are correct in supposing that the males of the 
Stylopide are useless and that development is invariably 
parthenogenetic, it may be pointed out that such a condition 
of affairs is not altogether without parallel according to the 
results of recent researches. In the Rhizocephala (18) 
degenerate Cirripedes parasitic on various Decapod crustacea, 
we find that in certain genera, e.g. Sacculina and 
Peltogaster, the parasites are hermaphrodites which propa- 
gate themselves by a continuous round of self-fertilization. 
Nevertheless, degenerate larval males are found, often in 


4.4.6 GEOFFREY SMITH AND A. H. HAMM. 


numbers up to twenty or thirty, fixed round the mantle 
opening of about 80 per cent. of young Sacculine, and 
these larval males are entirely degenerate and useless in the re- 
production of the species. Other genera of the Rhizocephala, 
e.g. Sylon, are purely female, and reproduce entirely by 
parthenogenesis, and have thus got rid of the marked dis- 
harmony (to borrow Metchnikoff’s term) which characterises 
the sexual economy of Sacculina and Peltogaster. 
Another instance of sexual disharmony has been described by 
Maupas in certain free-living Nematodes of the genus 
Rhabditis, where again the majority of the individuals are 
self-fertilizing hermaphrodites, but a certain number of males 
are still produced which are useless for reproduction. In 
other species these males have been entirely eliminated. 

These instances of undoubted disharmony, or imperfection 
of adaptation in sexual economy, should make us pause 
before we assume that the males of the Stylopide must still 
be functional in those species in which they occur in 
considerable numbers. 


2. Tur EFFECT oF THE PARASITES ON THEIR Hosts. 


We will consider first the effect of the parasites on the 
internal reproductive organs of the hosts. 

In the case of twenty female Andrena nigrownea, of 
which four carried male Stylops puparia and sixteen 
female Stylops, it was found in every case that the ovary 
was very greatly reduced in size and was incapable of pro- | 
ducing mature ova. The appearance of such reduced ovaries, 
as compared with that of normal ovaries, is shown on 
Pl. 33, figs. 9 and 10. 

No marked difference was observed in cases where male 
parasites were present from those in which female parasites 
were concerned. 

It was found, therefore, without exception, that stylopisa- 
tion brought about a reduction in size of the ovary and 
complete sterility. 


STUDIES IN THE EXPERIMENTAL ANALYSIS OF SEX. 447 


This result is in agreement with the observations of 
Pérez (1). 

In the case of fifteen male A. nigroznea, of which four 
carried male puparia, ten female Stylops, and one had a 
male and a female parasite, it could not be observed that the 
presence of the parasites in any case had exerted any effect 
on the development of the testes or their ducts. The figures 
given on Pl. 33, figs. 11 and 12, show the male reproduc- 
tive apparatus in normal and stylopised males, and it will be 
seen that there is no reduction in size in the stylopised 
individual. In order to test whether the testes of the 
stylopised males produce ripe spermatozoa, it was found 
necessary to examine bees early in the year soon after their 
emergence from the burrows, since both normal and stylopised 
individuals later in the year were generally found with the 
vesicules empty of spermatozoa. 

If, however, stylopised males are taken early in the year, 
it is possible to show that their testes and ducts are in the 
same condition as normal males, and that abundance of ripe 
spermatozoa are present in the large vesicles which lead from 
the three testicular tubes on each side into the vas deferens. 
The section on PI, 33, fig. 18, through the three testicular 
tubes and the vesicle of one side of a stylopised bee, shows 
the presence of abundant spermatozoa in the vesicle. ‘The 
testicular tubes in this section are more or less empty with a 
rather ragged degenerate epithelium, but this appearance is 
due to the fact that spermatogenesis is over, and is equally 
to be noticed in normal males. 

This absence of any effect of the stylopisation on the male 
internal organs is on the whole in agreement with what other 
authors have found, though Pérez records some cases of a 
one-sided damage being inflicted on the testes by the 
parasite, and Theobald (7) is inclined to believe that the 
damage may be considerable. Perkins (5), on the other 
hand, both for males and females, tends to minimise the 
effect of the parasites on the internal organs, and records the 
fact that stylopised males have been taken in copula with 


448 GEOFFREY SMITH AND A. H. HAMM. 


stylopised females, showing that the sexual instincts may still 
persist in stylopised individuals. 

There can be no doubt from our own observations, and 
from the general consensus of opinion, that the female bees 
suffer far more serious reduction in their ovaries than the 
male bees experience in the case of their internal reproductive 
organs. 

The reason for this difference in effect appears to us fairly 
obvious. The testes of the male bee are exceedingly minute 
structures, about a hundredth part of the ovaries in size, and 
they, therefore, require a small fraction of the nutriment 
which is demanded by the ovaries. The presence of the 
parasite, therefore, while cutting off a large part of the 
necessary nutriment from the ovaries, does not succeed in 
depriving the minute testes of the small amount of nourishment 
which they require, and hence they are able to attain their 
normal development, though the ovaries are seriously 
starved. 

We may now proceed to the effect on the external 
characters. 

The males and females of A. nigroznea differ, firstly 
from one another in their external genital armature which 
consists of a complicated copulatory apparatus in the males 
and of an ovipositor in the female. After examining a long 
series of stylopised males and females, we are unable to find 
any reduction or abnormality in these structures as the result 
of stylopisation. In this respect we are not in complete 
agreement with Pérez (1), who reports a marked reduction in 
the development of the genital armature as the result of 
stylopisation in several cases. We do not doubt that this is 
correct, but the effect in any case is a comparatively slight 
one, and there is never the slightest difficulty in at once 
recognising the male and female stylopised individuals by the 
genital armature which is always typically developed, though 
it may be in certain cases somewhat reduced in size. 

Another important secondary sexual character, affecting 
the hard chitinous structure of the bee is found in the 


, 


STUDIES IN THE EXPERIMENTAL ANALYSIS OF SEX. 449 


antennz, which are 13-jointed in the males and 12-jointed in 
the females, of all Andrena (see Pl. 35, figs. 14-15). 

This character in our experience and in the experience of 
all other observers is quite unaffected by stylopisation, the 
infected individuals always having the number of joints 
typical of their sex. Pérez again is of opinion that in certain 
cases the relative length of the joints in infected individuals 
is slightly altered, but here the effect is admittedly very 
slight indeed, and the figures given to illustrate the effect do 
not appear to us to bear out the contention. 

A marked distinction between the sexes of all Andrena 
is found in the structure of the femur and tibia of the hind 
legs, which are thin and not markedly hairy in the male, but 
in the female are greatly enlarged to form the scopa or pollen- 
collecting apparatus. The condition of the normal male 
(A. nigroenea) is shown on Pl. 34, fig. 16, of the infected 
wale, in Pl. 34, fig. 17, of the normal female in figs. 18 and 
19, and of two stylopised females in figs. 20 and 21. 

These figures bring out the fact, which we have found 
invariably in A. nigrownea, that as the result of stylopisa- 
tion the male does not acquire in any degree the scopa of the 
female, while the scopa of the female is always to some extent 
reduced in size by the action of stylopisation. 

We also find that stylopised females never carry any pollen 
on their scope, in marked distinction from the normal 
females, the majority of which are found with their scope 
plastered with pollen as shown in fig. 18. The stylopised 
females have evidently entirely lost the instinct for collecting 
pollen, though they still continue to visit the burrows. Of 
the hundred or so stylopised females examined not a single 
individual had pollen on it, but we are not in a position to say 
that such a thing cannot ever occur, as there is certainly a 
great degree of variation in the intensity of the effects of 
stylopisation in different individuals and species. 

We have found that stylopisation affects the punctuation of 
the chitin of the abdomen to a certain extent, though the 
effect can only be appreciated by examining a good series of 


4.50 GEOFFREY SMITH AND A. H. HAMM. 


normal and infected individuals together. If we look at a 
series of normal males and females together, we shall notice 
that the males reflect the light more brightly than the females, 
owing chiefly to the less degree of punctuation and hairiness 
of the abdomen. } 

The stylopised males, on the other hand, tend to have the 
abdomen dull, very much as in the female, and this appears 
to be due to the deeper and more frequent punctuation on 
the abdomen, and not to a greater hairiness. The stylopised 
females do not appear to be affected either in punctuation or 
hairiness. 

We have now dealt with the most important secondary 
sexual characters which concern the structure of the hard 
chitinous parts, and it will be recognised that the effect 
exerted by stylopisation is small and consists in a reduction 
of certain sexual characters, and never in a real assumption of 
characters proper to the opposite sex. The most constant 
and striking effect is the reduction of the scopa in the female 
and the loss of the instinct for collecting pollen. Comparing 
these effects with the effect of Sacculina on the secondary 
sexual characters of Inachus (14), it will be admitted that 
the complete inversion suffered by the males of Inachus has 
no parallel in the bees modified by stylopisation, so far as 
structure is concerned. 

‘here remains for consideration, however, a very important 
character which may undergo a very complete inversion as 
the result of stylopisation. In certain species of Andrena, 
e.g. A. chrysosceles and A. labialis, the female has the 
ordinary black clypeus, but the male has a yellow or white 
one (see Pl. 35, figs. 22, 23, 25 and 27). Pérez discovered 
that as the result of stylopisation the female might assume 
completely or in part the coloured clypeus of the male (see 
Pl. 35, fig. 26), while the male might undergo considerable 
retrogression and lose a great part of the yellow colouration 
(fig. 28). Pérez makes it clear and has personally informed 
us that this remarkable effect is by no means invariable and 
that very frequently stylopised males and females of A. 


STUDIES IN THE EXPERIMENTAL ANALYSIS OF SEX. 451 


labialis may exhibit their proper clypeus colouration with- 
out any modification. It is therefore necessary in this case 
to be able to examine a long series of infected individuals, and 
it would appear that other observers have not been fortunate 
in securing such aseries because no confirmation of Monsieur 
Pérez’s discovery has hitherto been published. We are 
fortunately in a position to produce confirmatory evidence in 
the case of A. chrysosceles. PI. 35, fig. 22, depicts the 
head of a normal female specimen, while fig. 23 shows the 
head of a normal male with the coloured clypeus. Fig. 24 
shows the head of a stylopised female, taken at Sandford 
near Oxford by one of us (A. H. H.), which has developed 
the coloured clypeus of the male in a typical manner. ‘This 
specimen was parasitised by a male Stylops. It may be stated 
that long series of normal A. labialis and chrysosceles 
have been examined, and that in no case has any assumption 
of the coloured clypeus by the female or vice versd been 
observed apart from the effects of stylopisation. 

This acquisition of clypeus colouration by the female is by 
far the most striking alteration brought about: by stylopisa- 
tion, because it really amounts to a true acquisition of a 
positive character belonging to the opposite sex, and not to a 
mere negative suppression of characters that should normally 
be developed. ‘lhe alterations in the hard parts and the 
blackening of the clypeus in the males, can all be interpreted 
as mere negative suppressions, but the acquisition of the 
yellow clypeus by the stylopised females is in a different 
category. In the majority of Andrena the clypeus of both 
sexes is black, so that the loss of the yellow colour in the 
stylopised males of A. labialis may be considered a mere 
repression, but not so the acquisition of the yellow colour by 
the female. 

Before discussing these results there are two points 
which merit attention. Since the Stylops parasites are 
of separate sexes, it appeared possible that the sex of the 
parasite might have an important imfluence on the effect 
exerted upon the bee. For instance it might be found that 


452 GEOFFREY SMITH AND A. H. HAMM. 


only females stylopised by male Stylops could develop the 
white clypeus characteristic of the normal male bee. Of 
course if such a contention could be proved it would have a 
most important bearing on the theoretical interpretation of 
how the effect is brought about. It would suggest, in fact, 
that the male Stylops exerted a specific male influence on 
the bee, and the female Stylops a specific female influence. 

In answer to our inquiries Monsieur J. Pérez has kindly 
sent us some of his specimens of A. labialis, which satisfac- 
torily settle this point. Among these specimens there are 
two female A. labialis parasitised by female Stylops 
which show a considerable amount of white colour on the 
clypeus, and there is also a male A. labialis parasitised by 
a male Stylops, which shows avery marked reduction of the 
white colour on the face. 

Dr. R. C. L. Perkins has also sent us two valuable instances 
bearing on this question, viz., two females of A. labialis, 
with their faces coloured as in the male, both of which are 
parasitised by a single female Stylops. These instances are 
abundantly sufficient to demolish the view that the sex of 
the parasite has any determining influence on the effect pro- 
duced on the secondary sexual characters. It is probably 
true that the presence of a male Stylops has a more generally 
damaging effect on the bee, but there is no evidence of the 
male parasite exciting a specifically male effect and of the 
female exciting a female effect upon the host. 

The theoretical importance of this fact will be given its 
due weight in a later paragraph. 

The second point to which attention may be called is the 
great amount of variation exhibited by Andrena and other 
insects in their reaction to strepsipterous parasites. This 
variation does not only subsist as between different species of 
hosts, but also as between different individuals of the same 
species of host. Wheeler (12), who has made a most detailed 
and exhaustive examination of the effect of Xenos on the 
wasps Polistes came to the conclusion that the parasite 
had no definite effect on the secondary sexual characters of the 


STUDIES IN THE EXPERIMENTAL ANALYSIS OF SEX. 4038 


host, and it is clear that his conclusion is perfectly correct. 
The same may probably be said of the effect of Eleuchus 
on the Homoptera. 

We have already seen that no other observer has apparently 
described the effect of Stylops on the clypeus colouration of 
certain Andrena, noticed by Pérez, until we came across 
the case of A. chrysosceles published here. These appa- 
rently contradictory results have led to much confusion, but 
to anyone familiar with the facts of parasite castration in 
other branches of the animal kingdom they will occasion no 
surprise. In cases where the effects of the parasite lead to 
the most startling and complete inversions, as in Sacculina 
on Inachus and Peltogaster on Eupagurus (14), there 
is always a certain small percentage of individuals which 
remain almost completely unaffected by the presence of the 
parasite, while in other cases, such as Sacculina on 
Carcinus, the effect of the secondary characters is often 
nil, and never consists in more than a slight approximation 
of the male to the female type. Such variations then, 
whether due to differences in individual or specific suscepti- 
bility or to some casual event in the history of the disease, 
are to be expected, and should warn the observer not to draw 
conclusions without examining a long series of infected 
individuals. 


(5) Discusston or Resvtrs. 


If we compare the effects of stylopisation with those of 
Sacculina on Inachus, we shall recognise that they are 
much slighter and less radical in the former than in the 
latter case. Thus in the internal reproductive organs stylopi- 
sation only causes a reduction in size of the ovary, and 
preveuts ripe ova from being produced, while the testes of 
the male are practically unaffected. ‘The presence of Saccu- 
lina, on the other hand (18), may occasion the complete 
destruction of all the internal genital organs, with the excep- 
tion of some remnants of, germinal epithelium, while infected 


A454, GEOFFREY SMITH AND A. H. HAMM. 


male crabs may be induced by the Sacculina to produce 
ripe ova in their testes. 

Correspondingly the effects of stylopisation on the second- 
ary sexual characters are comparatively slight, and amount 
to no more than a reduction of certain characters, such as 
the scopa of the female, while in the case of Sacculina the 
whole morphological structure of the male crab may be 
eutirely converted to the female state. 

In one case, however, that of the colour of the clypeus, the 
female bee when stylopised may assume the positive male 
character. 

It is clear, therefore, that the reaction of the bee to the 
Stylops does not go so far as that of the crab to Sacculina, 
either internally or externally, and whereas in sacculinisation 
we are forced to the conclusion that the Sacculina exerts 
an active feminising influence on both sexes of infected crabs, 
in the case of stylopisation it is sufficient to hold that the 
action here consists merely in an arrest of development inci- 
dent on the cutting off of a certain amount of nutriment from 
the ovaries, and to a less extent from the testes. In sacculi- 
nisation we have argued (18) that the Sacculina roots, by 
demanding a certain type of nutriment, viz. fat, stimulate a 
certain type of metabolism in the crab, which is characteristic 
of the adult female when maturing its ovaries, and that this 
internal change of metabolism brings in its train all the 
deep-seated changes in the internal and external genital 
structures. he Stylops, on the other hand, does not, 
initiate such wide-spreading changes; it stops short at 
abstracting a certain amount of nutriment from the blood, 
and so causes a merely quantitative alteration in the develop- 
ment of the internal and external genital organs. It has 
been pointed out that the Stylops parasite does not appear to 
be taking up any special nutriment from the blood of the 
host, but rather to receive the nutriment from the blood, 
ready-made, and thus it would not be expected to stimulate 
any special line of metabolic changes. Further, the Stylops 
always receives its oxygen from the outside air, while the 


STUDIES IN THE EXPERIMENTAL ANALYSIS OF SEX. 455 


Sacculina roots are living amerobically, and must split 
off their oxygen from the blood of the host. This implies a 
more intimate relation between the metabolism of host and 
parasite in the case of Sacculina. 

We thus see that the effects of stylopisation may be inter- 
preted as due toa merely quantitative abstraction of nutriment 
normally destined for the reproductive glands, and that this 
abstraction brings in its train a reduction in size of these 
glands, especially the ovaries, and a corresponding reduction 
in the development of some of the secondary sexual 
characters. 

It may be urged, however, that this explanation does not 
apply to the assumption of the yellow clypeus by the female, 
as this is a positive male character. We have an excellent 
analogy for this case in sterile female birds (15) which, either 
as the result of operative ovariotomy or else of ovarian 
disease and atrophy, may assume male plumage to a very 
marked extent. The assumption of the yellow clypeus by 
stylopised female bees with reduced ovaries seems to us 
exactly parallel to the assumption of cock’s plumage by 
female birds with ovaries either atrophied or removed by 
operation. It seems that in both cases the mere atrophy or 
suppression of the ovary is sufficient in both cases to induce 
the development of certain male characters in the colouration. 

It is not necessary, therefore, to ascribe a special mascu- 
linising influence to the Stylops parasite in the same sense as 
one must ascribe an active feminising influence to the 
Sacculina roots. The masculinising influence resides in 
the female bee itself, just as in the female bird, and is 
called into activity by the mere suppression of the ovarian 
function. 

In this manner we may look upon the acquisition of the 
yellow clypeus by the female as due to the same cause as the 
other alterations brought about by stylopisation, viz. to 
the mere quantitative cutting off of nutriment from the ovary, 
and not to any specific or qualitative action of the parasite, 
as in the case of Sacculina on Inachus. 


456 GEOFFREY SMITH AND A. H. HAMM. 


It appeared possible to us at one time that a qualitative 
action might account for the assumption by the female of the 
male clypeus in the following manner: It might be possible 
that this assumption by the female only followed when she was 
parasitised by a male Stylops, which might exert a specific 
male influence on the host. This supposition is not confirmed 
by the facts, as the presence of a female Stylops can equally 
bring about the assumption of the male clypeus by the 
female. 

The fact that the sex of the parasite has no influence on 
the effect exerted on the host is in reality a strong confuta- 
tion of the idea that the effects of parasitic castration are due 
to a specific internal secretion produced by the parasite. 
For if such a secretion were produced by the parasite, we 
should certainly expect that the female parasite would 
produce a female internal secretion and the male a male one, 
whereas we find that the parasites of both sex exert a similar 
effect. This is perfectly intelligible if we suppose that the 
parasites of both sexes act on the host merely by cutting off a 
certain amount of nutriment from the gonad, a process which 
reacts more profoundly on the female than on the male, owing 
to the larger size of the ovaries and the larger demand made 
by them on the nutriment in the body. 

The peculiarity of the case of Sacculina consists in the 
fact that the roots of the parasite happen to demand an 
excessive supply of the same sort of nutriment which the 
ovary of a normal female crab requires, and so bring 
about a series of profound metabolic changes leading to the 
feminisation of the host. 

The result of all the above considerations is to show that a 
parasite may act on the sexual characters of its host in two 
ways. Firstly, it may simply take up a certain amount of 
nutriment from the blood so as to deprive the gonad of its 
proper supply and lead to its partial atrophy, but without 
bringing about any deep-seated alteration in the metabolism 
or stimulating any special set of metabolic changes. ‘The 
abstraction of this nutriment, by depriving the blood of its 


STUDIES IN THE EXPERIMENTAL ANALYSIS OF SEX. 457 


proper supply, may lead to the atrophy of the gonad and to 
the reduction of the secondary sexual characters, and even to 
an inversion of certain secondary sexual characters, e. g. the 
colour of the clypeus in Andrena. Secondly, as in the 
ease of Inachus parasitised by Sacculina and of 
Eupagurus parasitised by Peltogaster, the parasite in 
obtaining its food from the blood of the host, may set going 
a special set of metabolic changes in the host, and these 
changes may result in diverting the metabolism of the host 
to the female state, so that the host assumes female 
characters throughout and the infected male may even be 
induced to produce ova in its testes. This second type is not 
a mere passive inhibition like the first, but an active reaction 
explicable, as has been shown, on the basis of an immunity 
reaction (14). 

Another point which emerges from this study is that the 
bee, Andrena, belongs to the same category as many birds 
(Pheasants, Fowls, Ducks, Ostriches, etc.), (15) in that a mere 
atrophy of the ovary is followed by an assumption of a posi- 
tive male secondary sexual character. In all these animals it 
would appear that the normal ovary exerts an inhibitory 
action preventing certain male characters from emerging, and 
that when the ovarian influence is removed or interfered 
with the stimulus is given for the development of these 
characters. 


SUMMARY. 


(1) From a study of the anatomy and life history of 
Stylops, it appears that despite the existence of active 
winged males, fertilisation cannot occur and development is 
always parthenogenetic. 

(2) The parasite obtains its oxygen from the outside air by 
means of tracheal openings on the cephalo-thorax, and it does 
not possess any special absorptive organs for taking up a 
special kind of food from the host. Nutrition appears to 
take place by simple filtration from the host’s blood through 
the very thin skin of the parasite. 


458 GEOFFREY SMITH AND A. H. HAMM. 


(3) The effect of the parasite on the internal genital organs 
is slight, as compared with the effect of Sacculina on 
Inachus, and leads to a reduction in the size of the ovaries 
to about quarter the normal size, while the testes are usually 
unaffected. The ovaries of stylopised bees never produce 
ripe ova, but the testes generally produce normal ripe 
spermatozoa. 

(4) The effect on the secondary sexual characters is again 
slight as compared with that of Sacculina on Inachus. 
The external gonapophyses are usually unaltered, or they may 
be slightly reduced in size ; the antenne are unaltered. The 
scopa of the parasitised female is generally reduced in size, 
and she never or very rarely collects any pollen. The 
punctuation on the abdomen of the male may be increased. 

(5) The most striking effect occurs in certain species (e. g. 
A. labialis and chrysosceles) in which the male normally 
has a yellow clypeus and the female a black one. Stylopisa- 
tion in those cases may lead to the female assuming a yellow 
clypeus as in the male, while the male may lose the yellow ~ 
and acquire a partially black clypeus. 

This acquisition of the yellow clypeus by the female is the 
only change which can undoubtedly be interpreted as a 
positive acquisition of a secondary sexual character proper to 
the opposite sex. ; 

(6) This effect may be brought about by male or female 
Stylops indifferently, the sex of the parasite having nothing 
to do with the nature of the effect exerted. 

(7) The effects of stylopisation may be ascribed to a merely 
quantitative abstraction of nutriment from the gonad, leading 
to its partial atrophy, and not to a qualitative alteration of 
the metabolism such as is brought about by Sacculina. 
This also applies to the assumption of the yellow clypeus by 
stylopised females, on the analogy of the assumption of male 
plumage by many female birds as the result of simple 
ovariotomy or ovarian atrophy. 


10. 


11. 


12. 


13. 


14. 


15. 


16. 


17. 


STUDIES IN THE EXPERIMENTAL ANALYSIS OF SEX. 459 


LITERATURE. 


. Pérez, J—* Des effets du Parasitisme des Stylops sur les Apiaries 


du Genre Andrena,” ‘ Actes Soc. Linn. Bordeaux,’ vol. xl, 1886. 


. Siebold, Th. v.—‘ Uber Strepsiptera,” ‘Archiv. f. Naturg., Jahrg. 9. 
. Dale, C. W.—* Stylopide,” ‘Entom. Monthly Mag.,’ p. 50, 1892. 


. Eaton, A. E.—‘‘ Notes on Elenchus tenuicornis,” loc. cit., 


p- 250. 


. Perkins, R. C. L.—* Stylopised Bees,” loc. cit., p. 7. 

. Saunders, E.—‘ Elenchus tenuicorius,” loc. cit., p. 249. 

. Theobald, F. V.—“ Stylopised Bees,” loc. cit., p. 40. 

. Brues, C. T.—* A Contribution to our Knowledge of the Stylopide,” 


‘Zool. Jahrb. Abth. f. Anat.,’ xviii, p. 241, 1903. 


. Perkins, R. C. L.—‘ Report of Work of the Experimental Station 


of the Hawaiian Sugar Planters’ Association,’ ‘‘ Leaf. Hoppers 
and their Natural Enemies,” Part iii, “Stylopide,” Honolulu, 
1905. 

Nassonow, N. V.— Xenos rossii, seine Anatomie und Entwick- 
lungsgeschichte,” ‘ Bull. Univ. Warsaw,’ 1892. ‘On the Meta- 
morphosis of the Strepsiptera,” ‘Warsaw University News,’ 
No. 7, 1892. “On the Morphology of Stylops melitte,” 
‘ Warsaw University News,’ Nos. 8 and 9, 1893. 

Pierce.—* A Revision of the Strepsiptera,” ‘Harvard Museum Bull.,’ 
1906. 

Wheeler, W. M.—<‘ The Effects of Castration in Insects,” ‘ Journ. 
of Exp. Zool.,’ vol. viii. 

Smith, G.—*The Rhizocephala,” ‘Naples Monographs, No. 29, 
1906. 


“Studies in the Experimental Analysis of Sex,’ Parts 
7 and 10, ‘ Quart. Journ. Mier. Sci.,’ vols. 57 and 59. 

and Mrs. Haig Thomas.—* Sterile and Hybrid Pheasants,” 
‘ Journ. of Genetics,’ vol. iii, 1913. : 


Meinert, F.—“ Contribution a l’histoire naturelle des Strepsiptéres,” 
‘Bull. de ’Acad. Royale des Sciences de Danemark,’ November, 
1896. 


Muir, F.—“ Notes on Some Fijian Insects,” ‘Report of Work of the 
Experimental Station of Hawaiian Sugar Planters’ Association,’ 
Bulletin No. 2, 1906. 


4.60 GEOFFREY SMITH AND A. H. HAMM. 


EXPLANATION OF PLATES 32-35, 


Illustrating Mr. Geoffrey Smith’s and Mr. A. H. Hamm’s 
“Studies in the Experimental Analysis of Sex.” Part 
II.—* On Stylops and Stylopisation.” 


LETTERING. 


b. Brood-passage. 6. v. Dorsal blood-vessel. bu. Bulb on vas 
deferens. c. Supporting cells. ch. Chitinous investment. ep. Modi- 
fied epithelium of brood-passage. g. Gut. gn.1. Brain. gn. 2. Sub- 
cesophageal ganglion. gn. 3. Abdominal ganglion. m. Mouth. md. 
Mandible. n. Ventral nerve. nu. Nurse-cells. o. Opening of brood- 
passage. ov. Ova. te. Testes tubes. tr. Tracheal tube. vd. Vas 
deferens. v.e. Vasa efferentia. ves. Vesicula seminalis. 


Fig. 1—Abdomen of Andrena nigrow#nea ? with heads of three 
Stylops melittze 2, protruding in natural position. 

Fig. 2—Stylops melittw 2, from ventral surface, removed from 
bee. 

Fig. 3.—Stylops melitte 2, in sagittal section. 

Fig. 4.—Ditto, transverse section through middle of body. 


Fig. 44.—Section through developing egg with first polar body (p. 1), 
blastomeres (b/.), and follicular epithelium (f/f). 


Fig. 5.—Triungulin larve, from brood-passage, ventral view. 


Fig. 6.—Stylops 2, adult, with body full of triungulins ready to 
hatch. 


Fig. 7.—Ditto, triungulins removed, showing five trumpets attached 
to epithelium of brood-passage. Two trachex spread out laterally. 
Fig. 8—Stylops melittx, ¢, dorsal view; abdomen telescoped. 


Fig. 84.—Copulatory apparatus of S. melittez, lateral view. B. 
Penis. 


Fig. 9.—Ovary of normal Andrena nigroenea 2°. 

Fig. 10.—Ovary of Stylopised A. nigro#nea ¢. 

Fig. 11.—Testes and ducts of normal A. nigroxnea @. 

Fig. 12.—Testes and ducts of Stylopised A. nigroxnea J. 

Fig. 13.—Section through testes, ducts, and vesicula of one side of 
Stylopised A. nigroznea, showing ripe sperm in vesicula. 

Fig. 14.—Antenna of normal A. nigroenea 2°. 


Fig. 15.—Antenna of normal A. nigroewnea d. 


STUDIES IN THE EXPERIMENTAL ANALYSIS OF SEX. 461 


Fig. 16.—A. nigrownea, normal ¢. 

Fig. 17.—Ditto, stylopised ¢, showing unaltered legs and gonapo- 
physes. 

Fig. 18.—Ditto, normal ?, carrying pollen on scope. 

Fig. 19.—Ditto, normal ? , showing scope without pollen. 

Fig. 20.—Ditto, stylopised 9, showing reduced scopw, otherwise 
unaltered. 

Fig. 21.—Ditto, stylopised 2, showing reduced scope, otherwise 
unaltered. 

Fig. 22.—Andrena chrysosceles, head of normal 9. 

Fig. 23.—Ditto, head of normal ¢. 

Fig. 24.—Ditto, head of stylopised 9. 

Fig. 25.—Andrena labialis, head of normal 9, after J. Pérez. 
ig. 26.—Ditto, head of stylopised ?, after J. Pérez. 
Fig. 27.—Ditto, head of normal ¢, after J. Pérez. 
ig. 28.—Ditto, head of stylopised ¢, after J. Pérez. 


voL. 60, PART 3.—NEW SERIES. 32 


- 


THE RAT-TRYPANOSOME, I'RYPANOSOMA LEWISI. 


4.63 


The Rat-Trypanosome, Trypanosoma lewisi, 
in its Relation to the Rat-Flea, Cerato- 


phyllus fasciatus. 


By 
E. A. Minchin 


and 
J. D. Thomson. 


With Plates 36—45 and 24 Text-figures. 


ConTEN's. 


Part I—IntTRODUCTORY . 

(1) Personal Narrative . 

(2) Notes on the Flea, Cer Roca IheL ae fageratae 
(a) Anatomy, Methods of Dissection ; 
(b) Notes on the Parasites of the Flea p 
(c) Notes on the Histological Structure of the Suonmedh 

of the Flea é : 
(3) Technique 


Part I].—TuHE DEVELOPMENT OF T. LEWISI IN THE FLEA. 
(1) General Introduction 
(2) The Developmental Series 
A. The Stomach-Phase 
(a) The Extracellular rey pRnGsGrien 
(b) The Intracellular Multiplication of the Tr ypano- 
somes ‘ 
Appendices to ne Siomdck Phase, 
(i) The Occurrence of the Intracellular 
Multiplication 
(ii) The Type of Cell Attacked by the 
Trypanosome 5 ; 


voL. 60, PART 4.—NEW SERIES. 33 


545 


547 


4.64 E. A. MINCHIN AND J. D. THOMSON. 


PAGE 
iii) The Relation of the Trypanosomes to 
the Cells : 548 
(iv) The Effects of the Tey parioseaiee on 
the Epithelial Cells . : 549 
(v) The Relation of the Ty panoeoraes 
Infection as a whole to the Stomach 551 


B. The Migration to the Rectum : ; . 568 

c. The Rectal Phase ; - oD 

(a) The Transition to the Crithidial For . 570 

(b) The Established Rectal Phase : ei eis. 

(3) The Degenerative Series. : . 592 


Appendices to the Meyelopinent: 
(i) Previous Investigations on the Deve- 
lopment of Trypanosoma lewisi 597 
(ii) On the Possibility of the Occurrence 
of Sexual Phenomena in T. lewisi. 603 


Part JII.—ExXxpPEeRIMENTAL STUDY OF THE PROBLEMS OF 


TRANSMISSION AND DEVELOPMENT . = 605: 

(1) Introduction : : : : . 605 

(2) General Problems . : : . 609 

(3) Problems of Special Nate ‘ ; . 659 
BIBLIOGRAPHICAL REFERENCES . : : Be OS: 
DESCRIPTION OF PLATES. E ‘ 3 211682 


PART I. INTRODUCTORY. 


(1) Personat NarRATIVE. 


In this memoir we give a detailed account of investiga- 
tions which have occupied us intermittently, with many 
interruptions, during the past five years. When we 
undertook this task our object in view was to work out as 
fully as possible the life-history and mode of transmission of 
a trypanosome, so that in at least one species of these 
important parasites its relation to the invertebrate. host 
might be as thoroughly known as, for instance, the relation 
of the malarial parasite to the mosquito, thus furnishing a 
standard with which the life-histories of other species of 


THE RAT-TRYPANOSOME, TRYPANOSOMA LEWISI. 465 


trypanosomes might be compared and contrasted as they 
become known. How far we have succeeded in our task 
must be left to our readers to judge. 

The species which we selected as the subject of our in- 
vestigations was Trypanosoma lewisi, the common 
parasite of rats which is apparently of world-wide distribu- 
tion. ‘This species offers many advantages for such a study. 
It is common in London and easily procured when required ; 
its vertebrate host, the rat, is a mammal of small size, the 
domesticated variety of which lives well and breeds rapidly 
in confinement, is inoffensive, and is easily handled; its 
invertebrate host, the rat-flea, is also easy to keep in 
captivity and is extremely prolific, and it is of a size which, 
though it increases to some extent the difficulties of manipu- 
lation, has a great advantage that the material to be searched 
and studied microscopically is confined within a _ small 
compass!; and finally, by no means the least of the 
advantages of working with T. lewisiis the fact that it is 
non-pathogenic to its natural host and cannot live at all in 
human blood. 

Since there is no difficulty whatever in obtaining the 
vertebrate host in abundance, either in the clean (i. e. non- 
infected) or infected condition, our first care was to obtain a 
stock of the invertebrate host, the flea. This we succeeded 
in doing from rats trapped in the open near Mr. Gurney’s 
Laboratory at Sutton Broad, Norfolk. Fifty specimens of 
Ceratophyllus fasciatus were obtained in this way in 
the autumn of 1908, and were kindly identified for us by the 
Hon. N. C. Rothschild, and with these fleas a breeding-cage 
was stocked and a flea-farm started. The cages used were of 
the type used by the Plague Commission, as figured in the 
‘ Journal of Hygiene,’ vol. vi, pl. iv. A rat was kept in the 
cage to feed the fleas, and they were left to themselves. 
Karly in 1909 one of us (H. A. M.) went to Rovigno for some 

1 An advantage which those will appreciate who have had practical 


experience of searching for trypanosomes through many centimeters of 
the digestive tract of the tsetse-fly, for instance. 


466 E. A. MINOHIN AND J. D. THOMSON. 


three months, during which time the fleas were left to breed 
under the care of an assistant, whose duties consisted of 
attending to the rat and replacing it if it fell ill or died. 
When the cage was examined after Haster it was found to be 
swarming with fleas, and onr work began in May, 1909. We 
have worked ever since then with the fertile progeny of the 
original fifty fleas from Norfolk, and have never added 
further to our stock from without. The fleas breed so fast 
that it is often necessary to keep their numbers down, other- , 
wise they take too much blood from the rat and affect its 
health. Fresh breeding-cages have also been started, and 
during the greater part of the time that we have been at 
work we have kept two cages constantly going, one in which 
the fleas are fed always on a clean healthy rat, and another 
in which an infected rat is always kept. We shall refer to 
these two cages as the non-infected and the infected breed- 
ing-cages respectively. As will be shown below, the stock of 
fleas with which we have worked all along was fortunately 
quite free from any natural infection with leptomonad or 
other flagellate parasites. Thus we have been saved from a 
fertile source of confusion and error, since we can be quite 
certain that any flagellates found in our fleas are stages of 
T. lewisi and nothing else. 

Although we cannot claim that in our work we have solved 
completely every problem presented by the transmission of 
the trypanosome and its development in the flea—a result 
which probably no man could achieve in a life-time—we 
think it now fitting that we should publish such results as we 
have obtained, after having done as much as we were able to 
do in the time and under the circumstances. We claim 
at least that we have not jumped to our conclusions; our 
note-books contain not only the records of many experiments, 
but also of the dissection and examination of over 1,600 fleas, 
and we have over 700 drawings of stages of the development 
of the trypanosome, from which those given in this memoir 
are a selection. It would, indeed, have been easier for us to 
have written a plausible and apparently complete account of 


THE RAT-TRYPANOSOME, TRYPANOSOMA LEWISI. 467 


the development of T. lewisi, full of positive statements, 
after one year of our work than it is after five years, during 
which we have been forced by the logic of facts to abandon 
or modify many of our earlier conclusions or beliefs. 

It is our pleasant duty at this point to express our thanks 
to those of our friends and colleagues to whom we are 
indebted for assistance. To Dr. Woodcock and Miss 
Robertson we are grateful for much advice, friendly criticism, 
and valuable suggestions. Our work could not have been 
carried out, certainly not in the time at all events, without 
the assistance of Miss Rhodes, who has not only drawn all 
the illustrations with a skill to which it is quite unnecessary 
for us to draw the reader’s attention (since the figures speak 
for themselves), but has also relieved us of a large part of 
the wearisome drudgery of searching through the microscopic 
preparations. Mr. George Kauffmann has been most helpful 
in every part of the investigation, not only assisting in 
making preparations, examining rats, and other similar 
duties, but more especially in carrying out intelligently and 
enthusiastically all the details of the experiments, in which 
his extraordinary skill and resourcefulness in controlling the 
wayward flea were invaluable. Dr. D. J. Reid has given us 
the benefit of his skill and experience in microphotography. 
From our colleagues of the Lister Institute, Dr. C. J. Martin, 
the Director, and Mr. Bacot, who have been themselves 
engaged in studying the transmission of plague by fleas, we 
have had many valuable hints and help in various ways. ‘To 
each and all of these we desire to express our cordial thanks 
and gratitude. 


(2) Notes on THE FLEA, CERATOPHYLLUS FASCIATUS. 


(a) Anatomy. Methods of Dissection. 


The fleas collected for dissection and examination were 
thrown, or allowed to hop, on to the surface of a small 


468 E. A. MINCHIN AND J. D. THOMSON. 


quantity of salt-citrate solution’ placed in a suitable glass 
capsule. The fleas are quite helpless on the surface of the 
liquid, and each flea that it is required to dissect can be 
picked off the surface of the liquid and transferred to a 
small drop of the same solution on a slide for further 
operation. 

The examinations of the fleas were usually conducted by 
both of us acting in concert. One of us worked with the 
dissecting - microscope, extracted the parts of the flea 
required, placed them on slides, covered them with glass 
slips, and handed them to the other, who proceeded to search 
- them carefully through under a microscope, using dry lenses 
of fairly high magnification (Zeiss D or apochromatic 4 mm.). 
In some cases one of us worked entirely alone, but it is 
difficult for one person to carry out satisfactorily both the 
dissection and examination of the flea; the various parts of 
the digestive tract often require prolonged and careful 
searching to find the flagellates, and if the operator be 
working single-handed, one preparation may dry up while 
he is searching through another. 

For the dissection of the flea? the following apparatus 
was used: A pair of fine needles mounted in wooden 
handles, a fine pair of forceps, and a dissecting-microscope, 
besides slides and coverslips. The needles used were 
sharpened on a hone, one to a sharp point, the other to a 
flat, chisel-like edge with rounded corners. ‘he pointed 
needle was the more useful for holding, the flat-edged needle 


1 Made up as recommended by Laveran and Mesnil, namely: 1 grm. 
of sodium citrate and 1 grm. of sodium chloride dissolved in 200 ce. of 
distilled water. This mixture appears to be most favourable for the 
examination of living trypanosomes. 

2 If an operation can be properly called dissection which consists 
in treating the flea as the Thracian women are said to have treated 
Orpheus: “Discerptum latos Juvenem sparsere per agros” 
(i. e. fields of the microscope, in this case). It need hardly be said 
that our object was not to study the anatomy of the flea, but to extract 
from its body those organs which might possibly harbour developmental 
stages of the trypanosome. 


THE RAT-TRYPANOSOME, TRYPANOSOMA LEWISI. 469 


for cutting. The dissecting-microscope used was a Zeiss 
binocular, with No. 4 eyepieces and the paired objectives 
F*, It is also necessary for the dissector to have at hand an 
ordinary microscope armed with a low power, since it is 
often difficult to distinguish the minute’ organs of the flea 
under the dissecting-microscope ; the intestine, for instance, 
if severed from its connections might easily be confounded 
with a portion of a Malpighian tubule. 

In the following paragraphs is described the method of 
procedure for making what may be considered an exhaustive 
examination of a flea for trypanosomes; it was not always 
necessary, however, to attempt so much, nor is it claimed 
that the entire operation was always successfully carried out, 
since both our knowledge of the flea’s anatomy, and our skill 
in extracting the organs required, advanced considerably 
during the progress of our investigations. 

The flea, as stated above, is picked up with a fine pair of 
forceps, holding it by its head, and placed on a slide (slide 1) 
in a drop of salt-citrate solution. The first operation is to 
cut off the head, which is not always easy if the flea be a 
lively one, in which case it is best to asphyxiate or drown the 
flea partially by holding it under water with the forceps for 
a short time. ‘lo decapitate the flea, hold it still by pressing 
the pointed needle across the thorax, and with the flat-edged 
needle cut across the head in the region of the eyes. ‘The 
severed head may then be removed to another slide (slide 2), 
covered with a cover-glass, and the contents of the pro- 
boscis examined ; but as the proboscis was never found to 
contain trypanosomes we ceased to trouble about it in our 
later studies. 

It is frequently the case that the flea has its rectum filled 
with faeces or with partially digested blood, and when this is 
so it happens commonly that the rectum empties itself by a 
violent contraction at the instant that the head is severed 
(sometimes also eggs are extruded) ; or if the evacuation does 
not take place at this point in the proceedings, it is very 
difficult to avoid squeezing out the contents of the distended 


470 E. A. MINCHIN AND J. D. THOMSON. 


Trxt-Fic. 1. 


Digestive tract of a female flea, dis- 
sected out and drawn with the 
M Eq. camera lucida at a magnification 
"of 60, reduced in the reproduction 
to 40. Theanterior part of the dis- 
section is seen in ventral view ; the 
rectum and its surroundings in side 
view. cs. @sophagus. prov. Pro- 
ventriculus. St. Stomach. Mt. a. 
Malpighian tubule of the anterior 
pair; that on the left side of the 
stomach is shown in its normal - 
position, that on the right has its 
distal limb pulled out and away 
from the stomach. Mt. p. Malpig- 
hian tubule of the posterior pair. 
int. Intestine. 7. p. The six rectal 
papille. R. The rectum. ¢.s. Ter- 
minal segments. an. The anus. 


THE RAT-TRYPANOSOME, TRYPANOSOMA LEWISI. 471 


rectum during the subsequent operation of opening the 
abdomen. In cases where feces are thus extruded the body 
of the flea is removed at once to another slide (slide 5), and 
the feeces left on slide 1 are covered with a slip and 
examined. 

Through the integument of the flea the stomach can be 
seen lying ventrally in the anterior ? of the abdomen, and 
often the rectum can be seen at the hinder end in the 
dorsal region. The ventral posterior and dorsal anterior 
part of the abdomen is seen to be occupied by a whitish 
mass, most conspicuous in the female, and consisting chiefly 
of the reproductive organs. 

The next stage in the proceedings is to open the body of 
the flea. This is done near the hinder end, at about the level 
of the fourth or fifth tergite of the abdomen. The body is 
held still with the pointed needle, with which the thoracic 
region is pressed down or speared, and with the flat-edged 
needle the body-wall is cut through dorsally and ventrally in 
the region indicated, and the hindermost segments of the 
abdomen gently detached in such a way as to separate the 
integumental portions without rupturing or tearing the 
internal organs. It is especially important, if it be desired 
to examine the contents of the body-cavity, that the 
digestive tract should not be in any way torn or punctured. 
By holding the anterior part of the body and pulling gently 
on the detached hinder part, the gut can be stretched out 
and seen in nearly its full length; the stomach, usually 
containing a greater or less amount of more or less digested 
blood, is seen projecting from the anterior part of the body, 
the rectum is contained in the detached hinder part, and 
stretching between the two is the intestine like a delicate 
white filament, exposed in its whole length, but more or less 
obscured by the fat-body, Malpighian tubules, and generative 
organs, especially by the large ovaries in the female; these 
organs render the female flea much more difficult to dissect, 
in spite of its larger size, than the less-encumbered male. 
The generative organs and as much as possible of the fat- 


472 E. A. MINCHIN AND J. D. THOMSON. 


body are now pulled out on to the slide and cut off from the 
body, care being taken not to injure the gut. The carcase of 
the flea, with the hinder part hanging on by the still intact 
intestine, is now removed to another slide (slide 4), and the 
extracted contents of the body-cavity on slide 3 can be 
covered with a slip and passed on for examination ; but so far 
as stages of T. lewisi are concerned, it is superfluous to do 
So, since they are never found in the body-cavity unless the 
gut has been punctured or ruptured. 

The next step is to divide the digestive tract into two 
parts, thereby severing completely the hinder part of the 
body from the fore-part. This is done at the point at which 
the Malpighian tubules are given off at the junction of the 
stomach and intestine, the region which represents the 
transition from the mid-gut, lined by endoderm, to the hind- 
gut or proctodzeum, lined by ectoderm. The Malpighian 
tubules are four in number in the flea; two of them run 
forward a short way on the wall of the stomach right and 
left, attached to it by fine tracheal tubes, and then turn 
backwards again with a sharp, elbow-like bend towards the 
dorsal side of the body ; the other two tubules run backwards 
parallel to the intestine and alongside of it towards the hinder 
end of the body. The posterior pair of the tubules are also 
bent on themselves towards their distal extremities, but not 
so regularly as the anterior pair. The gut is cut across with 
the flat-edged needle at the point of origin of the tubules, and 
if this be performed accurately one pair of tubules (the 
anterior pair) remains attached to the stomach, the other pair 
to the intestine ; sometimes, however, all four tubules remain 
attached to one or other of these organs. The hinder part 
of the body, with the intestine and rectum, is now removed 
to another slide (slide 5). The stomach is then pulled back- 
wards out of the anterior part of the body on slide 4, and 
with it come out also, continuous with its anterior termina- 
tion, the proventriculus and the cesophagus, these two parts 
representing the embryonic stomodeeum, lined by ectoderm, 
while the stomach represents the whole of the embryonic mid- 


THE RAT-TRYPANOSOME, TRYPANOSOMA LEWISI. 473 


gut. The proventriculus is lined by a thick chitinous cuticle 
prolonged into stiff, curved, pointed spines, densely planted 
and forming, apparently, a straining apparatus ; it is approxi- 
mately globular in form and usually contains blood. The 
cesophagus is a delicate tube, its walls composed of the 
chitinous cuticle internally and a delicate network of 
muscles externally ; it generally performs active move- 
ments, twisting from side to side, when freshly extracted. 
The two pairs of salivary glands are situated in the anterior 
region of the abdomen right and left of the stomach. Hach 
gland has the form of a simple oval pouch, the wall of which 
is composed of a single layer of large cells with very large 
nuclei. From each gland comes off a duct, which, after 
running a short distance, unites with the similar duct of the 
other gland of the same pair. (In one instance we have seen 
the two glands of one side of the body fused into one, but 
with their ducts quite separate; on the other side of the 
body there was a pair of distinct glands in their normal 
relations). ‘The common duct of each pair of glands then 
passes forwards alongside the gut through the thorax into 
the head, where it meets and joins with the corresponding 
duct from the other side of the body. The common salivary 
duct then runs a short distance and opens into the proboscis, 
doubtless on the hypopharynx as in other insects. The 
salivary ducts are recognisable at once under the microscope 
by their trachea-like structure, being lined by a thick cuticle 
which has ring-like thickenings; the rings are, however, some- 
what irregular and easily distinguishable from the very even 
and regular spiral thickening of the wall of a tracheal tubule. 
Externally to the cuticular lining the tubule is covered by an 
investing layer of protoplasm, of uneven thickness in different 
parts and containing fairly large nuclei at irregular intervals. 
The ring-like thickenings of the cuticular lining become less 
marked as the ducts approach their point of junction, and 
cease altogether before they unite ; the cuticular lining being 


quite smooth in the common duct and for short distances in 
the paired ducts. 


474 E. A. MINCHIN AND J. D. THOMSON. 


Not infrequently the salivary glands come out with the 
stomach when it is pulled out; more usually, however, they 
do not do so, but remain in situ. In such cases the anterior 
part of the body is removed to another slide (slide 6), and 
the stomach, left on slide 4, is teased up, covered, and handed 
on for examination. 

Now the dissection of the hinder part of the body, on 
slide 5, is proceeded with, in order to extract and separate 
the intestine and rectum. The rectum, situated dorsally to 
the accessory reproductive apparatus, penis or receptaculum 
seminis, is a fairly large pear-shaped organ, the stalk of the 
pear terminating in the anus. The slender intestine joins 
the rectum at its broad end, and in this region are situated 
the six conspicuous rectal papille, remarkable and very 
characteristic structures, the presence of a single one of which 
makes it easy to recognise even a small fragment of the 
rectum. Behind the papille the rectum has a thin wall, to 
which the crithidial stage of the trypanosome, when present, 
is usually found attached, sometimes in vast numbers. In its 
anterior part, the region of the papilla, the rectum has only 
circular muscle bands, between which are wide interspaces. 
In the hinder region, behind the papille, there are both 
circular and longitudinal muscle-bands; the latter can be 
traced forward to just behind the papille, at which point 
each band becomes rapidly narrowed to a tendon-like fibre, 
and at the same time the striations of the muscle disappear. 
The tendinous continuations can be traced forwards, in the 
living condition, for some distance, but we have not made out 
the exact points of their insertion. 

The intestine is characterised by a continuous coat of ring- 
like muscle-bands, with interspaces, arranged very regularly 
external to the epithelium. When the edge of the intestine is 
focussed under the microscope, the layer of circularly-disposed 
muscle-fibres is seen in optical transverse section like a string 
of beads. ‘I'he intestine is frequently seen to be performing 
active peristaltic movements, and it may be thicker in some 
parts than in others, owing to the contraction of the muscles. 


THE RAT-TRYPANOSOME, TRYPANOSOMA LEWISI. 479 


The rectum must be dissected carefully out of the hinder 
part of the body, so that it remains on the slide, free from all 
the adjacent organs or chitinous plates of the integument. 
The easiest way to do this is to make an obliquely longitudinal 
cut with the flat-edged needle so as to sever the ventral- 
anterior half of the hindmost segments, together with the 
genitalia, from the dorsal-posterior half containing the rectum 
and anus. The genitalia can then be removed and the rectum 
extracted without much difficulty. It requires some care to 
separate it from the anus without injuring it. When this has 
been accomplished, all unnecessary débris is cleared away. 
If it be desired to make separate examinations of the intestine 
and rectum, the intestine is cut off as close as possible to its 
junction with the rectum. ‘To effect this it is best to spear 
the rectum with the pointed needle and make the cut with 
the flat-edged needle; or the operation of cutting off the 
intestine may be performed before the rectum has been dis- 
sected out from the hinder part of the body. In either case, 
the intestine is removed to another slide (slide 7), and both 
rectum and intestine, on their respective slides (5 and 7) are 
teased up, covered, and passed on for examination. It is 
not difficult to tear the rectum into several pieces with the 
needles, but it is not so easy to tease up the intestine; it is 
too slender to make sure of splitting it lengthways, except by 
good luck and more or less accidentally, and itis necessary as 
arule to content oneself with cutting it transversely into two 
or three short pieces, the contents of which are generally 
squeezed out during the process. 

Finally there remain the salivary glands, on slide 6, in the 
portion of the carcase consisting of the thorax and fore-part 
of the abdomen from which the gut has been extracted. The 
salivary glands, as has been stated above, are lodged in the 
fore-part of the abdomen beside the stomach, and it is 
generally by no means difficult to extract them when the 
stomach has been removed. To do this it is best first to 
spear the thorax with the pointed needle, then insert the flat- 
edged needle into the abdominal cavity from behind, and rake 


4.76 E. A. MINCHIN AND J. D. THOMSON. 


out gently the contents of theabdomen. The salivary glands 
sometimes come out fairly clean, but more often they are 
embedded in fat-body, trachez, etc., from which they must 
be carefully freed as much as possible. In such cases they 
are sometimes a little difficult to detect under the dissecting 
microscope, but their position may be traced by their long, 
thread-like ducts. They are much smaller in the male flea than 
in the female. Another method which sometimes succeeds 
better in extracting the glands is to pull on the integument 
of the thorax with one needle and on that of the abdomen with 
the other. The body-wall then often tears across at the 
junction of the thorax and abdomen, and the salivary ducts 
are seen at once stretched out between thetwo. By continuing 
to pull the thorax forwards, the glands may be pulled out of 
the abdominal cavity and are seen hanging on to the back of 
the thorax, from which it is not difficult to detach them. By 
this method the glands may often be obtained very clean and 
free from encumbering fat and other tissue. When the 
glands have been extracted, other débris is cleared away 
and the coverslip is put on. The glands are very soft and 
are crushed immediately by the weight of a coverslip if 
there is no other tissue under it; but for examination of 
their contents this is not a disadvantage. 

In the foregoing paragraphs we have given a detailed 
account of a full examination of the flea, such as we practised 
in the earlier periods of our investigation. But when it 
became evident to us that the trypanosome, during its 
development in the flea, never strays beyond the limits of 
the digestive tract proper, we were able greatly to curtail 
the ritual of the examination and to omit entirely the proboscis, 
body-cavity, and salivary glands. It is also unnecessary, as 
a rule, to separate the intestine and rectum in the dissection. 
Consequently, our later examinations were reduced to (1) the 
excluded feces, if any, on the slide on which the flea was 
decapitated ; (2) the stomach, on a second slide; and (3) the 
rectum and intestine, on a third. 


Tt was no part of our task to make a special and detailed study of the 


THE RAT-TRYPANOSOME, TRYPANOSOMA LEWISI. 477 


anatomy of the flea, but a few points observed by us incidentally in our 
dissections may be noted briefly here. 

The nervous system, of which some beautiful dissections were made 
in this laboratory by Major Christophers, I.M.S., consists, as in insects 
generally, of (1) the brain or supra-csophageal ganglion-complex, 
sending off the peri-cesophageal connectives which pass on either side of _ 
the esophagus to connect with the foremost of (2) the three large 
thoracic ganglia, joined by connectives to form a series which passes on 
into (3) the abdominal chain of ganglia. It isa very difficult operation 
to dissect out the brain and the first two thoracic ganglia, but it happens 
very frequently that in the ordinary dissections of the flea the third 
thoracic ganglion and the abdominal chain of ganglia are exposed entire 
and in continuity. It is then seen that the abdominal chain consists 
of a series of small ganglia terminated posteriorly by a larger ganglion ; 
and further that in the male there are seven smaller ganglia, in the 
female only six, in the abdominal chain. The larger hindmost ganglion, 
from which nerves are sent off to the genitalia and rectum, evidently 
represents a fusion of several ganglia equivalent to the more anterior 
smaller ganglia. Consequently it is seen that the concentration and 
fusion of ganglia at the hinder extremity of the ventral chain has 
proceeded a step further in the female than in the male. 

The genitalia consist, in the male, of a conspicuous pair of testes, 
situated dorsally in the abdomen, and a pair of filamentous glands 
(prostates?) not unlike Malpighian tubules at first sight, but of slightly 
smaller calibre, and differing entirely in histological structure. There 
is no separate seminal vesicle, but each testis is a tightly convoluted 
tubule, the lower end of which is dilated to contain the ripe spermatozoa. 
Ducts from the testes and prostates unite to form a median Ductus 
ejaculatorius, which opens into a large penis of very complicated 
structure. In the female the two ovaries occupy practically the same 
position as the testes, but take up much more space and extend forwards 
to the most anterior limits of theabdomen. Each ovary consists usually 
of four egg-tubes or ovarioles, but in one specimen that we have 
mounted as a permanent preparation there are five ovarioles on each 
side. The ducts of the ovarioles unite to form the paired oviducts, and 
these unite in their turn to form the common oviduct. Ventral to the 
common oviduct lies the unpaired receptaculum seminis, consisting of a 
brown, chitinous capsule of a peculiar shape. The main body of the 
capsule is spherical, but gives off a curved, horn-shaped diverticulum, 
ending blindly. The horn-shaped portion has its concave curve turned 
towards, and connected by striped muscles with, the spherical part of 
the capsule. Aslender duct of great length, and much convoluted near 
its origin, arises from the spherical part of the capsule, and runs back 
to open probably into the distal extremity of the oviduct or into the 


478 E. A. MINCHIN AND J. D. THOMSON. 


genital vestibule. The spherical part of the capsule and duct of the 
receptaculum are surrounded with unicellular glands, thickly clustered 
round the capsule and the convoluted portion of the duct, but thinning 
out and becoming smaller towards the distal end of the duct. The 
receptaculum, dissected out, stained and mounted for the microscope, 
is a singularly beautiful object. It usually contains a dense mass of 
spermatozoa. 

The heart is frequently seen in dissections at the hinder end of the 
body as a delicate filament, which by its own contractions twists and 
lashes itself about. Under the microscope it appears a delicate tube, 
beset towards the hinder end by the pericardial cells which are attached 
to it oneither side, right and left, and are crowded together towards the 
hinder end, but occur more sparingly towards the middle region and are 
wanting in the anterior third of the heart. The ostia appear to be con- 
fined to the posterior region of the heart, but we have not made out their 
exact number or arrangement. For the pericardial cells, see Minchin 


(1910). 


(b) Notes on the Parasites of the Fleas. 


In a former publication one of us (E. A. M., 1910) has described some 
parasites found in our stock of fleas. The most important was a form 
to which the name Malpighiella refringens was given, occurring, 
as the generic name implies, in the Malpighian tubules of the flea. 
Since that time this infection seems to have died out entirely inour fleas, 
and we have not seen any Malpighiella in the fleas dissected by us 
for the last three years or more. Why this parasite should have died 
outin our fleas it is impossible to say, but it may be remarked that no 
conditions could possibly be more favourable for contaminative infection 
from flea to flea (whether from adult to adult, or larva to larva, or adult 
to larva, or vice versa) than those in our breeding cages, where vast 
numbers of fleas in all stages of development are herded together in a 
confined space. Consequently the disappearance of Malpighiella in 
our cages rather indicates that the fleas do not acquire infection with 
this parasite by the contaminative method. 

In the publication referred to, numerous yeast-like bodies were 
described and figured from the digestive tract of the flea. Since then 
we have found organisms of this kind abundantly in smears of the 
salivary glands (text-fig. 24, p. 642). 

In the larve of fleas that we have dissected and examined from our 
cages we have found the gregarine Agrippina bona (Strickland, 
1912). 

The cysticercoids of tapeworms are found not infrequently in the 
fleas. Nicoll and Minchin (1911) described two species of cysticercoids 


THE RAT-TRYPANOSOME, TRYPANOSOMA LEWISI. 479 


from our fleas, representing Hymenolepis diminuta and another 
species of the same genus. We have found the same two species 
frequently, and also have in our possession specimens of a third species 
notidentified. The cysticercoids appear sporadically, and are sometimes 
quite common for a period, and then are not found again for a long 
time. This uncertainty in their occurrence is quite intelligible, since 
their appearance must be caused by the introduction into the cage of a 
rat infected with tape-worms, which doubtless infects a large number of 
the larve that later become adult fleas. 


The point upon which we wish to lay special stress is the 
absence in our stock of fleas of any flagellate parasites, and 
more especially of the leptomonad described by Swingle (1911) 
under the name Herpetomonas (Leptomonas) pattoni. 
We have been at great pains to convince ourselves upon this 
point. In the first place we dissected at various times about 
eighty! fleas from the non-infected breeding-cage without 
finding any flagellates of any kind in them, while flagellate 
parasites occur in a very large percentage of those known to 
have fed upon infected rats, though not in all, since the 
trypanosome often fails to establish itself in the flea, and 
even when the insect has been fed on a rat with trypanosomes 
swarming in the blood, they often disappear completely from 
the digestive tract of the flea within twenty-four hours of its. 
having fed. 

We give here three tables (A, B (1), and B (2)) showing the 
results of dissections of fleas from our stock which had been 
put upon infected rats, and so had had the chance of 
acquiring an infection of T.lewisi. Fleas do not always 
feed, however, when given the opportunity to do so, 
especially in cold weather,’ and if the fleas are dissected and 
examined within tweuty-four hours after having been put ou 
the rat (the fleas in all cases having been kept hungry for 


1 As a matter of fact we dissected far more than this, all with nega- 
tive results, but we have not kept exact records of more than seventy- 
nine. 

? Note especially the twenty-one fleas of November 11th, 1911, in 
Table A, of which eighteen did not feed. This was due to a sudden 
cold snap, the first breath, so to speak, of winter. 


vot. 60, PART 4.—NEW SERIES. B4 


A480 E. A. MINCHIN AND J. D. THOMSON. 


two or three days previously to being put on), it is quite easy 
to distinguish those which have been fed from those which 
have not availed themselves of the opportunity of doing so. In 
this way useful controls are obtained for determining whether 
the fleas contained any flagellate infection before being used 
for putting on the infected rat. 


Taste A.—Fleas Examined within Twenty-four 
Hours after being put on an Infected Rat, to 
show the Numbers that had or had not Fed, and 
the Numbers of those that had Fed but in 
which no Flagellates were Found. 


Number of fleas | Number of fleas which ap- 
Date onwhich| Time since#| Number of | apparently not peared to have fedand 


the fleas were] fleas put on | fleas put on |fed and contain- which contained : 
examined. | infected rat. the rat. ing no flagel- | (a) Stages of (b) No 
lates. T.lewisi. flagellates. 

5:vii:’710| 6 hours 3 i Z 0 
(Oe wai TE) 12) ,, 10 2 8 0 
Bene is) 15 .., 14 3 a! 0 
5.3 Yen eat Eel ces 4 1 3 0 
D6cwe 11 18: © ,, 12 il 7 4 
Pordic cls 8, i, ad 3 5 1 
S0ea-da | 18 ;, 5 at 4 0 
Gea lio| 1S' Vie) 3 5 0 
4:71:°13) 20 ,, 14 113, 2 0 
10: vi:10} 24 ,, 4 2 2 0 
6:vii:710| 24  ,, 3 2 1 0 
Viens | 24 2, 5 A: 3 1 
15:vi:11| 24 ,, itl 6 4 1 
262ix<11| 24 ° ,, 7 a 8 2 
30:ix:711| 24 _,, 22 2 18 2 
Seed Used is 13 il 12 0 
6:x:711 | 24 ,, 14 9 1 1 
exue 11) 24 45, 1) 2 11 4 
Uae | a 21 18 2 1 
7 xed 24. |, 1a? 5 10 2 
24:xi:711| 24 ,, 20 2 12 6 
Hepaneg il Bh 9 12 3 7 2 
IDPS cba l| Gis 6 1 5 0 
2:vil:°12| 24 ,, 14 8 6 0 
20:vi:"11| 24 ,, 14 3 8 3 

Total . F = uzeo 92 167 30 


Percentage . 7 3100 31°83 57°79 10°38 


THE RAT-TRYPANOSOME, TRYPANOSOMA LEWISI. 481 


From Table A it is seen that of 289 fleas which were put 
on infected rats, 92 (31°83 per cent.) had not fed and con- 
tained no flagellates, 167 (57°79 per cent.) had fed and 
contained T’. lewisi, and 30 (10°38 per cent.) had fed but 
contained no flagellates. 

In addition to these negative data we have had the 
opportunity of comparing our stock of fleas with another 
stock which was actually infected with Leptomonas 
pattoni. When one of us (H.A.M.) was in Paris in 
January, 1913, he was very kindly presented by Dr. 
E. Chatton, of the Pasteur Institute, with some living fleas 
(Ceratophyllus fasciatus) from a stock infected with 
Leptomonas pattoni. These fleas were brought back to 
London and a fresh breeding-cage colonized with them. The 
fleas were left to breed for a year, during which time the rat 
in the breeding-cage was changed frequently, but none of 
the rats put in acquired any trypanosome-infection. When 
the fleas were examined at the beginning of 1914, they had 
multiplied enormously, and were found to contain Lepto- 
monas-infections. We did not keep any exact records of 
our dissections of the Leptomonas-fleas, but, roughly 
speaking, about 50 per cent. of the fleas contained teeming 
infections. The leptomonads appear to establish themselves 
in the fleas as readily as does T. lewisi, perhaps more so, 
since, as will be seen from Table B (1), barely more than 
14 per cent. of our stock of fleas contained swarming infec- 
tions when exposed permanently to infection with T. lewisi 
in the infected breeding-cage, and Table B (2), if we count 
only those known to have fed on an infected rat, not less than 
six, not more than fourteen days previously, gives but a 
slightly higher percentage (21°19). 

We may conclude, therefore, from a comparison of our 
stock of fleas with those bred from Dr. Chatton’s stock 
infected with the leptomonad, that, had our stock also been 
infected with leptomonads, we should not have failed to find 
fleas containing leptomonads in those fed on clean rats in 
the first place, and secondly, that the percentage of fleas 


482 E. A. MINCHIN AND J. D. THOMSON. 


Taste B.—Summary of the Condition of Fleas known 
to have Fed on Infected Rats. (1) Fleas taken 
at Random from the Infected Breeding Cage. 


Trypanosomes. - 
Caleee He 2.2 ree | eames of the 
mescee: ) f 
| alee: None. Scanty. Swarming. 
18 ie MOTs oe |e 1 1() Infection produced by 5 fleas 
of which 3 were dissected 
| | (Table J). 
23:11:710} 4) 1) 2 1 (s) 
Sornr= tO Ga! S13 2 (s) 
Deives LO | 6 0|3 3 (Lr, 2s) 
8:iv:710} 2 0 | 1 (si) 1 (si) 
10:iv:’710| 2 taal: | 0 
22:iv:710| 4 33 |) | 1(s) 
3:v: 10 4A. 4/0 0 
95:vii:710| 5 ales 0 
27:vii:’10| 3 12 1 0 
5:vii: 10) 5 ee || e33 0 One positive (Table K). 
6 :ix:°10 3 33, || | 0 
voix: 10 | 4 ae 0 
4-1x3°10| 3 0/2 | 1(s) 
tocix 103) 3 () |} 33 | 0 
paca: LOi) 6S bale 1 (s) One positive (Table K). 
mois LO) 1; 0 0 Negative (Table K). 
Q4:ix:"10)| 1 120 | O 
96:ix:°10| 2 110 | 1 (r) 
act AO) 3 Sao 0 
2ecam< 10) 9S 3/0 0 
29 :ix:710| 3 210 1 (ir) 
3:x: 10 2 2:1 0 im 
6:x:’10 2 Ot 1 (r) 
10:x:710 | 4 IL |) 1 (r) 
Usese sO) |p as 33. |) 0) 0 All negative (Table K). 
WAR Sst E OM SS |e 3) "0 0 One positive (Table K). 
h 6 hear Ve ae ibe 1 (x) All negative (Table K). 
LS ace OS Spe i 0 iS (Table K). 
24:x:710| 3 210 ma) e (Table K). 
Bos ee LOU see | 0 | 0 One positive (Table K). 
Ave xa SAO | es 2)1 | 0 (Table K). 
8:xi:710 | 3 Pee: 0 All negative (Table K). 
AS xa) es PAN 1 (x) One positive (Table K). 
16:xi:‘10| 3 a | 1(@) All negative (Table K). 
i Ms cra 2/0 | O Both negative (Table K). 
22:xi:°10| 2 2) 0 0 =: (Table K). 


THE RAT-TRYPANOSOME, TRYPANOSOMA LEWISI. 483 


Trypanosomes. 
(fens | Experimental infectivity of the | 
dissected.) | flea. | 
Fleas..None.| Scanty. Swarming. 
z meal ne |e 
2.08) 5 ered C0) Pil Ee at EC) 0 
sues 10) 2) 2) 0 1 (r) 
earce lO.) 2 OFT 1 (s) 
b:xi: 10) 5 1 | 4 0 Injection of sal. gl.negative.| 
gexue 10) 29) 0 | 2 i) 
foxa= 109) - 2° | ~ 0) 2 0 
Sexans 105) 72) 227) 0 0 Both negative (Table K). | 
isoxas tO} 3) 21 0 All negative (Table K). | 
|14:xii:710| 3 Pee ie a 0 
115:xii:°10} 6 5 | 1 (sr) 0 Fleas infected rat239, stom- 
| achs rat 257 (Table I). | 
pea eom 9:5) IO 0 Fleas positive (Table 1). 
hae it} 6 4 | 2 (r) 0 Stomachs injected, positive 
(Table I). 
260: 11} 6 5 | 1 (ir) 0 Negative (Table I). 
Zor bls! -6* 601-0 0 Fleas infected rat 251,| 
| stomachs rat 261 (Table 
| 10; 
eae uh | or! 5 10 i) Fleas positive (Table I). | 
eae LE | 10°) 21 6 2 Stomachs injected, 2 posi- | 
tive. 
Zone I 10 | 7 | 2 1 (r) Stomachs positive, recta | 
negative. | 
(cont: 13 | 8 8 | 0 0) 
13:10:7138| 4 0 | 1 (sr) 3 (2s, Isr) | 
14:ii1:713| 3 0 | 0 3 (1s, 2sr) | 
Peat 13 | 73 S320 0 | 
28:11:13). 9 5 | 4 (3s, Isr)| 0 | | 
Sin 1S || 7) 76 0 | | 
A:iv:713 | 9 4 | 0 1 (sr) 
7:iv:713 | 6 0 | 3 (sr) 3 (s) 
10:iv:713| 6 1 | 4(1s, 3sr)} 1 (sr) 
14:iv:°13) 2 | 0| 1(s) (al 
Total .|249 | 144 |70 35 
Percentage 100 o 78328711 14-06 


The batches marked * were fleas of which the stomachs and other 
organs were kept for injection into rats, and were therefore examined 
hastily and imperfectly. 

s = stomach; r=rectum; i= intestine. 


484 


E. A. MINCHIN AND J. D. THOMSON. 


Taste B.—Summary of the Condition of Fleas known 


to have fed on Infected Rats. 
Definite Periods. 


(2) Fleas fed at 


Age of infection 
in flea (approxi- 


mately). 


6 hours 


Date. 


No.of 
fleas. 


None. 


:710 


: 10 


Paid 


ie ld 


2711 


Sekt 


2711 


alo 
3° 1183 


13 


eels 
oak 
2710 


el. 


ra 


eal 
lull 
:711 
palit! 


Soll 
riggs! 
rg 
reee th 
ree 8 


Sal 


ay lik 


eZ 


ape IP 
a3 


el a 


: 710 
:710 
: 710 


All 


eral 


2D, 
2 2, 
5 lit 
LD, 


: 09 


: 09 
: 09 


bo 


OT OTS He DO OU CaF HB CO SO Ot 


7 
© Or Ot DH 


ot 
WokhAaLOOD 


eS oe 
LO DODO COR HDD OTH OD ODN HD HESIOD DOM NNN OSOHMBHMOMOS © 


Fleas containing Trypanosomes. 


Scanty. 


1 (si) 
0 


6 (4s, 2sr) 
10 (6s, 4sr) 
4 (1s, Isr, 2r) 
9 (7s, 2sr) 
8 (5s, 2sr, lr) 


4 (1s, 2sr, 1r) 
3 (sr) 
3 (r) 
1 (sr) 
12 (1s, 3sr, 8r) 
2 (1s, 1r) 


1 (s) 
0 


Swarming. 


3 (2s, Isr) 
1 (s) 

3 (Is, 2sr) 

6 (5sr, lr) 

3 (1s, 2sr) 
2 (sr) 

4 (1s, 3sr) 

3 (2s, Isr) 


2 (1s, Isr) 
8 (6s, Isr, 1r) 
8 (4s, 4 sr) 


4 (3s, 1sr) 
0 


6 (1s, 5sr) 
5 (1s, 4sr) 
5 (sr) 

5 (4s, Isr) 
0 


1 (s) 


6 (5sr, 1s) 
3 (1s, 2r) 
0 


0 
0 

( 
0 


2 (1s, Isr) 
0 


0 
1 (s) 


THE RAT-TRYPANOSOME, TRYPANOSOMA LEWISI. 


Age of infection 
in flea (approxi- 
mately). 


3 days 


bole 


10 
ll 
12 
14 


33 


33 


“ 
vy 


e 
3 ss 


Fleas containing Trypanosomes. 


Grand total 
Percentage 


Total of six days and over 


Percentage 


No.of 
fleas. 

None. Scanty. Swarming. 
on nei 2 (sr) 2 (1s, Ir) 
By |S 0 0 
4] 4 0 0 

10 | 6 | 4 (2s, Isr, 1r) 0 
10; 1 | 5 (1s, 2sr, 2r) 4 (1s, 3sr) 
12! 6 | 5 (1s, Isr, 3r) L@) 
16 | 4 8 (1s, 7r) 4 (1s, 3r) 
Gola 4 (sr) 1 (sr) 
8| 4 2 (sr) 2 (sr) 
3] 0 1 (sr) 2 (1s, Isr) 
10| 5 5 (1s, 4sr) 0 
12 9 i3 (sr) 0) 
12] 8 3 (r) 1 (v) 
2} 2 0 0 
Za O 0 2 (sr) 
3 | 2 1 (i) 0 
10} 6 1 (s) 3 (r) 
1h 27) 5 (x) 0 
403 1 (s) 0 
6] 2 2 (r) 2 (Isr, Ir) 
2 2 0 0 
5| 3 1 (r) i) 
4| 0 1 (sr) 3 (Isr, 2r) 
10) 3 2 (Is, Ir) 5 (2sr, 3r) 
6 5 1 (sr) 0 
6| 3 0 3 (r) 
(Auta 3 (sr, 2r) 3 (sr) 
6| 2 2 (Isr, Ir) 2 (sr) 
G3 1 (s) 2 (sr) 
oie 1 (sr) 1 (s) 
6 6 0 0 
6| 6 0 0 
Gr 2 3 (1s, 2r) 1 (sr) 
4| 2 2 (r) 0 
6] 3 2 (1s, lr) E(x) 
San 7 0 1 (7) 
2) 1 1 (@) 0 
Te ell 0 0 
2 2 0 0 
. 609 |230 293 156 
. {100 |37-77 36°62 25°61 
. (118 | 65 28 25 
23°73 21:19 


F bi 55°08 


* Stomachs only examined. 


485 


486 E. A. MINCHIN AND J. D. THOMSON. 


containing flagellates would have been far higher than is 
shown by our tables, in fleas exposed to infection by T. 
lewisi.! 


(c) Notes on the Histological Structure of the 
Stomach of the Flea. 


We shall have occasion, when describing the developmental cycle of 
the trypanosome in Part II below, to relate how the trypanosome pene- 
trates into the epithelial cells of the stomach of the flea and goes through 
a process of multiplication within them. It is a necessary preliminary, 
therefore, to understanding the effects of the parasites that we should 
preface our description of their development by some remarks upon the 
structure and contents of the flea’s stomach; and in the following 
section we give an account of our observations upon these matters, 
without claiming to have added anything to the scientific knowledge of 
insect histology. 

The histology of the digestive tract of insects has been the subject of 


1 Noller (1912), discussing the question of the leptomonas-infection of 
the fleas, remarks, p. 398, that since the larva of the flea acquires the 
infection, adult fleas bred in a cage can be infected, and that conse- 
quently “the arrangement of the experiments (‘ Versuchsanordnung’) of 
Minchin and Thomson, who used fleas bred in a rat-cage, does not 
correspond to the requirements (‘ Anforderungen’).” We are at a loss 
to understand to what this criticism applies or what are the ‘ Anforde- 
rungen”’ to which Noller refers. At the time Noller wrote we had 
published only our three preliminary reports. The first two of these 
(1910, 1911, 1) refer only to the transmission of T. lewisi by fleas, 
and it is sufficiently obvious that the presence of leptomonads in the 
fleas could not affect in any way the value or significance of positive 
results obtained in experiments on the transmission of the trypano- 
somes, since, ex hypothesi, the trypanosome and the leptomonad 
parasite are in no way connected. Our third report (1911, 2) gave an 
account of the intracellular multiplication of T. lewisi in the flea’s 
stomach, a discovery which Ndller himself has confirmed, and which 
also would be quite unaffected by the presence of leptomonads in the 
fleas. Ndller’s criticism appears to us, therefore, both premature and 
superfluous ; premature, because our stock of fleas was not, as a 
matter of fact, infected with leptomonads; and superfluous, because, 
even if the fleas had been infected with leptomonads, it would have 
made no difference to the experiments and observations which Noller 
criticises. 


THE RAT-TRYPANOSOME, TRYPANOSOMA LEWISI. 487 


numerous memoirs, and its general characteristics are very well known. 
It would be beyond the scope of this memoir to attempt to discuss this 
subject in detail or to cite the very copious literature dealing with it; 
but of recent works we may refer more especially to the very excellent 
monograph of Léger and Dubosegq (1902), who have studied the intestinal | 
epithelium of Tracheata from the same point of view as ourselves, that 
is to say, with the object of describing the changes produced in the 
epithelium by parasites (gregarines) attacking the cells, None of the 
insects studied by Léger and Duboseq, however, were of blood-sucking 
habit and the stomach-epithelium of the flea differs in a number of 
points from any of the epithelia described by the French authors. 

The wall of the stomach consists of the following principal layers, 
counted from within outwards (PI. 39, fig. 126): (1) the lining epithe- 
lium, (2) a layer of circular muscle fibres, and (3) a layer of longitu- 
dinal muscle fibres. In addition to these layers, which are very easily 
seen, there are to be found also, though by no means in every section, 
flattened epithelial cells external to, and in contact with, the lining 
epithelium, between it and the circular muscle-layer; they occur 
sparingly and far apart, but appear to represent an integral and perhaps 
primitive constituent of the wall of the mid-gut. Similar flattened cells 
are found here and there on the Malpighian tubules, the wall of which 
is similar in histological composition to the stomach-wall if the latter be 
imagined as reduced to the lining epithelium and the flattened cells 
alone, without the muscle-layers. 

We are concerned here only with the lining epithelium of the 
stomach, but it may be mentioned in passing that the circular muscle- 
fibres occur as bands or separate rings with considerable intervals 
between them, and consequently do not appear in every transverse 
section of the stomach. The longitudinal muscles are also separated 
from one another by intervals, a fact at once apparent in the transverse 
section, in which the muscles are seen cut across and in which it can 
further be seen in a well-preserved section, that each longitudinal 
muscle-fibre is connected to its neighbours by a delicate membrane, 
appearing asa fine line running between each adjacent pair of muscle- 
bands and forming a delicate sheath or investment round the whole 
stomach. The circular muscle-layer is continued on into the intestine, 
where it forms a continuous investment without intervals between the 
bands; the longitudinal muscles end at the pylorus posteriorly and 
start anteriorly behind the proventriculus, which has its own system of 
musculature, running for the most part in oblique bands arranged sym- 
metrically right and left. Hach muscle-band in the stomach-wall is a 
single, transversely-striated fibre, in which an occasional elongated 
nucleus is seen, embedded in a small quantity of protoplasm. 

The following account of the epithelium and contents of the stomach 


488 E. A. MINCHIN AND J. D. THOMSON. 


applies, unless otherwise stated, to sections of the stomach fixed with 
Flemming’s fluid and stained with iron-hematoxylin followed Eby 


TrxT-FIG. 2. 


Diagrammatic representations of sections of the epithelium of 
the flea’s stomach, to show the various conditions: a, to show 
a section through an epithelial crypt from which clear, 
regenerated epithelium is arising on all sides, while on one 
side three black, degenerated cells are seen (compare fig. 316, 
Pl. 44) ; b, to show the manner in which the border of the cells 
arises in relation to the gradually-developed separation of the 
cells from one another; c, to show the transition from the ordi- 
nary, columnar type of epithelium to the flattened type. 


Lichtgriin-picric (see the following section dealing with technique). If 
such a section through a number of stomachs—all taken from a batch 


THE RA'T-TRYPANOSOME, TRYPANOSOMA LEWISI. 489 


of fleas dissected at the same sitting and at the same interval of time 
after having fed on the infected rat, all preserved in the same way, all 
stuck on the same slice of liver, cut and stained simultaneously—be 
examined even inthe most cursory manner, very considerable differences 
are seen between the stomachs in one and the same microtome-section. 
These differences affect both the epithelium and the contents. The 
epithelium varies, in the first place, in the form of the cells, from 
flattened to columnar, and secondly, in the staining reactions of the 
cells. The contents of the stomach, that is to say the blood-débris, 
vary greatly in colour, staining in some cases deep opaque black, or less 
deeply in various shades of grey, in other cases, however, bright yellow. 

The variations in the form of the epithelial cells are to be ascribed to 
the differences in the degree to which the stomach is dilated by the 
ingested blood. In a gorged flea the distension of the stomach stretches 
the epithelium until the cells become thin and flattened ; but when the 
flea is hungry, or has taken ina small quantity of blood, or when the 
quantity ingested has become reduced by digestion and absorption, 
the epithelium resumes what may be considered its normal columnar 
form. Every gradation between the flattened and columnar conditions 
can be found in different sections or in different parts of one and the 
same section. 

The variations in the staining reactions of the epithelial cells depend, 
in the first place, on the age or senescence of the cells. It is a matter 
of common knowledge that the lining epithelium of the mid-gut of 
insects is continually being thrown off and regenerated. The ordinary 
epithelial cells do not multiply and no mitoses are ever found in them ; 
the centres of regeneration are the so-called epithelial crypts, each 
representing morphologically a small diverticulum of the epithelium in 
which the approximation of the cells usually obliterates the cavity and 
produces a solid, bud-like mass of cells (Text-fig. 2, a and PI. 44, tigs. 
314, 316). Whena flea’s stomach, containing a certain amount of ingested 
blood, is plunged into a fixative, the epithelial crypts are very easily 
seen with a hand-lens or with the naked eye as little opaque white spots 
in the semi-transparent stomach-wall, very conspicuous against the 
reddish-brown background of the stomach-contents. In the sections it 
is common to find mitoses in such a crypt, especially towards its fundus 
(Text-fig. 2, a). As the cells multiply they are pushed upwards to the 
general level of the epithelium and outwards from the crypt to replace 
the old epithelial cells which, having degenerated, are cast off from the 
wall into the lumen of the stomach, and are digested there. 

The young, freshly-regenerated epithelial cells have the cytoplasm 
clear, staining light-grey, and are relatively poor in granulations; the 
older cells, on the contrary, have the cytoplasm full of granules that 
stain very deeply, until finally the whole cell, including its nucleus 


490 E. A. MINCHIN AND J. D. THOMSON. 


becomes black and opaque. Consequently, the epithelia of different 
stomachs show very varied appearances. A recently regenerated 
stomach will show clear epithelium all round, and, according to the 
time that has elapsed since regeneration, there may be no detached 
cells in the lumen of the stomach, or there may be a certain number of 
detached black cells, or there may be still, here and there, isolated 
black cells or patches of such cells in situ in the epithelium or in pro- 
cess of being cast off from it. On the other hand, a stomach which is 
about to be regenerated shows very dark epithelium all round, and in 
places this may bein process of rejection and replacement from the 
crypts, in which the cells have clear cytoplasm. The condition of the 
epithelium may vary in different parts of the same stomach, and from 
what we have observed we have gained the impression that the regenera- 
tion proceeds usually from before backwards, so that the anterior part 
of the stomach is further advanced in degeneration or regeneration, as 
the case may be, than the posterior region. We find in our preparations 
all possible conditions of the epithelium in different stomachs in one 
and the same microtome-section, and we have not been able to establish 
any definite relation between the feeds of the flea and the regeneration of 
the epithelium, but we have not paid sufficient attention to this point to 
be able to state positively that no such relation exists; the differences 
seen in the epithelia of flea-stomachs examined at the same interval of 
time after feeding may be due to inequalities in the rate at which 
digestion proceeds. 

In addition to what appears to be the normal process of senile decay, 
in which the cells take up the iron-hematoxylin stain very deeply and 
become black and opaque, we have observed a second mode of degenera- 
tion, which we are inclined to ascribe to the action of the trypanosomes, 
since in all cases where it occurs in our preparations there are trypano- 
somes to be found in the stomach, and frequently in the degenerated 
cells themselves. In this second type of degeneration the black-staining 
granules in the cell diminish in quantity, without, however, disappearing 
entirely, while the cytoplasm of the cell stains yellow (Pl. 39, fig. 133; 
Pl. 40, fig. 140); hence we have generally referred to this condition in 
our notes as * yellow necrosis.” Inall the stomachs in which we have 
found it the blood-débris is also stained yellow, and it is often very 
difficult to make out the precise boundary of the necrosed cell-body, or 
to distinguish the cells from the débris when they lie free in it (Pl. 39, 
fig. 125), except by the presence of the nucleus and of a certain number of 
black granules in the cytoplasm of the necrosed cel]. Indeed, our first 
impression was that the yellow colouring matter of the blood had in 
some way penetrated into the cell and stained its cytoplasm, but there 
can be no doubt that this idea is an illusion and that the yellow colour, 
both of the blood-débris and of the necrosed cells, is due to the picric 


THE RAT-TRYPANOSOME, TRYPANOSOMA LEWISE. 491 


acid in the Lichtgriin-picric staining combination, though it is, of 
course, possible that the substance, whatever it may be, which stains 
yellow in the blood-débris may have become infiltrated into the dead 
cells and given them their peculiar staining properties. 

The variations in the staining reactions of the contents of the 
stomach are more difficult to explain. There appear to be two types of 
staining after the use of the iron-hematoxylin-Lichtgriin-picric com- 
bination, one in which the contents stain in various shades of grey up to 
black, with a greenish tinge, and one in which they stain a bright 
lemon-yellow. It is not possible to bring these two types of staining 
into one series, as there is no transition between them; the grey-black 
and the yellow types occur side by side in different stomachs in one and 
the same microtome-section, and each stomach shows either the one or 
the other condition through the whole series of sections. So far as our 
observations permit us to generalise, the grey-black series represent 
the normal stages of the digestion of the blood; the yellow reaction 
appears to be due to some abnormal condition. 

The blood ingested by the flea is very soon affected by the digestive 
action of the stomach, and the red corpuscles cease to be recognisable 
within a few hours after digestion. In the middle period of digestion, 
that is twenty-four hours, or thereabouts, after feeding, the blood has 
become thick, viscous and brick-red in colour, and contains immense 
numbers of irregularly shaped grains of all sizes, but for the most 
part coarse and large. Towards the end of digestion, forty-eight 
hours or so after feeding, the stomach contents are fluid and watery, 
dark brownish-black in colour, and the grains are much diminished in 
number and in size. 

In the sections, taking first the grey-black series, the blood in the 
earlier phases of digestion (eighteen to twenty-four hours) usually con- 
sists of densely-packed grains and spherules, varying in size from 
very coarse to very fine, and staining intensely black. Between the 
grains there is visible a coagulated albuminous matrix, stained 
greenish with the Lichtgriin-picric combination. The stomach-contents 
fill the whole section and adhere closely to the epithelial cells, 
penetrating down between them when they have the columnar form, 
but in the centre of the section there is generally a clear patch, 
circular in outline, which is seen to owe its clear appearance to the 
fact that coarse grains are absent, and it consists only of the albuminous 
matrix with finer granules. Hence the digestion, or more probably the 
passage backwards towards the rectum of the indigestible remnants, of 
the blood-débris appears to proceed from the centre of the section— 
that is to say, from the axial region of the stomach—towards the 
periphery. 

As the digestion proceeds, the grains in the débris become smaller 


492 E. A. MINCHIN AND J. D. THOMSON. 


and stain less deeply ; consequently the stomach-contents stain grey, in 
varying shades of darkness, while the matrix still shows the greenish 
hue. In stomachs thirty-six hours after feeding the contents of the 
stomach are generally greatly diminished in quantity, and are absorbed 
in the centre of the lumen, leaving a clear space of variable form, while 
round the periphery the greenish-grey débris adheres close to the 
epithelium. Leucocytes, especially the polymorphonuclear forms, can 
be recognised in the blood twenty-four hours after feeding, but at 
thirty-six hours we have not found them. Owing to the lighter tints of 
the stomach-contents at thirty-six hours, the trypanosomes free in the 
blood-débris can be seen more easily, in contrast with the earlier state 
of affairs. 

In the yellow stomachs the contents appear at first almost uniform, but 
on close examination they are seen to consist of closely-packed granular 
substance, all of which, both granules and matrix, is coloured by the 
picric acid in the staining combination used. The first point that 
strikes one immediately is that the contents in such stomachs are large 
in quantity and fill the whole stomach, or show but a slight amount of 
absorption towards the centre of the section, even at thirty-six hours, 
when the contents of, the grey-black stomachs are considerably 
diminished. The epithelium of the yellow stomachs may vary from 
the flattened to the columnar form, but the normal cells stain grey or 
black, in sharp contrast with the yellow contents. 

It seems obvious from these data that the yellow stomachs represent 
an abnormal condition; we have endeavoured, not very successfully, to 
find a relation between this condition and either the presence of 
trypanosomes, on the one hand, or the state of the epithelium on 
the other. 

In the yellow-staining stomachs which we have studied we have found 
trypanosomes to be present in the stomach in every case except one, and 
in that case there were attached clumps of crithidial forms imme- 
diately behind the pylorus, showing that the stomach-phase was over. 
But on the other hand, we have found the grey-black condition of the 
contents in well-infected stomachs also, showing at least that, if the 
yellow condition is in any way due to the parasites, they do not always 
produce that effect. On the other hand, in those cases in which we 
have found no trypanosomes at all, either in the stomach or outside it, 
the contents are always in the grey-black condition. A significant 
circumstance is, perhaps, the fact that we have only found the “yellow 
necrosis ’ of the cells in stomachs with yellow-stained contents. 

As regards the condition of the epithelium, we have found the yellow 
condition of the contents associated (1) with epithelium black all round 
and in process of being cast off, or (2) with epithelium mostly clear, but 
with black patches of cells in situ or detached ; in one such stomach the 


THE RAT-TRYPANOSOME, TRYPANOSOMA LEWISI. 493 


first condition is found in the anterior half, the second in the posterior. 
We can state, therefore, that in our experience the yellow-stained 
contents occur only in stomachs about to be regenerated, or in process 
of regeneration, or very recently regenerated. But, again, we have 
found the grey-black condition in stomachs that appeared also to have 
undergone regeneration very recently, which makes it difficult to 
correlate this condition with the process of regeneration. The question 
of the significance and cause of the yellow-staining condition of the 
stomach-contents must be left an open question at present; the data to 
hand do not suffice for drawing decisive conclusions, and it would lead 
us too far to attempt further investigations upon this problem. On the 
whole, however, it seems at least probable that we are dealing with an 
abnormal state of the digestive processes towards which the trypano- 
somes are a contributory cause, if not the sole one. 

As already stated, the different conditions of the stomach-contents 
described above are those seen after staining with iron-hematoxylin and 
Lichtgriin-picric. After the use of Giemsa’s stain the colour of the 
contents differs considerably in different cases. 

Most of our sections stained with Giemsa were fixed with Maier. In 
those in which the trypanosomes were best stained and show the flagella 
clearly and sharply the grains and spherules of the débris are coloured 
for the most part orange-pink, especially in those stomachs in which 
the digestion of the blood is further advanced (figs. 109, 113, Pl. 38) ; in 
the earlier stages of digestion many of the larger grains and masses in 
the débris are stained deep purple, making the contents of the stomach 
more opaque. In one of our series preserved in Flemming, consisting 
in all of seven slides, the first six were stained with the iron-hematoxy- 
lin-Lichtgriin-picri¢ combination, the seventh with Giemsa’s stain; on 
this seventh slide there are sections of four stomachs, two of which, on 
the other six slides, show grey-black contents, while the remaining two 
have the contents yellow in colour. In the Giemsa-stained slide the 
blood-débris shows a coloration very different after Flemming to that 
which it shows after Maier, being stained a bluish-green tint. The 
stomachs of the yellow type are slightly more blue in tint than those of 
the grey-black series, but otherwise the difference between them is but 
slightly marked. 

Having now described the chief variations in the conditions of the 
stomachs and their contents, or at least those differences which are 
obvious upon the most cursory inspection of the sections, it remains 
to give a more detailed account of the epithelial cell. In any given 
stomach the cells show great individual variation in form and structure, 
but, nevertheless, it is not possible to divide them into distinct classes. 
There are no special glandular or secreting cells, as described by Léger 
and Duboscq in other insects, and all the cells of the general epithelium 


494, E. A. MINCHIN AND J. D. THOMSON. 


of the stomach of the flea are to be regarded as equipotential, the 
differences visible between them being merely the expression of varying 
physiological conditions in relation to their changing environment, 
on the one hand, or to their constitutional vigour or senescence, on the 
other. Hence it is possible to give a generalised description of thé 
cells, beginning first with the normal, healthy cell and dealing afterwards 
with the changes it undergoes in the process of degeneration. 

The epithelial cells are produced, as already stated, in the “ crypts of 
regeneration,’ which have been described in various insects by Léger 
and Duboseq. In the flea these structures appear usually as solid, bud- 
like cell-masses that often project outwards from the wall of the stomach 
to a considerable extent (PI. 44, figs. 514, 316) beyond the level of the 
muscle-layers, which pass on either side of them. Internally the crypts 
do not rise up beyond the general level of the epithelium. The closely- 
packed cells of the crypts show distinct limits, and do not form a 
syncytial mass of protoplasm, as described by Léger and Duboseq 
(1. ¢., p.410) in the larva of Anthrenus verbasci, for example. At 
the fundus of the crypt mitoses are often found, sometimes in two cells 
simultaneously in the same crypt; in other cases a]l the nuclei are in 
the resting state. Doubtless the crypts have periods of active multipli- 
cation, alternating with periods of repose, as in other insects. The 
crypts are often seen to be marked off from the general epithelium by 
slender dark cells, the “cellules de recouvrement” described by 
Léger and Dubosegq (1. c., Pl. II, fig. 2, ¢.7.; p.388). The crypts appear 
to have the monopoly of cell-production in the stomach of the flea. We 
have not found basal cells, “cellules de remplacement,”’ in the 
general epithelium. 

By multiplication and increase in their numbers the cells are pushed 
outwards on all sides from the crypt to take their place in the general 
epithelium (PI. 44, fig. 316, and Text-fig. 2a). The young epithelial cells 
seen in the immediate neighbourhood of the crypts are columnar 
cells, roughly rectangular in form, and generally about twice as high 
as they are broad. The lateral boundaries of the cell are approxi- 
mately parallel, and each cell is in contact with its adjacent neighbours 
for its whole length. The free, apical surface of the cell is convex, and 
on this side is developed a very distinct, thick border, at first covering 
only the upper surface of the cell, which projects like a dome towards 
the lumen of the stomach. 

The further development in the form of the cell consists in an 
extension of the upper free surface, brought about by the cells 
becoming free and separated from one another at their sides, first at 
their apices and then downwards along almost the whole length of 
the side of the cell, till finally each cell is connected with the adjacent 
cells only by a narrow isthmus at its base (Text-fig. 2b). As the cell 


THE RAT-TRYPANOSOME, TRYPANOSOMA LEWISI. 495 


becomes free the border develops also on the exposed surface, so that, 
instead of being confined to the apex of the cell, it extends down the 
vertical sides also (Pl. 40, fig. 156). This process of separation between 
the cells has an obvious significance in connection with the process of 
flattening which they undergo when the stomach is dilated after 
feeding ; it can be regarded as an adaptation to the blood-sucking habit. 
When the flea gorges itself each cell is so stretched that its tallest part 
in the vertical direction is scarcely thicker than the nucleus, which 
bulges out the middle part of the cell in an even curve towards the 
lumen of the stomach, while towards the periphery the verticle height 
of the cell diminishes to the isthmus connecting it with its neighbours 
(Text-fig. 2c). As the cell resumes the columnar form the nucleus 
remains at or near the base, as a rule, and the cytoplasm of the cell is 
heaped up over it. In the extreme columnar form the apex of the 
cell is generally slightly expanded, the middle region more narrowed, 
so that spaces are left between adjacent cells, into which a con- 
siderable quantity of blood-débris penetrates (Pl. 39, fig. 126). The 
nucleus is usually situated at the base of the cell, but occasionally 
towards the apex (PI. 40, fig. 136). The border clothes the whole free 
surface of the cell, whether flattened or columnar, and is of considerable 
thickness over the apex and the sides, becoming thinner as it approaches 
the isthmus, but in the columnar form of the cell, when its apical region 
is expanded, the border may be thinner, as if stretched, at the apex of 
the cell (Pl. 40, fig. 144). 

The blood-débris has a great tendency to adhere closely to the border, 
so much so that the border is often more sharply marked off from the 
cell-contents within than fromthe blood-débris without, in the sections, 
but places can be found occasionally where the blood-débris has split 
away from the epithelium, leaving the border distinct and sharp. The 
border appears usually homogeneous and refringent, though in some 
preparations indications are seen of a vertical striation, as if it were 
composed of little darkly-stained rods, placed at right angles to its two 
limiting surfaces, and separated by intervening substance of lighter 
colour (Pl. 39, fig. 129, and Pl. 40, fig. 142). After sublimate-fixation 
the border is colourless, but when stained with iron-hematoxylin the 
blood-débris adhering to it often hinders the extraction of the stain 
and at these spots it remains black ; when the hematoxylin is extracted 
it tends to take up the green from the Lichtgriin-picric mixture (Pl. 40, 
fig. 147). With Giemsa after sublimate-fixation it stains a pinkish- 
yellow. After Flemming-fixation the border is yellowish, as if tinged 
by the chromic acid in the mixture, and when this fixation is followed 
by the Giemsa-stain the border is coloured green (PI. 38, figs. 99-103), 
There is no “ bordure en brosse,” or palisade of stiff rod-like cilia, 
external to the border, as in many insects. The condition in the flea 


voL. 60, PART 4.—NEW SERIES. 35 


496 E. A. MINCHIN AND J. D. THOMSON. 


more resembles that figured by Léger and Duboseq for Scolopendra 
(1. ¢., pl. vi). 

The border is evidently a fairly tough structure since in teased up 
stomachs examined fresh, the borders of cells are often seen quite empty, 
but retaining their shape, like shells. 

The nucleus of the epithelial cell calls for no special comment: as 
can be seen in our figures, it is rounded or oval, with the typical 
structure seen in tissue-cells, namely, a distinct membrane, a reticulum 
containing chromatin-grains of various sizes, and one or more nucleoli 
which stain black, like the chromatin, after iron-hematoxylin. Mitoses 
of the usual type are found commonly in the crypts of regeneration, but 
we have never seen the slightest evidence of nuclear division in cells 
forming part of the general epithelium outside the crypts. 

The cytoplasm of the epithelial cell varies at different ages. In the 
youngest cells bordering the crypts the cytoplasm appears more or less 
homogeneous and finely granular in all parts of the cell; it stains light 
purplish-grey or grey-black after iron-hematoxylin, bluish-purple after 
Giemsa, and no coarse granulations are to be seen. In the fully 
developed cell the cytoplasm has undergone local differentiation ; round 
the nucleus, in the basal half of the cell, it has a denser texture, but 
above the nucleus, in the apical region, it has become of looser con- 
sistence, more spongy, so to speak, in appearance, with irregular spaces 
(Pl. 39, figs. 126, 127, Pl. 40, figs. 136, 144), containing fluid in the living 
condition, and transversed by strands of protoplasm disposed irregularly. 
The more the apical part of the cells is expanded the more watery its 
contents appear. Sometimes the apical region appears almost empty 
in the sections, with only a few traces of cytoplasm close under the 
border and at the sides. It is in this region in which the stages of the 
trypanosomes are most often found, and into which the parasites first 
penetrate. 

In addition to these changes in the cytoplasm, numerous grains and 
enclosures of various kinds make their appearance in it. A detailed 
study of these granulations would require a lengthy investigation, an 
expenditure of time and trouble, that would go beyond the scope 
and objects of this work. We must confine ourselves to a brief 
summary of the appearances seen in our sections, without attempting 
to give physiological explanations of the various conditions seen. Itis 
obvious that the bare observation that a granule is stained black by 
iron-hematoxylin or red by Giemsa’s stain does not permit very far- 
reaching conclusions as to its nature or function in the cell; bodies of 
most diverse properties might agree to this extent in their reactions. 

The first granulations to appear are minute grains which, whatever 
the fixation, Flemming or sublimate, stain black after iron-hematoxylin 
and red after Giemsa. They are seen at first chiefly at the sides of the 


THE RAT-TRYPANOSOME, TRYPANOSOMA LEWISI. 497 


nucleus, between it and the cell-wall and extend up the sides of the cell 
close under the border. Scanty at first in the apical spongy part of the 
cell; they are soon deposited in this region also, appearing often in con- 
siderable numbers and varying in size from small granules to conspicuous 
_ grains, and even large masses (PI. 40, figs. 136, 137.) The larger grains 
are seldom homogeneous, but appear as rings, black or red, as the case 
may be, with clear centre, apparently hollow (Pl. 38, fig. 99). Those 
of still larger size show, especially after Giemsa, darker and lighter 
parts disposed in various ways; inside the peripheral deeply-stained 
shell there may be darker grains or patches. After iron-hematoxylin, 
however, the whole mass may be opaque black, but usually shows lighter 
inner portions. The extent to which these granulations are developed 
varies in different stomachs, doubtless in relation to their secretive or 
absorptive activity at the moment of preservation. When a number of 
stomachs are cut in the same block, one stomach all through the series 
may show the cells clear and very free from granulations, while another 
stomach shows nearly every epithelial cell loaded with coarse grains in 
its apical region. 

The red grains, as they may be termed from their distinctive reaction 
to Giemsa’s stain, appear to be always present in greater or less quantity 
in the fully-developed cells of every stomach. In addition there are 
often found, lodged in the apical spongy region of the cell, masses of 
relatively large size which do not retain the iron-hematoxylin stain 
firmly throughout their substance, and consequently appear for the 
most part light grey in colour after this stain (Pl. 40, figs. 145, 146); 
after Giemsa they are either scarcely stained at all, appearing a sort of 
neutral tint, or they are coloured bluish-purple in various shades, some- 
times very deeply, with streaks and blotches more reddish in tint 
(P1. 38, fig. 102). These masses vary considerably in size and contour, 
and show differentiation of their substance into lighter and darker parts. 
With superficial examination they often simulate the intracellular 
stages of the trypanosomes to a remarkable degree, especially in the 
living condition, when they are often very conspicuous ; for a long time 
we confused them with the spheres in the freshly teased-up stomachs, 
and spent much time watching them in the expectation, never of course 
fulfilled, of seeing them perform the characteristic movements. After 
we had made smears of stomachs in which these bodies were abundant, 
without finding any intracellular stages of the trypanosomes in such 
preparations, we came to the conclusion that these motionless spheres 
(as they appeared to be) were merely cell-products, and referred to them 
in our notes as “ pseudospheres.” Even in sections the pseudospheres 
often mimic the true spheres and might be confused with them at first 
sight, but only by an inexperienced observer who had never seen the 
actual intracellular stages of the trypanosome in the epithelium. The 


498 E. A. MINCHIN AND J. D. THOMSON, 


idea occurred to us at one time that some of the pseundospheres might 
possibly be degenerated stages of the trypanosomes, destroyed, and in 
process of absorption, within the epithelial cells into which they had 
penetrated; but we have found no decisive evidence for this. It is 
most probable that the pseudospherés represent secretion-masses 
formed by the cell itself. 

In some of the stomachs. especially in those preserved about twenty- 
four hours after feeding, there are to be seen dense and very conspicuous 
accumulations of coarse grains in the epithelial cells immediately below 
the border (PI. 38, fig. 98, Pl. 40, fig. 147). The grains in question are 
especially distinct after fixation with Maier’s fluid; they are more 
difficult to make out in the stomachs fixed with Flemming. The grains 
resemble very closely those of the blood-débris adherent to the border 
external to the cell, so much so that the first impression gained is that 
the débris has been absorbed into the cell through the border. It is 
easy to imagine this after iron-hematoxylin, which stains both these 
granules and the débris very black after sublimate-fixation (fig. 147); 
but the Giemsa-stain colours the grains within the cell differently from 
the débris (fig. 98), and when the digestion of the blood has gone 
beyond a certain point the grains inside the cell may be stained much 
darker with iron-hematoxylin than the grains in the blood-débris. It 
is improbable that the coarse grains of the débris would pass bodily . 
through the border, which is to all appearances a dense, tough 
structure ; but it is probable that these grains are formed in the cell in 
direct relationship with the process of absorption of nutriment from 
the blood. 

Amongst the enclosures of the epithelial cell must be mentioned 
finally peculiar yellow grains which occur with great frequency in some 
stomachs, not at all in others. Their presence or absence is in no way 
connected with that of the trypanosomes, and they occur both in 
normal as well as in degenerating cells, though perhaps more 
abundantly in the latter. In the Flemming-iron-hematoxylin sections 
these grains have a brownish-yellow tint, often with a darker shell 
(Pl. 40, fig. 141). They vary in size from small granules up to the large 
grains reaching as much as 13 in diameter (fig. 142). Their tint also 
varies in depth, being usually much lighter in the larger grains. With 
Giemsa, after Flemming, they are stained bright green (Pl. 38, fig. 103), 
probably as the result of a blue stain (azure) imposed upon their original 
yellow tint.!. These yellow bodies are very similar to, probaby identical 


1 A similar result is seen in the chitinous spines of the proventriculus 
in sections stained with Giemsa; the cuticle at the base of the spine is 
stained red, but that of the spine itself, from near the base to the tip, is 
coloured emerald green. The unstained spine is yellow in tint. Com- 


THE RAT-TRYPANOSOME, TRYPANOSOMA LEWISI. 499 


in nature with, the enclosures characteristic of the pericardial cells. As 
one of us has described elsewhere (E.A.M., 1910), the pericardial cells of 
the flea may beso crammed with yellowish-brown grains and spheres that 
the cell becomes visible with the naked eye as an opaque black spot 
through the integument of the living flea. In some of our stomach- 
sections there are also casual sections of pericardial cells which have 
been pulled out of the flea together with the stomach, so that we have 
had the opportunity of making a direct comparison between the yellow 
grains in the epithelial and pericardial cells. The yellow grains are 
probably an excretory product, eliminated by the flea under certain 
physiological but apparently normal conditions, and elaborated either 
in the epithelial cells, to be cast out into the lumen of the stomach, or 
in the body-cavity, to be taken up by the pericardial cells. 

We come now to the process of cell-degeneration which occurs in the 
effete, senile epithelial cells. This process is very different in the flea’s 
stomach from that described by Léger and Duboscq in various insects, 
none of them of blood-sucking habit. It is described by these authors 
asa‘ Dégénérescence mucoide,” an infiltration of the cells with 
mucoid substance. In the flea’s stomach the process appears to be more 
of the nature of a fatty degeneration, combined perhaps with a mucoid 
infiltration. 

In our sections fixed with Flemming’s fluid and stained with iron- 
hematoxylin the intensely black, often perfectly opaque, degenerated 
cells, which are seen frequently detached completely or in process of 
detachment from the epithelium, are very distinct from the clear, 
lightly-stained cells originating from the crypts of regeneration and 
taking the place of the degenerated cells (Pl. 44, fig. 516). In some 
of our sections the stain has been over-extracted: the trypanosomes 
have become ghosts, faintly visible only to the practised eye, the nuclei 
of the epithelial cells are pale, and even the blood-débris has had its 
usually intense black stain reduced to a shade of brown; but the black 
grains and masses in the epithelial cells remain as black as ever, showing 
that they do not owe their colour to the stain but to the fixation, that 
is to say, to the osmic acid in the Flemming’s fluid. Such preparations, 
spoilt for other purposes, are very useful for showing the gradual 
process of deposition of the blackened grains. First they appear as fine 
granules round the nucleus, near the base of the cell (Pl. 40, figs. 143, 
144). Next, other, and for the most part larger masses, are deposited in 
the cytoplasm above the nucleus. The cell then becomes gradually 
filled up with black grains from below towards the apex; often an 


pare also the green stain of the border, mentioned above, after Flemming 
and Giemsa, evidently due also to the super-position of a blue dye upon 
a yellow ground. 


500 E. A. MINCHIN AND J. D. THOMSON. 


empty space is seen at the apex, immediately below the border (PI. 40, 
fig. 138), but finally this, too, is filled up and the whole cell becomes an 
opaque black mass (fig. 139). 

Very instructive is one of our series preserved in Flemming, in which 
there is one stomach in which nearly all the epithelium is degenerate. 
The sections of this stomach are spread over seven slides, six of which 
were stained with iron-hematoxylin, while the seventh, on which are 
sections through the hindmost region of the stomach, was stained with 
Giemsa. On this slide the degeneration is not so far advanced as in 
the more anterior region of the stomach, and in different parts even of 
the same section the following conditions are to be found: (1) Cells of 
normal type, with clear cytoplasm containing a few red granules 
(Pl. 38, fig. 99) ; (2) cells with cytoplasm of a darker bluish-purple tint, 
with many more red granules and amongst them a few coarser grains 
intensely black in colour (PI. 58, fig. 100) ; (3) cells in which both the 
red and the black grains, but especially the latter, are greatly increased 
in number, leading up to (4) opaque black cells in which nothing can 
be focussed clearly. The black grains, it is obvious, can only owe their 
colour to the action of the osmic acid in the fixation, and must there- 
fore be of a fatty nature. On the other hand there is also a marked 
increase of the red grains in the degenerating cells, indicating, perhaps, 
that in addition to deposition of fat, there is also a tendency to mucoid 
infiltration, as described by Léger and Duboseq. The darker tint of 
the cytoplasm, in so far as this is not an optical effect due to crowding 
of the grains, indicates that it becomes impregnated with the substances 
produced in the process of degeneration. 

The deposition of the fat round the nucleus in the first instance indi- 
cates that the nucleus takes an active share in the process, and this is 
borne out by the fact that the nuclei themselves become very dark in 
the degenerating cells and are sometimes quite opaque.! 

In sections of stomachs fixed with sublimate mixtures the blackening 
of the degenerating cells seen in the Flemming-fixed sections is con- 
spicuously absent, so that at the first glance it is difficult to pick out 
the senile portions of the epithelium. More careful study of the subli- 
mate sections shows that here the degenerated epithelium is dis- 
tinguished from the regenerated by its pale, empty appearance, owing 
to the fat-grains having entirely disappeared, leaving empty spaces to 
mark their former position. This is best seen in stomachs fixed in 
sublimate-acetic, since, after sublimate-aleohol mixtures (Maier’s and 
Schaudinn’s) the cells are often much deformed and shrunk. In a 
favourable spot it is seen that the young cells, freshly produced from 


Léger and Duboseq have noted also that the mucoid substance is 
deposited first in the nucleus. 


THE RAT-TRYPANOSOME, TRYPANOSOMA LEWISI. 501 


the crypts, have denser cytoplasm filling the cell throughout, except in 
the apical expanded portion of the cell; the cytoplasm stains deep 
grey or neutral tint after iron-hematoxylin and shows relatively few en- 
closures. The senile cells, on the contrary, are full of cavities, so that the 
cytoplasm has a spongy appearance throughout the cell and not merely 
in its apical region; and scattered through the spongy cytoplasm are 
grains, fine or but moderately coarse, which are stained black after iron- 
hematoxylin, red after Giemsa. 

The difference between the senile cells after the two methods of 
fixation is easily explained if the grains deposited in them are principally 
fat. In all the sections alike the fat has been dissolved away during the 
process of imbedding in paraffin. In the Flemming-fixed sections, how- 
ever, each fat-grain has reduced the OsO, to metallic osmium, and 
consequently is represented in the sections by a black mass, a model of 
the fat-globule in metallic osmium. In the sublimate-fixed sections no 
such reduction takes place, and the fat-globule is represented by an 
empty space; only the mucoid grains (if we are right in calling them 
so) remain in the cytoplasm, stained red or black according to the 
stain used. 

It should be mentioned finally that after sublimate-fixations the 
blood-débris is stained very much blacker by iron-hematoxylin, and 
holds the stain much more tenaciously than after Flemming-fixation. 
This is especially true of that part of the débris which penetrates down 
between adjacent epithelial cells, and which often remains jet-black 
after all the rest of the débris has become pale in tint. In consequence 
the cells of the columnar epithelium in sublimate-fixed sections are 
often seen to be separated by black masses, which careless observation 
might confuse with the black stain of the degenerated cells after 
Flemming-fixation, especially when, as often happens in such sections, 
the main mass of the débris has shrunk away from the epithelium into 
the centre of the stomach-lumen. Such a mistake could only be made, 
however, with powers too low to discern that the black masses are 
between the cells and not in them. 

The degenerated cells are thrown off bodily into the lumen of the 
stomach, which often contains great numbers of them in the blood- 
débris. There they are doubtless digested and absorbed along with the 
other contents of the stomach. Légerand Duboscq described a process 
of mucoid degeneration in which the entire cell, having a remarkable 
and deceptive resemblance to a gregarine, is engulphed by a basal cell ; 
ultimately the latter also degenerates, and is thrown off with the cell 
it has taken in (1. ¢., p. 451). We have seen nothing of this sort in the 
flea, in which basal cells do not occur in the epithelium of the stomach. 


502 E. A. MINCHIN AND J. D. THOMSON. 


(5) TECHNIQUE. 


We have already described above our methods of dissecting 
the flea and extracting from it the organs which it is required 
to examine for the presence of stages of T. lewisi. Here 
we propose to describe the methods by which the trypano- 
somes, when found, were preserved as permanent preparations 
for microscopic study. 

The organs of the flea, extracted in the manner described 
above, are at once examined carefully under the microscope 
for the presence of trypanosomes in their various phases of 
development. When trypanosomes were found in any of the 
internal organs, after note had been taken, or sketches made, 
of their forms, position, and other points of interest, we pro- 
ceeded to make permanent preparations of them. For this 
purpose the coverslip is carefully raised up, by means of the 
pair of fine needles that were used in the dissection of the flea, 
lifted off, and dropped at once with wet surface downwards 
into a suitable fixative. The slide is then handed to the 
collaborator or to an assistant, who places it bodily into a tube 
containing a small quantity of four per cent. solution of osmic 
acid. In the tube the slide remains about ten to fifteen 
seconds, tightly corked up, in order to fix the trypanosomes 
with the vapour of osmic acid. Subsequently the slide is 
fixed with absolute alcohol for about fifteen minutes and 
stained with Giemsa’s stain in the usual manner. 

For the fixation of the coverslip-films we used, in the earlier 
periods of our investigation, either Schaudinn’s fluid (corro- 
sive sublimate, saturated solution in distilled water, 100 c.c.; 
absolute alcohol, 50 c.c.; glacial acetic, a few drops) or 
sublimate-acetic (Hg Cl, saturated in H,O, 95 volumes; glacial 
acetic, 5 volumes). Both these fixatives gave results about 
equally good; it is difficult to choose between them. Latterly, 
however, we used only Maier’s modification of Schaudinn’s 
fluid (distilled water, 200 c.c.; absolute alcohol, 100 c.c.; 
sodium chloride, 1:2 grm.; HgCl,, 10 grm.), since this 
appeared tous to give better preservation, and, in particular, 


THE RAT-TRYPANOSOME, TRYPANOSOMA LEWISI. 503 


less shrinkage of the bodies of the trypanosomes, than the 
others. The fluid being put into a large watch-glass, the 
coverslip is dropped into it with the film downwards. ‘The 
coverslip usually sinks in the fluid and then rests on its 
corners on the rounded bottom of the watch-glass, so that 
the film itself escapes any friction or injury. The coverslips 
are left in the fixative from ten minutes to half-an-hour or 
longer (the exact time appears to be immaterial), and are 
then passed through 50 and 70 into 90 per cent. alcohol, 
where they can be kept until it is convenient to stain them. 

The coverslip films were stained almost invariably with 
Heidenhain’s iron-hzmatoxylin, using 35 per cent. iron-alum 
solution and } per cent. hematoxylin-solution, both in dis- 
tilled water. ‘he film, after having been brought down 
through graded strengths of alcohol (80, 70, . . . 10 
per cent.) to water was left about twenty-four hours in the 
iron-alum, then as long in the hematoxylin. Immediately 
before using the hematoxylin-solution a few drops of a 
saturated watery solution of lithium carbonate was added 
to it, drop by drop, until the solution, when shaken up, 
was a bright claret-red colour. After the film had been 
twenty-four hours in the hematoxylin-solution the differ- 
entiation of the stain was carried out under control by the 
microscope in a weak (light brown) watery solution of iron- 
alum. When differentiation was complete the film was washed 
in a current of tap-water for at least twenty minutes, then 
rinsed in distilled water and brought up through graded 
strengths of alcohol to absolute alcohol. At this stage the 
coverslip was usually dipped for a moment into Lichtgriin- 
picric solution (Lichtgrin, 1 grm.; picric acid, 4 grm.; 
absolute alcohol, 100 e.c.), then washed again in absolute 
alcohol, passed through xylol, and mounted in pure xylol- 
balsam on aslide. The Lichtgriin stain must be used very 
rapidly, as it stains intensely. 

In this way two preparations were obtained of the contents 
of each organ—one on the coverslip, the other on the slide— 
and as a rule trypanosomes were found more or less 


504. E. A. MINCHIN AND J. D. THOMSON. 


abundantly on both of them, so that it was possible to 
compare corresponding phases of the development prepared 
by distinct methods of technique. It is very important, 
however, that the operation of removing the coverslip and 
fixing the films should be performed very rapidly and 
expeditiously, in order to avoid any drying taking place. 
The coverslip is particularly liable to dry, since the film of 
liquid that adheres to it is very thin; the slide, on the 
contrary, does not dry so quickly. A coverslip that has 
dried before fixation is quite useless for staining by the iron- 
hematoxylin method; the trypanosomes acquireacharacteristic 
shiny appearance, as if they had been glazed, and when the 
stain is extracted in order to differentiate the preparation, it 
does not come out of the cytoplasm evenly, but gives a 
blotchy appearance, with no sharp differentiation of the 
nucleus or flagellum. It sometimes happens that a coverslip- 
fim may be otherwise satisfactory, but may have dried 
slightly at or near the edges, thus affording opportunities for 
comparing the effects of desiccation on the trypanosomes 
with the condition of others that have never been dried. It 
is then seen that, in addition to the defective staining 
already described, the trypanosomes are flattened and 
distorted in various ways. 

The fragments of tissue in the dissection adhere, for the 
most part, to the coverslip ; it is not possible, however, to 
make out anything of trypanosomes which remain within the 
organs in film-preparations, and it is therefore necessary to 
tease up the organs well, after dissecting them out, in order 
to set free the trypanosomes. In the case of those phases 
which are attached to the gut-wall many remain so attached 
even when the wall is teased up, but a certain number are 
usually set free. When such forms are seen in the fresh film 
they should be dislodged, as far as possible, by tapping 
gently on the coverslip with a needle. 

In some cases a coverslip-film which had been stained with 
iron-hematoxylin was unmounted by dissolving the Canada 
balsam in xylol, after the trypanosomes on it had been 


THE RAT-TRYPANOSOME, TRYPANOSOMA LEWISI. 505 


studied and drawn, and the hematoxylin-stain completely 
extracted by placing it for twenty-four hours in a 3} per 
cent. solution of iron-alum. The coverslip was then washed 
for an hour in a current of tap-water, and could then be re- 
stained by some other method—for example, Twort’s stain. 
Trypanosomes that had been already drawn after the 
hematoxylin-staining could then be drawn again after bemg 
stained in a different manner. This double staining did not 
seem to injure the trypanosomes in any way, but it is note- 
worthy that after re-staining with Twort’s stain they always 
came out a little smaller, when re-drawn with the camera lucida, 
than they had done previously after the hematoxylin-stain 
(compare figs. 260-63, Pl. 42, with figs. 260a—263a, PI. 38). 

When, as sometimes happened, the trypanosomes were so 
scanty on the coverslip as to require prolonged searching to 
find them, it was often very difficult to judge the right 
amount of extraction of the hematoxylin in the process of 
differentiation by means of iron-alum. Morever, a degree 
of differentiation which is sufficient for trypanosomes in 
the thinner parts of the film is insufficient for the thicker 
parts. Hence it was often necessary to unmount the pre- 
parations and differentiate them further, perhaps two or 
three times, before the right degree was attained. It is 
difficult to judge of the required differentiation by the 
fragments of tissue in the films, since the minute bodies of 
the flagellates give up the stain much more quickly than the 
relatively thick tissue-cells, and in a preparation in which the 
latter are satisfactorily differentiated the trypanosomes 
become mere ghosts, requiring to be re-staimed altogether. 
The counter-stain with the Lichtgriin-picric mixture was 
found to show up the cytoplasm and flagellum of the 
trypanosomes more clearly. 

However carefully the preparations have been made, it is 
often difficult to make out clearly and with certainty the 
structural details of some of the minuter phases of the life- 
cycle, and for this purpose the best optical apparatus was 
required, both as regards the objectives and the illumination 


506 E. A. MINCHIN AND J. D. THOMSON. 


used. All trypanosomes in the permanent preparations were 
drawn by Miss Rhodes, under our supervision, with the 
camera lucida at a constant scale of magnification which was 
as nearly as possible 3000 diameters in the case of the film- 
preparations, 2000 diameters in the case of sections. 

Our study of the development of T. lewisi in the flea 
was based principally upon the examination of films, made as 
described above, but it was found necessary also to cut 
sections both of the stomach and rectum of the flea. The 
following is an account of the technique employed by us in 
preparing sections of the stomach; the same applies to 
sections of the rectum, the only difference being that the 
stomachs were cut transversely, the recta longitudinally. 

The stomachs of which sections were cut were taken from 
fleas fed eighteen, twenty-four, or thirty-six hours previously 
on an infected rat; the fleas themselves had been collected 
from the non-infected breeding-cage and kept hungry for 
about three days before being put on the infected rat. The 
stomach in each case was carefully dissected out from the 
flea, if possible without puncturing or injuring the stomach, 
in a drop of salt-solution on a slide, and then plunged into 
the fixative by inverting the slide in such a way that the 
stomach alone, all other parts of the flea having been 
removed, was in a hanging drop. If the stomach was 
ruptured or punctured in the process of extraction it was 
not, as a rule, preserved, except perhaps as a smear after 
teasing it up. 

A number of different fixatives were tried, but the best 
results were obtained with Flemming’s fluid! and Maier’s 
modification of Schaudinn’s fluid, and especially with the 
former. After Flemming the histology of the stomach is 
extremely good in all details; the blood fills the whole section 


1 The strong solution, made up as follows: a gramme tube of osmic 
acid is broken into a clean bottle, and to it is added distilled water, 
50 ¢.c.; 1 per cent. solution of chromic acid in water, about 187°5 c.c. ; 
and glacial acetic about 12°5 ¢.c.; the whole allowed to mix and 
dissolve. 


THE RA'T-TRYPANOSOME, I'RYPANOSOMA LEWIST. 507 


andis not shrunk away from the wall, and the trypanosomes, 
both free and intracellular, are well-preserved both in 
structure and form, and they stain well either with iron- 
hematoxylin or Giemsa, especially the former. After Maier’s 
fluid the histology of the stomach-tissue is ndt so good; the 
cells are shrunk and the minute structure of the nuclei is 
deformed. It is evident from a careful study of the prepara- 
tions that the defects of Maier’s fluid are due to unequal or 
differential penetration of its constituents ; the alcohol evi- 
dently diffuses into the tissues first and produces the shrinkage 
and deformation of the nuclei; the sublimate does not get to 
the various tissue-elements until they have already been fixed 
in a defective manner by the alcohol. The blood-débris is 
also much shrunk after the Maier; while the greater part, 
sometimes the whole of it, contracts to form a central mass in 
the section, a certain amount remains usually adherent to the 
epithelium at the periphery, leaving an irregular empty ring- 
shaped space between the central and peripheral zones of the 
blood-débris. But to compensate for these disadvantages, 
the trypanosomes are extremely well-preserved and stain 
admirably with Giemsa’s stain; some of our stomach-sections 
prepared in this way are as clear and demonstrative, so far as 
the trypanosomes are concerned, as any smear or film- 
preparation ; in fact more so in the case of the large 
“spheres,” which do not suffer so much from the tendency 
to opacity which is so disagreeable a feature in the smears. 
One is here confronted with the extraordinary difference, 
familiar to everyone who has worked at trypanosomes, 
between the reaction of these parasites, and that of tissue- 
cells, to the ordinary fixatives and stains used in cytological 
technique. 

Whatever the fixative used, it was allowed to act for about 
an hour. ‘The stomachs preserved in Flemming were well 
washed in tap-water and then brought up through a series of 
alcohols of gradually increasing strength ; those preserved in 
Maier were transferred from it direct to 50 per cent. alcohol. 
In either case the objects were brought up to 90 per cent. 


508 E. A. MINCHIN AND J. D. THOMSON. 


alcohol and there fixed on liver preparatory to being 
imbedded for section-cutting. Amyloid human liver was 
used. A moderately thin slice of a block of liver preserved 
in alcohol was cut by hand with a razor wetted with alcohol, 
and floated into a shallow glass vessel with a flat bottom, 
placed on the stage of the dissecting-microscope, and contain- 
ing 90 per cent. alcohol to the depth of about a centimeter. 
The stomachs, taken up in a pipette of suitably coarse calibre, 
were placed on the slice of liver and carefully arranged side- 
by-side, their axes parallel to one another and similarly 
orientated, with their proventriculi all at the same level 
and all pointing in one direction, their pylori in the oppo- 
site direction. Then a tiny drop of glycerine and albumin 
solution, such as is used commonly for sticking sections 
on slides, was taken up on the point of a needle and 
caused to touch the surface of the alcohol immediately above 
the stomachs. The dense albumin-solution falls at once 
through the alcohol and spreads out over the stomachs on the 
liver; at the same time the glycerine is extracted and the 
albumin coagulated by the alcohol, with the result that 
the stomachs are stuck to the slice of liver. From six to 
nine stomachs were thus attached side-by-side on a slice 
of liver. As the stomachs, before being stuck on, are very 
liable to roll about or become shifted in position with the 
slightest disturbance or touch of the microscope, it was found 
best in practice to put them on not more than three at a 
time; that is to say, three stomachs having been arranged 
and fixed upon the liver, three more are then put on beside 
them. When the required number of stomachs have been 
stuck on, the slice of liver is trimmed with a scalpel into a 
rectangular form, in such a way that the longitudinal axes of 
the stomachs are parallel to the shorter sides of the rectangle; 
so that by cutting sections of the liver parallel to the longer 
sides of the rectangle the stomachs are all cut transversely at 
the same time. 

We have thought it worth while to describe the method 
of fixing the stomachs on liver, although no novelty is 


THE RAT-TRYPANOSOME, TRYPANOSOMA LEWISI. 509 


claimed for it, in some detail, as it may not be familiar 
to some investigators working on similar objects, and 
because it is a procedure which saves much time and 
trouble. In the first place, it is much easier to imbed a 
relatively large block of tissue than a number of separate 
tiny little stomachs, and the orientation of the objects can 
be made much more accurate. In the second place, a great 
economy of labour in the section-cutting and of space in 
the slides and preparations is effected. To have a number 
of stomachs cut in the same section diminishes the labour 
of looking through the preparations under the microscope, 
and the presence in the section of the slice of liver makes it 
much easier to go from one section to the next under the high 
power. Thirdly, with a little experience the liver itself 
furnishes useful guidance in staining the sections, especially 
by the iron-hematoxylin method; one soon learns what 
degree of extraction of the stain from the liver-cells gives 
the best results for the trypanosomes, so that the process of 
differentiation can be carried out under low powers of the 
microscope—a great advantage. And finally, since it may be 
assumed that all the stomach-sections contained in one and 
the same microtome-section have received exactly the same 
treatment, it is legitimate to ascribe the very considerable 
differences seen in different stomachs in the same section to 
constitutional or functional differences in the stomachs them- 
selves and not to varying local effects of the stain. 

The stomachs, after being fixed to the liver in 90 per cent. 
alcohol, were imbedded in the usual way in paraffin, with a 
melting point of about 54°C. Methods of celloidin-imbedding 
were tried, but yielded no advantages to compensate for the 
extra trouble, especially that of extracting the celloidin from 
the sections—an indispensable preliminary to staining them. 
The best thickness for the sections of stomachs was found 
to be 6m; with less than that the trypanosomes are too 

* One of us (EK. A. M.) first became acquainted with this method in 


1891 from fellow-workers in the Zoological Station at Naples, and has 
practised it constantly ever since. 


510 E. A. MINCHIN AND J. D. THOMSON. 


fragmentary. The recta may with advantage be cut thinner 
than 6, since the crithidial forms are very minute. 

Various methods of staining were tried on the sections, 
but the results of the trials were that we kept finally in 
practice to two methods only, namely, iron-hematoxylin 
(Heidenhain), followed by Lichtgriin-picric in absolute alcohol 
as a counter-stain, and Giemsa’s method. For the iron- 
hematoxylin method the sections were treated first as has 
been described above for the coverslip-films. The Lichtgriin- 
picric, which stains very rapidly, was merely washed over the 
sections for a moment and then washed off again with 
absolute alcohol. Giemsa’s stain was used, according to the 
published prescription, as follows: The sections have their 
paraffin removed, and are brought down to water in the 
ordinary way. They are then washed in tap-water and put 
into dilute Lugol’s solution (1 ¢.c. of Lugol to 25 e.c. of 
distilled water) for ten minutes. After this they are rinsed 
quickly in tap-water and put into a 0°5 per cent. watery 
solution of hyposulphite of soda for ten minutes. Next they 
are washed in a current of tap-water for five minutes or 
longer, and then put into the stain. The distilled water used 
to dilute the Giemsa-stain has to be neutralised in the way 
prescribed by Giemsa.t The sections were first placed in 
fairly strong stain—say, 1 drop of Giemsa to 1 c.c. of 
neutralised distilled-water—for about an hour, and then were 
left overnight in a weaker stain—1 drop of Giemsa to 4 or 5 c.c. 
of neutral distilled-water. The excess of stain is removed 
by rinsing in water, and after the excess of water has been 
drained off differentiation of the stain is carried out with 

1 A measured volume of the distilled water to be neutralised is taken, 
and to it are added a few drops of hematoxylin-solution (5 per cent. in 
distilled water), sufficient to tint it. Then a very weak solution (1 per 
cent. in distilled water) of potassium carbonate is added drop by drop, 
the water being well shaken after each drop has been added, and left 
for a minute or two, until the colour of the tinted water changes from 
yellowish-red to reddish-purple. In this way the number of drops of 


the carbonate-solution required for neutralising a given volume of the 
distilled water is known. 


THE RAT-TRYPANOSOME, TRYPANOSOMA LEWISI. 511 


different strengths of acetone mixed with xylol, beginning 
with 95 per cent. acetone used for a very short time, in order 
to dehydrate the sections and extract the stain, and ending 
with pure xylol, after which the slides are mounted in dammar 
or Canada-balsam. 

Of the two staining methods principally used, iron-hema- 
toxylin gave admirable results after Flemming, especially for 
the intracellular stages; for the extracellular trypanosomes 
this stain is not so satisfactory, owing to the fact that the 
blood-débris, especially in the earlier stages of digestion, 
stains very intensely with it and refuses to give up the stain— 
at any rate not until after it has been all extracted from 
the cells and parasites. Consequently trypanosomes free in 
the blood may be entirely obscured by the opaque, deeply- 
stained débris, and hence quite invisible. The black stain of 
the blood-débris is even more intense after Maier than after 
Flemming ; sections of stomachs fixed in Maier less than 
thirty-six hours after feeding are hopeless for the iron- 
hematoxylin staiu, so far as the free trypanosomes are con- 
cerned, and those fixed in Flemming are not much better. 
By Giemsa’s method, on the other hand, the trypanosomes in 
blood-débris are sharply differentiated and admirably shown; 
in the cells they are also good, better, perhaps, as “ show ” 
preparations, but not so precise in minute cytological details 
as by the iron-hematoxylin method. 

To sum up the results of our experience in the technique 
of stomach-sections, we recommend: (1) Flemming’s fluid, 
followed by iron-hematoxylin and Lichtgriin-picric ; and (2) 
Maier’s fluid, followed by Giemsa’s stain. These two methods, 
supplementing each other, may be relied upon to reveal all 
essential details of the intimate life of the trypanosome and 
of the disturbances produced by it in the tissues of the host. 


vou, 60, PART 4,—NEW SERIES.% 36 


512 E. A. MINCHIN AND J. D. THOMSON. 


PART II.—THE DEVELOPMENT OF TRYPANOSOMA 
LEWISI IN THE FLEA. 


(1) Generat Inrropuction. 


From the results of experiments, described further below, 
it is shown that fleas fed on rats infected with Trypanosoma 
lewisi do not become infective to rats again until a period of 
at least five or six days has elapsed from the time that the 
fleas first ingested blood containing trypanosomes. From these 
experimental data it may be inferred that the developmental 
cycle of T. lewisi in the flea requires a minimum of five days 
for its complete course. The conclusions drawn from the 
experiments are confirmed by direct observation, since it is 
found, as will be described presently, that the little stumpy 
trypanosome which is the final form of the development in 
the flea, makes its first appearance in the rectum of the flea 
about five days after the development begins. 

During the entire course of its development the trypano- 
some is confined to the alimentary canal proper of the flea, 
and is found in the stomach, intestine, and rectum; it is 
never found in the body-cavity (heemoccele), and by a series. 
of observations and experiments, which in our opinion are 
exhaustive (see below), we have convinced ourselves that the 
trypanosome does not penetrate into the salivary glands. It 
may, however, occur in the Malpighian tubules exceptionally, 
as the small crithidial form, characteristic of the rectal phase, 
attached to the wall of the tubes at or near their proximal 
opening into the proctodzeum. 

The developmental cycle can be divided conveniently into 
phases characteristic of the parts of the gut in which the try- 
panosomes are found, and we can thus distinguish a stomach- 
phase and a rectal phase. These distinctions are useful and 
natural, but their sharpness is blurred by not infrequent 
variations in the course of events; thus forms belonging 
normally to the rectal phase may sometimes be found in the 


THE RAT-TRYPANOSOME, TRYPANOSOMA LEWISI. 413 


pyloric region of the stomach, though the converse case of 
the typical stomach-phase occurring in the rectum is not 
found. We may consider these phases first in their normal 
and typical modes of occurrence, and deal with the variations 
subsequently. 

The stomach-phase (Pls. 36-39) is the first period of the 
development and is characterised by a peculiar mode of multi- 
plication on the part of the trypanosomes, which penetrate 
into the epithelial cells lining the stomach and there repro- 
duce themselves by a process of multiple fission. Hence in 
this period of the development free and intracellular forms 
can be distinguished. The stomach-phase is of short. 
duration, perhaps in some cases lasting not more than 
twenty-four hours, in others two or three days, in rare cases 
four or even five days, but probably always terminated by the 
second feed of the flea, counting as the first feed that by 
which the flea became infected. 

In the intestine the trypanosomes find, as a rule, no resting 
place, but merely pass through it on their way to the rectum. 
Hence, the forms found in the intestine are usually active, 
migratory forms which have completed the stomach-phase and 
are on their way to the rectum to initiate the rectal phase. 
Occasionally, however, forms similar to those characteristic 
of the rectal phase may be found attached to the wall of the 
intestine, especially near the pyloric opening. 

The rectal phase (Pls. 41 and 42) consists chiefly of small, 
often minute individuals, which are crithidial in structure 
and are attached by the tip of the flagellum to the wall of the 
rectum, where they keep up a continual multiplication by 
binary fission. The crithidial form of the development takes 
origin in the rectum and is first established there, but may 
migrate forward to the pyloric region of the stomach later on. 
When once established in the flea, the crithidial phase endures, 
probably, as long as the flea lives, and thus constitutes a 
permanent stock of the parasite, enabling the infectivity of 
the flea to be maintained without renewal of the infection. 
From the crithidial phase arise by modification of individual 


514 E. A. MINCHIN AND J. ‘D. THOMSON. 


crithidial forms the small trypanosome-forms by which the 
infection of the rat is brought about, and which are the final 
forms of the developmental cycle in the flea. 

By no means all the trypanosomes, however, which are 
taken up from the rat by the flea undergo the course of 
development sketched out briefly in the foregoing paragraphs. 
By experiment it is found that only a relatively small number 
of the fleas fed on infected rats become infective, apparently 
not more than one flea in four, on an average (see below) ; 
and these results are confirmed by direct observation. If a 
number of fleas are fed on a well-infected rat, trypanosomes 
will be found in the gut of all the fleas dissected and examined 
a short time after feeding; but the longer the interval between 
the feeding and the examination of the fleas the larger the 
proportion of the fleas in which the trypanosomes have dis- 
appeared or become very scanty, until finally trypanosomes 
will be found in but few (see Tables A and B (2) above). Pro- 
bably the percentage of fleas in which the trypanosome 
succeeds in establishing itself permanently may be taken, on 
the average as about 25 per cent. (see p. 663 below). It follows 
that in about 75 per cent. of the fleas which digest blood con- 
taining T’. lewisi the parasites die out altogether, and it is 
probable that in all the fleas a certain number of the ingested 
trypanosomes die off, since fleas that have been fed on a rat 
with trypanosomes swarming in the blood may exhibit a very 
scanty infection of the gut at any subsequent period. 

From these data it is to be expected that together with 
developmental forms of the trypanosomes, various stages in 
their degeneration would also be found in the fleas, at least 
during the first few days after the parasites were ingested 
by them, and this expectation is fully realised. Itis necessary, 
therefore, to recognise a degenerative series of forms (PI. 43) 
as well as a developmental series in the gut of the flea, and to 
distinguish carefully the two series from one another. Any 
particular flea, when dissected and examined, may present an 
extraordinary medley of different forms of the trypanosome. 
To distinguish between the different forms and to refer each 


THE RAT-TRYPANOSOME, TRYPANOSOMA LEWISI. 515 


form to its proper position and sequence in the series, whether 
developmental or degenerative, is our task, and it is no light 
one. When we were at an earlier stage in our investigations 
we did not recognise sufficiently the importance of the 
degenerative series, and consequently tried to interpolate 
degenerative forms into the developmental series, greatly to 
our own confusion. On the other hand it is necessary to steer 
very clear of a tendency to explain any form as degenerative, 
of which the developmental position is not immediately clear ; 
thus we were at first inclined to regard the peculiar recurved 
forms in the stomach as degenerative, until we discovered the 
intracellular multiplication and were thereby enabled to refer 
the recurved forms to their true position. 

In the problem of piecing together and reconstructing 
the sequence of the two series, developmental or degenerative, 
there is, to begin with, a known and fixed starting point for 
each, namely, the ordinary form of T. lewisi as it occurs in 
the blood of the rat. Further clues are obtained. by linking 
together, through gradual transitions, the forms seen in the 
fleas, but more especially by the study of “‘time-fed” fleas, 
that is to say, fleas dissected and examined at known periods 
of time after they have been fed on the infected rat. The 
part of the gut in which a given form occurs is a further 
guide as to its significance; and all data and conclusions 
obtained from observation are controlled and checked by 
the results of experiment, especially useful in determining 
the final form of thedevelopment. Guided by these various 
considerations we have arrived at the conception, set forth 
below in fuller detail, of the changes undergone by the 
trypanosome in the flea (see especially Pl. 45 and description). 
In our account we describe separately the two series which 
we regard as developmental and degenerative respectively ; 
but it must be pointed out that while these two series are 
very distinct and easily recognisable as a whole, certain 
forms or stages of the one series are sometimes very difficult 
to distinguish decisively from very similar forms belonging in 
reality to the other series. Consequently it is impossible to 


516 E. A. MINCHIN AND J. D. THOMSON. 

be free from doubt, occasionally, with regard to the place 
to be assigned to a particular specimen or type of indi- 
vidual. 

It remains only to be stated at this point that we adhere to 
the following nomenclature for the parts of the body of the 
trypanosome or crithidia: Blepharoplast for the basal granule 
of the flagellum; kinetonucleus for the smaller, tropho- 
nucleus for the larger, of the two nuclei. In order to save 
space we shall, however, use for the kinetonucleus the 
symbol » (plural nz) and for the trophonucleus the symbol 
N (plural NN). 


(2) Tue DEVELOPMENTAL SERIES. 


(4) The Stomach-Phase. 


The blood ingested by the flea passes in the first instance 
into the stomach, that portion of the digestive tract which is 
derived from the embryonic mid-gut or mesenteron, and 
which is lined by a layer of epithelium representing the true 
hypoblast or endoderm of the embryo. In the post-embryonic 
stages of the insect, this part of the gut is characterised by 
the absence of the chitinous cuticular lining secreted by the 
ectodermal! epithelium of the parts anterior or posterior to if, 
namely, the stomodzeum, comprising the pharynx, cesophagus, 
and proventriculus, and the proctodeum, comprising the 
intestine and rectum. The boundary between mid-gut and 
hind-gut is further indicated by the origin at this point of 
the Malpighian tubules. 

in what may be called a normal feed, the flea fills the 
stomach and proventriculus alone. It is not an infrequent 
occurrence, however, for some fleas to gorge themselves to 
such an extent that the freshly ingested blood not only fills 
the stomach completely, but overflows beyond it into the 
intestine and rectum ; we have observed this to happen most 
frequently in the case of female fleas, rarely in the case of 
males. In such cases some of the ingested trypanosomes 


THE RAT-TRYPANOSOME, TRYPANOSOMA LEWISI. 517 


may be carried on at once into the proctodzal regions of the 
gut, but all such trypanosomes degenerate, and need not be 
reckoned with in the developmental series, the first phases of 
which take place always in the stomach alone. 


(a) The Extracellular Trypanosomes. 


The trypanosomes introduced into the stomach very soon 
begin to undergo changes (PI. 36, figs. 1 and 2; Pl. 37, figs. 
47 and 48). The first change is probably purely physio- 
logical, since long before any alteration is observable in form 
or structure these ingested trypanosomes seem to have lost 
their power to infect when injected subcutaneously into clean, 
susceptible rats (see below, p. 634). The next change 
observed in these ingested trypanosomes may be seen on 
examining microscopically the contents of a flea’s stomach 
four to six hours after the first feed on an infected rat. A 
certain number of trypanosomes will then be seen to pass 
rapidly in a straight course across the field of the microscope 
with their flagella directed anteriorly. ‘lhe posterior third 
ot the body is held more or less straight and appears more 
rigid, as it does not share in the rapid undulations of the 
anterior end of the body. The movements of these trypano- 
somes thus contrast strongly with the sinuous, serpentine and 
wriggling rather than progressive movements characteristic 
of the trypanosomes in the blood. When not actively pro- 
eressing, the trypanosomes in the stomach have a tendency to 
attach themselves by the tips of their flagella to pieces of 
débris, to the wall of the stomach, or to the surface of any 
other firm body. The stiffening of the trypanosome-body is 
probably due to increased tension of the cytoplasmic contents 
produced by absorption of fluid from the ingested blood as it 
undergoes alteration in the process of digestion. As a result 
of absorption or imbibition of fluid, the body of the parasite, 
previously more or less distinctly flattened, acquires a cylin- 
drical and more rigid contour. If this explanation be correct, 
it follows that the first stimulus to developmental change is 


518 E. A. MINCHIN AND J. D. THOMSON. 


to be ascribed to differences of osmotic tension in the fluid 
medium, as has been shown experimentally by Miss Robertson 
(1911) to be the case in the development of the trypanosomes 
of fishes. 

In stained preparations most of the ingested trypanosomes 
show at first little modification from. the ordinary blood-try- 
panosomes. In rare instances the nuclei may be approximated 
(Pl. 36, fig. 2). Some show a darker staining-reaction of the 
posterior third of the body, which appears, from the back- 
ward position of 7 in such forms, to bea sign of degeneration 
beginning to set in (compare PI. 43, figs. 289, 290). 

In their free active state the trypanosomes in the stomach 
are never found to be undergoing multiplication by any form 
of fission, and it is doubtful if they undergo any develop- 
mental changes further than those described above, until 
after they have multiplied within the cells of the lining 
epithelium of the stomach. The multiplication of the stomach- 
phase takes place solely within these cells, and although, in 
strict chronological order, we should now describe the intra- 
cellular stages of multiplication, it is more convenient for 
purposes of description to divide the stomach-phase into 
“ free’ and “intracellular” stages, and to describe all the 
free developmental forms before describing the intracellular 
multiplication. 

It can be established by direct observation that the process 
of intracellular multiplication produces a long, free type of 
trypanosome which may be characterised by the term 
“ crithidiomorphic,” because while externally similar in form 
and movements to a large crithidial type of flagellate, it 
lacks, as a rule, the diagnostic structural feature of a true 
crithidial form, since only exceptionally is » found actually 
beside or in front of N (Pl. 36, figs. 3-11, Pl. 57, figs. 49-57). 
We shall now proceed to describe in more detail this late, 
free form of trypanosome. It must, of course, be understood 
that as several generations of the intracellular stage may 
follow each other in succession (see below), free and intra- 
cellular forms in all stages of development can be found 


THE RAT-TRYPANOSOME, TRYPANOSOMA LEWISI. 519 


together in the same stomach. We believe, however, that 
the typical crithidiomorphic type of trypanosome always 
follows an intracellular stage, that in its less developed form 
it is the direct product of intracellular multiplication, and 
that though in this form it may again enter an epithelial cell 
and multiply, it is, in the more advanced form, the highest 
developmental type of the stomach-phase and is destined to 
pass down the intestine into the rectum where, after under- 
going further modification, it initiates the characteristic 
crithidial rectal phase to be described below. 

In the living condition the crithidiomorphic form pro- 
gresses at a great pace in a straight line with the flagelluia 
directed anteriorly (“mouvement en fléche”’), in much the 
same manner as does the early stomach-form above described. 
It is, however, considerably longer and the posterior end is 
more rigid and swollen, often distinctly clubbed. Owing to 
the rapid motion and imperfectly straight body, the clubbed 
appearance is exaggerated in the living condition, but stained 
preparations also show that some of the trypanosomes are 
distinctly clubbed in shape. Like the early stomach-form, 
when not actively progressing the crithidiomorphic type has 
a strong tendency to attach itself by the tip of its flagellum 
to cells or débris, etc. Apart from its size the distinguishing 
characteristic between this and the earlier form is the marked 
approximation of the two nuclei, best seen in stained 


preparations (Pl. 36, figs. 7, 8, 10; Pl. 37, figs. 50, 51). 


It has been mentioned that while, as a rule, the long, free stomach- 
trypanosomes have n behind N, it is found in a few cases that they have 
the typical crithidial structure with » in front N. We have observed 
altogether but three instances in which such forms occurred in sufficient 
abundance to make them worthy of special note. The first and most 
striking was the case of a flea taken from a bell-jar in which a number 
of fleas had been kept for some time with an infected rat, so that the 
length of time since the flea had ingested the parasites was not known. 
The body-cavity of the flea contained a cysticercoid of Hymenolepis 
diminuta (vide Nicoll and Minchin, 1911). The intestine of this flea 
showed a peculiar malformation in the form of a globular pouch-like 
appendix, distended with red fluid, and due apparently to an obstruc- 


520 E. A. MINCHIN AND J. D. THOMSON. 


tion or strangulation of the intestine. The stomach, examined fresh, 
was seen to contain a great number of active trypanosomes, some of 
which were adhering together in couples, and in the intestine a clump 
of attached forms was seen near the origin of the Malpighian tubules. 
In the preparations of the stomach of this flea a great number of try- 
panosomes were found showing every possible gradation of structure, 
from forms similar to the ordinary blood-trypanosomes to a long 
crithidial type with n far in front of N (Pl. 36, fig. 11 and PI. 37, 
figs. 60-66). Many of these were found closely adherent in couples, just 
as had been seen in the fresh state, each such couple being composed of 
two crithidial forms in most cases, but sometimes of two ordinary 
forms (PI. 36, fig. 12, and Pl. 37, figs. 67, 68). In every couple seen the 
two individuals appeared quite distinct and showed no signs of actual 
fusion; one couple was found attached téte béche (as in PI. 43, 
fig. 310). In the preparation of the rectum and intestine (preserved 
together) a few similar large trypaniform or crithidial individuals were 
seen, and also a fair number of dwarfed, degenerative forms, but no 
couples. 
Special mention has been made of this flea because we were at first, 
and remained for some time, under the impression that the couples seen 
represented a true sexual fusion, and that we had discovered the sexual 
phase of the trypanosome. We have been quite unable, however, to 
confirm this notion or to find a similar state of things in any other flea 
of all those examined by us, and we now regard the state of things 
found by us in this particular flea as exceptional and abnormal, in rela- 
tion probably to the malformations noted by us in the flea itself. It is 
possible that the malformed condition of the intestine prevented, to some 
extent, the passage onwards of the trypanosomes from the stomach, and 
so caused an arrest of development in the parasites, in which the 
tendency towards the crithidial type of structure became realised to its 
fullest extent. The coupling of the trypanosomes must then be regarded 
as agglomeration due to abnormal and unfavourable conditions, though 
in no case were more than two trypanosomes seen adhering together. 
The second case in which the long crithidial forms were prominent 
was in a flea of a batch which had been fed onan infected rat three days 
before being examined and dissected. The fleas had been kept in an 
incubator at a temperature of 25° C. after the infective feed, and had not 
been fed again. In one of the fleas long crithidial forms with m in front 
of N, were fairly numerous, together with intracellular multiplicative 
stages, in the stomach (PI. 37, figs. 56,57); in the rectum one active 
trypanosome of the long stomach-type and a clump of degenerative 
forms were seen in the fresh state. 
The third case to be noted was in a flea of a batch which had been fed 
twenty-four hours previously on an infected rat. There was nothing 


THE RAT-TRYPANOSOME, TRYPANOSOMA LEWISI. 521 


special to note about this flea; the stomach contained many long active 
trypanosemes, with x and N closely approximated, or with nm in front of 
N (PI. 36, figs. 9,10), and also some dwarfed degenerative forms, but no 
multiplicative stages. In the rectum a few clumps of degenerative 
appearance and some developmental forms were seen. 

Besides these three cases which have come under our observation, in 
which the long crithidial type was conspicuously abundant in the 
stomach, we have noted the occasional occurrence of this type at 
various ages—twenty-four hours, forty-eight hours, and sixty hours— 
after the flea had fed on an infected rat. It is evident that it must be 
regarded as exceptional for the trypanosomes to reach the complete 
crithidial condition in the stomach, and that no special significance can 
be attributed to the crithidial form in this part of the life-cycle, 
although in rare instances and under special circumstances it may be 
abundant. 

It will be clear from the foregoing remarks that we are quite unable 
to agree with the statements of Swellengrebel and Strickland (1910): 
who, having examined two fleas one day after feeding on the infected 
rat and five others two days after the infected feed, describe the trans- 
formation of the long crithidiomorphie type of trypanosome into the 
long crithidial type as the normal and usual method of development in 
the stomach. On the strength of somewhat more extended experience, 
we consider the long crithidial form to be of highly exceptional occur- 
rence, both in the stomach and elsewhere, at so early a period of the 
development, as already stated; we can only explain the results of 
Swellengrebel and Strickland on the supposition that they were so 
unfortunate as to have chanced upon abnormal fleas, similar to the 
three cases described by us above, or that they may have regarded as 
crithidial forms the very commonly-occurring recurved forms, which 
they do not describe at all. 


In view of what is known with regard to both the later 
development of T. lewisi in the flea and the life-cycle of 
other trypanosomes in their invertebrate hosts, it is evident 
that the crithidial type of form and structure is the principal 
and most characteristic phase of the development, and that 
there is a pronounced tendency for the trypanosome to assume 
crithidial characters when taken up by the flea—a tendency 
which asserts itself more strongly after the trypanosomes 
have undergone multiplication in the cells of the lming 
epithelium of the stomach. So long, however, as the trypa- 
nosome remains in the stomach the atavistic tendency towards 


522 E. A. MINCHIN AND J. D. THOMSON. 


the assumption of the crithidial form (+) does not normally 


(or, at all events, usually) get beyond the crithidiomorphic 


form (=). Occasionally, nevertheless, the crithidial form 
n 


asserts itself, as it were, even during the stomach-phase ; more 
especially, perhaps, under the influence of any circumstances 
which tend to retard the development of the trypanosome and 
retain it in the stomach after it is ripe for passage into the 
proctodeum, but not infrequently even under conditions 
which cannot be asserted to be in any way abnormal. 


(6) The Intracellular Multiplication of the 
Trypanosome. 


As already stated, the multiplication never takes place in 
the free, active condition of the trypanosome, but only after 
it has penetrated into one of the large epithelial cells lining 
the stomach, within which it goes through a process of 
multiple fission to produce a number of daughter-individuals. 
which escape from the cell and pass back into the lumen of 
the stomach as free trypanosomes again. The whole pro- 
cess of intracellular multiplication, so far as it could be made 
out by observation of living trypanosomes in the stomachs of 
freshly-dissected fleas, was described by us in our preliminary 
report (1911) ; we had not then had sufficient time or oppor- 
tunity to make detailed studies, which present peculiar diffi- 
culties, of the multiplication in preserved and stained material. 
The ordimary smear-methods seldom permit any finer details 
to be made out of the trypanosomes within the cells, on 
account of the large size and thickness of the cells and con- 
sequent opacity of the preparation. It is only possible in 
smears to study the stages of multiplication set free by the 
rupture of the cells; but even of such specimens it is difficult 
to get perfectly satisfactory preparation for microscopic study 
of detail. With the method of fixation by vapour of osmic 
acid and subsequent coloration with Giemsa’s stain or other 


THE RAT-TRYPANOSOME, TRYPANOSOMA LEWISI. 523 


modification of the Romanowsky method of staining we have 
obtained occasionally very clear preparations of the later 
stages of the multiplication, but as a rule, the “ spheres ” 
with many nuclei take up the stain with such intensity that 
they become opaque masses showing nothing of the internal 
structure, although in the same preparations the free trypano- 
somes may be stained to perfection. If in such preparations 
the stain be cautiously extracted by means of acetone or other 
suitable media, it is possible to obtain specimens showing the 
nuclei satisfactorily, but then, as a rule, the flagella are 
invisible, having lost the stain completely, while the free 
trypanosomes or early stages of multiplication on the same 
slide have become mere ghosts or have vanished altogether, 
beyond the power of visual resuscitation by the most delicate 
and refined methods of microscopic illumination. Very often 
in such preparations only the kinetonuclei can be seen, the tro- 
phonuclei having disappeared. Inthe study of Romanowsky- 
stained preparations it was generally found necessary to begin 
by drawing all that could be seen, general outline, project- 
ing flagella, in the opaque, untouched preparations of the 
spheres, and then to perform a number of successive opera- 
tions of cautious extraction of the stain, examining the 
preparations after each such operation and adding to the 
drawing any fresh details of structure brought to hight. It 
was difficult, however, to control the extraction of so sensitive 
a stain with sufficient exactness to avoid losing the whole of 
it in an instant. The last state of the preparation was 
generally one which left it useless for purposes of demonstra- 
tion; always a disappointment to the microscopist and his 
friends. We have never succeeded in re-staining satisfac- 
torily preparations in which the Romanowsky stain has been 
over-extracted. 

By far the best and most instructive preparations of the 
intracellular multiplication were obtained in the coverslip 
preparations fixed in Schaudinn’s or Maier’s fluid and subse- 
quently stained by the iron-hematoxylin method, as described 
above. Only in such preparations was it found possible to 


524 E. A. MINCHIN AND J. D. THOMSON. 


control the stain so that in the largest spheres both nuclei 
and flagella were visible; even then, however, the tropho- 
nuclei were sometimes faint and difficult to make out clearly 
when the flagella were still sharp and distinct. Of one film 
in which the smear was thickly crowded with spheres of 
various sizes, some free, others still in the tissue, a very 
satisfactory preparation was obtained by staining with Mann’s 
hematoxylin, carried out with the friendly help of the inventor 
of the stain himself. The result was a very good “ show ” 
preparation of the multinucleate spheres, sharp and clear, 
even in the thick parts of the smear, and especially suitable 
for moderate magnification; the flagella, however, could not 
be made out. 

While the ordinary smear-methods presented special diffi- 
culties, very convincing and beautiful preparations of the 
intra-cellular phase. were obtained ir sections of fleas’ 
stomachs extracted carefully from the body and preserved in 
various ways; a full account of the technique employed is 
given above. Such preparations have the immense advan- 
tage of exhibiting the exact relations of the trypanosomes to 
the cells ; it is possible to look through every section of each 
series, to note every trypanosome, free or intracellular, 
occurring in each stomach, and to observe what each parasite 
was doing at the moment the stomach was preserved. On the 
other hand, for the study of the stages of multiplication, 
sections have the disadvantage that the parasites themselves 
are often halved or mutilated, so that any given specimen 
may be only a part or fragment of the whole body. Both 
smears and sections are therefore indispensable, and supple- 
ment each other in obtaining a complete picture of the course 
of events. 

So much for methods and technique; we proceed now to 
give an account of our observations. 

From Noller’s investigations on the development of T. 
lewisi in Ctenocephalus canis, it appears that! the 
intracellular multiplication begins about six hours after the 
ingestion of the trypanosomes by the flea. In our prepara- 


THE RAT-TRYPANOSOME, TRYPANOSOMA LEWISI. 525 


tions of a batch of fleas, of which the infection could not have 
been more than nine and a half or less than seven and a half 
hours old, we have found the recurved forms fairly commonly 
and also some of the rolled-up forms characteristic of the 
early intracellular stages (Text-fig. 23, p. 635). We have 
no other records of intracellular multiplication in our fleas 
earlier than twelve hours. The stages of the intracellular 
multiplication are to be found in all parts of the epithelium of 
the stomach, from close behind the proventriculus to the 
pylorus. 

Our investigations upon the intracellular multiplication 
contain, unfortunately, one gap which we have been unable 
to fill; we have not succeeded in observing the actual pene- 
tration of the epithelial cell by the trypanosome. Noller, 
however, has been so fortunate as to observe the process, and 
gives the following account of it. In a dog-flea which had 
sucked infected blood five hours and fifty-five minutes 
previously, he saw “a trypanosome, of which the pointed 
hinder end had already penetrated into an epithelial cell. 
The flagellum-bearing anterior end beat violently and 
incessantly, whereby the trypanosome penetrated further 
and further into the cell. After I had watched this 
spectacle for about five minutes the trypanosome, which 
had so far penetrated into the cell as far as the middle of 
its body, shot suddenly into the cell and stirred up the 
granular cell-contents by its lively movements. Since, how- 
ever, the cell was torn on the opposite side, the trypano- 
some soon shot out of the cell again.” Néller thus confirms 
the suggestions we made in our preliminary report (1911) 
with regard to the probable method in which the penetration 
of the cell is affected. 

We have frequently seen trypanosomes, not distinguish- 
able in the living state from the ordinary type, singly within 
cells; the first time we ever discovered the trypanosomes 
within the cells was just such a case, a single trypanosome of 
quite ordinary appearance, wriggling and squirming actively 
in the cytoplasm of an epithelial cell, in a flea which had 


526 BE. A. MINCHIN AND J. D. THOMSON. 


been fed twelve hours previously on an infected rat. Careful 
examination of the cell, at different foci of the microscope, 
convinced us, greatly to our astonishment, that the parasite 
was really within the cell and not above or below it. ‘This, 
and other observations, repeated subsequently upon trypano- 
somes of ordinary appearance, contained singly within epithe- 
lial cells, suggest that the trypanosomes in each such case 
had but recently penetrated into the cell; but the observation 
might also be interpreted to mean that the trypanosome seen 
was the last of a batch produced by multiple fission within 
the cell from which its sister-trypanosomes had already 
escaped. In the latter case, however, the trypanosome 
would probably be within a vacnole, as will be described 
presently. ; 

Observation of the free trypanosomes in the living state 
shows that, as already stated, they are extremely active, but 
have a great tendency to attach themselves by the tip of the 
flagellum to firm objects; to the wall of the stomach, to 
pieces of débris, even to the glass surface of the slide or 
coverslip when under observation. The study of sections of 
the stomach confirms this observation in an unmistakable 
manner; many trypanosomes of the long, stiff type are seen 
in the sections attached to the epithelial cells by their 
flagella. The attachment is not, as a rule, to the outer pro- 
jecting ends of the cells, but to their sides; the trypanosomes 
put their long flagella down between the epithelial cells and 
often adhere to the cell close to its base; it would appear as 
if the side of the cell, at least in its columnar form, is its 
vulnerable region. A still more striking point is that many 
of these trypanosomes attached to, but still quite outside the 
cell, have already assumed the recurved form. These obser- 
vations make it very probable that in some cases the trypano- 
some may first attach itself to the cell by its flagellum and 
then bore its way into the cytoplasm in some way. 

Several trypanosomes may penetrate independently into 
one and the same cell. We have frequently observed 
numbers of the parasites, from five or six up to a dozen or 


THE RAT-TRYPANOSOME, TRYPANOSOMA LEWISI. 527 


more, in different stages of multiplication, side by side in a 
cell both in the living condition and in sections (Pl. 38, 
figs. 112, 118; Pl. 39, fig. 130). The parasites lie in the 
cytoplasm usually in a distinct vacuole, produced, apparently, 
by the liquefactive action of the parasite on the cytoplasm of 
the host-cell. The trypanosomes are nearly always in a state 
of movement, a point to which we shall return again. It is 
not infrequent to observe a number of trypanosomes in the 
same vacuole, wriggling actively one over the other. 

The infected cell may become reduced simply to a bag 
containing fluid in which large numbers of trypanosomes, 
generally with their multiplication completed or far advanced, 
move actively. Such cells are found commonly in sections 
(Text-tig. 3); they are generally thrown off from the epi- 
thelium and lie quite free in the blood-débris, sometimes 
even in the centre of the lumen of the stomach; nothing 
remains of the cell-contents except a thin superficial layer 
of cytoplasm, under the cell-membrane, and the nucleus, 
adherent usually to the wall at some point. This condition 
obviously represents the last stages of the exhaustion and 
death of the cell, from which the trypanosomes will escape 
either by their own activity or by disintegration of the cell. 
We shall consider the effects produced by the parasites on 
the cells in more detail subsequently ; at present it will be 
more convenient to confine our attention to the development 
of the trypanosomes themselves. 

The study of the trypanosomes in the living cells, checked 
by the examination of preserved material, permits readily 
enough of the recognition of a number of well-marked stages 
in the process of multiplication : 

(1) Trypanosomes of quite ordinary appearance, which 
have apparently but recently penetrated into the cell, as 
already described. 

(2) Pear-shaped forms, with the flagellum continuing the 
stalk of the pear; the body of the pear is distinctly flattened, 
and therefore presents a contour which differs according as 
it is seen from the edge or from the flattened surface ; as the 

voL. 60, PART 4,—NEW SERIES. 37 


528 E. A. MINCHIN AND J. D. THOMSON. 


parasite is in constant motion within the cell it presents con- 
tinually different views to the observer. The body of the 
parasite also shows in life incessant “metabolic” changes of 
form, movements of an active protoplasmic body imperfectly 
restrained by the thin, vielding envelope or periplast. It is 
very easy to seein the living condition that these pear-shaped 
forms are produced by the body of the trypanosome being 
doubled upon itself, an interpretation confirmed by the 
examination of preserved specimens. When the stomach is 
teased up and examined fresh, many of the recurved pear- 
shaped forms are found swimming freely, with the flagellum 
forward, in the salt-citrate solution used in the dissection. 
If such a form be watched attentively, it is often seen to 
uncurl itself, straightening out the body and thus passing 
from the pear-shaped form to that of an ordinary trypano- 
some. In some cases a trypanosome which had been seen to 
unbend itself in this manner can be observed to curl up 
again, while swimming freely, and thus to assume or to lose 
the pear-shaped form several times in succession. 

In the fixed and stained preparations it is seen that the | 
trypanosome is bent upon itself in such a way that the 
posterior part of the body, containing the kinetonucleus, is 
closely applied to the anterior half of the body. ‘Thus a pear- | 
shaped body results in which N is lodged in the thickest part 
of the pear, at (Pl. 36, fig. 15) or near (fig. 13) the blunt end 
of the body; while is usually well in front of N, that is to 
say, nearer to the pointed end of the body (PI. 37, fig. 70). 
In some cases, however, 1 is close beside N (figs. 13-17), or 
even, exceptionally, behind WN (fig. 16). The variations in 
the positions of 7 and N are easily explained by their 
variability in this respect, already described, in the free 
trypanosomes, on the one hand, and on the other by varia- 
tions in the exact region of the body at which the bending 
takes place. As a general rule, the body appears to be bent 
between 7 and N, so that N lies a little way from the extreme 
posterior end and m in front of it (figs. 13, 17); but the point 
of greatest curvature may be in the region of N, which is 


THE RAT-TRYPANOSOME, I'RYPANOSOMA ILEWISI. 5929 


then at the hindermost extremity of the body (fig. 15), or 
even in the region of n, which is then at the extreme blunt 
end (fig. 16). In exceptional cases the bending takes place 
in front of N. In all cases the flagellum runs backwards 
along one side of the body, round the blunt posterior end, and 
forward, for a variable distance, to the basal granule or true 
blepharoplast situated close beside ». Thus the course of 
the flagellum, as a whole, may be compared to the letter U 
modified by making one arm of the letter much longer than 
the other. 

In some cases the distinction between the two limbs of the 
recurved body can be seen plainly in the fixed specimens 
(Pl. 36, fig. 15; Pl. 37, fig. 70), but in other cases no line of 
demarcation can be made out, and the applied portions of the 
body appear to have formed completely into a compact pear- 
shaped mass, leading on to the stage next to be described. 

The recurved forms differ remarkably in size, and from 
a comparison of these forms with one another, with the free 
trypanosomes, and with later stages of the intracellular 
development, there can be no doubt that the initial stages of 
the lite of the trypanosomes within the cells is accompanied 
by a pronounced diminution in the size of the flagellates (see 
especially Pl. 36, figs. 13-17, and compare them with figs. 1-10 
of free trypanosomes, and figs. 21-34 of later stages, on the 
same plate). How this shrinkage takes place it is difficult to 
say; probably the cytoplasm of the flagellate gives up a 
large amount of watery fluid and so diminishes in bulk, 
while becoming at the same time correspondingly denser in 
texture, a change which would account for the intensity with 
which the intracellular forms take up the stain and the 
consequent opacity which they acquire, as already noted. It 
is necessary, however, to exercise caution in estimating the 
size of the forms in preparations, since there is no doubt that 
they vary owing to differences in fixation. Thus, Pl. 36, 
fig. 14 shows a specimen from the same slide as fig. 13, but 
the former is from a part of the film which appeared to have 
dried before it was-exposed to the action of the osmic vapour ; 


530 E. A. MINCHIN AND J. D. THOMSON. 


its large size, light colour and the elongation of N are all 
indications that the soft body had become flattened out by 
being dried before fixation. ‘The specimens may also become 
deformed in other ways; Pl. 37, fig. 72 is probably to be 
explained as representing a recurved form in which the 
flagellum has become torn away from the side of the body 
and so projects freely from the rounded posterior end. 

We were at first under the impression that the pear-shaped, 
recurved trypanosomes found free in teased-up stomachs 
examined fresh were forms that had been originally intra- 
cellular and had been set free by rupture of their host-cells. 
As stated above, however, examination of sections proved 
that these recurved forms may be extracellular in occurrence, 
attached to the epithelial cells or even free from them. It is 
evident, therefore, that the recurved form is not simply an 
adaptation to life within the confined limits of the cell. 

(3) Forms with rounded or oval body, derived from the 
pear-shaped recurved forms by a further contraction and roll- 
ing up of the body ; these are the forms which we described 
in our preliminary account as ‘ block-like” since the body 
often shows during life irregular contours, changing con- 
tinually owing to the active metabolic movement. ‘The 
flagellum, which runs in a U-shaped course in the recurved 
forms, acquires now an additional bend (compare especially 
Text-fig. 23, p. 655, hand 7; it usually rans round the outercon- 
tour of the rounded body and protrudes from it to a variable 
extent. In some cases a very considerable length of the 
flagellum is free (Pl. 37, figs. 77, 78), in other cases a very little 
(Pl. 36, fig. 28; Pl. 37, fig. 82), while in other cases again the 
flagellum is simply wrapped closely round the body (Pl. 36, 
figs. 29, 30, 35; Pl. 37, figs. 81, 84). The extreme length of 
the free flagellum seen in Pl. 37, fig. 76 is possibly due in 
part toits having become artificially detached from the body 
in the process of making the smear. 

The data in the foregoing paragraph have been obtained 
chiefly from the study of preserved specimens. In the living 
condition this stage appears as a small, rounded or oval body 


HE RAT-TRYPANOSOME, TRYPANOSOMA LEWISI, 531 


within the cell, usually in motion and showing a distinct 
flagellum. Sometimes, however, these bodies are quite 
motionless with no flagellum visible. In our preliminary 
communication (1911), we were unable to decide whether 
a flagellum was always present, and were prepared to admit 
that in some cases this stage might be a non-flagellated 
leishmania-like form. We had observed in one case that a 
body which had been for some time quiet and motionless 
within the host-cell became suddenly active, showing a dis- 
tinct flagellum. In all our permanent preparations, however, 
whenever flagella can be made out in the other stages or in 
the free trypanosomes, they can be seen to be invariably 
present at this stage also, and there can be no doubt that the 
motionless forms are those in which the flagellum is wrapped 
round the body. We are now convinced that non-flagellated 
leishmanial forms do not occur. We have the impression 
that the rolled-up trypanosome can wrap its flagellum round 
the body and pass into a resting, quiescent condition for 
a time, after which it can become active again by un- 
curling and setting free its flagellum, or at least a certain 
length of it, probably never quite the whole length. 

The body in this stage, as in the last, is actively metabolic, 
with constantly changing contours in life. When the rounded 
forms are set free by rupture of the host-cell they swim 
actively in the liquid, progressing with the flagellum directed 
forwards; they then resemble ordinary flagellate monads, 
and the observer might easily have the impression that he 
was watching some intruding flagellate derived from contami- 
nation of the salt-solution or from some extraneous source. 

In a few rare instances the rolled-up forms have been 
found in preparations to exhibit a central perforation or 
fenestration (Pl. 36, figs. 24, 30, 34), evidently produced by 
the trypanosome curling itself round so as to leave a central 
space. ‘his condition, when it occurs, is probably quite 
transitory, the plastic cell-body of the trypanosome fusing 
into a compact lump sooner or later. 

It seems probable that some of the rolled-up forms degene- 


532 E. A. MINCHIN AND J. D. THOMSON. 


rate at this stage and are absorbed; that is, at least, the 
only explanation we are able to offer for such minute forms 
as those shown in Pl. 36, figs. 43-45, which appear to be 
undergoing degeneration. 

The smallest of the rounded forms have each a single kine- 
tonucleus, trophonucleus and flagellum. Now they begin to 
grow in size, with concomitant multiplication of their nuclei 
and formation of daughter-flagella. The division of these 
various parts appears to go on much as in other trypano- 
somes, independently, but more or less synchronously. The 
division of » may be slightly in advance of that of N, or 
slightly after it: thus, stages are found in which m and N 
appear to be both in the same stage of division (PI. 37, fig. 79) ; 
or in which N appears to be in advance of m (PI. 36, figs. 28 
and 37) ; or with two distinct nv and N still in division (PI. 37, 
figs. 80, 82) ; or finally with n and N both completely divided 
(Pl. 36, figs. 33, 36). The division of 1 is dependent on, or 
connected with, that of the basal granule or true blepharo- 
plast, which may be regarded as representing the centriole 
or division-centre for x. The original flagellum does not 
divide, however, but remains attached to one of the daughter- 
blepharoplasts, from which it arises in close proximity to one 
of the daughter nn; and from the other blepharoplast a new 
flagellum grows out, at first a very fine and delicate structure 
and consequently very difficult to make out clearly or with 
certainty in the opaque body. In many cases in which 7 is 
divided completely no second flagellum can be seen, but it 
would not be safe, in view of the difficulties of technique 
presented by these objects, to conclude in all such cases that 
the formation of the new flagellum had not begun, since in 
other cases a very delicate line can be seen plainly growing 
out from the daughter-blepharoplast—that is to say, from 
the blepharoplast other than that from which the original 
flagellum arises (P]. 37, figs. 79-81). As a rule the daughter- 
flagella can be made out in the preparations stained with 
iron-hematoxylin, but not in those stained by Giemsa’s 
method ; in the latter case, as already mentioned, the body 


THE RAT-TRYPANOSOME, TRYPANOSOMA LEWISI. 533 
usually stains so intensely as to obscure completely the 
delicate daughter-flagella, which are imbedded in the mass 
of the cytoplasm, while the original flagellum runs for the 
most part on the exterior of the body; and if the stain be 
extracted sufficiently to make the body clear, it comes out of 
the growing flagella and leaves them invisible. 

We may infer, therefore, that as the division of the nuclei 
proceeds the formation and growth of new flagella follow 
hard upon the division of the blepharoplast and kinetonucleus, 
and that a new flagellum grows out from each blepharoplast, 
in close proximity to 7, quite independently of the original or 
parent flagellum, which remains unaltered. 

As a general rule the multiplication of the nuclei begins at 
the rolled-up stage, but this rule is by no means invariable. 
Sometimes the nuclei are found to have multiplied even 
before the trypanosome has taken on the recurved form 
(Pl. 36, figs. 18-20). Such forms might possibly be specimens 
which, after becoming recurved, have straightened themselves 
out again, but their appearance is that of trypanosomes in 
which nuclear multiplication has begun before change of 
body-form. Attention must also be drawn to the peculiar 
elongated forms with 2 mn and 2 NN, such as Pl. 36, figs. 
21-23; Pl. 37, fig. 74. At one time we were inclined to sus- 
pect that these forms, and also the unaltered trypanosomes 
with 2 nn and 2 NN, might be examples of fusion instead of 
multiplication (see below, p. 604) ; but we could find no definite 
evidence of there being fusions of two trypanosomes, and the 
fact that forms occur with 3 nn and 3 NN (fig. 18, pl. 36) 
makes it more probable that they are early stages of multi- 
plication. On the other hand the possibility cannot be 
excluded that in some cases accidental and purely plasto- 
gamic (non-sexual) fusion of intracellular stages may occur, 
and that such fusions may explain the enormous size of some 
of the spheres and later stages of multiplication (Pl. 36, 
fig. 42). 

The multiplication of the nuclei proceeds apace, and the 
duplication of 7 and N is followed by a stage in which the 


534 EK. A. MINCHIN AND J. D. THOMSON. 


body contains 3 nn and 3 NN (PI. 36, fig.40!; Pl. 37, fig. 84). 
This stage, which is of common occurrence, indicates that 
after the first division of 2 and N one of the daughter-nuclei 
in each case remains undivided, while the other divides again. 
This interpretation is supported by fig. 39, showing a specimen 
containing 3 nn and 2 NN, one of the latter being in process 
of division. <A later stage is seen in fig. 38,in which both 
mand N have multiplied to four in number. The further 
stages of multiplication are not easy to follow in detail 
owing to the difficulty of making clear in one and the same 
specimen the nn, NN, and flagella, but from a study of various 
preparations there appears to be no reason to doubt that the 
an and NN maintain their parallelism in division, and con- 
sequent equality of numbers, and that as the nn, or, to be 
more accurate, their blepharoplasts, divide, they continue to 
give rise to new flagella. Since the blepharoplasts are of 
different ages, as the result of successive divisions, the 
daughter-flagella given off from them are of different lengths ; 
the original or parent flagellum remains, however, distinct 
and recognisable, both by its length, its superficial position, 
and the sharpness with which it stains (PI. 37, fig. 92, etc.). 
In some cases the daughter-flagella also project freely from 
the surface of the body. This fact was observed in a living 
specimen, in which two small flagella were seen in addition to 
the principal flagellum, but it appeared to be a temporary 
condition, since later on only the principal flagellum could be 
seen, and still later that also disappeared. In one of our 
preparations also three daughter-flagella are seen in addition 
to the main flagellum, but unfortunately the specimen was so 
opaque that no details of internal structure, except the nuclei, 
could be made out. 

With the growth in size and the multiplication of the 
nuclei the ‘‘ block-like ” stages pass by insensible gradations 


1 The specimen shown in fig. 40 appears to have become dried and 
flattened out before fixation; its size is altogether abnormal for this 
stage; compare fig. 38. Fig. 26, Pl. 36, also abnormally large, is from 
the same slide as fig. 40. 


THE RAT-TRYPANOSOME, TRYPANOSOMA LEWISI. 535 


into: (4) The “spheres,” a term used by us to denote the 
final stages of the intracellular multiplication. They appear 
in the fresh preparation as relatively large masses, more or 
less spherical in form, within the cytoplasm of the epithelial 
cells; they can be distinguished generally as “tailed” and 
“tail-less.” At the very last the tail disappears in all cases, 
but this appendage, though commonly present in stages not 
full grown, sometimes cannot be made out in specimens that 
do not appear to be mature. A comparison of fixed prepara- 
tions and a consideration of facts already discussed in 
describing the “ block-like” stage leaves no doubt but that 
the tail-less spheres are derived from those earlier stages 
which have the original flagellum wrapped round the body, 
the tailed spheres from those which have the flagellum free. 
It is, however, evident even in the living state that the tail 
represents more than the original flagellum, since it is often 
of considerable thickness at the base and tapers gradually 
to a point. Hxamination of stained preparations shows that 
the tail represents in early stages the anterior end of the 
body, with the flagellum, of the original trypanosome (Pl. 36, 
figs. 37, 39; Pl. 38, fig. 110); and that in later stages the 
daughter-flagella may grow out parallel to the original 
flagellum, and so contribute to the formation of the tail 
(Pl. 37, fig. 86; Pl. 38, figs. 108, 111). But in other cases the 
anterior end of the original trypanosome may be retracted 
into the main mass of the sphere, round which the flagellum 
is wrapped more or less completely ; the sphere then either 
has no tail (Pl. 37, figs. 90-94) or, if a tail is present, it 
represents the flagellum alone (figs. 95, 110). 

The appearance and behaviour of the spheres in the living 
condition have been described by us in our preliminary com- 
munication. They are in a state of incessant movement, 
due both to the activity of the flagellum or tail, when 
present, and to internal commotions. The movements of 
the flagellum cause them to rotate within the cell in a 
jerky, irregular manner. When there are several spheres 
within one cell they can be seen to push and bump against 


536 E. A. MINCHIN AND J. D. THOMSON. 


the cell-nucleus or against one another, the impact of one 
sphere causing a movement in another, which shows that they 
come into actual contact. Internal causes of movement are 
the metabolic form-change, already described, in the earlier 
stages, and movements due to the independent motility of 
the daughter-trypanosomes in the ripe spheres. 

Even in the living condition it is not difficult to make out 
that a process of multiplication is going on within the sphere. 
First the nuclei and flagella can be seen to have multiphed, 
though the details of the process are naturally not clear in 
the living specimen. Later it is seen that the whole contents 
of the sphere have divided up into a number of distinct 
daughter-try panosomes, which are seen writhing and twisting 
over each other within the enveloping periplast like a bunch 
oft eels in a sack. The independent movements of the con- 
tained trypanosomes make their separate individuality clearer 
in the living state than in the preserved specimen. After 
the daughter-trypanosomes have become fully distinct and 
separate from one another a number of changes are seen in 
the sphere. The tail, if it was present, disappears, and the 
body acquires a perfectly spherical contour; this can only 
mean that the original flagellum, distinct from the beginning, 
has been drawn into the sphere and taken over by one of the 
sister-try panosomes resulting from the process of multiplica- 
tion. The metabolic form-changes, so marked a feature of 
the earlier stages, now cease, and the sphere shows only slight 
oscillating or trembling movements, due to the activity of the 
contained trypanosomes, moving restlessly within the com- 
paratively tense envelope of the sphere, representing the 
periplast of the original trypanosome. 

As the final moment approaches, the envelope of the sphere 
becomes more tense and rigid, its outline in optical section 
ever more nearly a perfect circle, until it bursts suddenly, 
setting free the contained trypanosomes. If the sphere 
bursts in its normal situation, that is to say within the 
cytoplasm of the host-cell, the lberated trypanosomes are 
seen moving actively for some time in the cell, from which 


THE RAT-TRYVANOSOME, TRYPANOSOMA LEWISI. 537 


they escape one by one, passing out into the lumen of the 
stomach. But in the examination of stomachs freshly teased 
up, it is common to find spheres set free by the rupture of 
their host-cells. The spheres can then be seen to burst with 
such suddenness that they appear to explode ; the impression 
given by the eye disappoints the ear, which misses. the 
almost-expected report of the explosion; the trypanosomes 
set free scatter in all directions. 

In one case a single cell was observed to contain three 
spheres, which were seen to burst successively; first one, then 
another, and then the third became resolved into free try- 
panosomes until the cell was full of active trypanosomes 
which appeared to be wriggling in a clear fluid. The cell 
retained at first its contours, but was apparently reduced to 
a spherical envelope, with no nucleus visible. Soon after the 
last sphere had become resolved the trypanosomes were seen 
to be escaping from a small area of the cell-wall, which then 
collapsed slightly; all the trypanosomes eventually found 
their way out at this part of the cell and swam off. In 
another case we timed our observations of a sphere and 
recorded the events in chronological order. It is not 
necessary to repeat the data already published, but they may 
be summarized as follows: A full-grown sphere was seen 
within a cell, moving and rotating by means of its flagellum, 
which was distinctly visible, and also exhibiting active 
metabolic form-changes; twenty-three minutes after it was 
first seen both the rotatory and the metabolic movements 
became much slower, and one minute later they ceased 
altogether, the alterations in the contour being slight, and 
due apparently to the very active movements of the contained 
trypanosomes; the flagellum had disappeared completely from 
view ; the contour of the sphere then became very tense and 
rigid, and it burst after having been watched for twenty- 
eight and half minutes, or about five minutes after the 
disappearance from view of the flagellum and the cessation 
of the metabolic movements. 

The trypanosomes set free are the long free stomach-type 


538 E. A. MINCHIN AND J. D. THOMSON. 


already described, characterised by their great length, their 
stiff, crithidiomorphic form, and the more or less pronounced 
tendency to approximation of their two nuclei. The number 
produced in a sphere seems to vary considerably in different 
cases. We have attempted to count those which we have 
actually seen liberated within cells by bursting of spheres, 
and came to the conclusion that about eight was the usual 
number of trypanosomes produced. It is, however, very diffi- 
cult to count accurately a number of trypanosomes which are 
at different levels in the cell, and, therefore, not all in focus 
at the same moment, and which at the same time are moving 
actively and changing their position. Moreover, the full- 
grown, tense spheres, which can be observed to burst, differ 
so greatly in size that there can be no doubt that the con- 
tained trypanosomes must vary considerably in number, and 
that if the average capacity of a sphere be eight trypano- 
somes, some spheres must liberate not more than four, others, 
perhaps twelve or sixteen, when they burst. These conclu- 
sions are confirmed by a study of fixed preparations; in the 
large spheres we find 8 (Pl. 37, fig. 86), 10 (Pl. 37, 
fie. 96), 13 (Pl. 36, fig. 41), and 14 or 15 (Pl. 37, fig. 94, 
and Pl. 36, fig. 42) nn and NN. From the results of 
counting the nuclei in a number of preparations, the 
most frequent number of nn in the largest spheres would 
appear to be 10nn; the number of NN cannot always be 
made out, but where they can be seen their numbers are 
equal to, or slightly less than, those of the nn, indicating that 
the nn divide a little in advance of the NN. On the other 
hand, some spheres with but few nuclei appear to be nearly 
mature, having daughter-flagella so long that the nuclear 
division must have become much retarded, or perhaps ended 
altogether (PI. 36, fig. 40, and PI. 37, fig. 95). The perma- 
nent preparations confirm, therefore, the conclusion that 
while the number of trypanosomes produced is commonly 
eight or ten, it may be less, or as many as fourteen or more. 
The intracellular growth and multiplication of the trypano- 
some must depend on an interaction between the host-cell and. 


THE RAT-TRYPANOSOME, TRYPANOSOMA LEWISI. 9539 


parasite, each a variable factor, and it is, therefore, not sur- 
prising if the result varies in different cases. Moreover, 
when a cell contains several parasites in different stages of 
multiplication, the cell may become exhausted before the 
younger individuals have time to grow to their full size. 

The length of time required for a complete generation of 
intracellular multiplication is a point we have been unable 
to determine, as it is impossible to keep the living object 
alive under observation long enough. Moreover, the living 
trypanosomes in teased-up stomachs soon become abnormal, 
however carefully sealed up between slide and cover-slip, so 
that no data derived from direct observation of this kind 
could be entirely reliable. Until physicists and opticians 
have invented some method whereby the epithelial cells of the 
flea’s stomach can be focussed under high magnification 
without dissecting or injuring the flea, it will not be possible 
to give a direct answer to such questions. Our observations 
cited above show that barely five minutes are required for a 
sphere to burst after the multiplication is complete, but that 
gives no clue to the length of time required for the growth 
and multiplication of the original trypanosome after it has 
penetrated the cell. From the fact, already noted, that the 
earliest intracellular stages were found by us about eight 
hours, and the earliest trypanosomes of the long stomach-type 
about twelve hours, after the flea had fed on an infected rat, 
it might be inferred that an intracellular generation required 
about four hours, but this computation cannot claim to possess 
any exactness. 

It is also not possible to state how many successive genera- 
tions of multiplication take place in the flea’s stomach. 
Noller states that at least two such generations, probably 
more, succeed each other. From the fact (to be described in 
more detail below) that the stomach-phase may be ended in 
some cases in eighteen hours, and in other cases may persist 
for four or even five days in unfed fleas (see above), it is 
evident that the number of generations of intracellular multi- 
plication in the stomach must be a highly variable quantity. 


54.0 E. A. MINCHIN AND J. D. THOMSON. 


The foregoing account of the stomach-phase has been based 
almost entirely upon the study of the trypanosomes in the 
living condition and in smears. It was, in fact, all written, 
except for a few subsequent additions, before we had studied 
any sections. It remains now to supplement the description 
given above with the additional data furnished by the study 
of sections of the stomach, whereby a much more exact idea 
can be obtained of the relations of the trypanosome to the 
epithelium of the stomach. 

When sections through a _ well-infected stomach are 
examined the trypanosomes are found usually to be both 
intracellular and extracellular in occurrence; that is to say, 
both free in the lumen of the stomach and contained within 
the cells of the epithelium. The extracellular trypanosomes 
occupy two different positions, speaking generally : they may 
be scattered throughout the blood-débris, including its 
most axial region, or they may be found only along the 
periphery of the sections, close to the epithelium and very 
often attached to it by the tip of the flagellum. 

There can be no doubt that the trypanosomes found 
scattered through the central part of the blood-débris are 
for the most part forms in process of migration towards the 
rectum, their stomach-phase having been completed, but not 
all of them are to be interpreted in this way, since in many 
cases they are seen to be in clumps connected together by 
the tips of their flagella, and are obviously degenerative 
forms. In many cases also an infected epithelial cell is 
thrown off into the lumen, and may be seen there, crammed 
with trypanosomes, the product of the intracellular multipli- 
cation, which break out of the exhausted cell, and may 
possibly find their way into other epithelial cells again. 
The cell itself may become reduced to a mere sac containing 
fluid, in which a huge number of trypanosomes are squirm- 
ing over one another (Text-fig. 3). I any of the proto- 
plasm of the cell still survives it may contain a few spheres 
and other multiplication-stages. But that the forms free in 
the débris are for the most part migratory forms is indicated 


THE RAT-TRYPANOSOME, TRYPANOSOMA LEWISI. 541 


by the fact that they occur with greater frequency towards 
the pyloric end of the stomach and sometimes only in this 
region, which may swarm with them, and in such cases they 
are always to be found in the intestine also. 


TEXT-FIG. 3. 


Exhausted cells containing intracellular stages of the trypano- 
some, from sections through the stomach of a flea thirty-six 

p : 4.000 ’ 
hours after the infective feed, x =a. : (a) A cellin process of 


being pushed off from the epithelium; the indented outline of 
the lower edge shows where it was in contact with epithelial 
cells in situ. Thecellis reduced for the most part to a sac of 
fluid, in which trypanosomes are swimming, but round the 
edge some of the cytoplasm still remains, and in it are seen a 
few multiplication-stages (sph.), and also a few coarse black 
grains of fatty nature. The nucleus of the cell comes in another 
section. (b) A cell completely detached from the epithelium, 
the cytoplasm reduced to a few irregular strands and con- 
taining: (1) the nucleus at the centre ; (2) a sphere to the right 
of it; (8) some coarse fat-grains, some of which have been 
extruded into the blood-débris. 


542 E. A. MINCHIN AND J. D. THOMSON. 


On the other hand the trypanosomes that occur in close 
proximity to the epithelium are for the most part forms about 
to attack it, as shown by the fact that many of them are 
actually attached to the cells, and also by the frequency, it 
might almost be said the predominance, of the recurved 
form in this situation ; at any rate, most, though by no means 
all, of those that are actually attached to the cells are recurved 
(Pl. 38, figs. 104, 105; Pl. 39, figs. 120-124). A certain 
number of recurved forms are also found, however, quite 
free from the epithelium, though close to it (Pl. 59, figs. 118, 
119). The question at once suggests itself whether that 
represents the state of things asit was in life. Itis impossible 
to extract a stomach from a flea without doing a certain 
amount of violence to it; and it is quite conceivable that the 
stresses and strains to which the wall of the stomach is sub- 
jected in pulling it out of the flea’s carcase might well cause 
trypanosomes, previously attached, to become detached from 
the cells. We can but take the evidence as it is presented 
to us in our preparations, however, where we find that some 
of the attached trypanosomes are recurved and some not, 
and that some of the recurved forms are quite free from, 
though always close to, the epithelium. It would follow 
from these observations that in some cases the trypanosome 
assumes the recurved form after, in other cases before, 
attaching itself to the epithelial cell. 

The attachment of the trypanosomes to the cells is peculiar ; 
they are seldom attached to the extreme apex of the epithelial 
cell, but usually to its side (Pl. 39, figs. 121-124). The long 
flagellum runs down in the interspaces between two adjacent 
cells, and its extremity is often seen to be adherent to one of 
the cells close to the isthmus connecting it to its neighbour ; 
the body of the trypanosomes usually projects into the lumen 
of the stomach above the general level of the epithelium, but 
sometimes the whole body is deep in between the cells. 
Sometimes two or three trypanosomes are attached side by 
side in such a situation. These statements apply to the 
usual columnar form of the epithelium with deep spaces inter- 


THE RAT-TRYPANOSOME, TRYPANOSOMA LEWISI. 543 


vening between the cells; when the cell is flattened the try- 
panosomes may be attached at any point, and in the case of 
the young cells, which have but recently been produced from 
the epithelial buds and in which each cell is in contact with 
its neighbours along its side, without interspaces, the trypano- 
somes are attached to the convex apical region of the cell 
(Pl. 38, fig. 104), which may even contain early intracellular 
stages (PI. 39, fig. 151). 

As stated above, we have never been able to observe in the 
living state the actual penetration of the epithelial cell by the 
trypanosome; and in our sections we have searched in vain 
for appearances which could be interpreted as trypanosomes 
caught in the act of penetration. We have not found any 
trypanosomes half in and half out of the cell, from which it 
may be inferred, probably, that the actual penetration is 
effected very rapidly. But attention must be drawn here to 
a condition which is probably of significance in connection 
with this point. Trypanosomes of recurved type are found 
with the tip of the flagellum attached close to the base of the 
cell and with the rest of the body so closely applied to the 
side of the cell that careful focussing is necessary to deter- 
mine that the trypanosome is still bodily outside the border 
of the cell and has not invaded it. Suchappearances (PI. 39, 
fies. 122, 123), suggests powerfully that the parasite, after 
attaching itself by the flagellum, forces its body inte the cell. 
But the imagination must not be given too loose a rein when 
interpreting in terms of living activity the appearances seen 
in the dead fixed preparations. 

The next stage is a recurved trypanosome within the cell 
(P]. 38, fig. 107). Trypanosomes of the ordinary form (i. e. 
not recurved) are often seen in cells, but it is doubtful 
whether these represent individuals that have but recently 
penetrated the cell; if so, they might be forms which, after 
having been recurved, have temporarily straightened them- 
selves out again, but it is on the whole more probable that 
they are trypanosomes which have been produced by the dis- 
ruption of a sphere and which are about to leave the cell. In 

vou. 60, PART 4.—NEW SERIES. 38 


544 E. A. MINCHIN AND J. D. THOMSON. 


support of the latter conclusion is the fact that we have seen 
trypanosomes of ordinary type only in cells that appeared 
more or less exhausted and liquefied, and especially often in 
detached cells. But, even if such forms represent recent 
arrivals in the cell, it is certain that they soon take on the 
recurved form. Consequently the recurved form may be 
taken in any case as the starting point of the intracellular 
development. 

The growth of the recurved form into the sphere, the 
multiplication of its nuclei and flagella, and its ultimate 
fission to produce a bunch of trypanosomes, have been 
described in detail above, and only afew points in the 
development of the intracellular phase need be noted here. 
In the first place, in all preparations satisfactorily stained, it 
is seen clearly that flagella are present throughout, and that 
non-flagellated or leishmanial forms do not occur. ‘This is so 
in all preparations stained with iron-hematoxylin and not 
over-extracted. In preparations stained by Giemsa’s method 
the flagella are sometimes not to be seen, but in successful 
preparations they are exceedingly clear and sharp; it is 
always, however, one thing or the other, that is to say, we 
never find in one and the same preparation some parasites 
showing flagella and others none; either the flagella are 
invisible throughout the preparation, or are to be seen in 
every trypanosome, whether inside or outside the cells. 

In sections the parasites are naturally very often in a 
mutilated condition, especially the large spheres, which may 
go through two or more sections. Consequently, it is not 
possible to be quite certain always whether a given sphere is 
tailed or not. But if the tail happens to he in the plane of 
the section it may be shown very clearly, and then it can be 
seen that in some cases the tail is simply the original flagellum 
of the parent-trypanosome, projecting freely, while in other 
cases the tail includes the original flagellum with a_pro- 
longation of the trypanosome-body into which daughter- 
flagella may grow out, alongside of the parent-flagellum 
(PI. 38, figs. 108, 110). In some cases, especially when the 


THE RAT-TRYPANOSOME, TRYPANOSOMA LEWISI. 540 


host-cell is degenerating and breaking up, the flagellum of a 
sphere may project out from the cell (Pl. 39, fig. 152). 

In concluding the description of the intracellular multi- 
plication, we note that Chatton and Leger (1912), have 
described what they term a process of “agglutination and 
cytolysis” in Leptomonas drosophile, and suggest that 
the intracellular multiplication seen by us in the case of 
Trypanosoma lewisi in the stomach of the flea may be a 
phenomenon of the same order as that observed by them 
in Leptomonas drosophile. We cannot accept for a 
moment the suggestion that the behaviour of IT’. lewisi in 
the flea’s stomach is anything but a quite normal process of 
multiplication. Itis not incumbent upon us to criticise the 
statements of the authors, otherwise we might be tempted to 
inquire why they should think it necessary to interpret what 
they have seen in their Leptomonas as agglutination and 
cytolysis rather than as multiplication; but however legiti- 
mate such an interpretation may be for L. drosophile, we 
cannot admit it for the similar phenomena observed by us in 
T. lewisi. 


APPENDICES 10 THE DESCRIPTION OF THE STOMACH-PHASE. 


Having given in the foregoing pages an account of the intracellular 
multiplication of the trypanosomes in the stomach, it remains to 
discuss the relations of parasite to host and vice versa. Under this 
theme are comprised (1) the frequency of occurrence of the intracellular 
multiplication in fleas exposed to infection, (2) the type of epithelial cell 
attacked by the parasite, (3) the relation of the intracellular stages to 
the infected cell, (4) the effects of the parasites on the cells, and (5) 
the relation of the infection as a whole to the stomach. The first of 
these? problems is a statistical one, and is best studied in fresh or 
preserved film-preparations; the remaining four questions are best 
studied in serial questions. 


(1) The Occurrence of the Intracellular Multiplication. 


The following is asummary of our observations on the occurrence of 
the intracellular multiplication from twelve hours onwards, not count- 
ing,those fleas that were used for sections : 

At 12 hours approximately, after the fleas had fed on an infected 


546 E. A. MINCHIN AND J. D. THOMSON. 


rat, intracellular multiplication was found in one flea out of ten 
examined : 

At 15 hours, in 1 flea outof 15. 

At 18 a 2 fleas ,, 15. 

At 20 . lflea ,, 2. 

At 24, ~: 40 fleas ,, 240. 

At 36 5 OTE 5. 

At 48 55 Dk Klee 20. 

At 60 rs none hs 2: 

At three, four and five days after infection, we only found trypano- 
somes in the stomachs of fleas that had been kept isolated and starved ; 
in those that had been allowed to feed again on a clean rat no trypano- 
somes of any kind, free or intracellular, were found in the stomachs. 
Counting, therefore, only starved fleas, we found intracellular stages : 

At 3 days, in 5 fleas out of 16. 
IAG TSS os, 1 flea A 
At 4 _ 2 eas, 1 

Later than five days we have never found them.! 

The total of allthese observations is that we have found intracellular 
stages of multiplication in the stomachs of sixty out of 365 fleas 
examined from twelve hours to four days after infection, or approxi- 
mately 16°5 per cent. There is no doubt, however, that this figure falls 
below the actual percentage of fleas in which the trypanosomes succeed in 
establishing themselves inthe stomach and passing through their intra- 
cellular phase when ingested by the flea, for at least two reasons. In 
the first place, many of the 240 fleas dissected at twenty-four hours 
after the infective feed were examined very hurriedly and cursorily, the 
object being to find fleas in which the intracellular stages were swarm- 
ing, for the purpose of making permanent preparations of these stages, 
and when they were not found quickly in the teased-up stomachs 
examined in the fresh state, the search was abandoned ; consequently in 
many cases in which the intracellular stages were scanty they may well 
have been overlooked, and would probably have been found after more 
prolonged search.? In the second place, as regards the fleas examined 


oon 


1 Since this page was written we have found intracellular stages in a 
flea which had been fed on an infected rat five days previously, and 
since then had heen kept starved. 

2Compare especially the figures given below (Text-figs. 4-12) of 
stomachs reconstructed from serial sections; it will be seen that there 
may be instances in which only one or two small patches of epithelium are 
infected, and in such cases it would be a lucky accident if the infection 
were discovered in looking at the freshly teased-up stomachs, even after a 
prolonged search. 


THE RAT-TRYPANOSOME, TRYPANOSOMA LEWISI. 547 


at later periods than twenty-four hours, it is highly probable that in 
many of these the intracellular multiplication may have been completed 
and this stage passed. In one of our series of sections of a stomach 
preserved thirty-six hours after infection the stomach itself contained 
no trypanosomes of any kind, but two clumps of typical crithidial forms 
were found attached in the intestine behind the pylorus (PI. 42, fig. 273); 
in this case the stomach-phase must have been over, assuming that it is 
a necessary and essential part of the developmental cycle. In Table M it 
will be seen that in several cases fleas dissected at two or two and half 
days after the infective feed had no trypanosomes in the stomach but 
developmental! crithidial forms in the rectum. The absence of intra- 
cellular multiplicative forms in the stomach of a flea examined more 
than twenty-four hours after the infective feed is, therefore, no proof 
whatever that this phase has not been passed through in that flea. 

When due weight is given to both these considerations, namely the 
probability that scanty infections may be overlooked in hasty examina- 
tions of stomachs in the fresh conditions, and the fact that the 
stomach-phase may be passed through rapidly and ended very soon, 
there can be no doubt that the true percentage of fleas in which it must 
take place must be considerably higher than the figure, 16°5, derived 
from actual observation. We should probably not be far wrong in 
putting it above 20, a result agreeing fairly closely with the percentage 
of fleas found by direct experiment to become infective after having fed 
on an infected rat (see below, p. 663), and thus supporting our belief 
that the intracellular multiplication is necessary for the trypanosome to 
establish itself in the fiea, and is an indispensable part of the life-cycle ; 
a belief which is not, however, under the circumstances, capable of 
direct proof or verilication. 

It is at least evident from the figures given that the investigator must 
not expect to find the intracellular multiplication in more than a frac- 
tion (1/6 or 1/5) of the number of fleas examined which have fed but once 
onan infected rat ; and that heis most likely to find them from eighteen 
to thirty-six hours after the trypanosomes have been ingested by the 


flea. 


(2) The Type of Cell attacked by the Trypanosome. 


With regard to this point we have to note, in the first place, that we 
have never seen intracellular stages of the trypanosomes in the cells of 
the epithelial crypts. In this we are in agreement with Noller (1912), 
who has also studied sections of the flea’s stomach. Extracellular 
trypanosomes are sometimes seen attached to the exterior of cells 
which are passing out from the crypt, but never inside any cell which 
is not definitely a part of the general epithelium. It is possible that 


548 E. A. MINCHIN AND J. D. THOMSON. 


the separation which, as described above (p. 494), takes place between 
the epithelial cells until they are connected only by the isthmus at 
the base, makes it easier for the trypanosomes to penetrate into 
them, since the appearances seen in sections suggest that the attack 
is usually made on the side of the cell (Pl. 39, figs. 120-124); the 
occasional, though rare occurrence, however, of intracellular stages 
in quite young cells (Pl. 39, fig. 131), shows that the trypanosomes can 
penetrate into epithelial cells before the separation between them has 
developed. 

The degenerated epithelium, full of fatty deposits which blacken 
after treatment with osmic mixtures, is also not attacked by the try- 
panosomes. It is true that intracellular stages may sometimes be 
found in a degenerated cell, but in such cases the cell has been thrown 
off from the epithelium and the parasites are in the condition of large 
spheres ripe for breaking up into trypanosomes (PI. 39, fig. 155) ; or of 
masses and clumps of trypanosomes that have evidently been liberated 
by the recent disruption of a sphere (Pl. 38, fig. 114). It may be sup- 
posed that in these cases the cell was attacked by the trypanosomes 
before degeneration had set in; the process of degeneration may have 
been hastened by the action of the parasites. 

The trypanosomes attack by preference the fully-formed, but still 
young and vigorous cells, which may contain granules of the normal 
type and even yellow bodies, but no fatty deposits ; cells which may be 
well characterised as adolescent in type, and which stain a clear, light- 
grey with iron-hematoxylin after Flemming-fixation. Itisin such cells 
that the earlier stages of the intracellular multiplication are to be 
found ina flourishing condition and often in considerable numbers; but 
if the trypanosomes are numerous the cell soon becomes exhausted. 


(3) The Relation of the Trypanosomes to the Cells. 


The intracellular parasites are found most frequently in the apica 1 
expanded region of the cell, where the cytoplasm is usually of a loose 
spongy texture (PI. 39, fig. 126). Sometimes, however, the trypanosomes 
are seen below the nucleus and occasionally there may be parasites 
lodged above and below the nucleus in the same cell (PI. 39, fig. 134). 
Since the position of the nucleus in the cell is subject, as has been 
pointed out above, to variation, no special significance attaches to this 
point. It may be noted, however, that when a sphere is situated, as 
regards the principal mass of its body, above the nucleus, the tail of 
the sphere may run down to the base of the cell (Pl. 38, fig. 108). 

In epithelium of the flattened form the intracellular parasites are 
lodged beside the nucleus, which is often pushed to one side of the cell. 


THE RAT-TRYPANOSOME, TRYPANOSOMA LEWISI. 549 


(4) The Effects of the Trypanosomes on the Epithelial 
Cells. 


In describing the pathological effects of the trypanosomes, we may 
begin with the cases where a cell is attacked by a few, not more than 
two or three, parasites, and has been but little affected by them (PI. 39, 
figs. 126, 127). It is seen that the parasite is lodged in a distinct space 
or vacuole in the cytoplasm, a vacuole which is not to be interpreted as 
an artefact due to shrinkage from the action of reagents, but as pro- 
duced by the liquefactive or absortive action of the parasite on the 
cytoplasm, since the vacuoles containing the intracellular parasites can 
be seen very clearly in the living cell. Similar effects produced by Toxo- 
plasma gondii in the peritoneal cells of the mouse have been 
described by Miss Pixell.! In the living condition. as stated above, the 
trypanosomes move freely by means of their flagella in the liquefied 
cytoplasm, and when there are many parasites in the same cell they 
can be seen to jostle one another and even to bump against the cell- 
nucleus. Owing to the liquefaction of its cytoplasm the cell becomes 
empty and exhausted, and in its final stage is reduced to the condition of 
a sack, containing fluid, which may be crammed with trypanosomes 
produced by the process of multiplication within the cell; remnants of 
the cytoplasm may persist on the wall of the sac, and the cell-nucleus 
is also to be found at some point, generally adherent to the wall. Often 
these exhausted cells contain fatty deposits and also the brownish- 
yellow bodies described above. 

The effect produced on the parasitised cell asa whole is a more or 
less distinct hypertrophy, more obvious when the contained parasites 
are present in large numbers (PI. 38, fig. 114). Then the cell often 
becomes very much enlarged, but this enlargement appears to be 
brought about in many cases not solely by the hypertrophy of the 
individual epithelial cell, but also by the fusion of distinct cells ; at 
least this is the only way in which we are able to interpret the large 
multinucleate cells, containing numerous parasites in all stages of their 
multiplication, that are frequently to be found. Mitoses, or any other 
forms of nuclear multiplication are never seen in the epithelium, and 
the very large size of these multinucleate cells (Pl. 39, fig. 128 ; com- 
pare Pl. 44, fig. 313), certainly indicates cell-fusion having taken place. 
We have also seen multinucleate epithelial cells not containing para- 
sites: but such cases might very well be those in which the intra- 
cellular multiplication was ended and the daughter-trypanosomes 
produced had escaped from the cell. 

As the cells are exhausted and destroyed by the parasites there is a 
pronounced tendency for them to be thrown off from the epithelium. 


1* Proc. Roy. Soc.,’ (b), Ixxxvii, p. 73, (Pl. ix, figs. 3-6). 


DdO E. A. MINCHIN AND J. D. THOMSON. 


In the sections infected patches of epithelium are of frequent occurrence, 
in which the infected cells are seen bulging far out from the general 
level of the epithelium into the lumen of the stomach, giving sometimes 
an appearance like a bunch of grapes (PI. 44, figs. 313 and 315). Insuch 
protuberances every cell, as a rule, contains numerous parasites in all 
possible stages of multiplication, in progress or completed. In one of 
our series there is a stomach in which no parasites are to be found 
within the epithelial cells in situ in any part of the stomach-wall, but 
in the fore part of the stomach there are very numerous intracellular 
parasites, all contained in cells completely detached from the epithelium. 
On the other hand very large patches of epithelium may be found 
infected without the cell being thrown off; in one of our series there is 
a stomach, which, in one part shows infected cells the whole way round 
the section, except in the epithelial crypts (Text-fig. 9), but there is very 
little detachment of epithelial cells. It is evident that the expulsion of 
the infected cells from the epithelium is a measure partly of the inten- 
sity of the infection, i. e. the number of parasites contained in each cell, 
and partly of the length of time during which the parasites have been 
acting on the cells; it is not till the cells are becoming exhausted and 
incapable that they are ejected from the epithelium. 

From a study of the extracellular trypanosomes, which occur in close 
proximity to the epithelium, and more especially those actually attached 
to it, it is evident that the extent of the epithelial areas simultaneously 
attacked may vary very greatly in different cases. Sometimes only an 
attached trypanosome is to be seen here and there, scarcely so much as 
one in each consecutive section on the average; this corresponds to the 
frequent occurrence of solitary epithelial cells, or very small groups of 
them, containing parasites. Sometimes, on the contrary, all the epithe- 
lium on one side of many consecutive sections will be seen to have 
trypanosomes adherent to it, while on the other side of the same 
sections not a trypanosome is to be seen (compare Text-figs. 4-12). 
What determines the extent and distribution of the attacks it is impos- 
sible to say, but at least one necessity is probably the presence of 
regenerated “adolescent ” epithelium. 

It should be mentioned finally that in one of our series there is a 
stomach which in its hinder region is almost entirely denuded of epithe- 
lium. Vast numbers of trypanosomes are seen in all parts of the 
stomach-lumen, but swarming most thickly near the wall, on which only 
the epithelial crypts remain intact. Remnants of the general epithe- 
lium occur here and there in the form of broken-down cells containing 
the final stages of the intracellular multiplication, but for the most 
part the epithelium is gone altogether. In the anterior region of this 
stomach, trypanosomes are less abundant and the epithelium is in 
process of regeneration. It is surely a fortunate circumstance for the 


HE RAT-TRYPANOSOME, TRYPANOSOMA LEWISI. 551 


insect host that its epithelial crypts of regeneration are immune to the 
attacks of the trypanosomes ; were it not so, it is hardly credible that 
the flea could survive the extensive destruction of the epithelium that 


may occur with an intense infection. 


(5) The Relation of the Trypanosome-Infection, as a 
Whole, to the Stomach of the Flea. 


This point can best be made clear by describing a few examples, 
which will give a more graphic picture of the variations seen in different 
fleas than can be obtained froma general description. It must be stated, 
in the first place, that, as already pointed out above, in stomachs of fleas 
preserved eighteen, twenty-four, or thirty-six hours after having been 
fed on a well-infected rat, the trypanosomes may have entirely dis- 
appeared, having been digested with the blood and failed to establish 
themselves. Consequently, in many of our series of sections, some or 
all of the stomachs contain no trypanosomes at all. On account of the 
ereat expenditure of labour and time involved in searching through a 
complete series of sections of a stomach we have not attempted to com- 
pile any statistics of the numerical proportion of infected to non-infected 
stomachs, as we have done in the case of stomachs teased up and 
examined fresh, the latter being a method much less laborious, though 
not so exact, for determining whether a stomach contains trypanosomes 
or not. 

In the examples we are about to give we deal only with those 
stomachs in which trypanosomes have been found, and in order to 
exhibit at a glance the state of things in each stomach we have made 
reconstructions of a certuin number of stomachs in the following 
manner : 

(a) To show the distribution of the cells containing intracellular 
stages, the stomach is imagined as cut open from the intestine to the 
proventriculus along a line which corresponds to the southernmost 
(lowest) point in the transverse sections, but which is a purely arbitrary 
line so far as the flea is concerned and does not correspond to any 
definite anatomical plane of the insect. The wall of the stomach, sup- 
posed to have been cut open along this line, is further imagined as laid 
out flat. Of each stomach reconstructed in this way a diagrammatic 
sketch was made on paper ruled in millimeter squares to a definite 
scale, namely, 1 mm. to each transverse section of 6 » in thickness, and 
in the sketch three meridians are put in with dotted lines, the middle 
one to represent the northernmost (uppermost) point of the sections, 
while the meridians to the left and right represent respectively the 
extreme western (left-hand) and eastern (right-hand) points in the 
section. After making a diagrammatic sketch in this manner the series 
of sections was searched through and all infected cells were mapped out 


552 Ek. A. MINCHIN AND J. D. THOMSON. 


in the reconstructions, being represented by little circles placed in the 
line corresponding to the number of the section, and in the relation to 
the meridians that indicate its position in the section. For instance, if 
an infected cell is found in the twenty-seventh section, to the north-east 
of the section, the circle representing it is put into the diagram 27 mm. 
from the anterior end of the stomach and midway between the meridians 
N. and E. 

In this way a graphic representation, accurate to scale in the longi- 
tudinal direction, is obtained of the distribution of the infected cells in 
the epithelium. It is important to remember that each little circle 
represents not merely a single intracellular stage of the trypanosome, 
but an infected cell which may contain many such stages. 

(b) To show the distribution of the extracellular trypanosomes the 
diagram is supposed to represent the unopened stomach, and the 
trypanosomes that are in close proximity, or attached, to the epithelium 
are put in along the sides, while those scattered free in the débris are 
represented dispersed through the diagram, in the position corre- 
sponding to the number of the section in which they occur; when very 
numerous, however, it is impossible to represent them accurately, and 
they are merely crowded in as thickly as possible in the diagram. 

It is very important that the reader should understand clearly that in 
each of our diagrammatic reconstructions of the stomachs two 
distinct conventions are comb ined, as it were superposed one 
on the other. Originally our intention was to have given two diagrams 
for each stomach, one reconstructed according to method (a) to show 
the distribution of the intracellular parasites, the other according to 
method (b), indicating the occurrence of the extracellular trypanosomes. 
To have given two diagrams for each stomach, however, would not only 
have taken up much space, but would not have shown the state of 
affairs so graphically as when the two diagrams are combined into one 
and no confusion can arise if it is clearly understood that two different 
methods of reconstruction are combined in each figure. The fact that 
the attached trypanosomes are represented adherent to the sides of the 
diagram, for example, does not mean that they are attached only along 
the southern meridian ; they are attached at any point in the section, 
but had they been indicated in the diagram in the exact meridian in 
which they occur the eye would not have distinguished them from the 
free forms.! To indicate the exact position of the attached trypano- 


1 It was suggested by a friend that the attached trypanosomes might 
have been distinguished from those that are free by drawing at one end 
of them a little transverse line or semicircle, to indicate the attachment 
to the cell; but where the trypanosomes are thickly crowded, as in 
Text-fig. 10, it would have been impossible to distinguish clearly the 
free and the attached forms in this way. 


THE RAT-TRYPANOSOME, TRYPANOSOMA LEWISI. 553 


somes it would have been necessary to have shown the attached and the 
free trypanosomes in separate diagrams; either to have put in the 
attached forms in the diagram showing the infected cells, according to 
method (a), or to have constructed three diagrams showing the infected 
cells, the attached trypanosomes, and the free trypanosomes on each 
diagram separately. 

The reconstructions were made, as has been stated, on a definite 
scale as regards the length of the stomach, namely, on the scale of 
, or 167 approxi- 
mately. In the reproduction, however, the diagrams have been reduced 
to two-thirds of their original size, and appear, therefore, at a magnifi- 

sa 


¢ 


cation of Is oF 111 approximately, with the exception of figs. 3 and 


1 mm. to 6 p, that is to say at a magnification of 


11, which it has been necessary to reduce to one-half on account of the 
very large size of the stomachs, and which, therefore, appear at a 
magnification of about 83°5, while fig. 9 has been reduced to one-third. 
The diagrams show the great variations in the size of the flea’s stomach. 
The small stomachs are males, the large ones females. The number of 
sections in a series cut through a small stomach was generally about 
125, in a large stomach from about 220 to 300. This corresponds in a 
remarkable way to some observations we had made and recorded long 
previously to cutting the sections—on the feeding of the fleas. We 
found that a male flea took about a minute and a quarter to fill its 
stomach, afemale about two minutes and a quarter. 

The stomachs reconstructed by us were all from fleas fed twenty-four 
or thirty-six hours before being dissected and preserved. We will begin 
with the fleas of thirty-six hours, because we possess a very good series 
of this period, which has been studied very carefully by us, and which, 
having been preserved in Flemming’s fluid and stained with iron- 
hematoxylin, shows the state of the epithelium and the stomach-con- 
tents particularly well. 

(a) Thirty-six hours after feeding.—A batch of nine stomachs, 
all preserved at the same time, stuck on the same piece of liver and cut 
in the same block of paraffin. Owing to an accident to the block the 
two stomachs near one end of the liver were damaged; the anterior 
half of one of these stomachs and a small part of the anterior end of 
the other were broken away. In what remained of these two stomachs 
no trypanosomes of any kind were found, and the same was the case 
with one of the other stomachs in this series. There remain six 
stomachs to be described. 

(1) (Text-fig. 4). The stomach is a large one, going through about 
260 sections, corresponding to a length of about 156 mm., exclusive of 
the proventriculus. The blood-débris is large in amount and stained 


5 


a E. A. MINCHIN AND J. D. THOMSON. 


TEXxT-FIG. 4. 


me 
: 
Ro 
nfe? ; 
foe 


{ 


§ 
a 
Xi 
Xf, 
§ { ' 


w 
is 
ot 
‘ 
; 


Reconstruction of a stomach in the manner described in the text, 
to show the distribution of the intracellular and extracellular 
trypanosomes. The portion left blank is where some of the 
sections were accidentally destroyed. No meridians have been 
put in because, in this instance, all the intracellular trypano- 
somes were in detached cells. 36hours. Magnified about 83°5. 


THE RAT-TRYPANOSOME, TRYPANOSOMA LEWISI. 559 


yellow. The epithelium is for the most part flattened and very black, 
especially at the anterior end, where also there are very many detached 
cells. There are also a great many yellow necrosed cells, either in the 
form of extensive patches in the epithelium in situ, or of detached 
cells loose in the blood-débris. 

Extracellular trypanosomes abundant, especially round the detached 
cells in the anterior region. In the posterior half of the stomach they 
are much fewer, and almost all peripheral in position or attached to the 
epithelium. 

Intracellular trypanosomes very abundant in the anterior fourth, but 
only in detached cells, none in the epithelial cells in situ. Passing 
backwards they become scarcer, and in the posterior half of the stomach 
none are found. 

Since the intracellular parasites in this stomach are found only in 
detached cells, the meridians have been omitted in the reconstruction, it 
being impossible to refer the infected cells to their proper position in 
the stomach-wall. 

Interpretation.—This is a stomach in which the epithelium is in 
the final stages of degeneration, especially anteriorly, but regeneration 
has not begun. There has been an extensive intracellular infection in 
the most anterior region, represented now by advanced stages in the 
detached epithelial cells, and swarms of free trypanosomes are attaching 
themselves in preparation for a fresh attack. There are also clumps of 
degenerative forms free in the blood. 

(2) The stomach goes through about 220 sections corresponding to a 
length of 152mm. Blood-débris in moderate quantity, stained yellow ; 
epithelium clear, granular, containing many grains and pseudospheres. 
No trypanosomes in any stage found in the stomach, but in the intes- 
tine, behind the pylorus, in a large clump of attached crithidial forms. 

Interpretation.—A stomach which has recently been regenerated 
and in which the stomach-phase of the trypanosomes is completed and 
past. 

(5) (Text-fig. 5). <A large stomach going through about 260 sec- 
tions. Blood-débris about half absorbed; stains grey. Epithelium for 
the most part clear, columnar, very granular, with coarse grains and 
pseudospheres and many yellow bodies. There are a fair number of 
detached cells, some of them showing yellow necrosis, scattered through 
the stomach; in the posterior region there are some black degenerated 
cells still in situ or in process of detachment. 

Extracellular trypanosomes very scarce in the anterior half of the 
stomach ; in the posterior half they are more numerous and all attached 
to the epithelium or peripheral in position. 

Intracellular trypanosomes are found only in the anterior half, where 
there are scattered patches, variable in extent, of infected epithelium. 


HIN AND J. D. THOMSON. 


C 


A. MIN 


EH. 


5. 


TEXT-FIG. 


of sections 


series 


The 


Reconstruction of another stomach. 


Magnified 


36 hours. 


began just behind the proventriculus. 


about 111. 


7 


55 


~ 


TRYPANOSOMA LEWISI,. 


7 
= 


YPANOSOME 


vVAT-TR 


[ 


THE 


6. 


TEXT-FIG. 


Boi Caer etn, aa bee ES 4 ° 
S65 G0 
3 ‘ 
SS, r) 
\ eg 
\ eS 
\ == 
i] eas 
i : 
a ponte Sarees CaS Rah yan SE Ay ok ee hee oe 1 gaia! aera el ee 
lige lipealale Rh ARO keg rae an = ei le oS Sel eos aietevetel catato Sia /e in: < arr ener See ee 
' 
! 
| —— 
sia 
PINE ° 
Sons 89 
-- ) 
Bao ° ba) 
at a aes oo Ly 
ariel tee Cars Q 
Sein pee ind RROD Go OSS setae 
aes i Seen pr raeswsn | cas fees Reamer - —- -- Mp9 iS . 
i) C) 96 
0200 6 Ct) 
a) 


Magnified 


36 hours. 


another stomach. 38 
about 111. 


of 


Reconstruction 


558 E. A. MINCHIN AND J. D. THOMSON. 


Interpretation.—A stomach very recently regenerated, the pro- 
cess scarcely complete in the posterior region. In the anterior half 
practically all the trypanosomes have penetrated into the epithelium. 
In the posterior half an attack is beginning, but all the parasites are 
still extracellular. 

(4) (Text-fig. 6). A large stomach going through 225 sections. Blood- 
débrisas in last. Epithelium as in last, but less granular and with no 
yellow bodies ; no black cells in situ, but a few detached. 

Extracellular trypanosomes very few in number, scarcely more than 
twenty in the whole series; all attached to the epithelium or close to it. 

Intracellular trypanosomes in two fairly extensive patches of infected 
epithelium, one about the middle of the stomach, one close to the 
pylorus. In one of these patches the cells are badly affected ; in the 
other they are more normal. 

Interpretation.—A feebly-infected stomach, recently regenerated, 
in which almost all the trypanosomes are intracellular. 

(5) The stomach runs through 217 sections. Blood-débris large in 
amount; stained yellow. JEHpithelium: (a) in the first 130 sections 
mostly very dark, but interspersed with clearer patches of cells; many 
detached cells, some black, some yellow; (b) in the hinder 2, approxi- 
mately, of the stomach the epithelium is mostly clear, not very 
granular, with a few patches of black cells in situ,and many detached 
cells, black or yellow. 

Extracellular trypanosomes all confined to the anterior region, in 
rather scanty numbers for the most part peripheral in position and 
often attached. 

Intracellular trypanosomes found only in the posterior region; three 
patches of infected epithelium, one fairly large, running through forty- 
six sections, the cells considerably modified, and two small patches; also 
a detached cell containing a large sphere in section 132 (rather far 
forward). 

Interpretation—A stomach in which regeneration is just be- 
ginning in the anterior region and is fairly advanced posteriorly. In 
the anterior region the presence of necrosed cells indicates that there 
has been a recent attack, but at present there are only extracellular 
trypanosomes ; in the posterior region there are fairly extensive patches 
of infected epithelium and no extra-cellular trypanosomes. 

(6) (Text-fig.7). The stomach runs through about 240 sections. Blood- 
débris large in amount; yellow. Epithelium clear, columnar,and with 
many granules, and in places yellow bodies; interspersed are a few 
black or yellow cells, singly or in patches, in situ, and detached cells 
are also fairly numerous. 

Extracellular trypanosomes numerous; nearly all close to the 
epithelium and many attached along the whole length of the stomach. 


THE RAT-TRYPANOSOME, TRYPANOSOMA LEWISI. 


TEXT-FIG. 7. 


x ~ 
‘ 
eS : < 
ee. 
< tae 
av ; 
! i a 
| 
/ I 
H r 
1 ‘ = 
iPass 
! ; 
: ee 
~~ ! : 
! 
1 ! 
{ : 
‘ 
' roe of oie 
Ie L - 
we ' 3 
“3 I 
acre : 
! 
{ 1 : oe 
Deets ! ! : ; 
Re I ! I 
; 1 | ee 
: = 
co 1 pri 
; \ 
{ reo Feiss ; ae 
of \ : u a 
{ rom 0°0%_ 0 : 
} ; SO a il f 
we x ? , 
$ ie 
Ln ! al 2 
1 
\ ‘ ! = 
\ ! i “S 
at \ ; 
[ 
oi : i 
- ; I 1 an 
\s rs i ‘ ; a 
\ / 
\ al y Im 
| i , 
? | 
i ! : 
! nm 
! 
I See s 
| ' ; 
\ ; : 
- F 1 as 
Le \ i 
if : 4 
1 
1 i 
\ fi 
\ 
7 : 
: P ‘ 
n 1 
( : 
\ ; ! 
\ y 4 
: I ! 
1 ' / 
: 1 ‘ 
/ 
\ 1 
i] means 


2COT € Z c 4 ‘ 
Re sie f of = ( he st ™m ich. Owing to the exlg 1 Ss of 
S} ace t he hinder I al t; of the stomac h, in Ww hic h there were no 

y | c osomes 5 Ss t off . 36 hour Ss M g fi d | ; 
ti JANLOSO es ha been ¢ u . agni eda out ite 


VOL. 60, PART 4.—NEW sERIEs. 59 


E. A. MINCHIN AND J. D. THOMSON. 


560 


Intracellular trypanosomes practically confined to a fairly extensive 


patch of epithelium about the middle of the stomach. 


TEXT-FIG. 8. 


dentally 


The gap in the middle of 
were acci 


ach. 


stom 
the figure shows where three sections 


Magnified 111. 


24 hours. 


Reconstruction of another 
destroyed. 


Interpretation—A stomach which has undergone regeneration, 
which is scarcely completed. A previous intracellular generation of 


iven rise 


trypanosomes, indicated by the presence of necrosed cells, has g 


THE RAT-TRYPANOSOME, TRYPANOSOMA LEWISI. 561 


to numerous free trypanosomes that are attacking the regenerated 
epithelium, and have penetrated it and established themselves in the 
cells ina few places. 

(b) Twenty-four hours after feeding.—(7) (Text-fig. 8). 
From a series preserved in Maier’s fluid.!’ A small stomach, running 
through about 150 sections. In the anterior quarter, or thereabouts, 
of the stomach are a fair number of attached forms. Behind this 
region begins a very intense intracellular infection which diminishes 
somewhat in the posterior fourth, except for an extensive patch close to 
the pylorus. Through all this region there are practically no attached 
trypanosomes, but there are a few scattered through the débris, which 
become more numerous towards the pylorus, and are found passing 
down the intestine, where some of them are attached to the lining. 

Interpretation.—The intracellular multiplication is at its 
height and the trypanosomes are beginning to pass on to the rectum, 
but at the extreme anterior end a fresh attack is developing. 

(8) From the same series as the last. A small stomach running 
through about 154 sections. In the anterior twenty sections the epithe- 
lium is columnar, normal, or in places tending to degenerate, with no 
intracellular trypanosomes, and very few free trypanosomes in the 
sections. For about forty sections behind this there is an intense infec- 
tion of the epithelium, which in places is largely destroyed or broken 
up, and there are great numbers of extracellular trypanosomes, 
some scattered through the débris, others in close proximity to the 
stomach-wall. In the next fifty sections, approximately, the epithelium 
is almost entirely destroyed; only the crypts remain, with here and 
there a few broken-down cells still adherent to the wall and containing 
late or completed stages of the multiplication; there are also in this 
region enormous numbers of extracellular trypanosomes, mostly in close 
proximity to the wall of the stomach, but also scattered through the 
débris, in which there are, in addition, masses of detached, broken- 
down cells, many of them containing large spheres or clumps of trypano- 
somes. In the twelve sections behind this region the number of 
trypanosomes and detached cells diminishes rapidly, and in the last 
thirty-two sections there are no trypanosomes, though still a few 
detached cells. 

Interpretation.—A stomach with an extraordinarily intense infec- 
tion, which, in the middle region, has destroyed the epithelium com- 
pletely. The intracellular multiplication has produced vast numbers 
of trypanosomes, which are not beginning as yet to migrate backwards 


1 In the case of stomachs (7) to (15) the slides of the series were 
stained alternately with the iron-hematoxylin-Lichtgriin-combination 
and by Giemsa’s method. 


Di da, Ws THOMSON. 


CHIN AN 


fo | 
© 
ie) 


Tpxt-FIG. 2. 


0°29 ‘ 
OR 9.29 
a090 Op?D 

Do 


1010S 


‘og of sect 


The serie 


nach. 


stor 


j of another 


truction 


Recons 


n 

“= 

t=] 

fo} 

lon] 

ew 

= 

fon | 

eS 

iw 

is) 

= 

ne) 

eal 

(eb) 

a 

es 

~~ 

n 

3S 

Pol 

fas] 

oa 

Nn 

a) 

rob) 

_— 

on 

re 

na 

— 
re 

aor 

or 

~ 

aL 

jee 

a 

qo=_ 

A 3 

os 

orn 
o 

=n 

oe | 

Aa 

m & 

orcs 

Relist 

ee 


THE RAT-TRYPANOSOME, TRYPANOSOMA LEWISI. 563 


to the rectum, judging from their absence in the posterior region of the 
stomach. 

(9) From the same series as the last. A large stomach running 
through about 234 sections. .Epithelium almost everywhere columnar, 
a few flat cells; very few detached. 

Extracellular trypanosomes fairly numerous towards the hinder end, 
mostly peripheral and many attached, a few scattered in the débris. 
No intracellular typanosomes found. 

Interpretation.—A stomach in which one or more generations of 
intracellular multiplication have been completed and a new attack on 
the cells is beginning. 

(10) (Text-fig. 9). From the same series as the last. A large stomach, 
but unfortunately not quite enough sections were mounted and the 
series ends before the pylorus. Blood-débris large in amount every- 
where. 

In the anterior half of the stomach there are a few extracellular 
trypanosomes scattered through the blood-débris, very few attached ; 
there are also some intracellular parasites, for the most part scanty, 
but in one place there is a fairly extensive patch of infected epithe- 
lium. 

In the posterior half there is a very intense infection of the cells, 
which in many sections is seen all round the section, or interrupted only 
by the epithelial crypts. In this region, extracellular trypanosomes 
are practically absent; one clump, apparently of degenerative forms, 
was found, and in the hindermost region free trypanosomes, scattered 
in the débris, begin to appear abundant in the pyloric region. 

Interpretation.—Over a large extent of the stomach the intense 
infection of the epithelium has absorbed, so to speak, all the free try- 
panosomes except the degenerating forms. 

(11) From another series preserved in Maier’s fluid. A stomach of 
moderate size, running through about 180 sections. 

Extracellular trypanosomes swarming in the posterior region of the 
stomach and in the intestine; towards the anterior end they diminish 
in number, but are to be found right up to the proventriculus. 
Almost all these trypanosomes are free in the débris, very few are 
attached. 

No intracellular trypanosomes found. 

Interpretation.—A stomach in which the intracellular multipli- 
cation is practically ended and the trypanosomes, present in large 
numbers, are passing down to the rectum. 

(12) Text-fig. 10). From the same series as the last. A very large 
stomach, running through about 320 sections, corresponding to a length 
of nearly 2 mm. 

Extracellular trypanosomes occur in vast swarms along the whole 


564 E. A. MINCHIN AND J. D. THOMSON. 


length of the stomach; most of them are peripheral in position 
occurring all along the east (right-hand) side of the sections, and many 


TrExt-Fic. 10. 


Reconstruction of a very large stomach. 24 hours. Magnified 
about 55°5. 


of them are attached. Trypanosomes also occur, however, scattered 
through the débris in all the sections, but more especially in the 
posterior half of the stomach, and swarms of them are seen in the 


THE RAT-TRYPANOSOME, TRYPANOSOMA LEWISI. 565 


TExt-FIG. 11, 


; 1 
‘ ( 
t ‘ 
) 1 
' ‘ 
“4 1 
‘ 1 
' 
' { 
1 t 
' i] 
' 1 
' { 
i] bo wd a 
LY t 
1 
i] 
' 

Ss 1 
Tae ‘ 
fi - Se ( 
i. =O { 
aca \ 
Wy os / 
ie Me i 
' o 
' { 
U { 
( ° 
Hl x, y 
{ ° 1 
I : { 
t 
1 
F ( 
! { 
( 1 
( 

{ ! 
1 1 
u ! 
( 

1 ( 
1 

' 

t 

t 

' 

{ 

t 

t 

( 

1 

' 

1 

‘ 

1 

i 

| 

1 

! 

( 

t 

‘ 

1 

\ 

t 

| 

t 

i 

' 

1 

' 

t 

‘ 

' 

1 

1 

' 

' 

‘ 

‘ 

' 

1 

1 ‘ 
Sites 
See S 
aoe 


Reconstruction of another stomach. The anterior part of the 
figure, from just behind the proventriculus, has been cut off 
owing to exigencies of space, but there were no trypanosomes 
in the part that has been cut off. 24 hours. Magnified 111. 


566 E. A. MINCHIN AND J. D. THOMSON. 


pyloric region passing down the intestine. Many of these free forms, 
whether central or peripheral in position, are in clumps and are possibly 
degenerative. 

Intracellular trypanosomes are not found in the posterior half of the 
stomach, but towards the middle a few detached infected cells are 
found here and there. In the anterior region there is a very large 
patch of infected epithelium in situ and also two smaller patches. 

Interpretation. — Intracellular multiplication is proceeding 
actively in the anterior region, and a new attack:on the epithelium is 
preparing all along one side of the stomach. At the same time, trypano- 
somes are passing down in large numbers towards the rectum and there 
are many degenerative clumps. 

(13) (Text-fig. 11). From another batch preserved in Maier’s fiuid- 
A large stomach running through 255 sections. 

Extracellular trypanosomes very scanty, practically confined to 
posterior half of stomach, all attached except in pyloric region, where 
there are a few scattered freely in the débris. Many of the attached 
trypanosomes are in clumps, perhaps degenerative. 

Intracellular trypanosomes only in anterior region, very scarce. 

Interpretation.—The stomach-phase seems to be maintaining 
itself with difficulty in this flea; the trypanosomes are few in number 
and largely degenerative, but a few are passing backwards. 

(14) Text-fig. 12). From the same batch as the last. A large 
stomach, running through about 225 sections. 

In the anterior third, approximately, of the stomach extracellular 
trypanosomes are abundant, both free in the débris and attached to 
the epithelium, but there are no intracellular stages. 

In the region behind this, approximately the middle third or fourth 
of the stomach, there are no extracellular trypanosomes, either free or 
attached, but a very extensive intracellular infection is found, confined 
for the most part to the right side of the sections, where almost the 
whole of the epithelium, with the exception of the crypts, has been 
attacked; the reconstruction shows very well the clear spaces which 
represent the situations of the epithelial crypts. 

In the posterior third of the stomach a remarkable condition of things 
isfound. On the right side of the sections there are still numerous 
intracellular trypanosomes, but no attached extracellular parasites 5 
on the left side of the sections there are very numerous extracellular 
trypanosomes, attached or in close proximity to the epithelium, but 
none intracellular. Scattered through the débris are many free try- 
panosomes which are also passing down the intestine. 

Interpretation.—A stomach in which intracellular multiplication 
is still proceeding actively and from which trypanosomes are passing | 
down towards the rectum. 


567 


ified 


gni 


WISI. 


’ 
4 


TRYPANOSOMA Lk 


rl 
ot) 


RAT-TRYPANOSOME 


° 04% 2080 , 3 
pole? 9 88 a a 
20°. 08 oo 
oo 
oe50 5 ees a 
20n5 3° O52 o%o 
oo OTL» 
oe 2 ° oe i ee 
° ° ° 80 6° 00.09 Po 
¢ os 20987 9270 GP ne 
, 0° 0°, 907905 5 y) ° os 
WASH © 5% a0, 0.09 00%? 95%, _ 3° 00 
20259, 00 %05.%0° 0 
Sf Fae ° Baeza cer ePave vee vee Se toe 
ee ee bp aed aeneee Soap oe, 
Semore Pale Oa Loo) 9 020 9,2 > roe 
wy ¥ oe 
OIE s meat 259.00 SOP ©” Pos 5 "a0. 0 9, 0 0,00" © 
2" [os EY & nm 3 0° Gero 0020 ° F020 25 2 0 8 Oleo 9% 
¢ ° ° ~o. ° ° 
La iy nhs, nw chee iP Ne: Deeg PO EPS OR IOS =5O.G pone SOUO SS 
ye oe ey a 0005 Co 4 SR 2 0.0% Poo Qvet “_ 
~ 5 & ° ° 9 BG gydag 8 COON 8° GY 
‘ye? q eo ea aS ed) OO RU OY RS roe eGo 
eu fo c) ° TO) a MR ee Se GS, 
i, 2 Benes 2.059200 00 


sof UD} Oe 7S 483 Bak so 280 
fa a Whe ah ao 98g Bee Nagler F 
: Se 50.9 coos, 0 Raids titel gO cM CctleS is casa momaterats en UF 
gy ra 52 2% Beh ai SOTA aay TN Sorseeet ate NIA Eric 


z ww 
I 
—-t-—-._ 9 
anes 
GRR 
We AD ee 
Sam e&y$ 
et 
eee! 
Ste 
ao 
ae 
oat 
aa 
g g 
PS ms 
5 
~ & 
cant 
ee 
on 


TERxT-FIG. 12. 


Mcrae 
par San a foe Tt oe et ae a FE Ca Nee SSaunigice Pica dating hy Perea ys eye een pee 
Tae waco ton Fe 2h 48 aaa DIG eee 
Yarketeite ome CEES gp Ree ae ened 
Tae é cay mn eal Ratan is ate Joie > oN Seer LN, 
Wr oe oe OOD TERE SNE 
"Rr ees Fara ore pec nies ze a PPh Re SENN a BUN 
Ges Bene L ny aes: 
: DEAER ose --------- Oe be Sea Tadley ott 
See fe | ie Sa UNA So Be : 
$ v 25365 nb ° 
re eee a) Ag fa 
a 2p a 
a aire 35°00 90 om °. ig 


Ma 


24 hours. 


about 83°5. 


~ 
Uj 


THI 


Reconstruction of another stomach. 


5658 E. A. MINCHIN AND J. D. THOMSON. 


(15) From a batch preserved in sublimate-acetic. A large stomach 
which passes through 241 sections without quite reaching pylorus. 

About 10 extracellular trypanosomes, most of them attached to the 
epithelium, are found in the hindermost sections. 

No intracellular trypanosomes are to be found. 

Interpretation.—A stomach in which the intracellular multi- 
plication is either completed or almost inhibited. 


(8) The Migration to the Rectum. 


As stated above, the trypanosomes occurring in the intestine 
are usually in transit from the stomach to the rectum, and only 
exceptionally attach themselves to the intestinal wall or 
undergo further development there. We have been able on 
more than one occasion to observe the actual passage of the 
trypanosomes along the intestine. In a flea which had fed 
on an infected rat about twenty-four hours previously, the 
stomach, with a considerable length of the intestine attached, 
was dissected out on a slide partly teased up and covered 
with a glass slip; the posterior part of the stomach, however, 
with the intestine attached, was left intact. Close to the 
pylorus the stomach contained fluid in which a great number 
of brown granules were suspended, coarse granules of fecal 
appearance evidently representing indigestible residue of the 
last meal. These granules could be seen to be in a state of 
violent commotion, more violent than could be explained 
merely by Brownian movement, since they were being churned 
and stirred in every direction; but although there could be 
little doubt but that the movements were due to the activity 
of trypanosomes, the flagellates themselves could not be 
seen clearly in the opaque fluid through the stomach-wall. 
Soon, however, a bolus of fluid passed through the pylorus 
into the intestine and passed down it by peristaltic action, 
pushed onwards like a bead, until it reached the cut end of 
the intestine and was extruded from it. It was then seen at 
once that the fluid contained, in addition to the fecal granules, 
a swarm of excessively active trypanosomes, long and relatively 
slender forms with great powers of rapid forward progression. 


THE RAT-TRYPANOSOME, TRYPANOSOMA LEWISI. 569 


They began at once to spread in all directions in the fluid ; 
but at this moment the coverslip was picked off with needles 
and dropped instantly into Schaudinn’s fluid; after it had 
been stained and mounted it was found that, by a piece of 
good luck, the stomach and intestine had remained sticking 
to the coverslip, and that round the cut end of the intestine 
were several trypanosomes of the long stomach-type (Pl. 42, 
figs. 203, 204), evidently some of those that had been seen to 
pass down the intestine. In two other fleas, fed respectively 
twenty-four and eighteen hours previously on infected rats, 
we were able to confirm our observations on the passage of 
the trypanosomes down the intestine and to obtain prepara- 
tions of them (Pl. 37, fig. 50). 

It is unnecessary to give a detailed description of the 
migratory trypanosomes, since itis evident from the figures that 
they are simply of the long stomach-type already described. 
They are very active, and in form crithidiomorphic, with the 
hinder part of the body stiff and straight, sometimes slightly 
or even markedly clubbed and swollen. The two nuclei are 
more or less closely approximated, but 71s almost always well 
behind N. It is seen from this that the trypanosomes result- 
ing from the intracellular multiplication in the stomach may 
do one of two things ; they may penetrate again into epithelial 
cells and go through another generation of multiplication ; 
or they may collect in the pyloric region and be carried down 
the intestine (compare Pl. 45). The migration may begin as 
has been seen, as early as eighteen hours, but this appears 
to be rather exceptionally early, to judge from our observa- 
tions on the rectal phase; probably it does not usually begin 
till twenty-four or thirty-six hours. It continues, doubtless, as 
long as the stomach-phase lasts, and as already stated, we have 
found intracellular stages as late as five days in fleas not fed 
a second time; we may suppose, therefore, that the produc- 
tion of the migratory forms and their passage down to the 
rectum, may be going on continually, in some cases, until the 
second feed of the flea. In other cases, however, the multi- 
plication in the stomach probably comes to an end of itself, 


570 E. A. MINCHIN AND J. D. THOMSON. 


before the second feed, judging by the many observed 
instances in which, in fleas not fed a second time, the stomach 
may contain many long free trypanosomes, or the rectal 
phase may be weli established, without any intracellular 
multiplication occurring in the stomach. 


(c) The Rectal Phase. 


The starting point of the rectal phase is the long, active 
“‘ crithidiomorphic” type of trypanosomes already described, 
which migrates down the intestine. During its passage down 
the intestine the changes of form and structure which may 
have begun already in the stomach continue, so that by the 
time it reaches the rectum its posterior end is generally 
distinctly club-shaped. Arrived in the rectum, it very soon 
undergoes changes in form, habits and structure, and multi- 
plies by binary fission, giving rise ultimately to the typical 
forms of the established rectal phase, forms which, apart from 
other characters, are of much smaller size and bulk than 
those of the stomach-phase. We will discuss first (a) the 
transition from the initial to the established forms of the 
rectal phase, and then (b) the various types of modifications 
of the latter, culminating in the little trypanosome-form, 
which is the final stage of the development of the flea. 


(a) The Transition to the Crithidial Form. 


If the various processes of change in the initial stages of 
the rectal phase be analyzed, after study of both living and 
preserved material, and by comparing the starting point of 
this part of the development with its final result, we may note 
the following tendencies in the organism. 

In the first place it loses its intense activity and becomes 
more sluggish in movement, with a great tendency to attach 
itself by the tip of the flagellum ; under natural conditions 
the flagellate attaches itself to the wall of the hind-gut, but 
when under microscopic examination it can be seen to adhere 


THE RAL-TRYPANOSOME, TRYPANOSOMA LEWISI. 571 


firmly to pieces of débris of any kind, or to the surface of 
the glass slide or coverslip. When not attached in this way 
it progresses slowly with the flagellum directed forwards. 

Secondly, the body shortens and changes in form by the 
cytoplasm becoming concentrated towards the posterior end 
of the body. 

Thirdly, the flagellum becomes progressively shortened. 

Fourthly, the two nuclei, if still in their original positions, 
become transposed into the typical crithidial arrangement, 
with 2 close beside or in front of N, 

Fifthly, the nuclei, the flagellum, and finally the whole 
body are multiplied by division or reduplication. 

The order in which these different 'processes of change 
have just been stated is in no way to be taken as indicating 
their chronological order of succession in the development, 
since they take place more or less independently and, as it 
were, at different rates of acceleration in different indi- 
viduals ; the result is consequently the production of a great 
number of forms which at first are rather bewildering and diffi- 
cult to arrange in a series. The difficulty of tracing in detail 
the transition from the initial stage of the rectal phase to the 
established crithidial condition is increased by the fact that 
the early transitional stages are extraordinarily rare and diffi- 
cult to find in the permanent preparations; a fact which 
indicates that the transition takes place very rapidly and is 
completed very quickly. One explanation of this rapid 
change may perhaps be found in the great diminution in size 
which is brought about in this part of the development. 
Leaving out of consideration for the moment any structural 
or other changes, it is at least quite clear that the 
large individuals which come down from the stomach 
initiate a series of generations of multiplication by fission 
culminating in forms perhaps not more than a tenth the size 
of the initial forms from which they are derived. Conse- 
quently it is probable that the successive divisions of the 
body follow one another at first with extreme rapidity and 
without intervening pauses to allow the daughter-individuals. 


572 E. A. MINCHIN AND J. D. THOMSON. 


to grow to the size of the parent, as would happen in the 
normal multiplication such as takes place in the established 
crithidial stage. Another explanation for the rarity of the 
transition-stages may be the possibility that of the long 
trypanosome-forms which come down from the stomach to the 
rectum, only a small number may go through their metamor- 
phosis into the typical rectal phase and the greater number 
may degenerate or be carried out of the flea. Whether this be 
true or not, it is not necessary to suppose that all those which 
pass down from the stomach do so in a single swarm; it is 
more probable that they dribble down from the stomach, so 
to speak, in larger or smaller bands or singly, a supposition 
which would also account for the small number found at any 
one time in the rectum undergoing their metamorphosis. 

A further difficulty which may arise in distinguishing the 
forms of the transition from trypanosome to erithidia, and in 
assigning them to their proper position in the series, is the 
fact that the body may become artificially broadened in pre- 
parations which have been allowed inadvertently to dry up 
the least bit before fixation. Deformed specimens of this kind 
can berecognized by their flattened appearance and consequent 
even staining of the body, which in a properly preserved 
specimen should be thicker and more opaque in the axial 
region than towards the edges of the body; the trophonucleus 
in flattened specimens is often transversely elongated ; and, 
further, the process of drying seldom affects a single 
specimen, unless it is very isolated or near the edge of the 
film, but, if it has taken place at all, the effects of desiccation 
are apparent over at least a considerable area of the slide or 
coverslip. It is, therefore, not difficult with a little practice 
to detect the specimens which have become broadened arti- 
ficially ; and in any case the process of drying does not affect 
the length of the body or flagellum to any appreciable extent. 

It is a result, doubtless, of the rapidity with which the 
transition is effected that we have not been able to come to a 
perfect agreement of opinion between ourselves as to certain 
points of the development during this transitional period, 


THE RAT-TRYPANOSOME, TRYPANOSOMA LEWISI. 573 


namely, the exact stage in the process of change of form at 
which the first division of the initial rectal form (that is to 
say, the long club-shaped forms that. come down from the 
stomach) takes place, and consequently the type of binary 
division, whether equal or unequal, which initiates the 
whole series of generations in the rectal phase. Before 
we discuss this doubtful point, we may first classify the 
various types seen in the initial transitional stages of the 
rectal phase, for which it is simplest to take the types of 
body-form as the basis of classification. Bearing in mind the 
considerations of technique that have been discussed already, 
and being careful, therefore, to eliminate all cases where 
there is reason to suspect artificial deformation of the 
specimens, we can recognise the following series of forms, 
each of which is a stage in the progressive shortening and 
broadening of the body by concentration of the cytoplasmic 
substance in the posterior third of the original slender 
trypanosome : 

(a) Slender forms, differing but little from the long 
stomach type, and evidently but recently arrived in the rectum. 
The flagellum is still long, and N is usually in front of n 
(Pl. 42, fig. 208) ; sometimes, however, the reverse is the case 
(PI. 42, fig. 209). 

(b) Forms in which the hinder region of the body, contain- 
ing the two nuclei, become swollen and club-shaped (Pl. 42, 
fig. 206), and the anterior part correspondingly attenuated, 
until the body as a whole becomes more or less tadpole- 
shaped (Pl. 41, fig. 150; Pl. 42, figs. 205, 207). The flagellum 
at this stage is usually long, but sometimes distinctly 
shortened (PI. 41, fig. 158; Pl. 42, fig. 210).! The nuclei are 
usually still in their original relative positions, but are some- 
times crithidial in arrangement (PI. 42, fig. 211). 

(c) Forms in which the concentration of the body-substance 


' How the shortening of the flagellum takes place is not clear, but it 
is worthy of note that in smears there are often found broken pieces of 
flagella near the specimens, as if the flagellum had become brittle and 
had broken off (PI. 41, figs. 158, 160). 


574 EK. A. MINCHIN AND J. D. THOMSON. 


in the posterior half or third of the body has proceeded so far 
that the body is nearly half as broad as long, and the anterior 
prolongation which forms the undulating membrane is much 
reduced in length and very slender. Examples of transitions 
from the last stage to this are seen in Pl. 41, figs. 151, 152; 


TExt-FIG. 13. 


Diagram to show the possible modes of transition from the 
stomach-phase to the rectal (crithidial) phase: a, slender form 
newly arrived from the stomach (compare PI. 42, fig. 208, ete.) ; 
b, early club-shaped form (compare PI. 42, fig. 205, etc.) ; ¢, later 
(more swollen) club-shaped form (compare PI. 41, fig. 149) ; 
ec, unequal division of ¢ (hypothetical); d, large pear-shaped 
form (compare PI. 41, figs. 154, 155, etc.) ; dd, division of d (com- 
pare Pl. 41, fig. 179); e, forms resulting from the initial division 
of the rectal phase (compare PI. 41, figs. 153, 157, 164, ete.). (x 
about 2000.) 


the complete realisation of this stage is seen in PI. 41, figs. 149, 
154,162; Pl. 42, fig. 212. From a comparison of the figures 
it is seen that the position of the nuclei varies considerably ; 
n may still be at or near the hinder end (PI. 41, fig. 162), or 
may be close beside N (PI. 41, figs. 149, 154; Pl. 42, fio. 212). 
The hinder end may be pointed, or quite round. The 


THE RAT-TRYPANOSOME, TRYPANOSOMA LEWISI. 575 


flagellum may still be a fair length, or quite short (Pl. 41, 
fig. 162). 

(d) Forms in which the body is contracted into a pear- 
shaped mass, nearly as broad as long. The stalk of the pear 
is formed by the flagellum together with a slight projection 
of the body representing the anterior prolongation, now 
usually reduced to its smallest limits (Pl. 41, figs. 155, 156, 
161; Pl. 42, fig. 215). The nuclei are usually transposed, 1 
being beside or in front of N. The length of the flagellum 
varies within wide limits represented by PI. 41, figs. 155 and 
156. 

(e) Forms which are distinguished from the preceding stage 
chiefly by their smaller size. Here again, however, caution 
is necessary in referring a given specimen to its place in the 
series, since the apparent size in permanent preparations may 
be affected considerably by technique. In the preparations 
fixed on the slide with osmic vapour and stained by Giemsa’s 
method the trypanosomes always appear considerably larger 
than those on the coverslips fixed wet with sublimate 
solutions, stained with hematoxylin and mounted without 
drying in canada balsam (vide Minchin, 1909). Conse- 
quently, different standards of sizes are required for inter- 
preting preparations made by these two different methods of 
procedure, and preparations made by the one method must 
not be compared directly, without making due allowance for 
the differences in result, with those made out by the other 
method. Nevertheless, after giving due weight to these 
considerations, it is not difficult to distinguish forms which 
are about half the bulk of the stages already described 
(Pl. 41, figs. 153, 159, 160, 164; Pl. 42, fig. 218). Such forms 
are, without doubt, individuals derived from at least one, 
possibly more than one, division of the initial form, a conclu- 
sion supported by the occurrence of such forms in pairs, 
possibly as the result of division recently completed (PI. 41, 
fig. 157; Pl. 42, fig. 216a). 

The question which must remain open at present is, in 
which of the four stages of changes of form (a), (b), (c), or 

vou. 60, PART 4, —NEW SERIES. 40 


576 E.' A. MINCHIN AND J. D. THOMSON. 


(d) (see Text-fig. 15), does the initial division of the rectal 
phase take place? 

If, as is possible, and as one of us (J. D. T.) thinks probable, 
the first division takes place in the club-shaped form as soon 
as the concentration of the protoplasm is completed and 
when (as in Text-fig. 15, c) the two nuclei lie close together, 
it may well be the case that the products of division would 
then be markedly unequal in appearance owing to the reten- 
tion of the old flagellum by one of the two daughter-indi- 
viduals; that is to say, the division would be of the type 
found in similar club-shaped forms in the trypanosome of the 
gold-fish both in cultures and in the leech, and in bird- 
trypanosomes both in cultures and in the mosquito, and which 
has also been seen in an early culture of I’. lewisi itself. 
On this view the swollen, pear-shaped forms of stage d would 
have to be interpreted as forms subsequent to, and the pro- 
ducts of, the initial division. In spite of much searching, 
however, through our preparations, we have been unable to 
find actual examples of club-shaped forms showing division 
markedly unequal in appearance. 

If, on the other hand, as one of us (EH. A. M.) believes, it is 
the most usual state of things for the organism to continue 
the process of contraction and broadening of the body until 
it has reached the condition of stage d, it is probable that 
the flagellate would then divide by the type of binary fission 
characteristic of the subsequent generations of the crithidial 
phase, that is to say, producing two daughter-individuals that 
are equal, or not markedly unequal in bulk, and differing 
only in that one of them has the flagellum temporarily shorter 
than that of the other. Pl. 41, figs. 179, 180, and Pl. 42, 
fig. 216, may then be interpreted as examples of the initial 
division, and Pl. 41, fig. 157 and Pl. 42, fig. 2164, as pairs 
resulting from the initial division recently completed. 


- The problem of the initial division of the trypanosomes in the rectum 
is one which involves more than the question as to the type of form, 
club-shaped or pear-shaped, in which the division takes place, or the 
question whether the products of division differ in visible characters of 


THE RAT-TRYPANOSOME, TRYPANOSOMA LEWISI. 577 


bulk, structure, or appearance; it raises a much deeper and more 
fundamental problem, namely, whether the division which initiates the 
rectal phase is an equating division which gives rise to two equi- 
potential daughter-individuals or a differentiating division which pro- 
duces inequipotential forms. If the division-products are equipotential, 
then visible differences between them of any kind would be merely 
temporary and of no significance for the future behaviour and destiny 
of the sister-individuals, which would be true twins ; all such differences, 
however pronounced, would be immaterial for the development as a 
whole, and the same would apply to the parent individual, whether club- 
shaped or pear-shaped. If, on the other hand, the daughter-individuals 
are inequipotential, in the sense that the smaller of the two division- 
products becomes the starting point of an indefinite number of genera- 
tions of small crithidial individuals, while the larger is merely a “ parent ” 
which, though it may divide in the same manner to produce small 
crithidial forms several times in succession, does not ultimately develop 
further, but drops out, as it were, of the direct line of the development 
when its powers of reproduction are exhausted; if they are inequi- 
potential in this sense, it follows that the observed difference between the 
two daughter-individuals would have an important significance and that 
they would not then be true twins, and further, that the club-shaped form, 
assuming that this is the form in which division takes place, would then 
be a developmental form of special and peculiar significance in the life- 
cycle and not merely a stage in the change of form leading to the pear- 
shaped stage. We must be content, unfortunately, with enunciating 
these possibilities, without being in a position to decide between them ; 
owing to the fact, to which reference has been made above, that the 
extreme rarity of transitional forms in our preparations has supplied us 
with insufficient material for a decisive judgment. 


The type of the initial division and the exact point at which 
it occurs in the series of progressive form-changes of the 
rectal phase must be left at present an open question, unfor- 
tunately ; but this much may be stated positively about the 
process of division in the rectal phase in all cases, whether in 
the initial or subsequent stages. No multiple division occurs 
henceforth in the developmental cycle, but the parasites settle 
down to a course of simple binary fission continued indefi- 
nitely, and always taking place in the lumen of the gut, never 
within cells. The process of fission is initiated by division of 
the blepharoplast or basal granule of the flagellum, but the 
flagellum itself does not divide; the original or parent 


578 E. A. MINCHIN AND J. D. THOMSON. 


flagellum remains attached to one of the two daughter- 
blepharoplasts and a new flagellum grows out from the other 
blepharoplast. The division of » follows hard upon that of 
the blepharoplast, then N starts its division, and finally the 
whole body divides ; of the two daughter-flagellates produced, 
one has the original flagellum, the other has to grow a new 
one, which it may not do, in some cases, until after complete 
separation from its twin sister. In any case, one of the two 
products of division has a much shorter flagellum than the 
other, irrespective of any difference in bulk between the two. 

As already stated, while these changes of form and pro- 
cesses of multiplication are going on the flagellate is also 
undergoing a change which, though a very small thing in 
itself, produces nevertheless the characteristic morphological 
distinction between the trypaniform and crithidial types; 
namely, the approximation and final transposition of the two 
nuclei, so that 2 comes to lie beside, or even well in front of, 
N. The approximation of the two nuclei begins with the 
first alteration and changes of form, but the exact point in 
the development at which the change of position of the two 
nuclei is completed is subject to great variation. Very 
exceptionally, as has been described above, the transposition 
of the nuclei may be complete in the stomach itself, but more 
usually, it may be said normally, the change does not take 
place until the flagellate reaches the rectum. Any of the 
forms distinguished above as a, b, c, d or e, may be found 
with the nuclei, but slightly or completely approximated, and 
finally in the clumps of developmental forms such as that 
represented in Pl. 41, figs. 182-184, evidently consisting of 
forms which have completed several, at least two or three 
generations of fission since arriving in the rectum, all stages 
in the process of transposition of » or N are to be observed. 
Since the small forms of the established rectal phase show 
invariably the typical crithidial arrangement of the nuclei, all 
that can be said is that the change in position of the nuclei 
is effected earlier or later, but without fail, in the transition 
from the stomach-phase to the rectal phase. 


THE RAT-TRYPANOSOME, TRYPANOSOMA LEWISI. 579 


The successive generations of the transitional rectal phase 
cannot be distinguished with precision. ‘The size to which 
the individuals are diminished after a given number of 
generations is determined by two variable factors during 
those generations, namely, the rate of individual growth and 
the frequency with which multiplication takes place. All 
that can be said is that the rectal forms diminish in size by 
repeated division until they reach a minimum size which 
is attained when growth and multiplication balance each other 
more or less evenly; and that in the rectal infections of 
recent origin the average size of the individual flagellates is 
slightly larger, as a rule, than in the old-established infections. 


(b) The Established Rectal Phase. 


In its fully-established and definitive condition the rectal 
phase consists chiefly of small crithidial individuals which 
multiply by binary fission, attached to the wall of the gut 
by their flagella or by the anterior (flagellar) extremity of 
the body. The crithidial stage appears normally, if not 
- invariably, to take origin in the rectum in the manner described 
in the preceding section, and this part of the digestive 
tract is its most usual and characteristic habitat ; consequently 
we have made use of the term “rectal phase” for this part 
of the development, in spite of the fact, presently to be dis- 
cussed in more detail, that the crithidial and other forms of 
this phase may be found not uncommonly in other parts of 
the gut, more especially in the intestine close behind the 
pylorus, but also in the stomach itself. 

In one flea we have found the rectal phase well-established so 
early as eighteen hours after the infective feed (Text-fig. 14), 
and in another flea we found a few typical examples of this 
phase at twenty-four hours, but both these examples are 
abnormally early cases of the fully-developed crithidial phase, 
which, in our experience, is seldom completely established 
before thirty-six or forty-eight hours after the infective 
feed. 


580 E. A. MINCHIN AND J. D. THOMSON: 


The extent to which the crithidial infection is developed in 
different fleas varies greatly, from a swarming infection cover- 
ing the wall of the rectum like a pile-carpet, especially in the 
region behind the rectal papilla, to a condition in which a 
few scanty flagellates are to be found only by careful search- 
ing; in either case, however, the crithidial stage represents 


Text-Fic. 14. 


@) 
saa 
© 
Mac 
Uo. 


Various forms from the rectum of a flea eighteen hours after the 
infective feed, a, b, c and d are probably degenerative forms, 
the rest developmental; e and f, early rectal forms, g—/, hapto- 
monads (x 2000). 


the permanent stock of the parasites in the flea, multiplying 
continually and indefinitely and maintaining itself, in all 
probability, as long as the flea lives. We have succeeded in 
keeping a single flea alive for nearly three months, and during 
that time seven rats were infected by it (see p. 640 below). 
This permanent infectivity can only be explained by the 
establishment of the crithidial stock in the rectum and its 
continued multiplication. Each crithidial individual may do 
one of two things : it may multiply by binary fission to produce 
two crithidial forms (multiplicative phase); or it maybe 


THE RAT-TRYPANOSOME, TRYPANOSOMA LEWISI. 581 


transformed gradually into small stumpy trypaniform indi- 
viduals (final propagative phase). Hence the flagellates ot 
the rectal stock may be classified, broadly-speaking, into 
crithidial, transitional and trypaniform individuals. 

The forms that occur in a well-established rectal infection 
are very varied, and it is a somewhat difficult task to classity 
them and to assign to each and every form its due place in 
the developmental series. ‘l'o obtain a general notion of the 
various types met with, it is best to examine fleas some eight 
or nine days after they have fed once on an infected rat and 
subsequently on clean rats, so as to obtain the rectal phase in 
its typical condition, free from admixture of earlier develop- 
mental or degenerative forms. Such fleas may have no 
flagellates at all in their rectum or may present every grada- 
tion between an extremely scanty and a swarming infection. 

If a rectum containing numerous flagellates be examined 
fresh after having been dissected out in salt-citrate solution, 
opened up by tearing it with fine needles, and then covered 
with a coverslip, the majority of the flagellates are seen 
attached to the wall of the rectum, chiefly in the region 
behind the rectal papilla. Many of them will have become 
detached as the result of the dissection, and will be seen 
floating about singly or adhering together in larger or smaller 
clumps. If the preparation be sealed up carefully and 
watched for some time many of those that lie about singly 
will attach themselves to the glass of the slide or coverslip, 
and it can be seen very clearly in all cases, whether in the 
forms still attached to the rectal wall or in those set free in 
clumps or singly, that the attachment is by the tip of the 
flagellum or by the flagellar extremity of the body in those 
that have no free flagellum. The flagellates may be attached 
to the rectal wall in such numbers and so closely crowded 
together that they resemble a furry lining or pile on it. Seen 
in profile they appear in serried ranks, each in contact with 
its neighbours; seen in surface view they present the appear- 
ance of a honeycomb, each flagellate in optical transverse 
section, showing outlines nearly polygonal as the result of 
mutual pressure. In the living condition the attached 


082 EK. A. MINCHIN AND J. D. THOMSON. 


crithidias are motionless for the most part, but occasionally a 
given individual performs a kind of jerky nutating move- 
ment ; the body, remaining attached, sways rapidly from one 
side to the other, with a slight curvature of the axis. After 
bending, first to one side and then to the other, in this way 
the animal remains quiescent for a time, but when there are 
a large number of the flagellates, there is scarcely a moment 
in which one or another, or several at once here and there, 
are not performing these movements. 

In addition to the attached forms, with the flagellum for 
the most part very short or wanting altogether, there are 
usually a certain number of free forms. Some of these are 
crithidial in form, with the undulating membrane feebly 
developed and scarcely, if at all, recognisable, and with a 
distinct free flagellum, often quite long. The function of 
these flagellated forms appears to be that of migrating in 
order to colonize other parts of the wall of the gut. When 
a clump of attached forms has been multiplying actively 
at one spot it doubtless tends to become overcrowded, at 
least in its central part. Then probably one of two things 
may happen; some of the crithidias may become transformed 
into the final trypanosome-form and detached from the clump; 
or a certain number will develop flagella, remaining 
crithidial in type, migrate to another part, and attach them- 
selves again.! 

Besides the crithidial forms with long flagellum there can 
be seen free forms, also with a flagellum of variable length, 
but with a distinct undulating membrane running along the 
whole or greater part of the length of the body, which 
appears more flexible and performs sinuous undulatory 
movements more or less distinctly. These forms are the 
ttle trypanosomes, or transitional stages in their develop- 
ment, the forms which constitute the final infective stage of 


‘One of us has seen in the living rectum a crithidia with a long 
flagellum become detached from the rectal wall, swim actively across the 
rectal cavity, with its flagellum directed forwards, and attach itself 
again to the wall on the side opposite to its former attachment. 


THE RAT-TRYPANOSOME, TRYPANOSOMA LEWISI. 583 


the cycle in the flea (so-called ‘‘ metacyclical trypanosomes ”’ 
of Brumpt). 

We can thus distinguish in the rectum of the flea three 
principal types of individuals, each of which varies con- 
siderably in size and details of form, and between which are 
to be found every possible transition, namely : 

(1) The attached or haptomonad! form, which is_ the 
multiplicative stage of the rectal development. 

(2) The free or nectomonad! form. 

(3) The final trypanosome-form. 

The many variations of these three types and the transi- 
tions between them cause the rectal phase of the develop- 
ment to present a variety of form which is at first very 
bewildering, but which becomes easily intelligible if the 
classification into the three types given above be used as 
the key. It may be noted here that it is very common, 
in the development of other trypanosomes, for the crithidial 
phase to exhibit great variation in size, form, and structure. 
Compare the works of Chagas on the development of 
T. cruzi in the bug, and of Miss Robertson (1911, etc.) on the 
development of the trypanosomes of fishes in leeches. 

We will now proceed to describe the variations of these 
types and the transitions between them in more detail. 

(1) ‘he haptomonad or attached type is the phase in which 
multiplication by binary fission takes place ; consequently its 
variations of form are related mainly to the function of 
multiplication. Leaving out the forms which are actually in 
process of division, we find that chief variations in form are 
seen: first, with regard to the shape of the body, whether 
relatively slender, with the hinder end sharply or bluntly 


' Woodcock (1914) has proposed the useful term “ haptomonad” for 
the attached phase of a Crithidia or Leptomonas, the so-called 
“ gregariniform”’ individuals of Léger, but of which the resemblance 
to a gregarine is not very striking, not at least in the developmental 
phases of Trypanosoma lewisi. To denote the locomotor crithidias 
with long flagella we propose the term ‘‘nectomonad,” i. e. swimming 
monad, as a correlative to haptomonad, fixed monad. 


584. E. A. MINCHIN AND J. D. THOMSON. 


pointed, or stouter, with rounded hinder end, or finally ovoid 
or even globular in form ; secondly, with regard to the extent 
to which the flagellum is developed. 

The typical haptomonad, when not preparing for division, 
has the body spindle-shaped or pear-shaped with the thickest 
part anteriorly in front of the nuclei, and the hinder end 
more or less sharply pointed (PI. 41, figs. 174, 176,193; Pl. 42, 
fig. 252). The two nuclei are usually close together, situated 
either about the middle of the body or nearer to the hinder 
end; 2 is either just in front of N or close beside it. The 
cytoplasm has a great tendency to stain very dark by any 
method, especially towards the hinder end. After Giemsa 
the body has a purplish-blue tinge ; after Twort’s combination 
of neutral red and Lichtgriin it is seen to be full of very fine 
granules, stained red and scattered irregularly (Pl. 38, figs. 
260a-263a) ; from these reactions it is evident that the 
opacity of the body is due to the deposition in the cytoplasm 
of very fine “ chromatoid ” grains, probably of the nature of 
volutin. The cytoplasm is usually free from coarse granula- 
tions, which are, however, present occasionally. 

In preparation for division the hinder end of the body 
begins to swell and to become rounded at the hinder end, 
while the nuclei are shifted more posteriorly. In consequence 
the body becomes pear-shaped, but in the opposite manner to 
that previously described, since now the thickest part of the 
pear is the posterior end, while the anterior extremity of the 
body is narrowed, with the flagellum representing, as it were, 
the stalk of the pear (PI. 41, figs. 168, 188; Pl. 42, figs. 231, 
247-249). In other cases the body becomes evenly ovoid or 
even globular in form (PI. 42, figs. 243-247). 

The process of division calls for no special remark (see 
Pl. 42, figs. 218, 214, 226-2380, 249, 253, 266). The 
blepharoplast or basal granule of the flagellum divides first, 
one of the daughter-blepharoplasts retaining the old flagellum 
attached to it, while the root of the new flagellum begins to 
grow out from the other daughter-blepharoplast (PI. 42, fig. 


: 
o 
5 


252, lowest specimen). Following the blepharoplast, n divides 


THE RAT-TRYPANOSOME, TRYPANOSOMA LEWISI. 585 


next and after that N. Next the body is constricted into 
two, beginning from the anterior or flagellar end (Pl. 42, 
fig. 253). We have never seen any but binary fission of the 
crithidial forms ; multiple fission does not occur in any form 
at this stage. We have also sought without success for 
multiplication by endogenous budding (the so-called “infec- 
tive granules” of Balfour and others). In Leptomonas 
pattoni of the flea we have found very clear instances of 
apparent endogenous budding (PI. 42, figs. 281, 282-284), and 
Pl. 42, figs. 268, 269 are rather suggestive of a similar process 
occurring in the haptomonads of T’. lewisi, but further proof 
of it is lacking. 

The flagellum in haptomonad forms is extremely short, and 
as a rule is reduced practically to its intra-cellular root or 
rhizoplast, the projection beyond the limits of the cell-body 
being very slight or quite imperceptible. At the point where 
the rhizoplast comes to the surface at the anterior apex of the 
body there is usually a fairly large and distinct, but often ill- 
defined patch of substance which stains like the flagellum, 
that is to say, red, after Giemsa’s stain, black or greyish- 
black after iron-hematoxylin, and green with 'l'wort’s stain, 
and which appears to represent a secretion produced by the 
flagellate, a sort of cement by which the animal adheres to 
the wall of the rectum (PI. 41, figs. 174, 178; Pl. 42, figs. 223, 
226, 243-245, 260, 262, etc.). 

In many cases the rhizoplast fails to reach the surface of 
the body, or may even, very exceptionally, be absent altogether 
(Pl. 42, figs. 241-243, 267). ‘he body is then always ovoid or 
globular in form. In such cases we have true leishmanial 
forms, which appear to owe their origin to very rapid multipli- 
cation of the ordinary haptomonad type; multiplication is so 
rapid that one of the two daughter-individuals resulting from 
binary fission has no time to form completely its new flagel- 
lum or even the rhizoplast, before being split off from its 
twin sister and beginning to divide again. This view receives 
support from the extremely small size which is commonly a 
feature of these leishmanial forms. In no case have we seen 


586 E. A. MINCHIN AND J. D. THOMSON. 


anything that could be interpreted as encystation or encapsu- 
lation of the leishmanial forms; they appear to represent a 
purely trophic and multiplicative phase. 

The haptomonad forms occur usually, as has been said, 
attached to the cuticle lining the rectum; in surface views of 
the rectal wall, living or preserved, they may be seen attached 
singly, in clumps, or in a continuous carpet-like layer, and 
exactly the same is found in microtome-sections of preserved 
recta (Pl. 44, fig. 317; Pl. 42, fig. 277). In sections of a well- 
infected rectum the haptomonad forms are seen in a long, 
continuous line, like soldiers on parade. But, both in living 
and preserved recta, in film preparations or in sections, free 
clumps of haptomonads are also found, in which the indi- 
viduals all have their flagella directed pomand a certain point 
(Pl. 42, fig. 274). In spite of careful scrutiny it is not possible 
to detect any body or particle of débris at the centre to 
which the monads are attached ; they appear to adhere simply 
to one another by their flagellar extremities. The question 
at once arises whether these free clumps are a natural or an 
artificial condition. If they were only seen in teased-up recta 
one could have hesitation in ascribing their detachment to the 
manipulation, but they are found also in sections of recta. 
It is, of course, impossible to dissect out and preserve a rectum 
without subjecting it to great stresses and strains which might 
detach the monads, but it is remarkable how tenaciously they 
adhere to the wall. In one of our series of sections of a 
rectum it can be seen that it has been badly torn in getting 
it out ; part of the torn wall has curled right back and turned 
inside out. ‘The tear goes right through an attached carpet 
of crithidias which have nevertheless remained adherent to 
the wall, even on the part that has curled back, giving at a 
first glance the erroneous impression that the monads are 
attached to the exterior of the rectum. They must, therefore, 
be attached very firmly to the wall, which is intelligible when 
it is recognised that the crithidial forms are not the ripe, 
propagative stages of the cycle and that if they were carried 
to the exterior with the faces they would be lost. For this 


THE RAT-TRYPANOSOME, TRYPANOSOMA LEWISI. 587 


reason alone it seems highly probable that the free clumps of 
crithidial forms represent either clumps artificially detached 
by manipulation, or an abnormal condition of the flagellates 
detrimental to their future welfare. 

As transitions between the nectomonad and haptomonad 
phase we would expect to find both stages of the development 
of the nectomonad into the haptomonad, and stages of the 
development in the reverse direction. It is of course almost 
impossible to say, by simple inspection of a transitional form 
in a permanent preparation, in which direction it is develop- 
ing. We are inclined, however, to interpret as transitions 
from the nectomonad to the haptomonad the more slender 
forms, with short flagella and hinder ends pointed or but 
slightly blunted, such as Pl. 41, figs. 175,176; Pl. 42, figs. 
220, 252; and as transitions in the opposite directions the 
rounded or broad pear-shaped forms with flagella of various 
lengths, such as PI. 41, fig. 166, Pl. 42, figs. 231-234, 251, 265. 
In some forms of the latter type the distal ends of the flagella 
are very thin, much thinner and more delicate than the proxi- 
mal portions (PI. 42, figs. 246, 250, 268, 269), and we interpret 
this appearance as indicating that the flagellum is in process 
of growth rather than of regression in length, for the reason 
that a similar condition ‘is seen in the flagella of forms 
transitional to the final trypanosome-type (PI. 42, figs. 255- 
257), in which the flagellum is not likely to be in process of 
shortening. 

In the series which we interpret as transitional from the 
haptomonad to the nectomonad type we find globular forms 
with flagella of considerable length (Pl. 41, fig. 166; Pl. 42, 
fig. 234), and, since we have not observed such forms 
swimming freely in the rectum, we conclude that the hapto- 
monad, while attached, first develops its flagellum to a 
considerable length, and then acquires the elongated form 
of body, before becoming detached fromthe wall and set free. 

‘ Comparable, for example, to the sponge-larve, which attach them- 


selves to the surface-film of the water instead of becoming fixed to a 
firm object, and which in consequence perish inevitably—E. A. M, 


588 E. A. MINCHIN AND J. D. THOMSON. 


(2) The nectomonad or free type of crithidial flagellate has 
a more slender body, in its fully-developed form about five 
times or more as long as its greatest breadth, but with great 
variations in its relative proportions (Pl. 41, figs. 190, 194-197 ; 
Pl. 42, figs. 217, 235, 236, 254). The body is usually spindle- 
shaped, pointed at both ends, its thickest part at the level of 
or slightly behind, the middle point of its length, and 1m is 
usually well in front of N. When it swims the flagellum, 
directed forwards, is thrown into even sinuous undulations 
which begin at the tip and run backwards, in contrast to the 
type of movement so often seen in free-living flagellates, in 
which the proximal two-thirds, or so, of the flagellum is held 
stiff and straight, while the distal third performs lashing 
movements which drag the body forward. We have not 
found the fully-developed nectomonad type undergoing 
multiplication by fission, unless Pl. 42, figs. 253 and 266 are 
to be so interpreted. 

(3) The final trypanosome-form appears to rise in most 
cases from the haptomonad type, with which it is usually 
found closely associated in preparations; compare Pl. 41, 
fig. 202, of a section through the intestine close behind the 
pylorus ; the trypanosomes are seen with their posterior ends 
projecting above the level of the serried ranks of the hapto- 
monad crithidias, as if they were pushed upwards by the 
development and growth of their flagella. It is possible, 
however, that the final forms may sometimes arise from the 
nectomonad type, and that such an origin explains the 
occurrence of the slender forms of the trypanosomes, the 
stout forms being derived from the haptomonads. Forms 
such as Pl. 42, fig. 287, are perhaps to be interpreted 
as transitional from the nectomonad type to the final 
trypanosome-form. 

The essential feature in the origin of the final form from the 
erithidial form, of whatever type, is the transposition of the 
two nuclei, x, and N. Both nuclei move backwards usually, — 
but N only for a short distance, while » passes N and goes to 
the posterior extremity of the body (Pl. 42, figs, 238, 239, 


THE RAT-TRYPANOSOME, TRYPANOSOMA LEWISI. 9589 


255-259, 270). In some cases, especially in the slender 
forms, 2 stops short of the extreme posterior end of the 
body (fig. 259), but in the stumpy forms » becomes quite 
terminal in position, as a rule (Pl. 41, figs. 199, 200; Pl. 42, 
figs, 259, 271). Further characteristic of the final stage is 
the relatively large size of both n and N, and the faint stain 
that N usually takes in the permanent preparations. In 
many cases N appears distinctly elongated in the longitudinal 
direction (figs. 199, 259). With the displacement backwards 
of w and of the attachment and origin of the flagellum, the 
undulating membrane becomes correspondingly extended and 
lengthened. 

The occurrence of stout and slender forms of the final 
trypanosomes has been mentioned already, and was pointed 
out by Swellengrebel and Strickland (1910) ; but it is a fact 
somewhat difficult to explain. It may be, as already sug- 
gested, that it is simply due to difference of origin, the 
slender forms arising from the nectomonads, the stout forms 
from the haptomonads. On the other hand, it may be that 
the trypanosomes, when ingested by the rat, become exceed- 
ingly active in order to find their way from the digestive 
tract into the blood, and that the slender forms in the flea 
represent merely the precocious assumption of a type of 
structure which belongs strictly to a later period of the life- 
cycle. These are the only suggestions we can offer at present 
in explanation of the two forms. 

We have never in any case seen the final trypanosome- 
form dividing, but it is stated to do so by Swellengrebel and 
Strickland (1910), who, after having examined one batch of 
thirty-seven infected fleas, have been able to figure no less 
than three examples of a process of division that we have 
never been able to find in all the many hundreds of infected 
fleas we have dissected and examined. For our part we 
agree with Brumpt (1915), that these “ metacyclical trypano- 
somes,” as he proposes to call them, are “ phases d’attente ” 
which do not multiply further in the flea. 

- With the development of the final trypanosome-form the 


590 E. A. MINCHIN AND J. D. THOMSON. 


cycle of T. lewisi inthe fleais ended. It only remains to say 
a few words with regard to the occurrence of the rectal phase 
in regions of the gut situated further forwards than the 
rectum. It is by no means an infrequent occurrence to find 
clumps and carpets of various forms characteristic of the 
rectal phase attached in the intestine and even in the stomach.! 
In the intestine they occur most frequently at the upper end, 
close behind the pylorus. When they occur in the stomach 
they are probably always attached towards its hinder end, 
near the pylorus. Hence the two chief situations of the 
crithidial forms, when occurring outside the rectum, may be 
designated briefly ‘‘ pre-pyloric ” and ‘ post-pyloric.” 

Two possibilities present themselves at once to the mind 
with reference to these extra-rectal crithidial infections ; first, 
that the infection of the stomach or intestine is a direct one, 
brought about by forms which have attached themselves 
there immediately after completing their stomach-phase, 
without having ever travelled further back in the digestive 
tract; secondly, that the infection has been brought about in 
an indirect manner by forms which have migrated forwards 
from the rectum. 

So far as post-pyloric intestinal infections are concerned, 
we have some evidence that the infection may be sometimes a 
direct one; in one of our series of sections of the stomach of 
a flea that had fed thirty-six hours previously to being pre- 
served, there are two large clumps of crithidial forms attached 
close behind the pylorus. Probably in such cases those 
attached in the intestine represent but a small numerical 
proportion of those that migrated backwards from the 
stomach, the majority having passed on down to the rectum, 
while a few have stuck, as it were, higher up. 

With regard, however,to the pre-pyloric crithidial infections, 
we have no evidence of direct infection taking place, but all 
our data indicate that such infections of the stomach are 


1 Since a certain length of intestine was usually cut off with the 
stomach it is possible that many of the crithidial forms found by us in 
our stomach-films were really post-pyloric in situation. 


THE RAT-TRYPANOSOME, TRYPANOSOMA LEWISI. 591 


brought about indirectly, and the same is probably true, in 
most cases, of the post-pyloric infections of the intestine. In 
the first place we have no record of the occurrence of crithi- 
dial infections in the stomach (pre-pyloric) earlier than seven 
days after the first infective feed of the fleas; but at later 
periods than this we have so many records of such infections 
in the stomach that, had they been in all cases brought about 
directly, we should have expected to have found crithidial 
forms in the stomach during the period when such forms are 
being established, that is to say, from about thirty-six hours 
and five days or so, which we have never done. Secondly, 
the evidence furnished by experiment 39 (see below, p. 634), 
indicates very strongly that the final infective forms of the 
life-cycle were first produced in the rectum on the fifth day 
and were there also on the seventh day, but had migrated 
forwards to the stomach on the tenth day. 

It seems, therefore, most probable that in the majority of 
cases at least, the pre-pyloric and even the post-pyloric 
infections are the secondary results of a migration forward 
from the rectum of crithidial forms previously established 
there ; and since neither the haptomonads nor the final 
trypanosome-forms appear capable of undertaking such 
migrations, it must be the nectomonads, which are obviously 
active locomotor forms, that are responsible for such migra- 
tion. We have performed some experiments from which it is 
clear that the migration forwards is dependent on conditions 
of nutrition in the flea and that starvation favours a forward 
migration of the nectomonads towards the stomach. 

In many cases the forwards migration of the flagellates 
leads to the rectum being quite deserted by them. ‘This is 
well shown by the following instance, by no means an isolated 
one of its kind in our experience, but very typical. A flea 
was taken from the infected breeding-cage and put by itself 
on a clean rat for three days, from the 19th to the 22nd of 
September; 1t was then recovered and dissected. The 
stomach-preparations were found to contain a considerable 
infection of the typical rectal phase (T'ext-fig. 17, p. 627), 

vou. 60, PART 4,—NEW SERIES. Al 


592 E. A. MINCHIN AND J. D. THOMSON. 


but no flagellates of any kind were found inthe rectum. The 
rat became infected, and first showed trypanosomes in its 
blood on September 28th. The age of the infection of the 
flea was not known, but the crithidial stock seems in this case 
to have died out in the rectum and to have established itself 
exclusively in the pyloric region. 

We have also, though rarely, seen the attached crithidial 
form in the proximal portions of the Malpighian tubules. 

On the other hand we have never seen in our rat-fleas 
(Ceratophyllus fasciatus) infections such as are 
described by Noller (1912), and Wenyon (1913), in the 
dog-flea, where both the rectum and the intestine are 
described as being carpeted along their whole extent with 
the crithidial phase; though we have seen such infections 
in fleas harbouring the Leptomonas. It is a fact which 
seems at first strange, but is probably very significant, that, 
as we have pointed out elsewhere (p. 610), the rat-flea is not 
so efficient a host for the rat-trypanosome as other species of 
fleas which do not usually or of choice feed upon rats; from 
which circumstance it would appear as if the rat-flea has 
acquired a certain degree of natural immunity to the trypano- 
some of the rat which other fleas do not possess. 

For a general summary of the development of Trypano- 
soma lewisi in the flea, see Plate 45, and the description 


of it (p. 691). 


(8) Tue DEGENERATIVE SERIES. 


Trypanosomes undergoing degenerative changes may be 
found in either the stomach or rectum during the first few 
days after the flea has fed for the first time on an infected 
rat. They are most abundant in batches of fleas examined 
during the first twenty-four hours after feeding. After this 
period trypanosomes may have disappeared altogether from 
the gut of the flea, and after thirty-six hours degenerative 
forms are of infrequent occurrence. In some cases, however, 
degenerative forms may be found in the rectum much later 


THE RAT-TRYPANOSOME, TRYPANOSOMA LEWISI. 593 


than the first day, namely, up to three, four, or even five days 
after the infective feed. The degenerative forms of late 
occurrence are probably to be interpreted as individuals 
which have become degenerative after having developed 
in a normal manner for a longer or shorter period. The 
trypanosomes which begin to degenerate immediately after 
being ingested by the flea probably do not last long beyond 
twenty-four or thirty-six hours, usually not so long. Fleas 
that have fed on an infected rat whose blood is swarming 
with trypanosomes often show no trace of the parasites in 
any part of the gut by twenty-four hours. The majority of 
the degenerative trypanosomes that are found in the fleas 
are those that begin to degenerate immediately after being 
taken up from the rat. 

There is no essential difference between the degenerative 
forms found in the stomach and the rectum. We may, there- 
fore, give a general description of the forms of the degenera- 
tive series without taking special note of their provenance. 

In direct contrast to the changes undergone by the develop- 
mental forms in the stomach, the principal sign of degenera- 
tion is a progressive diminution in size, more especially in the 
length of the body. The trypanosome gradually dwindles 
and wastes away, beginning at the flagellar end, during 
which process the flagellum becomes converted progressively 
into a fluffy mass, which frequently shows a tendency to stain 
blue or bluish with the Giemsa stain, instead of the normal 
red (Pl. 48, figs. 294-296, 308). Meanwhile N is pushed 
backwards towards . The displacement of N does not 
appear to be due to any active migration on its part, but to 
be the purely passive consequence of the dwindling of the 
anterior part of the body, whereby it is forced backwards. On 
the other hand n shows no tendency whatever to move for- 
wards, but may do one of two things: it may remain where it 
is, or be shifted backwards only to a slight extent, in which 
case the hinder end of the body retains the sharp point 
characteristic of the trypanosome in the blood (PI. 43, fig. 301); 
or it may pass back towards the extreme posterior end, and 


594 E. A. MINCHIN AND J. D. THOMSON. 


even become terminal in position, in which case the hinder 
end becomes bluntly pointed or even rounded (Figs. 290, 291, 
308). If at the same time the body becomes broadened out 
posteriorly, as sometimes happens, the result is a form which 
may mimic very exactly the small stumpy trypanosome which 
is the final form of the development (Figs. 299, 306, 307). 

The trypanosomes that undergo this process of degenera- 
tion show a great tendency to adhere together in clumps 
attaching themselves to one another by the tips of their 
flagella (Pl. 45, figs. 304, 308; Pl. 44, fig. 311). The adherence 
in this way of the degenerative forms must be distinguished 
clearly from the process of agglomeration which T. lewisi 
undergoes so readily when placed in unfavourable circum- 
stances. Agglomeration takes place by the hinder ends of 
the trypanosomes and more especially by their nn, as Laveran 
and Mesnil have shown, and as a result of it the trypanosomes 
tend to form rosette-like clusters, in which the flagella radiate 
outwards. True agglomeration of this kind can also occur in 
the flea under special circumstances, as will be described 
presently. But in the degenerative clusters the conditions 
are precisely the opposite to agglomeration, since the flagella 
are directed towards the centre of the cluster, while the hinder 
ends of the trypanosomes radiate outwards. The tendency 
of the degenerative forms to adhere in clumps must be inter- 
preted as an expression of the general tendency (perhaps it 
might be termed instinct) of the trypanosome to attach itself 
by the tip of the flagellum to firm surfaces when in the body of 
the flea, a tendency very pronounced in all developmental 
forms, excluding the final stage of the cycle. 

Clumps and masses of very considerable size are formed by 
the degenerative forms adhering together in the manner 
described. Towards the centre the clumps often show a 
cement-like substance, which stains pinkish-red with Giemsa. 
The final stages of the degeneration are small forms, which 

1 Manteuffel (1909) has already drawn attention to the distinction 


between rosettes, with flagella directed inwards, and true agglomera- 
tion. 


THE RAT-TRYPANOSOME, TRYPANOSOMA LEWISI. 595 


represent simply the hinder ends of the original trypano- 
somes. ‘I‘hey are usually sharply or bluntly pointed (PI. 45, 
figs. 302-304), or may be rounded off (fig. 306). The large 
clumps of these little degenerative forms in the rectum are 
often very difficult to distinguish in the living state from 
the clumps of developmental crithidias. The degenerative 
clumps, however, generally occur loose in the cavity, while 
the true crithidias are attached to the wall of the rectum, 
though in the process of dissection the latter often become torn 
away from the wall. Whena loose clump of this kind consists 
entirely of forms with pointed hinder ends it is probably 
degenerative. The true crithidial clumps always have a con- 
siderable number of forms with rounded hinder ends, 
especially in the early periods of the establishment of the 
rectal phase, when the hinder ends of the crithidial forms 
are almost alwaysrounded. The degenerative forms, carefully 
examined, show a certain extent of undulating membrane 
running down the side of the body to n, which is situated 
behind N, while in the typical haptomonad phase the 
flagellum is reduced to the rhizoplast which comes off close 
to » and terminates at the surface of the body, n being 
situated beside orin front of N. Finally, it should be noted 
that the degenerative forms never multiply by division at any 
time. Nevertheless, in spite of all these distinctions, which 
are more easily perceived in permanent preparations than in 
the living state, it is sometimes difficult to pronounce decisively 
as to the nature of a given individual, whether degenerative or 
developmental, in preparations of the rectum; but as a rule 
there is no difficulty at all. 

A remarkable fact is the occurrence of recurved forms 
amongst the degenerative forms in the rectum, of a type 
essentially similar to the recurved trypanosomes occurring in 
the normal developmental series in the stomach (Pl. 43, 
figs. 297, 298). The recurved forms are often seen in the 
clumps of degenerative trypanosomes. ‘The occurrence of 
such forms in the rectum may perhaps be interpreted as an 
abortive effort on the part of the trypanosomes that have 


596 E. A. MINCHIN AND J. D. THOMSON. 


passed on prematurely into the rectum to go through a 
development similar to that which they undergo normally in 
the stomach, but which, in all probability, would be impossible 
in the rectum, where the cuticular lining would doubtless be 
an effective bar to the penetration of the epithelial cells by 
the trypanosome. It is possible that some of the recurved 
forms in the stomach may also degenerate without ever 
succeeding in penetrating the cells; the curious forms such as 
PI. 43, fig. 305, are very probably to be explained as recurved 
forms in process of degeneration. It would be difficult, how- 
ever, as a rule, to distinguish between degenerative and 
developmental trypanosomes in the recurved condition in the 
stomach ; but in the rectum all such recurved forms must be 
regarded as abortive and destined to degeneration. 

As has also been mentioned above, trypanosomes of de- 
generative type are found in the rectum on the third and 
fourth days after infection. Such forms may, in some cases, 
differ but little from ordinary blood-trypanosomes, and are 
then to be interpreted, probably, as trypanosomes ingested at 
a later feed, which have passed on to the rectum; but in 
other cases they may be forms which are undergoing de- 
generation after having developed normally in the stomach. 
They are found, not infrequently, mixed with true develop- 
mental forms in clumps, into which they have probably 
intruded themselves (Pl. 41, fig. 183). It is necessary to be 
careful not to confuse them with early forms of the rectal 
phase in which n is still behind N; such forms can be 
distinguished by their greater stoutness and bulk, and by the 
fact that 1 has generally migrated forwards to some extent 
(Pl. 41, figs. 181-187). 

True agglomeration very rarely occurs in the flea, but we 
have found it in its most typical form (Pl. 43, figs. 309, 310) 
in fleas of a batch, the record of which was as follows: The 
fleas, twelve in number, had been fed once on an infected rat 
in the usual way, and three days later they were fed again on 
a rat, the object being to test the influence of a second feed 
of clean blood on the persistence of the stomach-phase 


THE RAT-TRYPANOSOME, TRYPANOSOMA LEWISI. 597 


(see p. 664 below). By mistake, however, the fleas were fed 
again on an infected rat instead of a clean rat. The next 
day (four days after the first infective feed) the fleas were 
dissected and examined. In every flea the trypanosomes of 
the second feed could be recognised, quite unaltered from the 
blood-form, and in most cases agglomerating in pairs, threes, 
or rosettes composed of many individuals; they were found 
in the stomach in every flea and in a few in the rectum also, 
where in one case degenerative forms were noted; in some 
fleas these trypanosomes were very numerous ; in others they 
were scanty and had evidently undergone reduction in 
number. ‘lhe trypanosomes of the first feed had disappeared 
in nine out of the twelve fleas, while in the remaining three 
they were represented by developmental forms of the usual 
type in the rectum. Agglomerating trypanosomes of the 
second feed were found both in fleas in which those of the 
first feed had persisted and in fleas in which they had 
disappeared. 

From this observation it would appear that when a flea 
has once had an infective feed, its digestive tract, and more 
especially its stomach, acquires properties which cause the 
agelomeration and probably also the degeneration of try- 
panosomes taken in at later feeds, alike whether those 
ingested at the first feed have succeeded in establishing 
themselves in the, flea or not. 

Reference has already been made above (pp. 531-532) to 
intracellular forms which appear to be undergoing degenera- 
tion after having penetrated into an epithelial cell (Pl. 36, 
figs. 45-45). 


APPENDICES TO THE DEVELOPMENT. 


(1) Previous Investigations on the Development of 
Trypanosoma lewisi. 


The first who attempted to follow out the development of T. lewisi 
in its invertebrate host was Prowazek (1905), who studied the develop- 
ment in the rat-louse, Hematopinus spinulosus, and since those 


598 E. A. MINCHIN AND J. D. THOMSON. 


who followed immediately after him in similar investigations also made 
use of the louse, it is simplest to deal first with all those works in which 
the development in the louse is studied. Since we have not ourselves 
studied the development of this insect, we are not in a position to con- 
trovert the statements made, but it is legitimate for us to compare the 
forms and stages described with those which we have found in the flea, 
and, on the ground of such comparisons, to criticise the interpretations 
given by the authors. 

According to Prowazek, the general course of the development in the 
louse is as follows: The flagellates are first to be found in the stomach, 
where they do not collect at particular spots, but swim freely everywhere 
in the ingested blood. In the stomach the processes of maturation and 
fertilization take place. At the second feed of the louse the parasites 
are forced down to the end of the mid-gut and finally come to rest in 
the hind-gut, for the most part near the Malpighian tubules, but also in 
other parts. Resting stages are to be found on or between the cells of 
the mid-gut, and more especially at the beginning of the hind-gut. 
From the fact that the parasites disappear from the hind-gut, it is 
inferred that they can pass through the epithelium of the hind-gut. 
[This conclusion is by no means warranted by the observation on which 
it is founded; it is also, in our opinion, extremely improbable that the 
flagellates could penetrate through the chitinous cuticle lining the hind- 
gut.| In this way the parasites are stated to pass into the blood-stream, 
then into the larynx [ sic], and so finally back into the vertebrate host 
when the louse feeds. [It is not clear whether this statement is founded 
on observation, or simply on the analogy of the statements made by 
Schaudinn with regard to Trypanosoma noctuex; the author 
never succeeded in obtaining an infection of the rat by means of the 
louse.| No trypanosomes or their resting stages were found in freshly- 
deposited feces. 

The author describes in great detail, with figures, various appearances 
interpreted by himas maturation, fertilisation, and even parthenogenesis. 
As all this part of the work is in the highest degree unconvincing and 
appears to consist of forced theoretical interpretations of degenerating 
forms which were occasionally seen to undergo agglomeration, it is not 
necessary to do more than refer to the figures given by Prowazek. Pl. II, 
fig, 32, purports to show the “ fertilisation ” in the living, and PI. III, 
figs. 38 and 39, in the stained condition, while fig. 40 represents the 
“ ookinete,” a non-flagellated form with a single nucleus. From the 
ookinete a crithidial form is stated to arise in the manner described by 
Schaudinn (PI. ITT, figs. 41-43 and 45). An active multiplication follows 
[but Pl. III, fig. 55, which is given as an example of the division, is simply 
an unaltered blood-trypanosome which shows the commonly-occurring 
abnormality of possessing two NN; compare our Text-fig. 15}. 


THE RAT-TRYPANOSOME, TRYPANOSOMA LEWISI. o99 


The author appears to regard most of the crithidias as involution- 
forms, which form “agglomeration-stars.” Such a star is shown in PI. IIT 
fig. 54 [which represents a typical clump of normal developmental 
crithidias, similar to our PI. 41, figs. 182-184]. Inaddition to the crithidial 
involution-forms there are found other much smaller forms, wedged in 
between the cells and with the flagella completely absorbed. The 
crithidial involution-forms may also degenerate into small non-flagellated 
forms (fig. 50). No special inoculative or final form of the development 
is described. 

Baldrey (1909) professes to have confirmed the development described 
by Prowazek, including even the process of maturation and fertilisation. 


Taxr-ries 15, 


Trypanosome with two VN from the stomach of a flea eighteen 
hours after the infective feed. Such forms with two and even 
three NN are quite common in the blood of the rat and in the 
gut of the flea at early stages of the development; compare 
Minchin (1909), p. 803, Pl. 21, fig. 6, Pl. 22, fig. 74, and Pl, 23, 
fig. 84; they haye nothing to do with reproduction of the try- 
panosome by fission. (x 2000.) 


He gives two text-figures, one showing “male” and “female’’ forms 
and * copulation,” the other showing “ ookinetes ”’ and crithidial forms 
of the typical nectomonad type. He states that the ookinete recon- 
structs a flagellar apparatus and divides rapidly to produce crithidial 
forms, which, by repeated division become smaller and smaller, pass 
into the body-cavity, thence to the suctorial mouth-apparatus and so 
infect the rat. The complete cycle takes from eight to ten days. 
Rodenwaldt (1909) studied the development of T. lewisi in the 
louse in order to meet the criticisms of Patton, and succeeded inci- 
dentally in proving that Patton’s “Crithidia hematopini” is a 
mythical and non-existent species. On the first day of the development 
he found in the louse both unaltered forms of T.1lewisi and long forms, 


600 E. A. MINCHIN AND J.. D. THOMSON. 


which he described as “ Lanzettformen.”’ [The latter, from the figures 
given (PI. 1, figs. 7-12) are clearly the same as our long “ crithidiomor- 
phic” stomach-forms.] He also found trypanosomes alleged to be 
dividing (figs. 5, 6, and 12) [but these, again, are simply forms with two 
or three NN, such as occur frequently in the blood of the rat; see 
above]. On the second and third days he found the same forms, but 
a larger proportion of the Lanzettformen. In one louse, however, he 
found crithidial forms developed on the third day.. On the fourth day 
he found forms with a short flagellum or none at all (as shown in his 
figs. 13, 14), which he compares with the ookinetes of Prowazek and 
Baldrey; they are stated to bend their bodies without changing their 
place. On the fifth day long crithidial forms appear (compare his figs. 
15-21), and on the sixth and following days smaller crithidial forms in 
rosettes, and also, but more rarely, leptomonad forms (figs. 23, 27). 
From the tenth day there were found (1) small non-flagellated forms 
(figs. 31-34) ; (2) stout flagellated forms (figs. 35-39); (3) a few stout, 
non-flagellated forms, “ookinetes”’ (figs. 40-47); and also forms regarded 
as representing copulation of gametes (figs. 48,49). After twenty days 
slender, sporozoite-like forms were found (figs. 52-58) in the gut, never 
in the body-cavity. Rodenwaldt did not succeed in producing infection 
by means of lice. 

Breinl and Hindle (1909) describe the development in the louse mainly 
as follows. The ingested trypanosomes first show characteristic 
changes in the nucleus, of which the karyosome divides and the division 
products move to opposite ends of the nucleus. N and then become 
approximated and a division takes place. Some of the trypanosomes 
show about this stage a reduction of the cytoplasm, producing tadpole- 
like forms with a swollen head and the rest of‘the body reduced to a 
long flagellum (see their figs. 5-9); at this stage the two nuclei take 
on the crithidial (“Herpetomonas-like”) arrangement. The crithidie 
multiply by division and become “agglomerated in great clusters with 
the flagellum always directed inwardly” (fig. 18). [These clusters 
appear to be simply developmental clumps of crithidial forms; as 
pointed out above, clumps with the flagella directed inwards, whether 
of degenerative or developmental forms, are not instances of true 
agglomeration.] Round forms [haptomonads ?] are also found (figs. 
32-36). The alleged conjugation was not confirmed. 

[We cannot help remarking that all the stages of the trypanosome 
figured by Breinl and Hindle, even the crithidial clumps, present an 
extraordinarily sickly and degenerative appearance; we venture to 
think that anyone who compares their figures with ours will agree to 
this statement. | 

Swellengrebel and Strickland (1910), have had the advantage over 
previous authors that they were able to compare the stages in the louse 


THE RAT-TRYPANOSOME, TRYPANOSOMA LEWISI. 601 


with those occurring in the flea; their memoir is illustrated by numerous 
figures, for which, however, they appear not to claim great exactness, 
since they refer to them as “diagrams.” They find that ‘the develop- 
ment in the louse is a very irregular one and is not to be compared with 
that which takes place in the flea,’ The first changes are that n wanders 
in the direction of N, producing finally a crithidial form, slender or 
club-shaped (diagrams xv and xvi). From the crithidial forms arise 
large “ovals,” some of them without flagella and representing the 
“ookinetes” of former authors. Later ovals [haptomonads] and 
crithidiz [nectomonads] are found singly or in clumps (diagrams xvii 
and xviii). Degenerative forms were also seen, but no flagellates were 
found which could be identified with this small, final trypanosome-forms 
of the development in the flea. No conjugation was observed. 

From all these various works, only one positive fact emerges clearly, 
namely, that T. lewisi can develop in the louse into its typical crithi- 
dial phase. with both nectomonads and haptomonads. On the other 
hand, no intracellular multiplication has been observed, nor has it 
been proved as yet that the development can proceed so far as to 
produce the small trypanosomes which end the cycle in the flea. One 
form, apparently degenerative, occurs in the louse which we have not 
found in the flea, namely, a large oval form without a flagellum, the 
zygote or “ ookinete” of Prowazek and others. 

The first published works on the development of T. lewisi in the 
flea were those of Swellengrebel and Strickland (1910). We have 
already noticed above their statements with regard to the development 
of the stomach-phase and stated that we are quite unable to agree with 
the account given by them. On the third day of the development they 
find both long crithidial forms and large “ ovals” [stout crithidias of 
the haptomonad type] in the mid-gut (diagram vy). On the fourth day 
they state that the flagellates had all passed out of the stomach into the 
intestine. On the fifth day they found only round forms [hapto- 
monads| in the rectum (diagram vii). On subsequent days they found 
the various forms of the rectal phase, and on the eighth day they found 
the small trypanosome-forms, the final stage of the development, which 
the authors were the first to discover. In their diagram xvi, they give a 
summary of the development showing the following sequence of forms: 
(1) the normal blood-trypanosomes ; (2) a long crithidial form ; (3) a 
stumpy crithidial form with short flagellum ; (4) a haptomonad form ; 
(5) the same in process of division; (6) a form transitional to—; 
(7) a nectomonad; (8) a form transitional to—; (9) the final trypano- 
some-form. [In view, however, of the great differences seen in the 
development of the trypanosomes in different fleas, especially prior to 
the establishment of the rectal phase, it was somewhat rash to attempt 
to fix the order of events in so few as eighty-three fleas, and it may be 


602 E. A. MINCHIN AND J. D. THOMSON. 


remarked that the entire stomach-phase has practically been omitted 
from the cycle as summarised by the authors. | 

Swingle (1911) gave the following description of the cycle in the flea. 
He states that the trypanosomes remain but a short time in the stomach, 
but migrate to the intestine where important changes take place. The 
first change to be seen is a diminution in size, and at the same time N 
moves towards the posterior end of the body. Occasionally such forms 
degenerate ; in those that do not x moves forwards till it is close beside 
or in front of N, thus producing a true crithidialform. The individuals 
which do not change into the crithidial type curl upon themselves to 
form an oval rounded mass (figs. 15, 16) [apparently representing 
recurved forms]. Development of the crithidial forms may proceed 
along two separate lines which come to the same end; (1) they may 
“agglutinate” by the anterior ends forming rosettes (figs. 20, 21 
[representing typical early crithidial clumps]); or (2) they may form 
solitary cysts (figs. 22-30) [apparently representing typical examples of 
the degenerative series]. Other forms [apparently degenerative] are 
also described; but the haptomonad and other forms of the rectal 
phase are all referred: by the author to the form-series of the lepto- 
monad described by him as Herpetomonas pattoni; aconclusion 
which Swellengrebel and Strickland (1911-12), justly criticise, though 
they go too far in the opposite direction in suggesting that H. pattoni 
is a stage in the development of T. lewisi. 

Noller (1912), studying the development of T. lewisi in the dog-flea 
(Ctenocephalus canis), confirmed our discovery of the intracellular 
multiplication in the stomach and added some further details; he 
observed the penetration of a trypanosome into a cell five hours and 
fifty-five minutes after the ingested blood had been ingested by the flea 
and states that the trypanosomes go through at least two generations, 
probably more, of intracellular multiplication. Whether or not the 
trypanosomes establish a normal infection in the flea depends, in 
NOller’s opinion, upon whether they succeed in fixing themselves in the 
intestine or rectum, or not. As regards the multiplication of the 
attached forms in the end-gut, Noller finds that they can always be 
distinguished from the leptomonads by the possession of a typical 
undulating membrane and by undergoing a process of multiple fission ; 
neither of these statements accord in the least with our experience of 
the development of T. lewisi in Ceratophyllus fasciatus. 

It remains to mention that Swellengrebel and Strickland (1910) made 
some observations on the development of T. lewisi in Ornithodoros 
moubata and Cimex lectularius. In the former they got no 
development of crithidial forms; in the latter they found large 
crithidias but no development in the hinder part of the mid-gut. 


THE RAT-TRYPANOSOME, TRYPANOSOMA LEWISI. 603 


(2) On the Possibility of the Occurrence of Sexual 
Phenomena in T. lewisi. 


It has been seen from the foregoing summary of previous investiga- 
tions on the development of T. lewisi that Prowazek first, and after 
him Baldrey, Rodenwaldt, and Gonder, asserted that the development 
of T. lewisi in the louse begins witha process of fertilisation, of which 
the main features are stated to be as follows: Slender male and stout 
female forms of the trypanosome are differentiated; their nuclei go 
through a process of maturation and reduction, after which a fusion of 
the gametes takes place. The zygoteis described as an ‘ ookinete” of 
elongated, oval form, with no flagellar apparatus and with a single 
nucleus (synkaryon). The nucleus is then stated to divide into two by 
a heteropolar mitosis to produce the two nuclei of a trypanosome 
nand N, and then the locomotor apparatus, flagellum and undulating 
membrane, are formed. The result is a flagellate of crithidial structure, 
which proceeds to multiply actively by binary fission. 

It must be remarked here that Prowazek’s account of the “ ookinete ” 
and its development in T. lewisi was modelled in every essential 
detail on the account given by Schaudinn for Trypanosoma noctue. 
The fertilisation observed by Schaudinn, however, was not that of a 
trypanosome, but of Hemoproteus (Halteridium). It is a process 
of true fertilisation, which was first observed in vitro by Macallum, 
and its occurrence is not open to doubt. Schaudinn differed from all 
previous investigators in asserting that the ookinete (zygote) of 
Hemoproteus became converted into a crithidial flagellate, a state- 
ment which has never been confirmed, and seems never likely to be. 
There can be little doubt at the present time that in linking together 
the development of Hemoproteus noctue and of Trypanosoma 
noctuz into a single life-cycle Schaudinn fell into error. Prowazek, 
on the other hand, derived his “ookinete’’ in T. lewisi from the 
sexual union and fusion of two trypanosomes, so that in its alleged 
origin the ookinete of T. lewisi is of quite different nature from that 
of Hemoproteus noctuze. Prowazek is, therefore, the first 
investigator who claims to have seen sexual conjugation of trypanosomes 
in the invertebrate host. 

_ Later investigators of the development of T. lewisi in the louse 
have not confirmed Prowazek’s statements as regards the sexual phase, 
nor has anything similar been found in the flea. Those who have 
investigated the development of other trypanosomes in their invertebrate 
hosts have also failed altogether to observe sexual phases or sexual 
behaviour, in spite of much careful searching for phenomena to which 
their attention has been strongly directed. As stated above, Prowazek’s 
account of the sexual processes is most unconvincing, and the data he 


604 E. A. MINCHIN AND J. D. THOMSON. 


brings forward are quite inadequate to support the superstructure 
of theoretical interpretation built upon them. In short, the question 
of sexuality in trypanosomes may be summed up in the words of 
Miss Robertson (1912, p. 247): “ There is at present no sound evidence 
of conjugation in any trypanosome life-cycle so far worked out.” 

We have ourselves searched most carefully, but in vain, for sexual 
phases and syngamy in the development of T. lewisi. As stated 
above (p. 519), we found in one flea long crithidial forms adhering in 
couples in a manner very suggestive of true sexual behaviour, and 
believed that we had observed true syngamy. We were never able, 
however. to confirm this observation or carry it any further, and we 
are now convinced that the phenomena observed on that occasion 
were simply processes of agglomeration of abnormal forms of the 
trypanosomes in a malformed flea. When we discovered the stomach- 
phase we thought it very probable that the sexual processes might take 
place in this part of the developmental cycle, and we were inclined to 
interpret as evidence of sexual union some of those stages with two 
nn and two NN, such as PI. 36, figs. 19-23, which are certainly at first 
sight very suggestive of the fusion of two trypanosomes. We have no 
evidence, however, of any subsequent fusion of the nuclei, nor of any 
antecedent processes of nuclear reduction such as should be the pre- 
liminary to the process of syngamy. We are not able to arrange the 
figures of these stages in any series which would suggest a sexual 
process. In short we are not able to interpret these forms as anything 
but early stages of the multiplication of the trypanosome. 

On the other hand, it has been shown convincingly that the cycle in 
the invertebrate host effects a marked change in the properties or 
idiosyncrasies of the trypanosomes that have undergone it. Gonder 
showed that an arsenic-resistant strain of T. lewisi remained arsenic- 
resistant so long as it was transmitted from rat to rat by direct 
inoculation, but lost that property when transmitted by the. louse. 
Miss Robertson (1912) also found that strains of T. gambiense 
became changed in character when transmitted through the tsetse-fly, 
and remarks: “It seems clear that the cycle in the fly asa whole, 
whether conjugation actually occurs or not, has much of the biological 
significance of the process.” 

Those who believe that trypanosomes pass through sexual phases in 
their invertebrate host will be inclined to ascribe the changes in the 
properties of the parasite to the effects of the sexual process. At the 
present time it is not possible either to affirm or to deny, with certainty, 
that sexual processes occur. All that can be said with any approach 
to verisimilitude is that the change appears to be connected in some 
way with the metamorphosis of the trypanosome and its passage through 
a crithidial stage ; but proof is lacking that the crithidial stage follows 


THE RAT-TRYPANOSOME, TRYPANOSOMA LEWISI. 605 


upon, and is the product of, a sexual process. Attention may be drawn 
here to another possibility already indicated above, namely, that the 
crithidial phase may be initiated by a differentiating division into two 
inequipotential products, one of which is destined to be eliminated 
sooner or later from the direct line of the life-cycle. If this supposition 
is correct a possible explanation might be afforded for the renovating 
effects of the invertebrate cycle. In the present state of knowledge: 
however, such an explanation must remain hypothetical, and lacking 
objective foundation. 


PART IIT.—_ EXPERIMENTAL STUDY OF THE PROBLEMS 
OF THE TRANSMISSION AND DEVELOPMENT. 


(1) Lyrropucrion. 


THROUGHOUT our investigation of the relations of Try- 
panosoma lewisi to the flea, we have endeavoured, as far 
as possible, to make experiment and observation go hand in 
hand, employing the one method to check or throw light 
upon the results obtained by the other. In the following 
pages we set forth our results in a number of sections which 
arrange themselves naturally into two groups. One group 
(sections i-xv) embraces problems that deal with the complete 
cycle (including the passage of the propagative forms back 
into the rat) and with the establishment of T. lewisi in the 
flea, raising questions that are of general interest in the study 
of trypanosomiasis. The other group (sections xvi-xix) 
deals with some further problems that are of interest more 
especially in relation to the flea Ceratophyllus fasciatus 
and to ‘I’. le wisi itself under more or less special conditions in 
the flea. Hach section is headed by a proposition which it is the 
object of the experiments cited to establish. If we consider 
that the proposition is proved satisfactorily by our experi- 
ments, it is put in the form of a positive or negative state- 
ment; if, on the other hand, the problem stands in need of 
further proof, the heading of the section is expressed in 
interrogative form. 

The details of each experiment are given when it is cited, 


606 E. A. MINCHIN AND J. D. THOMSON. 


but a few general remarks upon our methods may be made 
conveniently at this point. We kept going two breeding- 
cages of the type used by the Plague Commission (see 
‘Journal of Hygiene,’ vi, Pl. iv). In one cage a clean rat 
was always kept to feed the fleas, in the other an infected 
rat ; these two cages are designated, in the account of our 
experiments, the non-infected and the infected breeding-cage 
respectively. From the former we could always obtain a 
plentiful stock of clean fleas when required, while the latter 
furnished infective fleas. The rats used were almost always 
white rats bred in captivity ; we found them as a rule docile 
and good-tempered so long as they were handled with the 
hands and not with forceps, and the operation of pricking 
their tails to obtain drops of blood, when required, did not 
arouse their resentment in the slightest. They live well in 
captivity, and weré none the worse for being exposed to the 
fleas in the breeding-cages, provided the number of fleas was 
not allowed to become too great. Many of them suffered, 
however, from a troublesome itch, caused by a minute 
Acarine, which is very difficult to get rid of. One of our 
breeding-cages became over-run by rat-mites, rendering it 
necessary to destroy it and start a fresh one. 

For our actual experiments we used in many cases, 
especially for experiments with small numbers of fleas, cages 
of special design in the form of a cylindrical tin-canister with 
the bottom closed in with tin, the top provided with a lid 
with a tightly-fitting rim. The canister was 10 in. high and 
6 in. in diameter. The top of the lid was made of strong 
wire gauze, to prevent the rat jumping out, and over that 
muslin-gauze was pasted to prevent escape of fleas. After 
these cages had been in use for some time, however, they 
tended to become rusty on the inside and then the fleas could 
climb up the tin easily. Consequently, it was found more 
suitable to use inverted bell-jars, each about 154 in. in height 
and 7in. in diameter. The bell-jars were supported each on 
a wooden block, or several together in a wooden crate. The 
open upper end of the bell-jar had a zine wire cover to 


THE RAT-TRYPANOSOME, TRYPANOSOMA LEWISI. 607 


prevent the rat jumping out, but it was not necessary to take 
precautions against the fleas escaping, because they are 
unable either to jump so high or to crawl up the smooth glass 
if kept clean. The bell-jars were cleaned out thoroughly 
once a week. 

The beil-jars were especially suited for experiments with 
single fleas or a small number of fleas. First of all clean 
saw-dust is put in the bell-jar to a depth of about 3 in., then 
the rat and the flea or fleas are put in. Since Cerato- 
phyllus fasciatus is a flea which does not live perma- 
nently on the rat but only goes on to it for food, and lives 
naturally in rat-burrows, the fleas were generally to be found 
without difficulty in the saw-dust, when it was necessary to 
recover them, but sometimes they were on the rat itself. In 
the latter case the rat was held over a deep bowl of 
enamelled iron and the flea disturbed. by blowing on to the 
fur of the rat, which has the effect of soon making the flea 
come to the surface of the fur. It was then captured, as a 
rule, by seizing it gently byfinger and thumb, an operation 
which must be performed rapidly and deftly, otherwise it 
burrows down into the fur and must be dislodged again. 
Sometimes the flea drops off the rat and falls into the 
bowl, where it can be recaptured easily. Our assistant, 
Mr. George Kauffmann, became exceedingly expert at this 
job, and if a flea could not be found by him it was safe to 
assume that it had died or been eaten. In our earlier experi- 
ments we used chloroform for recovering the fleas, but later 
we abandoned this method, often fatal to the rats. 

In some cases it was required to expose a large number of 
fleas—200 or so—to infection on an infected rat for a night 
or a day. For this purpose the bell-jar was also handy, but 
it was often found that the rat ate a great many of the fleas, 
sometimes as many as 100 or more in a single night. To 
prevent this a cylinder of wire gauze was made, of sufficient 
length to fit into the bell-jar in such a way that its two ends 
were closed by the glass wall of the jar, and of such a calibre 
as to allow the rat to walk forwards or backwards along it, 

vot. 60 PART 4,—NEW SERIES. 42 


608 E. A. MINCHIN AND J. D. THOMSON. 


but not wide enough to permit the rat to turn round or use 
its paws freely, and consequently hindering it from catching 
and eating the fleas. 

For the purpose of collecting large numbers of fleas from 
the breeding-cage the following method was found to be the 
simplest: Two glass capsules were used, each provided with 
a well-fitting lid, the one smaller, about 2} in. in diameter 
and 1{in. in height; the other larger, about 6in. in diameter 
and 3 in. in height. First of all, débris from the breeding- 
cage containing fleas in all stages of their development is 
scooped up with the small capsule and the lid at once clapped 
on. ‘Then the small capsule is placed in the large one; the 
lid of the small capsule is removed with one hand, and the lid 
of the large one put on with the other. The adult fleas in 
the small capsule then begin at once to jump out of it in 
every direction, and so fall into the enclosing large capsule, 
in which they soon collect on the side furthest from the light. 
When the fleas have swarmed out in this way the small 
capsule is removed and the débris contained in itis returned 
to the breeding-cage. The fleas in the large capsule can then 
be emptied through a glass funnel into a suitable receptacle, 
such as an Erlenmeyer flask. Or, if the large capsule be left 
to stand until all the fleas have congregated on the side furthest 
from the light, then by suddenly turning the capsule round 
through about 180°, so that the side which was furthest from 
the light is now the most illuminated, the fleas begin at once 
to move towards the opposite side ; and if then the small cap- 
sule be held in their way they can be made to jump into it of 
their own accord, and they can thus very easily be counted 
and disposed of as required. Ceratophyllus fasciatus is 
not a very good jumper and its trajectory is low. 

The fleas collected can be kept, if required, for a considerable 
time ; we found the best method was to put a little clean white 
sand, moistened with two or three drops of water, at the 
bottom of a flask. The fleas burrow down into the sand and 
appear to live comfortably. If they are to be kept any 
length of time the sand must be moistened again every 


THE RAT-TRYPANOSOME, TRYPANOSOMA LEWISI. 609 


two or three days. Like most blood-suckers, the flea can 
stand a prolonged fast. 


(2) GunrraL PRrosiems. 


(i) Trypanosoma lewisi is transmitted from Rat to 
Rat by the Rat-flea, Ceratophyllus fasciatus. 


It is not necessary that we should cite experiments 
specially to prove this proposition, since it is established by 
the experiments brought forward under the headings that 
follow, and it has been proved beyond all possibility of reason- 
able doubt by experiments already published by others as 
well as by ourselves. 


The agency of fleas in the transmission of T. le wisi was first demon- 
strated by Rabinowitsch and Kempner (1899), who succeeded in infect- 
ing clean rats by intra-peritoneal injection of teased-up fleas (species 
not stated) which had previously been fed on infected rats. In these 
experiments the trypanosomes appeared in the blood of the rats in six 
to eight days after the injection. The authors state that they were not 
able to find any stages of the trypanosome in the flea-débris which 
was injected. They also obtained positive results by placing fleas, 
previously fed on infected rats, upon clean rats; the trypanosomes made 
their appearance in the blood of the rats after two to three weeks. 
Their experiments with lice gave negative results. 

In spite of the experiments of Rabinowitsch and Kempner, the work 
of Prowazek (1905) on the development of T. lewisi in the rat-louse, 
Hxematopinus spinulosus, led to this insect being regarded as the 
true host of the rat-trypanosome, and no more experiments with fleas 
appear to have been undertaken until those of Nuttall (1908), who 
obtained positive infections of rats with fleas, using both Cerato- 
phyllus fasciatus and Ctenophthalmus agyrtes. Two years 
later we published accounts of a number of experiments, since when the 
rOle of the flea has been established beyond the necessity of further 
experiment upon the subject. 

We have confined our experiments throughout solely to the common 
English rat-flea, Ceratophyllus fasciatus, but it has been shown 
by Noller (1912) and Wenyon (1913) that the transmission can be 
effected by other species of fleas, namely, the dog-flea, Ctenoce- 
phalus canis; the human flea, Pulex irritans ; and the Indian 
rat-flea, Xenopsylla cheopis. It is indeed noteworthy that other 
species of fleas appear to be more efficient as true hosts of the rat- 
trypanosome than the species which in this country occurs habitually 


610 E. A. MINCHIN AND J. D. ''HOMSON. 


in association with rats, since Dr. Wenyon has informed us that in the 
fleas with which he experimented, the trypanosomes never failed to 
establish themselves and to go through their complete developmental 
cycle, while in Ceratophyllus fasciatus we found that only a small 
percentage of the fleas became infective (see below), and examination of 
the fleas showed that the trypanosomes establish themselves in a corre- 
spondingly small percentage (see p. 659). It would appear, therefore, 
that the flea which, more than any other species, is exposed in this 
country to infection by T. lewisi, has developed a certain degree of 
natural immunity to the parasite. From the experiments published by 
the authors cited it is probable that T. lewisi would undergo its 
development in any species of flea,and would be transmitted by it, pro- 
vided that the flea could be induced to suck the blood of an infected 
rat. The natural efficacy of any given species of flea in transmitting 
T. lewisi depends probably on the habits and tastes of the flea, and 
not on any specific ability to harbour the trypanosome. Brumpt (1913) 
has pointed out that all the trypanosomes of small rodents seem to be 
able to develop in fleas. 

A number of experiments have been performed by several investi- 
gators on the transmission of T. lewisi by means of the rat-louse, 
Hematopinus spinulosus. The first experiment with rat-lice 
(species not stated) was carried out by MacNeal (1904), who transferred 
“several” lice from an infected to a clean rat ; trypanosomes appeared in 
the latter after fourteen days. Positive results in experiments of this 
kind with rat-lice are reported by Nuttall (1908), Baldrey (1909), Breinl 
and Hindle (1909), Manteuffel (1909), and Gonder (1911). To judge, 
however, from the published accounts of these transmission-experi- 
ments, positive results were by no means frequent and were obtained in 
some cases at least with difficulty and by the exercise of great patience 
and perseverance, or by using large numbers of lice. Nuttall obtained 
one positive result in an experiment in which sixty lice were used ; two 
‘other experiments, in which fewer lice were used, were negative. 
Baldrey reports two experiments in which infection was obtained by 
means of lice; in the first, 100 lice were used, and the result is regarded 
by him as a case of direct mechanical infection, but for what reason is 
not at all clear; the second, in which ten lice were used, is interpreted 
as demonstrating a developmental cycle in the louse. Breinl and 
Hindle report three successful transmissions by means of lice, after 
carrying on numerous experiments for over a year. Manteuffel seems 
to have been more successful than most other experimenters in this 
field, though he does not record the actual number of his experiments 
or the proportion of those which were positive in result, but he states 
that infections with lice were “prompt and frequent ”; his method was 
to put infected rats, with lice on them, in the same cage with clean 


THE RAT-TRYPANOSOME, TRYPANOSOMA LEWISI. 611 


rats, and from his results he concludes that lice do not transmit the 
trypanosome longer than from three to five days after being removed 
from the infected rat, and that the transmission is effected by the act of 
blood-sucking ; if the first of these two conclusions be true, it would 
appear that the trypanosome does not succeed in establishing itself in 
the louse in the way it does in the flea. Gonder reports that after many 
fruitless attempts to transmit T. lewisi with definite numbers (80- 
100) of lice, he obtained six positive results in a series of fifty experi- 
ments, and eight positive results in another series of fifty, using greater 
numbers (gréssere Mengen) of lice; and he also brought about six 
infections by making emulsions of lice taken directly from an infected 
rat, the lice having been left on the infected rat for five, nine, eleven, 
thirteen, sixteen, and twenty-one days respectively, in these six experi- 
ments. On the other hand, Prowazek, who first described developmental] 
stages in the louse and claimed that this insect was the true host of T. 
lewisi, was unable to obtain experimental transmission ; Rodenwaldt 
obtained no positive results with numerous transmission-experiments ; 
and we also have obtained only negative results in any attempts that 
we have made to transmit T. lewisi by means of the rat-louse. 

It is evident from the results summarised briefly in the foregoing 
paragraph that transmission of T. lewisi can be effected by the rat- 
louse, but only with difficulty, and in a small percentage of cases. 
This is a great contrast to the ease and comparative certainty with 
which the trypanosome can be transmitted by fleas. We have always 
used our flea-cages as the simplest and easiest method of obtaining 
infected rats when required by ourselves or by our colleagues or friends, 
and not only have we infected rats with single fleas on many occasions, 
but we have even succeeded in infecting several rats successively with 
one and the same flea. We have no hesitation, therefore, in regarding 
fleas as the usual agency whereby T. lewisi is transmitted from rat 
to rat in Nature, a result brought about by the louse rarely and 
exceptionally. 

It should be noted that Brumpt (1913) has succeeded in infecting a 
rat with T. lewisi by inoculating it with the rectal contents of a bug, 
Cimex lectularius, fed on an infected rat thirty-eight and again six 
days previously. There is no evidence, however, that this insect 
transmits the infection naturally. 


(i) [he Transmission takes place by the Cyclical 
Method. Transmission by the Direct Method 
has not been proved to occur. 

These are among the conclusions drawn from experiments 
described in full detail in our preliminary report (1910). It 


612 E. A. MINCHIN AND J. D. THUMSON. 


is sufficient here to state that experiments “A” (20) and 
“B” (21) in our report were devised chiefly to separate 
“ direct” from “cyclical”? infection, a matter of primary 
importance at the time that these investigations were begun, 
and they show that in the individual cases cited (A; and B,) 
transmission was effected when all possibility of the direct 
method was excluded. Experiments ‘‘C” (22) and “ D ” (28) 
multiply such cases many times, and show further that fleas 
once infective retain the infection so as to infect a series of 
rats without themselves being exposed to fresh infection. 
Since then a number of experiments have been carried out 
by us, many of which are enumerated under the different 
headings which follow, and the sum-total of these experiments 
not only supports the conclusions in our preliminary report, 
but establishes beyond doubt that the rat-flea isa true inter- 
mediate host, that it can transmit the infection to other clean 
rats only after the developmental cycle has been completed 
within itself; that, in short, the infection takes place by the 
cyclical method; and that there is no evidence whatever to 
show that the rat-flea is capable of carrying the infection 
from one rat to another by whatis called the “direct” or 
“ mechanical” method. 


The term method in the phrase * method of transmission,” if used 
without qualification, should include comprehensively all that happens 
in the transmission of infection from one vertebrate toanother. In the 
transmission of trypanosomes the natural transmitting agent, when 
such is known, is a blood-sucking invertebrate of some kind. When 
the method is said to be “ contaminative” or “inoculative,’ trans- 
mission is viewed from the side of the invertebrate in its relation to the 
vertebrate, and the problems involved are particular, that is to say 
such as deal with modifications due to special cireumstances in those 
relationships, and are concerned at most with special groups of 
trypanosomiases rather than with trypanosomiasis in general. On the 
other hand when the method of transmission is said to be “ cyclical” 
or “ direct,” transmission is viewed from the side of the trypanosome in 
its relation to the invertebrate, and the problem becomes a general one, 
dealing with that phase of the transmission which is concerned with the 
life-history of trypanosomes as a group of parasites, and with the wider 
question of their double relationship to vertebrate and invertebrate, 


THE RA'T-TRYPANOSOME, TRYPANOSOMA LEWISI. 613 


bringing them into line with other known relationships among parasitic 
Protozoa in this respect. 

A great impetus was given to the study of trypanosomes by economic 
and other considerations arising out of the prevalence of tsetse-fly 
disease and sleeping sickness in Africa. It was long known that these 
diseases could be transferred artificially by direct inoculation of blood 
from a diseased to a healthy subject by means of a hypodermic syringe. 
Naturally, therefore, before much work had been done in this direction 
tsetse-flies known to be associated with the spread of these diseases 
were supposed to transmit them in this direct way. Bruce and others, 
experimenting with bred-out flies, proved the possibility of this taking 
place under certain conditions which, in the case of sleeping sickness 
at all events, were very unlikely ever to be fulfilled in Nature. Only 
with a swarming infection and by interrupted feeding could the 
disease be passed on directly from an infected to a clean animal 
with any approach to certainty, and even under the most favourable 
conditions in other respects, the longer the interval between inter- 
rupting the feed on the infected animal and continuing it on a clean 
’ animal, the less the chance of the clean animal becoming infected, 
until, with the lapse of about half-an-hour, it was just as certain 
that the infection would not take place. Moreover, however short 
the interruption between the feeds, an interposed partial feed on a 
clean animal rendered the fiy non-infective to a second clean animal. 
Later experiments showed that the contents of the stomachs of flies 
that had fed on an infected animal, if injected into a clean animal, 
could produce infection only up to about two days after the infective 
feed. The fly itself, however, could not be shown to act in any way 
resembling a hypodermic syringe, and the idea of ‘‘ delayed mechanical 
transmission’ never found support from feeding experiments. The 
conclusion to be drawn from all the earlier experiments on direct 
transmission seemed to be that when infection was obtained it was 
with ‘“‘fouled proboscis” before the blood in its lumen distal to the 
entrance of the salivary duct, and perhaps also on its external surface, 
had had time to dry, and that the conditions under which it was 
shown to be possible were never likely to be fulfilled in Nature, in 
the case of sleeping sickness, and in the case of tsetse-fly disease of 
cattle, far too seldom to account for the spread of the disease, while 
in no case could such a method of transmission account for the 
existence of fly-belts through which healthy domestic stock cannot 
pass. Other things being equal, the efficiency of an invertebrate as 
a transmitter of trypanosomes would be enormously increased if the 
invertebrate were a true intermediate host and not merely a‘ porter ” 
of the parasites from an infected to a clean subject, and to demon- 
strate beyond doubt that trypanosomes underwent an alternation of 


614 E. A. MINCHIN AND J. -D. THOMSON. 


generations was of primary importance in connection with the general 
trypanosome problem at the time that we undertook this investiga- 
tion, when it was being maintained by Patton and others that no 
trypanosomes went through a developmental cycle in the invertebrate, 
that all transmission of trypanosomes was direct, and that the crithidial 
forms found in blood-sucking invertebrates were all of them independent 
parasites of the invertebrate, having no connection with the trypano- 
somes or other parasites of the vertebrate. It was known that rat-fleas 
could transmit T. lewisi from infected to clean rats, and although 
transmission by fouled proboscides seemed quite out of the question, it 
was necessary to demonstrate beyond doubt that the rat-flea is a true 
intermediate host of T. lewisi, that it can transmit the infection to 
other rats only after the developmental cycle has been completed within 
itself, and that once infected it remains infective for a considerable 
time, so as to be able to infect a series of clean rats without itself being 
exposed again to infection. These points, which we believe concern the 
transmission of trypanosomes in general, and which may be taken as 
typical of the relations which trypanosomes as a group bear to their 
invertebrate hosts, as well as other points of more special interest (con- 
fined, it may be, to T. lewisi alone or to the lewisi group), are 
dealt with under different headings in what follows later. Although 
the practical cannot properly be separated from the scientific, the most 
interesting problem of the transmission from the scientific point of 
view is perhaps the way in which the trypanosome becomes established. 
There are considerable variations in the details of the cycles of 
different species or groups of trypanosomes in their natural hosts due 
to special conditions, but arising out of the very meaning of a cycle, 
and therefore common to all is the fact that until the cycle is com- 
pleted the invertebrate, though infected, is not infective. This may 
be of direct practical importance in special cases, and where that 
is so it is important to ascertain the length of time required for the 
completion of the cycle in each case. Of more general practical 
importance in questions connected with the spread of infection is the 
fact that the trypanosome does establish itself in such a way that the 
invertebrate remains infective for a long time without requiring to be 
exposed again to infection. 

(iii) The Trypanosomes make their Appearance in 
the Blood of the Rat Five to Seven Days after 
Infection; the Multiplication of the Trypano- 
somes in the Blood of the Rat come to an End 
‘Eleven to Thirteen Days after Infection. 


In order to establish with exactness the length of the 


THE RAT-TRYPANOSOME, TRYPANOSOMA LEWISI. 615 


incubation-period and the multiplication-period in the rat 
after infection, it is necessary that the rat should have been 
exposed to infection by the fleas for a short time; long 
exposure leaves too wide a margin between the possible 
maximum and minimum deducible from the actual data 
furnished by the experiment for the mean to be of any value 
im reckoning the length of the two periods in question. 
When the rat is removed from contact with the infected 
fleas it is further very necessary that all fleas should be 
removed from its skin. The rat is then kept in a flea-proof 
cage and its blood is examined daily in fresh, living films 
until trypanosomes are first detected in it, in order to deter- 
mine the duration of the incubation-period ; then smears of 
the blood are made and preserved daily and examined until 
the multiphication-period is found to be past and ended. So 
long as the trypanosomes are multiplying in the rat’s blood, 
they are of various sizes, some of the ordinary, normal size, 
others very small, and others again much above the normal 
size. Marked variation in the size of the trypanosomes is a 
sure sign that multiplication is proceeding, even when actual 
division-stages are so scarce in the preparation that pro- 
longed search is necessary in order to find them. As soon as 
the multiplication is ended the trypanosomes are all of one 
type and size, allowing for slight individual variations that 
are not perceptible without careful measurement; to such 
trypanosomes; the normal form of T. lewisi and the sole 
form occurring in the blood when once the multiplication is at 
end, we shall refer always as “ ordinary.” 

We cite here a few examples from our series of experiments, 
choosing first (Table C), those in which the rats were exposed 
to infection for one day only, so that the periods of incubation 
and multiplication can be determined within a margin of one 
day. In our second table (D), we quote those instances in 
which the rats were exposed to infection for two days, so that 
a wider margin of possible error must be allowed for in 
calculating the two periods. In a third table (E), we shall 
give some results obtained with rats which were infected 


A. MINCHIN AND J. D. THOMSON. 


E. 


616 


* ~ST=E1 io ta) OT.:! ‘16 Ohh SL Olt, SOL Ole? 48 98 ss 

Y  BEOL eRe OL r “St OL.:! ‘FT Ol. tee OL.:t *9 SSL FS 

GO Flan Sie? lel) 20s Oles Deus OL! $06 OL. Po SL 891 i 

y SSIIT gee he) OL.?! *96 OL.?! *13 OT.:! ‘ST Ol t= "SI POL (a =) 8 

SSC La0L ee!) Cher at 60, | HX? 8Z 60, ? EX! BB 60.) UX! 03 TFI be 

Pe Aolstb bathing 60. ? HX! 9G 60. ° 1X20 | 60? HX! SL 60, } HX? ST SI = 

aol SS ORY, 60, IX! PZ 60. 2 HX! LT 60, : UX? 81 60.) UX! TL CSI . 

oe | Ca Eh oF 60, ' HX: 1B 60.) HX! FT 60.) 1X! OT 60, HX?8 FEL : 

ae ea ee Jc 60.) UX! BL 60. ? HX? SL 60.) 1X8 60.) HX: 9 Ic (0 =) 3 
skup T[-6 shkep 9-f 60. * TEX! BZ 60.) WX! LT 60,? UX! ST 60.) HX! TL OST (V =) 02 
ee evar ees SOneoT GIT herecuvonan “AROSE Roel tis arise ‘ON 

"Moloozuy of otnsodxg ,shvuq omT—'(_ Wavy, 

ie olait DOE GT, : WIA: 6 GT.) WIA § ST. ? TA * 63 GT, ? WA $8 [Lg CP 
shkvp [[-OL PeoRe ile Te 2) Sie Ili 4 ? 91 Li. Oo L418 OF 

ie 2 perp Fwy 60.) WX +13 60, ? UX? LT 60, 1X: OT SSI (dq =) &3 

<< SELL Steele) 60, ¢ HX + EZ 60.2 UX 5 OT 60.) HX! LT 60. HX! OT 621 (0 =) 3 

GEO pe 1-9 60.' UX #8 60.) UX 2 PT 60, UX8 60.) UX: Y StI (q =) Iz 
skep [[-0L sep 1-9 60,:MX *ST | 60? 1X PPT 60, UX? 8 60,2: UX: Z GCL (¥ =) 06 
is Sie eee Fae caesar soe eceadiny "ano ere eal aaa S Eides 16 ‘ON 


‘UOTJPOOFUT 


0} oinsodxgy s,Avq ouQ—') ATAV, 


617 


‘oTqGez oY} UL pozeys 


= wey} aarpave Lup v usaq oavyq Avr ‘porod WorPwoAypNUt oy} FO WorpvUrUttoy oy} JO Lo “‘pooyd Aroq} UL sotmosound sary 
= jo sountvedde ay} Jo sa}rp oy} Ajquenbesuoy ‘sep o}eute}][B WO ATWO pauIUtexs o19M SUY} POYIVUL SPLI OWT, x 
me 2 Coock i > OL GL.‘ WA + 9G Gi ese ¥ (lige Coll | #696 
= ; “GL x Oe rail BS Yo Glee G6 ¥ Che Te SE | eR 
= | (xem) séep gy | (xem) skepyy | TE? 8t :9L | kei =f {61 WO 9F9] TL? MHA :63 | ¥ST8 
2 ic FI-6 SOS DD eels [Eig oe ys I UIE? tere Wie 2b ees0e 
>, 2 OES ea ee eG IL. ‘4! ° 46 TL. * 42 +06 Th. * 4b *ST | 966 
a Se ie eh Wile Aen, Ti ee nl oe ATE OG US AR he *COG | 
4 Ame Eee) Pe Oar igs 2O (Sli, Biers oes Ele At Te Sees seloor | 
PARI ee Oiles: IL.‘ +86 Lele ie ie Sr eg ee lee ae Coam: 
5 -) sis6 ~ Ole ia rac Diets = hoe lee iets SSI Lees SRL #296 || 
ot ee EO ise) (Bl Gree es!) gee all [le AG se ieee Ae ¥696 | 
= Olas ey has ie Ae ey ee ae Grey ane eG Th? HE *€6 | 696 
F  TE9 vin MORSE. Wis Ae rere igs 2 ff [ls Pee IL TE. ' HE *26 | #88e 
S ee eel io) Gla OL.* = “96 OL Ee SG lige es tS] Oh 2S |) sakes 
= ae pe (lg) peer cleats O)auesaeeoa tes Ole ese Dige es 9 ste Olina ale SEG 
pA py clias! 4k =§ OL: = = 66 Oa Obes a (Migs 2s OIL #666 
oi we Lise ee) Oe =. 0s Ole 8c OS ece Ole to «C16 
cr | sup T[-8 skep 6 -9 (Dig MAES tl Olig ea OS Oh eG «LOG 
oA | = 7 as 
ae 

: t *pow 9 “maes “qe 
sa Pion as eae dowenratntnt qsay souroso wed Tee eee Saki “Qua wo and werd ra N 
i 


‘VoTq e[sarg e@ Aq poqoazuy syuy— A AAV], 


A. MINCHIN AND J. D. THOMSON. 


i 


E 


618 


IL 
IL 
IL 
GL 
skep [7 


ee 


*porod 
woTqpRorLay [OW 


ee 9 
ce 9 
ee 9 
“ec 9 
oe 9 


6 L 
skvp 6 


*potoed 
TOMBqnoUy 


“sBold yo worqookuy £q paqooyuy sqazy—al anay J, 


‘TOVULO}S = 8 fumyoot = 1 


oe 


(poqou you) 
Sea) 
Olathe e2G 


Meee 2 SE 

Th “SE 

Ties) 01 
*popuad 


WOTPBOTLALY [NAL 


Tl.? 1 +06 


TI. ? te +06 
yeas 
OL. * tH * 1¢ 
Lee 00 
Tl. #8 
tie 38 


1G ira 
Dene ek 
Lie 
OL. M+ ST 
[Eres 
ee lt 
LT. At * 66 


cree tee 
T8G roe 
193 | #8 
LOG « oi\) © are 
PIE 6E 
rat 68 
) 608 | 68 


“UOOS SAY 
soutosound Acq, 


4 SBI]y JO ‘ON 


‘worpoatur JO oFVQ 


| 


“yea JOON | quounted xg 


THE RAT-TRYPANOSOME, TRYPANOSOMA LEWISI. 619 


each by a single flea, in order to show that in many cases, at 
least, the maximum periods of incubation and multiplication 
which can be deduced from experiments under these con- 
ditions are not greater than those indicated by the experi- 
ments in which many fleas were used to obtain infection. In 
a fourth table (F) we give for comparison, the results obtained 
by inoculating rats with the stomachs or recta of fleas; in 
such cases the length of the periods of incubation and multi- 
plication can be determined with exactness, the moment of 
infection being known. 


The determination of the length of the multiplication-period in an 
infected rat is of practical importance for interpreting other experi- 
ments, since, when it has been determined, it furnishes a datum from 
which approximately accurate conclusions can be drawn as to the time 
at which the rats become infected, when the point is shown definitely by 
the details of the experiment. As regards the first appearance of the 
trypanosomes in the blood, they appear at first in such scanty numbers 
that it is very easy to overlook them, and they may often be reported 
absent when a more prolonged search would have detected their 
presence. Similarly, the trypanosomes at the end of the multiplica- 
tion-period may sometimes have been reported as “all ordinary” in a 
smear in which more careful searching might have led to the discovery 
of a few individuals above or below the normal size. Consequently, the 
errors of observation are such as tend to over-estimate the length of 
the incubation-period, and to under-estimate that of the multiplication- 
period, from the serutiny of the blood-films. On the whole, however, 
the results obtained in our experiments are very uniform and indicate 
an incubation-period of about six days, a multiplication-period of about 
twelve days. It is interesting to note that these results agree with 
those obtained in the case of rats infected artificially by inoculation, 
intra-peritoneal or otherwise, with blood from an infected rat. Since a 
syringe would inoculate far more trypanosomes than the rat would 
obtain from even a large number of fleas, it might have been expected 
that the rat would, so to speak, fill up quicker when infected by means 
of a syringe, and that consequently the multiplication-period would be 
correspondingly shorter. In our experience, however, the length of the 
multiplication-period remains approximately constant in all cases, 
whether the infection is effected by a syringe, by a large number of 
fleas, by a few fleas, or even by a single flea ; a fact which indicates that 
the length of time during which the trypanosome multiplies in the rat 


620 E. A..MINCHIN AND J. D. THOMSON. 


is not determined by the number of trypanosomes put into the rat, but 
by the mutual interaction of host and parasite. 

It may be noted here that some rats appear to possess a certain 
degree of natural immunity to infection with T. lewisi. A single 
instance which came under our experience will suffice to demonstrate 
this point. A rat was exposed to infection on June 9th and its blood 
was examined daily; on June 25th a few trypanosomes were first seen 
in the blood in scanty numbers, just as they are usually seen at their 
first appearance between the fifth and seventh days of the infection. 
The rat was then removed from contact with the fleas and kept apart ; 
but neither on the next day nor on any subsequent day were any try- 
panosomes to be found in its blood. This rat, therefore, contracted 
only a transitory infection which was late in its appearance and dis- 
appeared after one day ; had the trypanosomes been overlooked on that 
day the experiment would have been returned wrongly as negative in 
result. 


(iv) The Cycle of Development in the Flea requires 
a Minimum-of Five Days for its Completion. 


This point was dealt with in our preliminary communication 
(1910), in which we came to the conclusion that the incuba- 
tion in the flea was six or seven days. Our method of deter- 
mining this was, first of all to expose non-infected fleas to 
infection, by putting them on a well-infected rat, for but a 
single day, so that if the fleas afterwards produced an infection, 
the time at which they themselves became infected could be 
determined within a narrow margin, twenty-two hours in our 
actual experiment. The fleas were then placed in contact for 
three days with clean rat (1) which did not become infected ; 
after that for three days with clean rat (2), which also did not 
become infected; and then for two days on clean rat (3), 
which showed trypanosomes in its blood six days after being 
removed from contact with the infected fleas. Clean rat (3) 
was, therefore, infected by the fleas in the interval between 
the sixth and eighth day after the fleas themselves had 
acquired the infection ; consequently the infection in the fleas 
could not have been more than eight days old. 

Subsequent experiments performed by us have indicated a 
possible minimum of five days for the flea-cycle of the try- 


THE RAT-TRYPANOSOME, TRYPANOSOMA LEWISI. 621 


panosome. In experiment 39 (see below, p. 630) it is proved 
that the rectum of the flea, injected into the rat, can produce 
an infection as early as the fifth day, and in such fleas the 
examination of films shows the presence of the small trypano- 
somes which are the final form of the development in the flea. 
In experiments 26 and 28, undertaken in order to ascertain 
whether a rat, in which the trypanosomes are still in the 
multiplication-period, is capable of infecting fleas (see below, 
p. 657), the results obtained indicated a short incubation- 
period in the fleas. Thus in experiment 26, 127 fleas, after 
being three days (from 8: ii: 10 to 11: ii: 710) on the in- 
fected rat were put on rat 187 for another three days (from 
11 :ii to 14: ii). Rat 187 showed trypanosomes in its blood 
after five days (19 : ii), and the multiplication-period ended 
five days later (24: i1). Consequently the incubation-period 
in the fleas could not have been more than six days (8 : 11 to 
14: ii). In experiment 28, 137 fleas were put first on the 
infected rat for four days (15: ix: 710 to 19: ix: 710), and 
then were put for one day (19 : 1x to 20: ix) on rat 209 ; after 
this they were put on rat 215 and left on it. Rat 209 had 
shown no infection when it died nine days later (29 : 1x) ; 
rat 215 first showed trypanosomes on 30: ix, and the multiplhi- 
cation was ended 3: x. This result indicates that rat 215 
was infected about 21 : ix, in which case the incubation-period 
in the flea could not have been more than six days (15 : ix to 
91: ix). Since rat 209 showed no trypanosomes for at least 
nine days after being exposed to infection it was probably 
not infected ; so that the infection in the fleas was probably 
not ripe for at least five days (15 : ix to 20 : 1x). 


On the other hand we have instances, as already mentioned in our 
preliminary communication (1910), of an incubation-period in the flea 
apparently much longer than six days. Thus in experiment 19 a cage 
(J) was stocked with seventy-two fleas from the non-infected breeding- 
cage and an infected rat (No. 81, a wild black rat, naturally infected) 
was put with them for three days (20:ix :’09 to 23:ix:’09). Rat 81 
was then removed and rat 82, a clean, tame rat, was put in its place 
(23: ix) and left in the cage. Rat 82 first showed trypanosomes in its 
blood 29:x; the multiplication-period was ended about 4:xi, indi- 


622 E. A. MINCHIN AND J. D. THOMSON. 


cating that the actual infection of rat 82 took place about 23:x. In 
this case, therefore, the fleas did not produce an infection in the clean 
rat for at least a calendar month after their contact with the infected 
rat was interrupted. Such a result, however, permits of no conclusion 
whatever as to the length of the incubation-period in the flea; it merely 
demonstrates a point proved also by other experiments, namely that in_ 
fective fleas often fail to infect. We have put forward already (1910) 
one possible explanation for this, that a rat, which is comparatively 
immune to begin with, may resist infection for a long time, but its re- 
sistance may be overcome at last. Another possible explanation may 
be given by the method in which infection of the rat by the flea is now 
known to take place, namely by the rat licking off the moist feces of 
infective fleas that are deposited on its skin (see below, p. 648). It is 
evident that if the rat fails to lick off the feces while still moist, or if 
the infective flea does not defecate on the rat, no infection is brought 
about. A negative result of this kind is most likely to be attained when 
the number of infective fleas on the rat is very small, as seen in the 
large number of negative and small number of positive results in our 
series of experiments in which single fleas were used (see below, p. 661). 
That infective fleas in Cage J were rare is shown by the fact that 
between 23 :ix and 19: x thirty fleas from this cage were dissected and 
examined without finding a single one infected. On the other hand, 
when fleas are sufficiently numerous and have been well infected, posi- 
tive results are fairly certain (Experiment “C” of our preliminary 
report). 


(v) Transmission is never effected until the Deve- 
lopmental Cycle is completed; that is to say, 
until at least Five Days have elapsed since the 
First Exposure of the Fleas to Infection. 


We have found, as already stated, by direct observation, 
that the final form of the developmental cycle appears in the 
gut of the flea five days after the infective feed (see below, 
p. 630). We bring forward here a few instances to show that 
at least five days must elapse before the flea becomes infective, 
after having ingested trypanosomes from an infected rat. 


(1) Experiment 20.—A cage colonised with forty-four fleas that 
had been exposed to infection from 4: x : 09 to8: x: 09. 

Rat 93 put into the cage from 8:x tol2:x, i.e. during a period 
in which the age of the infection in the fleas could not have been less 


THE RAT-TRYPANOSOME, TRYPANOSOMA LEWISI. 623 


than two days old at the beginning nor more than eight days old at the 
end. Result negative. 

(The next rat used in this experiment died, but subsequent rats used 
showed that the fleas had become infective.) 

(2) Experiment 21.—A cage colonised with 157 fleas that had been 
exposed to infection from 11: x : 09 to 15: x: ’09. 

Rat 97 put in from 15 : x to 19: x, i. e. during a period in which the 
age of the infection in the fleas could not have been less than a few 
hours at the beginning nor more than eight days at the end. Result 
negative. 

(The next rat put in became infected, apparently about 27: x; age of 
infection in the fleas then between twelve and sixteen days.) 

(3) Experiment 22.—A cage colonised with 160 fleas exposed to 
infection from 24 : xi: 09 to 27: xi: ‘09. 

Rat 116 put in from 27 : xi to 30: xi, i. e. during a period in which 
the age of the infection in the fleas could not have been less than a few 
hours at the beginning nor more than six days at the end. Result 
negative. 

(The next rat put in became infected, apparently about 3 : xii; infec- 
tion of the fleas then six to nine days old.) 

(4) Experiment 23.—A cage colonised with 162 fleas exposed to 
infection from 7 : xii: 09 to 8: xii: ’09. 

Rat 125 putin from 8: xii to 11: xii, i.e. during a period in which 
the age of the infection in the fleas could not have been less than a few 
hours at the beginning nor more than four days at the end. Result 
negative. 

Rat 133 put in from 11 : xii to 18: xii, i. e. during a period in which 
the age of the infection in the fleas could not have been less than four 
days at the beginning nor more than six days at theend. Result nega- 
tive. 

(The next rat put in became infected, apparently about 15 : xii; infec- 
tion of the fleas then seven to eight days old.) 

(5) Experiment 24.—A cage colonised with fifty fleas exposed to 
infection from 7 : xii: 09 to 8: xii: ‘09. 

Rat 126 put in from 8: xii to it: xii, i.e. during a period in which 
the age of the infection in the fleas could not have been less than a few 
hours at the beginning nor more than four days at the end. Result 
negative. 

(The next rat put in became infected, apparently about 3:1:710; age 
of the infection of the fleas then between twenty-six and twenty-seven 
days.) 

(6) Experiment 25.—A cage colonised with seventy fleas exposed 
to infection from 31: xii: 09 to3:1: 10. 

Rat 152 put in from 3:ito6:i,i.e. during a period in which the 

vot. 60, part 4.—NEW SERIES. 45 


624. BE. A. MINCHIN AND J. D. THOMSON. 


age of the infection in the fleas was not less than a few hours at the 
beginning nor more than six days at the end. Result negative. 

(The next rat putin became infected, apparently about 6 or 7:1; the 
age of the infection at 7 : 1 was from four to seven days.) 

(7) Experiment 45, Batch C (see below).—Bell-jar colonised 
with thirty fleas exposed to infection from 22 : vii : 13 to 23: vii : 713. 

No infection produced in rat 370 put in fora period of two days, 
during which the infection in the fleas could not have been less than 
five days oldat the beginning nor more than seven days old at the end. 

No infection produced in rat 370a, put in for a period of one day, 
during which the infection in the fleas could not have been less than 
eight days old at the beginning nor more than ten days old at the end. 

(The next rat put in became infected.) 

In the previous section it has also been pointed out that in Experi- 
ment 28, rat 209 escaped infection when exposed to infection by 138 fleas 
during a period of one day, at the beginning of which the age of infec- 
tion in the fleas could not have been less than a few hours nor more than 
five days at the end. Rat 215 became infected by the fleas a day later, 
when the age of the infection in the fleas could not have been less than 
two or more than six days. 


Putting together the results of this and the last section, it 
is seen that fleas in which there is a possibility, from the data 
of the experiment, of the infection being more than five days 
old, may fail to produce infection, although the subsequent 
history of those fleas shows them to have been infected 
effectively, but no infections have been obtained in any 
experiment of which the data are incompatible with the infec- 
tion being at least six days old in the fleas that produced the 
infection. 


(vi) The Infection of the Rat is brought about by 
the Small Trypanosome-form which is the Final 
Form of the Development. 


This point is scarcely capable of direct proof, since it is 
impossible to be absolutely certain that when an infection 
has been produced no other forms of the developmental cycle 
in the flea were introduced into the rat except the trypano- 
some-forms. It can, however, be demonstrated in experi- 
ments planned for that purpose, that the trypanosome-forms 


THE RAT-TRYPANOSOME, TRYPANOSOMA LEWISI. 625 


are present when an infection is produced. Ndéller (1912) 
and Wenyon (1913), have shown that the trypanosome-forms 
were present in all cases in the feeces with which they infected 
rats per OSs. 

The following experimental results indicate that the try- 
panosome-form is the effective agent in infection. In Experi- 
ment 35 B twelve fleas were taken at hazard from the infected 
breeding-cage and put on clean rat 243 for five days (23 : ii: 
711 to 28:11: 711; Rat 243 was found to be infected on 3: 
ii:711). Ten of the fleas recovered were then dissected (the 
other two lost); of each flea the stomach was placed on one 
shde in a drop of salt-citrate solution, the rectum on another 
slide in another drop. Each stomach and each rectum were 
then teased up and examined microscopically in order to see 
if trypanosomes were present in any form; but since this 
examination had to be performed very rapidly and cursorily 
and without putting a coverslip over the drop, trypanosomes 
may have been often overlooked, when they were not present 
in abundance. Whether trypanosomes could be seen in the 
fresh specimen or not, each teased-up stomach was inoculated 
by means of a syringe into a separate clean rat; the rectum 
was only inoculated if trypanosomes were seen in it.! After 
the drop containing the teased-up stomach or rectum had 
been drawn up into the injecting syringe the film of moisture 
left on the slide was fixed with osmic vapour, stained with 
Giemsa’s stain, and carefully searched for trypanosomes. The 
following are the results obtained with each flea. 


Flea (1).—Nothing seen in the fresh stomach or rectum. Stomach 
inoculated into Rat 279. No infection produced. Nothing found in 
the preserved film. 

Flea (2)—As last, stomach inoculated in rat 280, no infection, 
nothing found in the films. 


1 The reason for the differential treatment of the stomach and rectum 
was because we believed, at the time, that infection was brought about 
by regurgitation of infective trypanosomes through the proboscis from 
the stomach, and also because the presence of trypanosomes in the 
rectum is not so easily overlooked as in the stomach. 


626 E. A. MINCHIN AND J. D. THOMSON. 


Flea (3).—As last, stomach inoculated into rat 281, no infection, 
nothing found in the films. 

Flea (4).—One sluggish stumpy form, which may have been erithi- 
dial or trypaniform, was seen in the fresh teased-up stomach; nothing 
seen in the fresh rectum. Stomach inoculated into rat 283, result 
negative. Nothing found in the preserved film of the stomach. 

Flea (5).—Nothing seen in the fresh stomach, numerous trypano- 
somes seen in the rectum. Stomach inoculated into rat 285, result 
positive. Rectum inoculated into rat 248, result negative. One try- 
panosome-form and one transitional form found in the preserved film of 
the stomach (Text-fig. 16, b and c). Nothing found in the preserved 
film of the rectum. 

Flea (6).—Nothing seen in the fresh stomach; a few forms, some 


TrExt-Fic. 16. 


S2C% 


Small trypanosome-forms from the stomach-films of fleas 5 and 7 
in Experiment 35 B, and flea 5 in Experiment 27 (see text). 
(x 2000.) 


stout and of crithidial appearance and some slender, apparently try- 
paniform, seen in the rectum. Stomach inoculated into rat 286, result, 
negative; rectum not inoculated. No films preserved. 

Flea (7).—Nothing seen in the fresh stomach or rectum. Stomach 
inoculated into rat 288, result positive. One trypanosome (Text-fig. 
16, a) found in the preserved film of the stomach. 

Fleas (8), (9), (10)—In each case nothing was seen in the fresh 
stomach or rectum. The stomachs were inoculated into rats 289, 262, 
263 respectively, results in each case negative. Nothing found in the 
preserved films. 


Summary.—In the case of two fleas out of the ten used, 
the stomachs, when inoculated into clean rats, produced an 
infection. The final trypanosome-stage was found in both 
the stomachs that produced infections, but in none of the 
remaining eight stomachs that produced no infection. 


La J 


Experiments 27, 29 and 32 were conducted in a different 


THE RAT-TRYPANOSOME, TRYPANOSOMA LEWISI. 627 


manner. Fleas taken from the infected breeding-cage were 
put each on a separate rat and left on it for three or four 


days. The flea was then recovered (if it could be found), 
dissected and examined. 


Experiment 27.—Flea (5), placed on rat 207 for three days 
(2: viii: 10 to 5 viii : °10) produced infection (see Table E). The flea 
dissected (5 : viii), and one large transitional form found in the slide of 
the stomach (Fig. 16, d). 


TExtT-FiG. 17. 


Various forms (haptomonad, a—d, nectomonad, e-h, transitional, 
il, and trypaniform, m and n), from the stomach-film of flea 
3, Experiment 29 (see text). (x 2000.) 


Experiment 29.—Flea (3) placed on rat 212 for three days, 
(19 : ix : 710 to 22 : ix : 10), recovered and dissected 22 : ix °10 (see Table 
E). Large clumps of attached forms were seen in the stomach and also 
free forms; nothing was seen in the intestine, rectum, salivary glands 
or proboscis. The preparations of the stomach showed crithidial, 
transitional and trypaniform types in abundance (Text-fig. 17). Rat 
212 became infected and first showed trypanosomes in the blood on 
28: ix. Four other fleas in the same experiment failed to infect their 
rats; in two of these fleas nothing was found, in the third a small 


628 E. A. MINCHIN AND J. D. THOMSON. 


Trxt-Fic. 18. 


4 
> 


mM. 


h. J: 


Various forms (haptomonad, a, b, nectomonad, ¢, d, transitional, 
e-h, and trypaniform, 7—m), from the stomach-film of flea 3, 
Experiment 32 (g) (see text). (x 2000.) 


TExtT-FIG 19. 


Various forms (haptomonads, a—d, nectomonad, e, transitional, 
f~, and trypaniform, j-m) from the rectum-film of flea 3, 
Experiment 32 (h) (see text). (x 2000.) 


THE RAT-TRYPANOSOME, TRYPANOSOMA LEWISI. 629 


attached clump was seen in the intestine, nothing in any other part of 
the flea. The remaining flea was not recovered. 

Experiment 32 (g).—Flea (3) placed on rat 241 for four days (5X1: 
10 to 7: xi: 10) produced no infection. In the stomach of the flea, 
dissected 8: xi, crithidial, transitional and trypaniform types were 
found (Fig. 18). 

Experiment 32 (h).—Flea(3) placed onrat 244 for four days (11 :xi-710 
to 15:xi 10) produced an infection (see Table E). The flea was 
dissected 15 : xi; nothing was found in the stomach, intestine or salivary 
glands ; the rectum showed a typical swarming “ pile-carpet ” infection 
with all the usual types of form (Text-fig. 19). 


TExt-FIG. 20. 


og ade 


Flea3, Experiment 32(i). a, haptomonad, and b, trypanosome-form 
from the rectum-film; cand d, haptomonads, e-g, nectomonads, 
from the film of the intestine (see text). (x 2000.) 


Can a 


Experiment 32 (i).—Flea (3), placed on rat 247 for three days 
(12: xi: 710 to 15: xi: °10), produced no infection. The flea, dissected 
and examined (16: xi), showed no trypanosomes in the stomach, but 
in the intestine were clumps attached behind the pylorus (Text-fig. 20, c), 
and the rectum contained a teeming infection (Text-fig. 20, a and b) of 
the usual types. 


The cases cited show that when infections were produced 
the final trypanosome-form was found either in the stomach 
or rectum; but it should also be mentioned that we have 
three instances in which the rat became infected under similar 
circumstances without our having been able to discover an 
infection of the flea, which must have been so scanty as to 
escape detection in our films. On the other hand the experi- 
ments also show that the infective form may be present in the 


630 E. A. MINCHIN AND J. D. THOMSON, 


flea without any infection resulting when the flea is on the rat 
for not more than four days. The failure of the flea to infect 
in such cases must be correlated with the casual nature of the 
contaminative method of infection by the fleas, evidently not 
so sure a method as that of inoculation. It will be shown 
further (Experiment 39, below) that the period at which the 
flea becomes infective coincides with the first appearance of 
the small trypanosomes in the rectum. 


(vii) The Final Infective Form of the Cycle is deve- 
loped first in the Rectum on the Fifth Day of 
the Developmental Cycle, but may appear later 
in the Stomach. 


In order to ascertain how soon the trypanosomes, ingested 
by the flea, attain to maturity in the different parts of the 
digestive tract of the flea an experiment (Experiment 39) was 
carried out in the following manner. A number of fleas 
(about one hundred) were collected from the non-infected 
breeding-cage, put into test-tubes, with clean sand, slightly 
damp, and kept there for four days (20:iv:711 to 24:iv:71]), 
in order that they should be properly hungry and ready to 
feed. The fleas were then (24: iv) put into a special flea- 
proof tin cage with a well-infected rat (No. 259).’ 

At regular intervals batches, each of ten fleas, were 
recovered from rat 259, kept in the test-tubes on sand, and 
dissected on the following day (to ensure that the fleas had 
not ingested blood containing trypanosomes for at least a day 
previous to being dissected). In the dissection of the fleas, 
the flea was first placed on a slide in a drop of salt- 
citrate solution and the proboscis removed by cutting through 
the head in the region of the eyes; the proboscis was then 
placed in a separate capsule in a small quantity of salt- 
citrate solution. Very often feces were extruded when the 

+ Rat 2459 was put in with a single flea on 22 : iii: “11; trypanosomes 
were first seen in the blood on 30: iii; multiplication was ended on 
4: iv; see Table E, p. 617 above. 


THE RAT-TRYPANOSOME, TRYPANOSOMA LEWISI. 631 


head was cut through. When this occurred the carcase of 
the flea was at once removed to another slide, and the 
extruded faeces examined microscopically. If trypanosomes 
were found in sufficient abundance the slide and coverslip 
were fixed and stained. The carcase of the flea was then 
opened in the hinder end of the abdomen, and the junction 
of stomach and intestine cut through behind the Malpighian 
tubes, after which the portion of the carcase containing the 
stomach was transferred to another slide. ‘Then the stomach 
and Malpighian tubules, together with the proventriculus and 
cesophagus, were removed together and transferred to a 
second capsule, and the proctodzum (intestine with rectum) 
to a third. In making these dissections some of the con- 
tents of the stomach or rectum escaped on to the slide (the 
organs being purposely punctured to allow some contents 
to escape, when necessary). The escaped contents were 
examined microscopically, and if trypanosomes were found in 
them in sufficient numbers they were preserved. 

After all the 10 fleas of each batch had been dissected in 
this way the 10 proboscides were inoculated into one clean 
rat, the 10 stomachs into another, and the 10 recta into a 
third. In each case the whole of the salt-solution in the 
capsule was injected also. he stomachs and recta were 
teased up as fine as possible, and the proboscides crushed up, 
before injecting them. 


The following are the actual injections performed; in the results 
stated, 0 signifies that the rat inoculated acquired no infection, + 
that it became infected. 

26: iv :°11.—The 10 fleas recovered on the previous day were dis- 
sected ; trypanosomes were seen in the stomach, rectum, and extruded 
feces of several fleas. All trypanosomes seen appeared to be of quite 
ordinary type: 

10 proboscides injected into rat 289: result 0 
10 stomachs ,F of 299) e eee 0) 
10 recta i - BUG oe 0 

97: iv :’11—The 10 fleas recovered on the previous day were dis- 
sected; trypanosomes were seen in the stomach of one, the stomach and 
extruded feeces of another, in the stomach and rectum of a third, and in 


632 E. A. MINCHIN AND J. D. THOMSON. 


the rectum and extruded feces, very abundantly, of a fourth; the 
rectal forms appeared pear-shaped or club-shaped in the living state 
(Text-fig. 21), but no post-crithidial trypanosome-forms were present : 


TEXT-FIG. 21. 


Various forms from the rectum and feces of a flea of the batch 
of 27:iv:’ll. Note that no final trypanosome-forms are 


present (see text). (x 2000.) 
10 proboscides injected into rat 301: result 0 
10 stomachs 2 eo eens pA arr AU) 
10 recta “ Se OUD a same 
28: iv:’11—The 10 fleas recovered on the previous day were dis- 
sected ; trypanosomes were seen in the stomachs of three, and in the 
stomachs and extruded feces of two others. The feces were preserved 


THE RAT-TRYPANOSOME, TRYPANOSOMA LEWISI. 633 


in one case where they were seen, but no trypanosomes were found in 
the preparations : 

10 proboscides injected into rat 304: result 0 

10 stomachs . eee U5 en me 0 

10 recta % ae iy ose f:.2 0 

29:iv:°11—The 10 fleas recovered on the previous day were dis- 
sected; trypanosomes were seen in the extruded feces of one, in the 
rectum and feces of another (abundantly) and the rectum of a third 
(abundantly) ; of the last two, preparations were made of the rectal con- 
tents, and there were found pear-shaped crithidial, transitional, and 
post-crithidiai trypaniform individuals (Text-fig. 22). 
TEXT-FIG. 22. 


boon (AL 
bark OC) 


Various forms from the rectum and feces of two fleas of the 
batch of 29:iv:’11, Experiment 39 (see text). Note the trypano- 
some-forms (last 4 figs. to the right, second row). (x 2000.) 


10 proboscides injected into rat 307: result 0 
10 stomachs oa ‘S ty Talis ~ o5.8 4) 
10 recta 5 3 pes Neh Mgt Se 
1:yv:’l1l.—The fleas recovered 29:iv were dissected; in one of 
them, a male, minute crithidial individuals were seen in the stomach 
contents, but not in the rectal contents; no preparation was made. 
10 proboscides injected into rat 310 : result 0 
10 stomachs . = alll Be se (O) 
10 recta Se Fe oli tes) 
4:v:'11.—The 10 fleas recovered on the previous day were dissecte d, 
trypanosomes were seen in the rectum of one, in the extruded feces of 
another; no preparation made. 
10 proboscides injected into rat 313: result 0 
10 stomachs re = SeolAr .s0 i 
10 recta By - olor: pee O 


634 E, A. MINCHIN AND J. D. ''HOMSON. 


Summary of Experiment 39.—None of the rats inocu- 
lated with organs of the fleas which had been exposed to 
infection two, three, or four days previously became infected. 
On the fifth and seventh days inoculation of the recta pro- 
duced infections, while the inoculations of the stomachs were 
negative. On the tenth day, on the other hand, inoculation 
of the stomachs gave a positive, that of the recta a negative 
result. It is seen, therefore, (1) that the fleas first became 
infective on the fifth day, when also the post-crithidial try- 
panosomes were first found in preparations of the rectum ; 
(2) that the developmental forms which produce infection 
were present in the rectum, but not in the stomach, on the 
fifth and seventh days; and in the stomach, but not in the 
rectum, on the tenth day. 


(vii) The Developmental Forms of the Trypano- 
somes in the Flea are not infective when 
inoculated into the Rat during a period ex- 
tending from a short time (half an hour?) 
after being taken up by the Flea until the 
Developmental Cycle is complete. 


After we had shown, in Experiment 39 (see above), that 
the trypanosomes in the flea are not infective to the rat after 
they have been in the flea for two days, and that they do 
not acquire infectivity for five days after being ingested, 
we instituted a number of experiments with a view of dis- 
covering how soon the ingested trypanosomes lose their power 
of intecting. 

In our first experiment a number of fleas collected from our 
non-infected breeding-cage and kept hungry for three days 
(7: x1:711 to 10: xi:711), were put on a well-infected rat at 
6 am. (10:xi:’11) and collected two hours later. They 
were then dissected in batches, the stomachs of each batch 
being placed together on the same slide, teased up in a drop 
of salt-solution, drawn up intoan injection-syringe and inocu- 
lated into a clean rat. After the drop had been drawn up 


THE RAT-TRYPANOSOME, TRYPANOSOMA LEWISI. 635 


into the syringe the film of moisture left on the slide was fixed 
with osmic vapour and stained with Giemsa’s stain, in order 
to get an idea of the modification, if any, which the trypano- 
somes had undergone. 


The following were the batches dissected and injected : 


Trxt-FIG. 23. 


cv. 
=r 


Trypanosomes from the stomachs of fleas of the batches of 
10: xi: 11 (see text). 


(1) Four fleas, injected into rat 323 at 11.55 a.m. The preserved film 
showed trypanosomes of the ordinary blood-type. 

(2) Four fleas, injected into rat 324at 12.55 p.m. The preserved film 
showed trypanosomes of the ordinary blood-type (Text-fig. 23, a, b). 

(3) Four fleas, injected into rat 325 at 1.30 p.m. The preserved film 
showed trypanosomes for the most part unmodified, with » approxi- 
mated to N, others dwarfed slightly as if beginning to degenerate 
(Text-fig. 23, ¢, d). 

(4) Four fleas, injected at 2.35 p.m. into rat 326. The preserved film 
showed trypanosomes of the ordinary blood-type (Text-fig. 23, e, f). 

(5) Four fleas, injected at 3.30 p.m. into rat 327. The preserved film 


636 E. A. MINCHIN AND J. D. THOMSON. 


showed trypanosomes mostly modified, some rather lengthened out, 
others recurved (Text-fig. 23, g, h, z). 

(6) Three fleas, injected at 4.35 p.m. into rat 327. The preserved film 
showed trypanosomes mostly unmodified,some rather long (Text-fig. 23, 
j, k). 


All the results were negative, since none of the rats 
became infected. It is seen from the times of feeding and 
injecting the fleas that none of the trypanosomes had been 
in the fleas more than ten and half hours (6 a.m. to 4.35 p.m.), 
or less than three and half hours (8 a.m. to 11.30 a.m.). 

After obtaining this result we made a number of other ex- 
periments, modifying slightly our method of procedure. Fleas 
taken from the non-infected breeding-cage were fed under 
observation on a well-infected rat and the time of feeding 
noted. The flea was recovered, dissected, and its stomach 
injected into a clean rat after being teased up in a drop of 
salt-citrate solution. As before, the film of moisture left on 
the slide was preserved in some cases. 

In some cases also some blood from the infected rat was 
inoculated subcutaneously into a control clean rat. The 
controls were not always positive, however, since the sub- 
cutaneous method of injection is notoriously less efficient 
than the intra-peritoneal method, when a small quantity of 
blood is taken direct from the rat. 


27: xi: ‘11.—Fleas fed under observation on an infected rat after 
having been kept hungry for three days. 

Fleas (1) and (2) fed at 11.35, injected into rat 329 at 12.10 (35 
minutes). The preserved film showed trypanosomes quite unmodified. 

Flea (5) fed at 11.45, injected into rat 329 at 12.20 (35 minutes). The 
preserved film showed trypanosomes quite unmodified. 

Flea (4) fed at 11.30, injected into rat 330 at 12.30 (1 hour). Try- 
panosomes in preserved film quite unmodified. 

Flea (5) fed at 11.50, injected into rat 330 at 12.50 (1 hour). Try- 
panosomes seen in the fresh stomach, but film not preserved. 

Flea (6) fed at 11.50, injected into rat 330 at 12.50 (1 hour). Film 
badly preserved. 

Flea (7) fed at 10.45, injected into rat 331 at 12.45 (2 hours). Try- 
panosomes in preserved film quite unmodified. 


THE RAT-TRYPANOSOME, TRYPANOSOMA LEWISI. 637 


Summary. 


Rat 329 was injected with 3 fleas 35 minutes after feeding. 


99 = . 
880) iy = oe a) al hour Ay ry 
ess: . 1 flea 2 hours after feeding. 


The results in all three cases were negative. No control 
rat was injected. 


8:xii: 711.—Fleas fed under observation on an infected rat (322). 

Fleas (1) and (2) fed at 11.40, injected into clean rat 324 at 12.26 
(36 minutes). Active trypanosomes seen in the fresh stomachs. 

Fleas (3) and (4) fed at 11°53, injected into rat 324 at 12.37 (44 
minutes). Trypanosomes seen in the fresh stomachs. 

Fleas (5), (6), and (7) fed at 12.4, injected into rat 324 at 12.42 
(38 minutes). Trypanosomes seen in the fresh stomachs. 

Fleas (8), (9), and (10) fed at 12.10, injected into rat 324 at 12.47 
(37 minutes). Trypanosomes seen in the fresh stomachs. 

Flea (11) fed at 12.20, injected into rat 324 at 12.55 (385 minutes). 


Summary.—Rat 324 inoculated with the stomachs of 
eleven fleas fed on rat 322 between 35 and 44 minutes pre- 
viously. Result negative. 

Control rat 323 inoculated at the same time with a small 
drop of citrated blood from rat 322. Result positive (found 
to be infected on 15 : xi). 


24:i:°11—Hight fleas, kept hungry for three days previously, were 
fed under observation on an infected rat (No. 392); the stomachs of six 
inoculated into clean rat 411. 

Flea (1) fed 10 a.m, injected 11 a.m. (1 hom). 


Flea (2) ., 10.10 a.m., injected 11.10 a.m. (1 hour). 
Flea () ,, LO0:I5 . a 11.14 ,, (59 minutes). 
Flea (4) ,, 10.20 ,, un EO) Gvhour): 
Flea (5) ,, 10:25 _., ae 2S 3 
Bieai(G) - 2029) 5 5 20) 


Fleas (7) and (8), fed 10.33 a.m. and 10.38 a.m. respectively, were 
dissected and fixed preparations made of them; numerous trypano- 
somes, quite unmodified in appearance, were found in them. 


Summary.—Rat 411 inoculated with the stomachs of six 
fleas, each of which had fed on infected rat 392 an hour 
previously. Result negative. 


638 E. A. MINCHIN AND J. D. THOMSON. 


Control rat 412 inoculated with a small drop of citrated 
blood from rat 392 at 12 noon. Result negative. (N.B.— 
Rat 412 put into the infected breeding-cage on 15:11: 713 
contracted an infection in due course and was therefore not 
naturally immune.) 


11: ii: ’°13.—Seven fleas fed on infected rat 402 under observation. 
The stomachs inoculated into clean rat 414. 
Flea (1) fed 10. 9 a.m., injected 11. 7 a.m. 


Flea) 1012, ap, dlalOees 
Fleas). .,10A7 ~,, Jer oleh yeas 
Pilea (4) ,, 10.25 ., au Ob yeep 
Hlea (5) ,,.1028 4, dee leo 
Eiea6) 1097 ., i HESS 
Hiea(7) ..1058- 5, . ss 


(In all cases the time of feeding was reckoned from the moment the 
flea withdrew its proboscis.) 

Some of the blood-was allowed to escape from the stomach of flea (7) 
and permanent preparations made; numerous trypanosomes quite 
unmodified in appearance were found. 


Summary.—Rat 414 acquired no infection; control rat 
415, inoculated with one drop of blood from rat 402, became 
infected (trypanosomes first seen, 21 : 11: 713). 


25 : ii :’138.—One flea was fed under observation at 10.45 a.m. on 
infected rat 415; the stomach was inoculated into clean rat 419 at 11.45. 
Results negative. (The other fleas refused to feed.) 

27:i1:°13.—Hight fleas fed under observation on infected rat 415; 
the stomachs inoculated into clean rat 419,in each case 59 minutes or an 
hour after feeding. Result negative. 

Control rat 414 inoculated with a drop of blood from rat 415. Result 
negative (rat 414 was later infected by being put into the infected 
breeding-cage.) 


It will be seen from the foregoing accounts that all experi- 
ments, in which infected blood ingested by a flea was mocu- 
lated into a clean rat from about half an hour onwards after 
ingestion by the flea, gave uniformly negative results. ‘The 
controls were sometimes negative, sometimes positive. 


THE RAT-TRYPANOSOME, TRYPANOSOMA LEWISI. 639 


(ix) The Flea, when once it has become infective, 
remains so for a considerable Length of Time. 


This point was dealt with in our preliminary account (1910), 
when it was shown in two experiments (“C” and “D 7’) that 
cages of fleas when once rendered infective continued to pro- 
duce infections for some time without being re-infected. 
Experiment 22 (‘‘C’’) was continued for some time after the 
publication of our paper, but on 28:1:710 an infected rat 
was unfortunately introduced into the cage, an oversight 
which vitiated all subsequent results, and the experiment was 
discontinued. During the period prior to this accident, how- 
ever, the experiment was not open to objection, and produced 
a result which may be summarised as follows. A cage 
colonised with 160 clean fleas, into which an infected rat was 
introduced for three days (24 to 27 : xi: ’09), and which 
produced the first infection between 30 : xi: ’09 and 3: xii: 09, 
continued to produce infections without being re-infected up 
to 24:1:710, a period of approximately 55 days. It is, 
therefore, a safe conclusion to aflirm that the infectivity 
persisted in some fleas, or at least one flea, for that period 
of time. 

Later a more exact experiment (Experiment 37) was carried 
out, in which single fleas were used. ‘To begin with, 10 fleas 
were taken from the infected breeding-cage and put each on 
a clean rat for four days; at the end of that time the 10 fleas 
were recovered and each flea put by itself in a separate test- 
tube in damp sand for six days. Meanwhile the 10 rats 
were examined daily, and in due course two of them 
developed infection, the remaining eight being negative. 
The eight fleas that gave negative results were returned to 
the infected breeding-cage. 

The two fleas that had been proved experimentally to be 
infective—henceforth known as Flea “A” and Flea “B”?— 
were then placed singly on a succession of clean rats. That 
is to say, each flea was placed by itself on a clean rat for so 
many days, then was recovered and placed on another clean 
rat for so many days, and so on. ‘I'he experiments with each 

vou, 60, PART 4,—NEW SERIES. 44 


640 


E. A. 


MINCHIN AND J. D. THOMSON. 


flea were continued until the flea disappeared: that is to say, 
until the flea could not be recovered when sought for. 
The details and results of Experiment 37 have been sum- 


marised in tabular form in our preliminary communication 


(1911), but since we have frequently had occasion to refer 
to the experiment in the present memoir, we think it worth 
while to reproduce the table already published. 


Taste G.-Summary of Experiment 37. 


Flea “A.” 


Trypanosomes. 
Flea on the rat. Result. |— = 
First seen in blood. Multiplication ended. 
14-18:1:711 — Q4:11:°11 Py san 91 (4L 
24: ii- il iil: iT: 0 
1-6: i117 711 + 15:11: 711 (Rat died, 
16: iii: 11) 
6-13 : 11:11 0 
13- hg :ii: Ll 0 
yee sir 211i _ Dif = Ws AL 31:i1:°11 
23-— see 0) 
yf eriell aie 3 lit + Div Ul 9:iv:’1l1 
Hes sree A 0 
5-10 av. Lt 0 
10- Deve al 0 
52.0 ave on 22 :iv:711 95 :iv: 711 
D0=Dbiiv eel + iboyeo- 1 Dewi wl 
95-29 :iv:711 i esyeevlal! 6:v:71l 
29 :iv-4:v:711 0 
4-9: v:°11 ? \(Rat died, 9:iv:"11) 
9215): vi: 1 0 
(l5:v ot ‘flea not recov ered. ) 
RVieane tae 
14-18 :i1: 711 = 94:1: "11 28:11:11 
94:ii-1: ii: 711 0) 
1-6 :11:711 0 
6-13 : 111: 711 0 
13=07 21 271 0 
pee oii 1 0 
23-202 Te + 30:10: 711 4’:iv : 711 
97 :ii-L:iv:’Ll + Siawiemlal 1 .Avieoiel: 
Neate 971i! 0 


5-10:iv: 711 
(LO cance* 


0 


11 flea not recovered.) 


THE RAT-TRYPANOSOME, TRYPANOSOMA LEWISI. 641 


From the tabular summary it is seen that flea “A” 
remained infective from about 15: 11:’11 to about 26: iv: 71] 
—that is to say, for a period of 70 days—and flea “ B” from 
about 16: i1: 711 to about 27 : 11: 711—a period of 40 days. 
We have no data for determining the maximum period of 
time during which a flea can remain infective. It is, perhaps, 
not improbable that a flea, once rendered infective, may 
remain so as long as it lives, but we have no knowledge with 
regard to the average longevity of the flea. Flea “A” (?) 
lived under our care from 14:11: 711 to 9: v :’?11—a period 
of 84 days; but we have no clue as to its age when it was 
first taken from the breeding-cage. 


(x) The Trypanosome does not penetrate into the 
Salivary Glands of the Flea, but is confined, 
during its whole Development, to the Diges- 
tive Tract. 

To prove this point we began by dissecting out salivary 
glands of fleas taken from the infected breeding-cage and 
examining the glands both in the fresh condition, by teasing 
them up or crushing them with a coverslip, in a drop of salt- 
citrate solution, and also by means of fixed permanent smears 
of the glands. In a large number of glands examined very 
carefully in this way many yeast-like organisms and other 
similar bodies were found, but never anything that appeared 
in the least like any possible stage of a trypanosome. 
Examination of fluid from the body-cavity gave negative 
results also. 

Thinking that we might have failed in these examinations 
to obtain an infective flea, we took fleas from the infected 
breeding-cage, put them singly on clean rats for three days 
or so, and then recovered and dissected them. The organs of 
the fleas were examined carefully in the fresh condition, and 
in some cases permanent preparations were also made of 
them and laidaside. If arat became infected subsequently, 
and so proved the flea put on it to have been infective, 
the preparations of that flea were searched very caretully. 


642 KE. A. MINCHIN AND J. D. THOMSON. 


In this way we were able to examine the salivary glands 
and other organs of fleas which had been proved experi- 
mentally to be infective. The following are the details of the 
experiments and observations; the sign + signifies that 
trypanosomes were found in the organs mentioned, while 0 
means that none were found: 


(1) Flea ($) put on rat 212 for three days (19 : ix : °10 to 22: ix 710, 
Experiment 29). Rat 212 found to be infected 28 : ix; multiplication 
ended 30: ix. Flea dissected 22: ix; proboscis 0, salivary glands 0, 
stomach + (Fig. 17), rectum 0, intestines 0. 

(2) Flea (¢) put on rat 223 for four days (10: x: 710 to 14: x:’10, 
Experiment 32a). Rat 223 found to be infected 17: x ; multiplication 


TEXT-FIG 24. 


Yeast-like bodies of various kinds from the salivary glands of a 
flea. They are shown in groups, as they were found in the 
preparation. (x 2000.) 


ended 22: x. Flea dissected and examined 14: x; no trypanosomes were 
seen in the proboscis, body-cavity, stomach, intestine, rectum or salivary 
glands ; but unfortunately no permanent preparations were made. 

(3) Flea (2) put on rat 233 for five days (20: x: 710 to 25: x: 710, 
Experiment 32e). Rat 233 found to be infected 28 : x; multiplication 
ended 31: x; scanty infection, with few trypanosomes. Flea dissected 
and examined 25 : x; no trypanosomes seen in proboscis, body-cavity, 
stomach, intestine, rectum or salivary glands; permanent preparations 
made of salivary glands, 0. (N.B.—Many yeast-like bodies in the 
salivary glands, see Text-fig. 24.) 

(4) Flea (¢) put on rat 238 for three days (1 : xi: °10 to 4: xi: “10. 
Experiment 32f). Rat 238 found to be infected 14: xi; multiplication 
ended 16: xi. Flea dissected and examined 4: xi; no trypanosomes 
were seen in the proboscis, stomach, intestine, rectum or salivary 
glands; permanent preparations made of the stomach and salivary 
glands, both 0. 


THE RAT-TRYPANOSOME, TRYPANOSOMA LEWISI. 643 


From the foregoing experiments it is seen that nothing 
which could be recognised as a trypanosome was found in the 
salivary glands of four fleas known to have been infective. 
In one of the four fleas the salivary glands were examined 
only in the fresh condition, but in the other three fleas the 
salivary glands were examined both in the fresh condition 
and in the permanent preparations. 

It will be remarked, however, that no trypanosomes were 
found in any part of the digestive tract, when examined fresh, 
in three out of the four fleas, although there can scarcely be 
any doubt that trypanosomes must have been present. Asa 
matter of fact, a scanty infection of the crithidial or final try- 
panosome-forms, small in size and sluggish in movement, is 
easily overlooked in the fresh preparations of the teased-up 
digestive tract, especially in an organ relatively so large as 
the stomach, which may be gorged with blood-débris greatly 
hindering and obscuring the examination. It has been our 
experience not infrequently that the smaller forms of the 
cycle have been found scantily in permanent preparations of 
stomachs in which nothing was seen in the fresh examination. 
This scarcely applies, however, to organs so minute as the 
salivary glands, the contents of which can be scanned com: 
prehensively in one field of the microscope. We have, there+ 
fore, cited the second flea in spite of the fact that no per- 
manent preparations of the salivary glands were examined. 

After having made the negative observations recorded 
above, the idea occurred to us that any form of the trypano- 
some found in the salivary glands would probably be a final 
stage of the cycle, destined to be inoculated by the flea 
through the proboscis into the rat; and that consequently the 
examination of fleas which had produced an infection recently 
would be inconclusive, since in such fleas the salivary glands 
might be purged of their infection, temporarily at least. 

We therefore carried out some experiments in which the 
object was to determine which organs of the flea contained 
the infective stages of the trypanosome, by dissecting fleas 
taken from the infected breeding-cage and injecting their 


644. E. A. MINCHIN AND J. D. THOMSON. 


organs separately into clean rats. Thus a batch of fleas, at 
least five in number, was taken from the infected breeding- 
cage, and all the fleas in the batch were dissected at the same 
sitting. The stomachs of all the fleas were put together in 
one watch-glass and the salivary glands in another. Each 
flea has four salivary glands (two on each side of the body), 
but we did not succeed in every case in dissecting out all 
four of these minute organs; sometimes only two or three 
were obtained, or even one only of the four (on foggy morn- 
ings) ; but in every case at least one of the four glands was 
obtained. In the case of the recta (cut off behind the pylorus 
and therefore including the greater part of the intestine 
as well), each was examined microscopically and only kept 
for injection if seen to contain trypanosomes. 

In our earlier experiment (Experiment 33) only the salivary 
glands were used for injection, with the following results : 


Taste H.—Experiment 33. 


) | ; 
No. of fleas | No. of salivary | | 
| ete: dissected. glands injected. SLE oo ) 
i | i. ae 
| 17 saa: 13 ul 34 | 948 | <i 
| Doan 10 5 18 248 0 

/ 


Hence fifty-two salivary glands, dissected out from sixteen 
fleas taken from the infected breeding-cage, failed to infect a 
clean rat when injected into it. 

Our later experiment (Experiment 34) was carried out more 
elaborately. In the first place each batch of fleas, after 
having been collected from the infected breeding-cage, was 
first put on a clean rat, as a control. Then the fleas were 
recovered from the control rat and dissected; the salivary 
glands and stomachs were then injected into two clean rats 
respectively, and the recta, if anything was seen in them, 
into a third. The details of Experiment 34 are shown in the 
following table: 


645 


THE RAT-TRYPANOSOME, TRYPANOSOMA LEWISI. 


OUNUTUME ATPRANGLU JOU ELOFOLOY]} OAM pu ‘syToUTTedx quonbosqns Url poyosur oUBdEd y POYAVU SPRe OUT, 


| 
ay CSGa| ae oa a 0) xS9G q 0 GOG |) cS aalilliee tre (ce tomer een sO eS 
ar SGa| se = aa Ir L9G 9 0) Gilt ele se meets Lene Mammen Nr le Mabe ly 
0 0co =O 896 il 0 ¥O96 9 0) woGe | FO MDs “OCMieh Sali b volige 
ar LkG 0 696 G aR LVG 9 0) Sho | roe lle hon nee OND ie swans 
0) * 1&6 | — on = an GPG a 0 rte yea MIE MGs ets cat TIULR OTN Oe sepals 
si 686 | —_ aS a ar LGG 9 0 9¢G | 2G OT. * HX‘ STOL.* HX + STOOL. HX «ol 9 
0 686 0 GGG L 0 LGG 9 0 9Ga| FS Ola t= *ScOl = “Sc0le = “ccd 
. . ‘yoqoal Ho \lenaqoel a hu eereet “qua 
teceeter joxyti09 “qnsoxy aes ie myo9 “‘qmsoy ( re stowrtions “‘V[USOYy OM spr eee beetemce oy corae otent or ae 
| jo ON | yo'on | Jo'ON | Jo'‘ON | JO'ON yo ‘ON, ae i SvOLA ‘ON 


‘pe quowutadxyf—'] aay, 


646 E. A. MINCHIN AND J. D. THOMSON. 


From the details summarised in the above table it is seen 
that no infection was produced by 150 salivary glands taken 
from forty fleas, divided into seven batches. In three of 
these batches, comprising eighteen fleas, from which sixty- 
eight salivary glands were obtained, both the control rats and 
the rats infected with the stomachs! became infected. Ina 
fourth batch, comprising five fleas from which nineteen 
salivary glands were obtained, the control rat was negative, 
the rat injected with the stomachs was positive. In a fifth 
batch, comprising five fleas from which fifteen salivary glands 
were obtained, the control rat became infected but the 
stomachs produced no infection ; the infectivity of this batch 
must have been in the recta, which unfortunately were not 
injected. Thus, omitting the two batches that gave negative 
results throughout and reckoning only with five batches 
proved to contain infective fleas, it is seen that 102 salivary 
glands obtained from twenty-eight fleas produced no infection. 
hese results convinced us finally that the salivary glands of 
the flea play no part whatever in the transmission or develop- 
ment of the trypanosome, and from this time we paid no 
further attention to them. 


(xi) The Rat can become infected by eating in- 
fected Fleas, but not until the Developmental 
Cycle of the Trypanosome in the Flea is com- 
pleted. 


The fact that rats can become infected by eating infected 
fleas must now be considered as well established. It was first 
stated in print by Strickland (1911), but was then already 
known to us both from experiments performed by ourselves 
and from others carried out by Dr. Nicoll at the Lister 
Institute (see our preliminary report, 1911). 

We carried out some further experiments to determine 
whether the rat could become infected in this way before the 

1It should be noted that the stomachs here included the post- 


pyloric upper end of the intestine, which, as shown above, is often the 
site where the trypanosomes establish themselves. 


THE RAT-TRYPANOSOME, TRYPANOSOMA LEWISI. 647 


developmental cycle of the trypanosome was completed in the 
flea. The experiments (lHxperiment 46) were carried on for a 
long period, each experiment occupying a week. ‘lhe following 
sample is typical of the whole series, all being performed in 
the same manner and each corresponding stage of the experi- 
ment being carried out on the same day of the week. 


Sunday, 20:iv:°13.—A large batch of fleas (batch 6) collected from 
the non-infected breeding-cage the Friday previously were put on a 
well-infected rat. 

Tuesday, 22:iv:°13.—About fifty fleas were collected from the 
batch (6) and fed to the “two-day rat,” clean rat 445. The method 
was to place the fleas, when collected, on the surface of water in a suit- 
able vessel; then each flea was carefully picked off with a fine forceps, 
and either it was decapitated on a slide in a drop of water with a needle; 
or its head was crushed with the forceps. Several fleas so treated were 
stuck into a pellet of damp bread and given to the rat, previously kept 
hungry fora time. Asarule the rats when fed in this way ate both 
the bread and the fleas readily and even greedily. 

Thursday, 26:iv:°13.—About fifty more fleas of the infected 
batch (6) were collected and fed to the “ four-day rat,” clean rat No. 448, 
in the same manner. 

Saturday, 24:iv.’13.—About fifty more fleas of the infected 
batch (6) were collected and fed to the “six-day rat,” clean rat 432, in the 
same manner. 

On the Sunday following afresh batch (batch 7) was exposed to infec- 
tion in the same way. On the Tuesday following the “two-day rat” 
(No. 445), being found not to have become infected, was fed again with 
about fifty fleas of batch 7. On the Thursday following the four-day 
rat (No. 448), not having become infected, was fed again with about 
fifty fleas of batch 7. On the Saturday following (8:v:'15) the six- 
day rat (No. 432), which then showed no trypanosomes in its blood, was 
fed with thirty-four fleas of batch 7. Rat 482 was found, however, to 
be showing trypanosomes in its blood when examined four days later 
(7:v:°13); it must have been infected by the fleas of batch 6. For 
batch 8, in the following week, afresh clean rat, No. 452, was appointed 
to be the new six-day rat; it later became infected by the six-day fleas 
of batch 12. 


‘'hese experiments were continued in regular routine in 
the manner described, from 16: 11:713 to 22:ii1:713, and 
from 20:iv:’13 to 12:vi:'13, in all thirteen weeks and 


648 E. A. MINCHIN AND J. D. THOMSON. 


thirteen batches. The results may be summarized briefly : 
No rat fed either with fleas exposed to infection two days 
previously, or with fleas exposed to infection four days pre- 
viously, became infected; on the other hand, two of the six- 
day rats fed with fleas exposed to infection six days 
previously became infected. It may be inferred, therefore, 
that rats cannot be infected by eating infected fleas until the 
infection in them is ripe, that is to say until the develop- 
mental cycle of the trypanosome in the flea is complete. 

(It may be mentioned here that when any fleas remained 
over from any of the batches used in this experiment they 
were fed under observation on clean rats on the Monday, 
Tuesday or Wednesday following, that is to say, eight, nine 
or ten days after they had been first exposed to infection 
(see p. 654, below).) 


(x11) Infection of the Rat is effected contamina- 
tively, by way of the Rat’s Mouth, by the Rat 
licking from off its Fur or Skin the Moist 
Feces of Infective Fleas containing the Final 
Propagative Form of the Cycle. 


This mechanism of infection was first demonstrated by 
Noller (1912) and fully confirmed by Wenyon (1913) by 
means of experiments which put the matter beyond all 
reasonable doubt. Without repeating the experiments of 
these authors, we tested their results by another method, 
namely, by exposing rats, muzzled and pinioned, to infection 
by a large number of infected fleas. 

For the purpose of our experiments, we made use in most 
cases of our infected breeding-cage, in which the fleas were 
swarming in great numbers, and in which an infected rat is 
kept habitually, so that the fleas ingest blood containing try- 
panosomes every time they feed. As a preparation for the 
experiment the infected rat was removed from the breeding- 
cage and kept apart, all fleas found on it being carefully 
cleaned off and returned to the breeding-cage. The breed- 


THE RAT-TRYPANOSOME, TRYPANOSOMA LEWISI. 649 


ing-cage was then left without a rat in it for a certain time, 
sometimes merely from morning to evening, in other cases a 
day or two, in order to induce hunger in the fleas. Then a 
clean rat was introduced into the cage for a single night or 
from morning to evening, after having been muzzled in the 
following manner: ‘I'he muzzle was a conical cap of fine wire 
gauze, with meshes too fine for a flea to pass through. The 
cap was large enough to cover the whole head, including the 
ears ; the opening of the cone had a broad rim or sleeve of 
soft cloth, and a draw-tape was run through the free edge of 
the sleeve, so that by pulling the two ends of the tape the 
sleeve could be drawn up as tight as required. When fixing 
the muzzle, it was slipped over the head and then the sleeve 
was tightened round the neck behind the ears; the two free 
euds of the tape were then passed downwards and forwards 
over the chest and backwards under each axilla; each end of 
the tape was given a single turn round the upper joint of the 
fore-leg of its side and then passed backwards and upwards 
to be tied to the other end of the tape from the other side of 
the body over the back of the rat. In this way, not only was 
the head of the rat muzzled so that it could not lick itself or 
eat fleas, but owing to the fore-legs being secured firmly it 
could not use them to tear off its muzzle, which the rat 
always makes violent efforts to do, and which, in spite of 
all precautions, it sometimes succeeds in doing. 

The clean rat, having been muzzled in the manner described, 
and exposed to the attentions of the fleas for a certain time, 
was removed from the breeding-cage, and before being 
unmuzzled it was subjected to a cleansing process which con- 
sisted of removing all fleas from it and of washing its fur all 
over thoroughly with a disinfectant (lysol, about 2 per cent.). 
‘he liquid used for washing usually became coloured reddish- 
brown from the feces deposited on the fur by the fleas. 
After the rat had been cleansed thoroughly, its fur was dried 
by holding it close to an ordinary electric lamp and then its 
muzzle was removed and it was allowed to lick itself as much 
as it wished, being kept apart from all fleas and its blood 
examined at regular intervals. 


650 E. A. MINCHIN AND J. D. THOMSON. 


As a variation of the above procedure the muzzled rat 
was not put into the infected breeding-cage in some 
instances, but into a bell-jar witha certain number of fleas 
taken from the infected breeding-cage and previously kept 
hungry. 

At first the experiments were controlled by putting 
unmuzzled rats into the infected breeding-cage, but it was 
found superfluous to do this, since in many of our actual 
experiments the muzzled rats succeeded in tearing off their 
muzzles and thus furnished improvised but very efficient 


controls. 


The following is a brief statement of the experiments (controls 
marked *) : 

(1) Rat 447, muzzled and put into the infected breeding-cage from 
4.30 p.m., 23 :iv:°18, to 10 am., 24:iv:°13. Examined from 28: iv to 
30: v, not infected ; inoculated later from wild rat and acquired infec- 
tion in due course. 

(2) Rat 449, muzzled, put into infected breeding-cage for night of 
25-26 :iv:°13. Examined up to 30:v, no infection; put into infected 
breeding-cage, 50: v, became infected in due course. 

(5) * Rat 450, muzzled, put into infected breeding-cage evening of 
30 :iv:°15: found next morning with its muzzle off; it became infected 
in due course. 


After the above-mentioned experiments had been performed 
in the manner described, namely, by cleaning and disinfecting 
the rat before unmuzzling it, a change of procedure was 
adopted. The rat that had been muzzled and put in the 
infected breeding-cage was not disinfected when taken out, 
but, after having been freed from fleas, its fur was merely 
dried thoroughly by holding it near an electric lamp before 
unmuzzling the rat and allowing it to lick its fur. This 
was in order to see whether dried feces would produce an 
infection when licked off. 


(4) Rat 467, muzzled, put into infected breeding-cage, 8 a.m. to 4 
p-m.,16:i1:°14 (the infected rat having been removed from the cage two 
days previously). Examined up to 5: iii, no infection. 

(5) Rat 468, muzzled, put into infected breeding-cage 10 a.m. to 
2p.m.,19:ii:°14. Procedure otherwise same as in last, no infection. 


THE RAT-TRYPANOSOME, TRYPANOSOMA LEWISI. 651 


(6) *Rat 469, muzzled, put into infected breeding-cage, 7 p.m., 
20:11:14; found next morning (8 a.m.) with muzzle off; became infected 
in due course. 

(7) Rat 470, muzzled, put into infected breeding-cage 11.50 p.m., 
23:ii:714, treated as 467, etc.; no infection. Muzzled and put into 
bell-jar with 200 hungry fleas taken from infected breeding-cage 9 a.m. 
to 2.30 p.m., 3: iii: 14; no infection. 

(8) * Rat 472, muzzled, put into bell-jar with 200 hungry fleas from 
the infected breeding-cage, 9 a.m., 26:i1:°14; found at 2.30 p.m. with 
its muzzle off; became infected in due course. 

From the data quoted, it is seen that when the experiments 
were carried out successfully, that is to say, when the muzzle 
kept on, the rat did not become infected, alike whether its 
fur was cleaned and disinfected, or merely dried, before it 
was unmuzzled. But in those cases in which the rat succeeded 
in ridding itself of its muzzle by its own efforts, it became 
infected in due course. 

It is evident that the sole effect of the muzzle is to exclude 
infection of the rat by way: of its mouth. ‘The way is still 
open for the rat to become infected through the skin, either 
(1) by the trypanosomes passing in through the puncture 
made by the proboscis of the flea, a possibility suggested by 
Noller; or (2) by the faeces being rubbed into wounds or 
abrasions on the skin; or (3) by the small trypanosomes in 
the feeces penetrating by their own efforts through the skin. 
But since all the results with rats muzzled efficiently were 
negative, it is evident that no infection per cutem by any 
of these possible ways occurred in these experiments and it 
becomes highly probable that it does not take place naturally 
in any of these ways. On the other hand, since the muzzle 
excludes only infection per os, it becomes also probable that 
the rats that succeeded in tearing off their muzzles were 
infected in this manner. 

It may be mentioned finally that an experiment (Hxperi- 
ment 43) was carried out in which the feces, collected over- 
night in a moist glass capsule, of fleas from the infected 
breeding-cage were injected under the skin of a clean rat, but 
the result was negative. 


652 E. A. MINCHIN AND J. D. THOMSON. 


(xiii) Can the Flea infect the Rat by inoculating 
the Trypanosomes into it through the Pro- 
boscis? 


The first to throw doubt upon the occurrence of this mode 
of transmission were Strickland and Swellengrebel (1910, 712), 
who made a number of attempts to infect rats by feeding 
infected fleas on them through gauze. Every such experi- 
ment gave negative results. Werepeated these experiments, 
and always with the same negative results. 

We also carried out a large number of experiments in 
which fleas taken from the infected breeding-cage were fed 
on clean rats under observation. Our course of procedure, 
in its latest and most highly elaborate form, was as follows: ! 

A certain number of fleas, collected from the infected 
breeding-cage, or fed on an infected rat more than seven days 
previously, or known experimentally to be infective, were 
kept in a flask containing some damp sand for three or more 
days, to make them hungry. When the experiment was 
about to take place the fleas that it was proposed to use were 
put each into a separate test-tube. A clean rat was prepared 
for the experiment by shaving a small region of its skin, usually 
on the belly, sometimes on the inside of the thigh (a favourite 
spot for fleas to feed). The rat was then held still by an 
assistant, with its tonsure upwards. ‘lhe test-tube was inverted 
on to the tonsure and at first kept pressed upon it; the flea was 
thus emptied out on to the shaved area of the skin. At first the 
flea runs round and round inside the circle formed by the rim of 
the test-tube, but usually comes to rest very soon, and inserts 
its proboscis into the skin. In some cases, however, the flea 
refuses to feed, and either continues to run about or remains 
perfectly still in one place, presenting a deceptive appearance 

1 We did not adopt Néller’s method of tethering the fleas, since it 
seemed to us better that the flea should be free and unhampered in its 
movements, and would then be more likely to feed in a natural way. 
Our colleagues, Dr. Martin and Mr. Bacot, who were doing experiments 
at the same time on transmission of plague, also found it unnecessary 
to tether the fleas. 


THE RAT-TRYPANOSOME, TRYPANOSOMA LEWISIT. 653 


of being engaged in feeding, but when recaptured and 
examined it is found to have been merely thinking. 

In fleas carefully watched it is not difficult to observe with 
a hand-lens the penetration of the proboscis into the skin, the 
rush of blood into the stomach of the flea, and the withdrawal 
of the proboscis when the flea is replete. From a number of 
fleas that were timed carefully it was found that the male flea 
took about 14 minutes to fill its stomach, the female about 

+ minutes. In all the fleas we have fed under observation 
we have never once observed the flea to defecate while 
feeding, though particular attention was directed to this 
point. Only in one case, when the fleaafter feeding succeeded 
in making good its escape into the fur, it was found to have 
defzecated there.! 

While the flea is feeding an assistant holds over it a paint- 
brush dipped in a thick syrupy solution of sugar and water. 
As soon as the flea has filled its stomach it withdraws its 
proboscis and makes a rush for the fur of the rat, but as soon 
as it does so the assistant dabs the paint-brush down on it 
and catches it. From the paint-brush, to which it sticks, the 
flea is put in water, which cleans off the sugar-syrup ; it can 
then be dissected or put back in the cage, none the worse for 
its adventure. If it succeeds in getting into the fur of the 
rat it must be recaptured, and any feces it may have deposited 
must be washed off with a disinfectant. 

Fleas that refuse to feed can either be put back in their 
test-tubes and given another chance on the following day, or 
dissected and examined as controls. 

In the majority of cases the fleas fed under observation 
were either fleas taken at random from the infected breeding- 
cage or fleas which had been fed on an infected rat at a 


1 Our experience of the feeding habits of Ceratophyllus fasci- 
atus does not agree in the least with the account given by NOller for 
Ctenocephalus canis. The habits of the two species are evidently 
quite different. The former lives in the burrows of the rat, and only 
goes on to the rat in order to feed, while the latter lives more or less 
permanently in the fur of the dog. 


654 E. A. MINCHIN AND J. D. THOMSON. 


definite time, and of which the age of the infection was known 
exactly. In a few cases we used fleas which had been put 
singly on clean rats and had produced an infection, and were 
therefore known to be infective. Thus rat 367a had a 
“known infective” flea fed on it on 26: vii: 712 and again 
on 29 : vil: 712, and another such flea was fed on the same 
rat on 3: villi: 712, but the rat did not become infected 
(Experiment 44a). 

We have in our notebooks records of 150 fleas fed on rats 
under observation in this way. Since the results were 
uniformly negative it is quite unnecessary to refer to them in 
further detail, but it may be of some interest to mention 
some cases in which the fleas were dissected and examined 
immediately after the experiment. These were fleas left over 
from the batches used in Experiment 46 (see above, p. 648), 
and were therefore, all of them, fleas in which the age of the 
infection was known. 


1] : iii : *13.—Six fleas left over from batch 3; infection of the fleas 
nine days old. Five fleas fed, in three of which no trypanosomes 
were seen, in a fourth the stomach contained an infection of crithidial 
forms, and in a fifth both stomach and rectum contained crithidial 
forms. (In a sixth flea, which had not fed, both stomach and rectum 
also contained crithidial forms.) 

12: 11: °13.—Four fleas of the same batch as yesterday; infection 
ten days old. All fed. In two nothing was seen, the other two had 
erithidial forms in the rectum. 

13 : iii : °"15.—Two fleas of the same batch as last; infection eleven 
days old. One, which fed, had a few crithidial forms in the rectum. (In 
the other, not fed, nothing was found.) 

17 : iii : °"13.—Four fleas of batch 4; infection eight days old. One 
which fed, had a swarming infection of the rectum and a few crithidial 
forms in the stomach. (Of the three which did not feed all had 
crithidial forms in the rectum, one in the stomach also.) 

18: iii : °18.—Three fleas of the same batch as last; infection nine 
days old. One, which fed, had crithidial forms in the stomach. (Of 
the two which did not feed, in one nothing was found, the other had 
scanty crithidial forms both in stomach and rectum.) 

19 : iii : °15.—Six fleas of the same batch as last; infection ten days 
old. Four fed. In one nothing was found; in two there were crithidial 
forms in the rectum only; in the fourth there were crithidial forms in 


THE RAT-TRYPANOSOME, TRYPANOSOMA LEWISI. 655 


the stomach only. (Two did not feed. In one of them nothing was 
found, in the other there were crithidial forms in the rectum only.) 

5:v:713—-Ten fleas of batch8; infection eight days old. Five fed. 
Of these three showed crithidias in the rectum only, one in the rectum 
and stomach and one was quite negative. (Five did not feed. Of these 
one had erithidial forms in the stomach only, one in the rectum only, 
one both in the stomach and rectum, and two were quite negative.) 

13: v 713.—Six fleas of batch 8, infection nine days old. One fed, 
five did not; the examination of all the six was negative in result. 

14: vy :°13.—Eight fleas of the same batch as last ; infection ten days 
old. Five fed, three did not; the examination in all cases was negative 
in result. 

19: v:°13.—Six fleas of batch 9; infection eight days old. Only 
one flea fed; all the six negative. : 

20: vy :°13.—Six fleas of the same as last; infection nine days old. 
Four fed, two did not; all six negative. 

26: v:°13.—Six fleas of batch10; infection seven and a half days old. 
One fed which was negative. (Of the five which did not feed, one was quite 
negative, three had crithidial forms in the rectum only, and one both 
in the stomach and rectum.) 

27: v : °13.—Six fleas of the same batch as last ; infection eight anda 
half days old. Of four that fed, two were quite negative, two had 
erithidial forms in the rectum. (The two that did not feed had 
crithidial forms, both in stomach and rectum.) 

2:vi: 13.—Six fleas of batch 11; infection eight daysold. Three 
that fed in two cases crithidial forms in the rectum only, the third was 
quite negative. (In the three that did not feed, two were quite negative, 
one had crithidial forms in the rectum only.) 

3: vi: 13.—Six fleas of same batch as last; infection nine days old. 
Two that fed were both negative. (In the four that did not feed, two 
had crithidial forms in the rectum only, one in the stomach only, one in 
both rectum and stomach.) 

9: vi: “13—Seven fleas of batch 12 ; infection eight days old. 
Three fed, two of which had crithidial forms in the rectum only ; one 
was quite negative. (All the four that did not feed had crithidial forms 
in the rectum, and one of them in the stomach also.) 


It is seen from the data that in many cases the fleas that 
fed on the rats contained copious infections in the rectum, 
the stomach or both; but not in a single case was any 
infection produced in the rats fed upon by the fleas. At the 
present time, therefore, the answer to the question posed at 
the head of this section must be a very decided negative. 

vot. 60, PART 4,—NEW SERIES. 45 


656 E. A. MINCHIN AND J. D. THOMSON. 


(xiv) Hereditary Transmission of the Trypanosome 
from Flea to Flea does not, in our Experience, 
take place. 


In order to obtain, if possible, fleas infected hereditarily, a 
clean, freshly-prepared breeding-cage was colonised with 648 
flea-larvee taken from the infected breeding-cage at various 
dates between 23 : ix : 710 and 26: x:’10. Adult fleas were 
first seen in the cage on the latter date, subsequently to which 
100 more larvee were added (3: xi: 710). 

Thinking, however, that the larve in the infected breed- 
ing-cage might possibly infect themselves directly from the 
feeces of adult fleas in the cage, we also colonised another 
clean breeding-cage with larve newly-hatched from eggs laid 
by fleas taken from the infected breeding-cage. The method 
was to take a certain number of fleas from the infected breed- 
ing-cage and keep them overnight in a glass capsule contain- 
ing a glass coverslip at the bottom. In the morning a certain 
number of eggs were usually found, some on the glass of the 
capsule, some on the coverslip. The fleas having been 
returned to the infected breeding-cage, the eggs were care- 
fully removed by means of a soft camel’s-hair paint-brush, 
slightly moistened, from the glass of the capsule and each egg 
was placed on a small piece of black paper, to which it 
adhered. At first the eggs, attached either to the coverslip 
or to the black paper, were put in the new breeding-cage and 
allowed to hatch there; between 1:x:’10 and 11: x: 710 
there were sixty eggs introduced in this way, but since they 
did not all hatch, another method was adopted. The eggs 
laid were kept in a glass capsule till they hatched, and then 
the newly-hatched larve were put into the breeding-cage. In 
this way 159 larve were introduced into the breeding-cage 
between 18: x:710 and 22: xi: 710. The time taken by 
the eggs to hatch varied between six to eight days in October 
and ten to twelve days towards the end of November (labora- 
tory-temperature). 

In the two cages colonised in this way clean rats were kept 


THE RAT-TRYPANOSOME, TRYPANOSOMA LEWISI. 657 


for a long time, but no infection was produced in either 
case, 


(xv) The Trypanosomes in the Blood of the Rat can 
render Fleas infective very soon after they 
make their First Appearance in the Blood, 
before their Multiplication-Period is over. 


In their paper on the life-history of Trypanosoma 
lewisi in the rat-louse, Breinl and Hindle (1909) state that 
they were unable to produce an infection of the louse when it 
was fed on the rat during the multiplication-period; they 
state that “during the first stages of infection, so long as 
dividing and segmenting forms were present in the blood, the 
trypanosomes taken up by the louse only degenerated.” 
Wishing to find if this was true for the flea also, we did two 
experiments (Experiments 26 and 28) in which a number of 
fleas were first fed on an infected rat during the multipli- 
cation-period of the trypanosomes. The fleas were then 
recovered and used to colonise a freshly-prepared flea-cage, 
into which a clean rat was put. In both cases the result was 
positive, showing that fleas can become infective after having 
fed on infected rats in which the trypanosomes are under- 
going multiplication. 

We then planned and carried out a more elaborate experi- 
ment (Experiment 40) to determine how soon a rat infected 
by fleas can infect fleas again. For this purpose rat 317 was 
placed for one day (16-16:v: 711) in the infected breeding- 
cage after the cage had been kept withouta rat in it for three 
days, to make the fleas hungry. Rat 317 first showed try- 
panosomes in its blood on 21:v; the multiplication-period 
was ended 25 : v or the following day. 

Meanwhile, a succession of eight flea-cages, numbered A to 
H, were colonised with clean fleas, taken from the non-infected 
breeding-cage, and rat 317 was put in each cage successively 
for one day, thus: 


658 E. A. MINCHIN AND J. D. THOMSON. 


Cage A, colonised with 70 fleas, 12: v; rat 317 put in 16-17: yv 


B ae aE Pe elae 7 17-18:v 
C D0 ne le: Wy 18-19:v 
D ze POO ~ 32: ne NG ay, 19-20:v 
E 4 Pipes lyk; 20-22: v7 
F ; peo eee. CUS iy, _ 22-93 :v 
G D0) 72. nl Qe 23-24:v 
H - ie OO “gss SoU em 24-25 :¥ 


After rat 317 had been taken out of each of the cages, a 
clean rat was put into the cage in its place and left in, with 
the following result : 


Rat 218 put in Cage A, 17:v; result 0 
», 264 Bi eb Say; pte 
. 266 a ow ee) 
» 267 3 oi) 20: ¥ aE) 
ODS Napoca anit 2 wr eee 
», 269 EF 23:¥ eed) 
SKY) - G 24:v . + Found to be infected, 
14: vi. 
A A - po) ME 25 iy, Be OY) 


From the above data it is seen that rat 317, exposed to 
infection 15-16:v, did not render any fleas infective before 
the batch that fed on it in cage G, 25-24: v, seven to nine 
days after it had been infected and two days after trypano- 
somes were first seen in its blood, ata time when the multi- 
plication of the trypanosomes was proceeding actively. The 
late appearance of the infection in rat 270 (the record of 
which indicates that infection took place about 6: vi, twelve 
days after it had been in contact with the fleas and thirteen 
days after the fleas themselves had been exposed to infection), 
indicates that only a small percentage of the fleas in cage G 
became infective. 

From these experiments we deduce that infection of the 
flea is not dependent on the presence of special propagative 
forms of the trypanosome produced late in the rat’s blood. 


THE RAT-TRYPANOSOME, TRYPANOSOMA LEWISI. 659 


(3) ProBLeMs OF SPECIAL NATURE. 

(xvi) The Trypanosomes succeed in establishing 
themselves in the Flea and rendering it 
infective to the Rat in only a Small Proportion 
of the Fleas (Ceratophyllus fasciatus) that 
ingest them. 


It has already been pointed out above, in the description of 
the developmental cycle of the trypanosome, that in the 
greater number of the fleas fed on infected rats the trypano- 
somes degenerate and die out completely, and that they 
succeed in establishing themselves in but a small percentage 
of the fleas. We have also tested this question experimen- 
tally by the method of taking fleas from the infective 
breeding-cage and putting these fleas on clean rats. At 
first we carried out such experiments by taking small batches, 
each of five or six fleas, from the infected breeding-cage and 
putting each batch separately on a clean rat. Eleven such 
experiments were performed with the results summarised in 
Table J, from which it is seen that three batches, each of five 
fleas, produced one infection, and that eight batches, each of 
six fleas, produced four infections. In each case, as in many 
of the subsequent experiments now to be recorded, the fleas 
were left on the rat three or four days on the supposition, 
which is probably correct, that a flea will become sufficiently 
hungry to feed in the course of three days, and on the further 
supposition, which has now proved to be incorrect, that the 
infection passes into the rat through the proboscis of the flea. 

The results of the experiments summarized in Table J are 
inconclusive, and could only permit of deductions approxi- 
mately exact if carried out in great number. When a batch 
gives a positive result there is no clue as to the number of 
infective fleas contained in it. Further, it has been brought 
home to us by subsequent experience that an infective flea 
often fails to produce an infection. Thus, comparing 'l'able J 
with Table I, it is seen that the batch of 5:1:711 failed to 
infect rat 231 (a rat infected subsequently in another ex- 


660 E. A. MINCHIN AND J. D. THOMSON. 


Taste J.—Experiments to determine the Percentage 
of Infective Fleas by taking batches of Fleas 
from the Infected Breeding-cage and putting 
each batch on a Clean Rat. 


l : | 
eat No. of | Pleas let S | tryps. fi Multipl. 
ment | Date, oan) ee | NG) | ee | 
| 
| 27 bis| 14:11:°10| 5 4 days ae 0 ms — saslane — i 
| ee ‘. 5 AT ye, 187 + | 21:i1:710| 2 : i 270) 
es . Sea Rt ee 188 | 0 a _ 
34 122:x1;°10) 6 3 239 0 — — | 
5 egsctig Iho) 6 3 239 + |19: xii: 710|22: xi: 710 
x 5:1:711 6 4, 231 0 — = 
(yang aul 6 AND as 241 SEP aero b yey 3 ali! 
erie UL 6 A 242 0 — a 
UMereeh | Al ks 250 0 —_ — 
7 6 A ss 251 + | 26:1:711 | 30:1:21 
| 5 -6 Alma: 252 se 25 c-cd 30 eae 


Compare also Table I above (p. 645). 


periment), but, nevertheless, the stomachs of this batch, 
injected into rat 245, produced an infection. 

In order to obtain results from which more exact conclu- 
sions could be drawn, we experimented in a different manner, 
acting under the advice of Dr. M. Greenwood, Statistician 
of the Lister Institute. We took batches of fleas from the 
infected breeding-cage and put them on clean rats, one flea 
on each rat. A large number of such experiments were 
carried out with the results summarized in Table K, from 
which it is seen that 115 fleas placed each on a clean rat 
produced but eleven infections, a percentage of 9°56 
approximately. 

It is evident, however, that the bare numerical results of 
the experiments summarized in Table K cannot be taken as 
final, for the following reasons : 

(1) The number of days given in the table as the time 
during which the flea was left on the rat is reckoned from 
the date the flea was put on the rat to the date on which the 
flea was sought for. But in many cases the flea when sought 


THE RAT-TRYPANOSOME, TRYPANOSOMA LEWISI. 661 


Tanne K.—Infections produced by Fleas taken at Random 
from the Infected Breeding-cage and put singly on 
Clean Rats. 


| j 
| leas left on the rat. | Date put on. | NESitis” | Tesults. nctal | positive. | of positive. | 
One day Social gots Ai) a an 
12: xii:710 | 3 = = ss 
18: xii: 710 2, | se —_ — | as 
Total . os ear — 6 aa 0 | 
Tispestdapend) Ohara slOn gress! (Vue bte te Be if 
19:ix:710 | 2 1 = a2 = 
10:x:'10 3 — aaa = a 
llesare all) | 2 1 a = i, 
182x022 105 | 3 | — —_ = = 
19:xi:710 3 a ne ay Ee 
Total . — | — —_— 17 } 3 15 | 
| 
Four days 19:ix:710 | 2 lees 2a fel Fe _ 
Heal OFexe: 71.0) 9 at ES | Ass bas 
Sie 710 3 —— — —_ = 
144:x210 | 3 == Si 
3:xi:°10 3; == a — — | 
| 11:xi:'10 2 i peed = — 
12:xi:'10 3 = = | = = 
29:xi:710 ! it == == | = =— 
4c oe | 3 | ee as = = 
28:11: ‘Aluil 9 — as = — 
Total .| — = = 36 4. 10 
Five days | 19:x:710 3 = =. as =! 
20:x.710 2 if = uel wee 
Total .| — = a 5 1 166 
Sie dase en ead icc hiro a ee Ae a1 
Total . | = 2 = We ee 0 
Ten days | 29:vii:711 Y 1 = — cae 
Total .| — — = 7 1) 485 
Sixteen days 9:1:12 | 9 1 ee ee aS 
13:1:°13 7 1 The = = 
Total . aa — 16 2 111 
Eighteen days 3:iv:13 5 = at is) = 
21:iv:713 a a a ae 
Total . — — eu a Wi ee 0 
Grand total . | -- = _ 104 tk + 956 


662 Er. A. MINCHIN AND J. D. THOMSON. 


for was not found, and it must be supposed that the flea died. 
If it died a natural death before going on to the rat the 
result would of course be negative whether the flea was 
infective or not. If eaten by the rat, an infective flea would 
probably produce a positive result. 

(2) The fact, now established, that the flea does not infect 
by the puncture of the proboscis, but contaminatively, renders 
the infection a very casual affair, especially when only one 
flea is on the rat, and the probability of the infection taking 
place is relatively low. ‘This is clearly shown by Experiment 
37, tabulated above (Table G, p. 640), in which one infective 
flea, put on seventeen rats successively, over a period of 
about three months, infected seven of them, and another flea, 
put on ten rats over a period of about two months, infected 
only three of them. An analysis of these results is given in 
Table L, from whith it is seen that twenty-seven exposures of 
rats to infection by two fleas known to be infective produced 
ten infections, equivalent to 387 per cent. of positive results. 


TasLtE L.—Infections produced by Two Fleas known 
experimentally to be Infective and placed 
singly on Clean Rats. 


; Negative Positive Total Total 
BUD GENE Ae results. results. negative. positive. 
Four days : : 4 4 
Five days. : F 10 6 
Six days E : oe 1 -- 
Seven days a 2 — 
17 10 


A side-light on the problem of the percentage of infective 
fleas might be obtained from the dissection and examination 
of fleas six days or more after they have been fed on an 
infected rat. From Table Bit is seen that, of 118 such fleas 
examined, fifty-three were found to contain stages of the 
trypanosome ; of these twenty-eight had a scanty, twenty-five 


THE RAT-TRYPANOSOME, ''RYPANOSOMA LEWISI. 663 


an abundant, infection. The value of these data, however, is 
somewhat uncertain, as an aid to the solution of the problem 
under consideration. 

From the foregoing summary of the data, it is evident that 
the computation of the percentage of fleas that become 
infective, of those exposed to infection, is a somewhat com- 
plicated statistical problem. We have submitted the data to 
Dr. Greenwood, who has kindly supplied us with the report 
appended below, from which it is seen that the percentage of 
infected fleas lies probably between 5°9 per cent. and 
45°7 per cent., the mean being 25°8 per cent. 


Report of Dr. M. Greenwood. 


The problem it is desired to solve is the following: 

Within what limits does the true proportion of infective 
fleas in the population of which those enumerated in Table K 
are a sample probably lie? 

Of 115 fleas left not more than eighteen days on clean rats, 
eleven produced infections. But all infective fleas do not 
produce infections, and, according to Table L, in twenty-seven 
trials with unquestionably infective fleas only ten produced 
infections in clean rats. Accordingly, it follows that the 
proportion of fleas in the first experiment which actually pro- 
duced infection must be divided by the proportion found in 
the other experiment to arrive at the ratio of potentially 
infective fleas, which is the quantity sought. This is— 

il 


EEE 0583 or 25°8 per cent. 

27 
But the two ratios from which this result is derived are each 
subject to errors of random sampling, and it is necessary 
to compute the “probable error” of sampling to which the 
final proportion is subject. Since the two proportions are 
entirely independent one of another, the square of the 


S 2 B- a q 2 A’ 
. - . = 2 ans Sp: 
standard deviation of their ratio is Tae at where A 


664 E. A. MINCHIN AND J. D. THOMSON. 


is the proportion observed among the 115 fleas, and s, its 
standard deviation and B and s, similar quantities in the 
case of the twenty-seven trials of the other experiments. 

Using this formula we reach 9°84 per cent. as the value of 
the standard deviation or ‘67449 times this = 6°64 per cent. 
for the “ probable error.” Taking the usual margin, three 
times the ‘‘ probable error,” the conclusion may be drawn that 
the real proportion of infective fleas is very unlikely to be 
beyond the limits 59 per cent. and 45°7 per cent. 

This is the conclusion which might, I think, legitimately be 
drawn from the two experiments, but two cautions must be 
had in mind. The first is that the number of trials in the 
second Experiment, 27, is rather small, and consequently the 
application of the customary theory of sampling errors must 
be made with hesitation. The second caution is that we are 
using twenty-seven trials with two fleas, not twenty-seven 
separate fleas, consequently, the two experiments are not 
strictly in pari materia. The two fleas used gave very 
different proportions of successes, and it might happen that 
were a larger number of definitely infective fleas used, the 
factor for division would be substantially modified. The 
above calculation can naturally give us no information on this 
point since it proceeds in terms of trials, twenty-seven trials 
with two fleas being assumed to be the same as twenty-seven 
trials with twenty-seven fleas, and that the differences do not 
depend upon the idiosyncrasies of the fleas, but upon the 
fluctuations of chance, the fleas being used simply as dice or 
counters. 


(xviil) Can the First Phase of the Development of 
the Trypanosomes, namely, the Intra-Cellular 
Multiplication in the Stomach of the Flea 
continue beyond the Second Feed of the Flea 
(counting as the First Feed that by which it 
became infected) ? 


With regard to this point, it should first be made quite 
clear that our observations on fleas examined during early 


THE RAT-TRYPANOSOME, TRYPANOSOMA LEWISI. 665 


periods of the development show conclusively that the try- 
panosomes may have disappeared from the stomach, and the 
rectal-phase may be well started, even so early as 18, 24, or 
36 hours after the first feed! A very clear case of this is 
the stomach mentioned above (p. 555), in which the infection 
was thirty-six hours old, and which was examined after being 
cut into a series of sections; no trypanosomes of any kind 
were found in the stomach, but immediately behind the 
pylorus were two attached clumps of quite normal crithidial 
type. On the other hand, in fleas not fed again after the 
infective feed we have found normal forms of the stomach- 
phase as late as three, four, or even five days after the first 
teed. From such observations it is evident that the stomach- 
phase is of very variable duration, for some reason, and that 
in some cases it is ended? long before the time when the flea 
would, under natural conditions, feed again, while in other 
cases it persists at least up to this time. 

The question, therefore, is not, “ Does the stomach-phase 
continue beyond the second feed ?” (since it is certain that it 
is very often ended before the second feed), but “Can it do 
so?” and with regard to the question so posed it must be 
pointed out that a single clear instance of the stomach-phase 
persisting beyond the second flea would suffice to give an 
answer in the affirmative with certainty; but so long as the 


1 We do not refer to those cases in which the trypanosomes had dis- 
appeared from the stomach by degeneration, and in which the 
rectum was either empty or contained only degenerative forms ; but 
only to those cases in which the presence of true crithidial forms in the 
intestine or rectum showed that the development of the trypanosome 
was following its normal course. 

2 Assuming, that is, that the intracellular multiplication is an 
essential part of, and takes place invariably in, the normal develop- 
ment of T. lewisi; we believe this to be the case, but we are unable 
to assert that it is so. It is at least within the bounds of possi- 
bility that the development may take occasionally a short cut, that 
is to say, that the trypanosomes may pass on to the rectum, and 
there establish the normal crithidial phase without undergoing intra- 
cellular multiplication in the stomach, 


A. MINCHIN AND J. D. THOMSON. 


E. 


666 


‘snore ApAeT SVIPTYYLAID 
‘shorotanu APR; SVIPIYFLID, 
“SUIq}ON 


“SUULLOF (4) [PIPIGIO MOF W 


‘Oo SuULyYO N 
‘oouvreodde 


| TeEPIygI4o JO suULOF [[eUIS Maz Vy | 

| SVIPIyIao [epuoeudopaaeq 

[ejueutdopoaop 

Afyueredde ‘sutoy ypems moy 
« 9[odpry ,, 

SULIOF Te Moy W 

“SULTON | 


ouo §=>puL 


“SULLOF [RIPTYAIAO TepUUTdopaaoqy 

‘[eyuout 

“dopeaop ATyuoredde ‘sutoy [peUTg | 
ey 

-ntouesop A{yuoIrdde ‘suLtofz TPeRUTg | 


“OUITIO NT 
“SULIOF 
aAlpe.ouiasop ATpUeAIedde ‘tees OMT, | 
OATP | 
-ntouasop A[yUAIVd de ‘suULLoFZ MOF W 


*OULTIO NT 


“BOT Wy oT4 url 


se 


es 
“se 


te 


GI * 


LETTOoR.1VUL 

pue ooif soutosourdéay Moy W es “ 

. . . . . . . | ee 

. . . . . . Sury40 NT ve se 
eM] [oo 

-vijut pur saif soutosourddy, ee e 

. . . . . . 7 +e ee 

. . . . oe ee “ 

| 
bt ini a ea skep } | GI. ' TA: 8G 

. OUITIO NT ee ee 

: ' somosourd&y vaay Moy -W as a 

. . . . . . SUIT}O N | ee ae 
Tepe 

“eur pur eetf somtosourd dy, S Ks 
aATPRABUaSep ATJUA1ed 

-dv ‘sommosourdday ysissnys mop y = We 

. . . . . . SUL ION se (in 
IR[N] [Io 

-eajur pur saif soautosourdd.ry, | : : 
‘S9ORyS AR[MPOoRAZUT AWRUT 

pure sourosouedAay @ATPOR MOF VW = ve 
$9.0r4s qe[NTpeowryut 

pue soutosourdday oaror Aur us th 
ed 44 [eyUeUdopaaep 

jo sowosourdday oaror mey YW) sep Al ape a ed 
. dhe . 
| 
: : _ bile “Boy oun. | oa 
“TORBULOIS JO SVUNqUOs/ UL TOTWOFUL | ) “eee JO". 
joosy || @ 1JD9SSTp JO 99Rq 


‘OUT T, puovsg vB 


souosouvd daz, 


pe JOU sve, y—wV 


a4} Jo osvyd-youmoyg oy44 Zo 


“pads (OATZOOJUL) 


4s.iy jo aquq 


V9IUDSIS Id oly uo Pee puovdg B JO VOUOUT FUT 94 ISO], 04 squowtiad xy— fy aTav J, 


667 


WISI. 


nl 
4 


THE RAT-TRYPANOSOME, TRYPANOSOMA LI 


“OULY4O N, 
‘wLoF [eJUOtUdOpOANp oud 


“OUT JO NT 

REID FET) 
‘OUIULIVAS SVIPIYFAO 
‘suLtoF pozurod Jo dumypD 
‘SULLOF (4) TRIPIQAD 

“SULLOF [RIPTYPAD, 

“SUIPTyyWo (4) Teyuoeurdopaaog 
“SULIOJ [RIPTYAO, 


‘SULOF poqurod poyorgyy Moz Vy 


‘SULLOF OATJLAOUBSOP MOF W 


“OUTTION | 
‘SULLOF LOPUITS 9ATJOR OMT, | 


“OUTTT}O N 


‘SULIOF (J) OATQL.AIOMOSep OMT, | 


‘SULLOJ (2) TIPO 


“SUITION 

“SBIPIYBIIO JO suIVMG 
“SuIYIO N 

"SULLOJ (4) DATZR.IOMIOSOG 


“SUrq ION 


"SULIOJ (4) oaTyR.LOTOSOq 
“SULIOF (7) OATZRIOUOSOG 
“SVIPIYALIO TeyuetIdopaaaq 
“OUITIO NY 

‘SUILOJ OATIR.IOTIS9(] 
“SUIT}O N 

*SULLOF OATZV.LOMO SOC 


pur 


~VAIQUL pus 


‘soutosourdday 00.7 


. . . . . 
. . . . oe 


*  SUITION 
epaypeo 


. . 


O0AF 


. . . . we 


* SuryjoN 
somosourd day Mot VW 
9 ‘ *  SUIYION, 
sdumypo peroryqe 
somosourdday — aa.tf 
p ; : * SUIYION 


sommosourd say Mot W 
seutosoundday SUOT MOF W 


; : : * SUITION 


soemosourd Any, | 


| 
| 
| 
| 


Suory 


soumosourd sy datz OATQOV | 


Woes atlosourdséay SuOT oUC) 


. . . . . oe 

* SUIYGON 
* SUILOF [[VUIS MOF W 

sommosourddny oatf MOF W 


; * SUIION 
soutosourd 
ILPNT[IOVAJUL PUL dVIF MOF W 


. oe 
. . . ve 

. . oo 
. . . . . ay 
. . . . . “ee 


SUL ION 


. . . . . 


TN 
mn 

a 
=-—) 

tga’ 


ve 


skup %z 
se 


skup 


6 
oe 


oe 


skup © 
se 


skup % 


(6) 


re 


Tice 


1 (exo 


Le: 


se 
HA: GT 
“6 
HA: TT 
IA : G 
ce 
IlA : 
te 
HA: g 


- 


Cc 


A. MINCHIN AND J. D. THOMSON. 


KE. 


668 


"‘SULIOF poyulog 

‘suLOF poqurod moj W 
SUMO 

"SUILOF (¢) OATZLAOUOSOpP MOF VW 
PRET TOAEO Beh, 

“SUIY}O N 

SUGIOS TELPIUAEL) 


"SULLOF (¢) OATPRAOTOSOCT 


“SUIY}O N 
‘SUUALOF [RIPIYYMIO, 
“OULTTJON 


"SULLOF [VIPIFIIO MOF VW 


“SULLOF (Z) OATPR.ALOUODOCT 
oe 


ee 


oe 


“SUTTON 
“SUUALOF [RIPIGFIAD 


ee 


ee 


“SULIIO N 


. . . . “cc 
. . . . “ee 
. . . . oe 
. . . . oe 
. . . . oe 
. . . . oe 
. . . . ee 
. . . . ee 
. . . . oe 
. . . . ce 
. . . 


* SuryqjON 
((4) oraopéd 


-ysod) surtoz podvys 

-1ved ][vus OM} 10 OUD 

. . . . ee 

. . . . i 

. . . . ee 

. . . . iad 

. . . . oe 

. . . . se 

. . . y . oe 

. . . . ee 

' ; : * SUITION 
*TPOVULOIS JO SFO} ULOL) 


mM 
m 

3S 

aS 
aia 

ise) 


skep ¢ 


“Boy OTL 
ur Uoooyur 
jo oo 


ee oe 
“ce ee 
“ee ‘ 
“ec ee 
Ti “ 
oe 

~ . Py 6 

GI. * HA: GL 
ee oe 
ee ee 
e ee 
“ee ia) 
ol.* H4: TT 


(‘ue QT-6) 


ee eT Gb eee 


“cc “ 
a7 Lay 
oe ee 
ve 6 
oe ee 
. . ee 
BI. WAL 
ee “ 
ee ee 
ia “e 
se ee 
be oe 


Glee Waly es 


GI, : WA: 8 
(qy.51u) 
GI, TA: 8 


Glee eco 


*pooy 
puodes Jo 09Rq 


“BOT JO 
WOTJIOSSIp JO 94B 


“OULL YT, puovdsg BV pow $BO[H— aA 


*(panuyywoo)— We ATAY,, 


“poog (9A1W00FUT) 
qsay Jo o9Rq 


A GIOD Gis EOS 


669 


THE RAT-TRYPANOSOME, TRYPANOSOMA LEWISI. 


‘T[@ UL JUOSGe OOM STLIOF-Yoruroys [Rody oxy 

—snay » ‘Ayourvu—surioj peyuemdopoaop pourezuoo ATurTEZA00 WINZodeA OY} YOY UT sesvo esoyy ATUO SuryuN0|— gq 
‘BOT OY} UL Aoyqos0q[v porvoddusip oavy 07 porvodde Loyy Yorys JO T] UI ‘sesvo GF 

UI PUNOF JOU O19 PUL ‘SasvO OF UI TPORUIOYS OY} UI PUNOF 00M somMOsoUrd Ay ‘oy PUODES V Poy JOU SvEy GF UT—V 


"AUVWIOS 


‘uoijisod ur o1top£d-qsod useq oavy AvUI Youutojs dn-posvo} oy} UI Woes esoy} pur “loqovavyo 
UI [LIPITJIWO elVM UES STUAOF OJ [[B OSV STYZ UI ‘voy T UL WNgoot pUL YOLTMOJs UI JOG Mees seMOosouLdé.y, 
‘ed £q pequourdojeaop jo ‘Ayure91e0 g wr ‘AT[NFZQuop T 
ur Sed4y oayvaouesep jo AT[NJZQnOp svop oseyy JO g UI ‘svoH ET UI WNyZooeL 04} Ul Ajuo uses seuosourd{ay, 
‘SvOf QT Ul Uses soumosourd4éy ON 
‘q SaIuag 
‘ad Aq [equeuwt 
-doyoaop Jo ‘ATUre7100 suUAOF VAS} JO G UT ‘AT[NFAZQnop suLIOF Woy} JO F UT poureyzWoo AF ¢ od fy oaAtyeaouesop Jo AT[NIZqQuop 
SUILOF SLA OSoy} JO G UI PoUTLJWO WHZod OY} ‘SvOH FT UL WNZoal pu YORUIOZS oT} UT YO Woes sourosourd day, 
‘sve 9 UL Youutoys oy} ur AyUO Woes somOosoURd AI, 
‘ad £y pequoutdoyoaop Jo ‘Ayure}.Ae0 g ut ‘AT[NFZQnop sesvo oy} Fo g ur t odéy 
aaTyRlouesep Jo ‘ATurezA00 ¢ UT ‘AT[NJIQNOp sesvo oy} JO P Ul ‘Svoy ST UL WNyZoor oy UI Ajuo woes soumosourdéay, 
‘SvO} [[ UI uees sourosourdiy ON 
‘VY saraag 


‘SLTNSHU HHL AO SISATVNV 


“suoryrarudoad guouvuated Jo woryeururexe ey} Aq 
patAyWoo JOU SvAL FI FLYY PUL UIezA100 oyMb you svM 07RzS YSO1J OY} UL Moos SUIIOF OY} JO SMNYVU olf} FO TOTPVUTUAEJOp 
94) FLY} SOYTUSTs 41 ‘TANZOaL OY} JO SPUOPMOO Of} SUTTAGUOD FUOUTOFLIS OY} OF POXTB St (4) Atonb v wey MA—'d ‘N 


"SULLOF (2) [PIPIYGLIO Moz VW | * : : : ip ss i es ee ‘Ze | 

“OULYIO NV . . . . we ee ‘ay ay a “1K | 

“SVIPIQALIO BNLOTOMM A : : ‘s ue % ss ee “Og | 

| ‘ . e oe oe Las Lad 6G | 

| - SMR Waa skep p | GI: 14: 80 | GE: i LS | Be! i HS | “8B | 

| -Surty}0 N - 5 5 c ‘ ‘6 | ‘“ | ‘“ 12 
‘sutog poqutod Moy VW | ° p ” ‘ Zuipjon | skep $e | ST,: WA: SL! Ble: WA: IL | GL: M4: 8 | 92 


670 E. A. MINCHIN AND J. D. ‘THOMSON. 


answer is in the negative, it cannot be regarded as certain, 
but only as possessing a greater or less degree of probability. 

In order to test this point we fed batches of fleas on infected 
rats and then divided each such batch usually into two batches, 
Aand B. Batch A in each such case was kept starved until 
it was examined; batch B was fed again before being 
examined. In cases where fleas of batch B were found on 
examination not to have availed themselves of the chance of 
feeding, they were reckoned in batch A. Sometimes the 
original batch was not divided, but treated as a whole either 
as an “A” (not re-fed) or “B” (re-fed) batch. 

The results of these experiments are tabulated in Table M, 
which seems at first sight decidedly in favour of the conclu- 
sion that the trypanosomes cannot persist beyond the second 
feed of the flea. It isseen that in forty-nine fleas not fed 
again after the infective feed, trypanosomes were present in 
twenty cases; in thirteen of the cases the trypanosomes were 
of the long stomach type and in eight cases intracellular forms 
were seen. On the other hand, in thirty-two fleas examined 
after having been fed a second time, the typical multiplicative 
stomach-phase was not present in a single instance. Unfor- 
tunately, the force of these figures is rather weakened by the 
fact that, of the fleas fed a second time it can only be asserted 
positively in seven cases that the rectum contained true 
developmental crithidial forms and that the developmental 
cycle was in these cases a “ going concern,” so to speak. 
While the figures make it probable, to a certain degree, that 
the stomach-phase, if it persists up to the time of the second 
feed, must come to an end then, this conclusion cannot be 
considered established and must remain a point for further 


ce 


investigation. 

If it be true that the stomach-phase cannot persist beyond 
the second feed, we may enquire how such a result is brought 
about. It is intelligible that a fresh meal of blood might 
sweep on all free, extracellular trypanosomes from the 
stomach towards the rectum, but this would not account for 
the disappearance of the intracellular forms. It has been 


THE RAT-TRYPANOSOME, TRYPANOSOMA LEWISI. 671 


mentioned above that in some insects the epithelium of 
the mid-gut is regenerated completely after each meal, and 
we stated further that we were not in a position either to 
affirm or to deny that, in the case of the flea, the regeneration 
of the stomach-epithelium takes place in regular correlation 
with the feeding. If it were so, however, it would become 
quite intelligible why a second feed should put an end to the 
intracellular multiplication in the stomach. 


(xvill) Starvation of the Flea during the Incuba- 
tion Period of the Cycle does not inhibit, nor 
does it necessarily retard, the Developmental 
Cycle of the Trypanosome in the Flea. 


Experiment 45.—A batch of about 250 clean fleas, having been 
collected from the non-infected breeding-cage and kept hungry forthree 
days, were put (22: vii:°13) on a well-infected rat, for about twenty-four 
hours. The next day 150 of the fleas were recovered and kept in a flask 
containing some damp sand. 

Two days later (25: vil), thirty of these 150 fleas in the flask were put 
into a freshly-prepared bell-jar A with aclean rat 368. The next day 
(26: vii) rat 568 was removed from bell-jar A, all the fleas on it being 
cleaned off carefully and put back into the bell-jar. Rat 368, examined 
regularly up to 26: viii, did not become infected. 

Two days later (28 : vii), clean rat 368a was put into bell-jar A, con- 
taining the fleas that had been in contact with rat 368. On the same 
day. thirty more fleas from the flask were put into bell-jar B with clean 
rat 369; another thirty in bell-jar C with clean rat 370; and another 
thirty in bell-jar D with clean rat 371. It will be remembered that the 
fleas in the flask had been exposed to infection for one day (22-23: vii), 
and kept without food since then; consequently, bell-jars B, C, and D 
were colonised each with thirty fleas that had been exposed to infection 
between five and six days previously and starved since then. 

The next day (29: vii), rat 368a was removed from bell-jar A and all 
fleas recovered from it put back into the bell-jar. Rat 368a did not 
acquire infection. The same day rats 369, 370, and 371 were removed 
from bell-jars B, C, and D, all fleas being recovered from them and put 
back into the respective bell-jars. Rats 369 and 370 did not become 
infected ; rat 371, on the other hand, became infected in due course (see 
Table C, p. 616). 

Two days later (31: vii), rat 368b was put into bell-jar A, rat 369a 
into bell-jar B, rat 370a into bell-jar C, and rat 371a into bell-jar D. 

VoL. 60, PART 4.—NEW SERIES. 46 


672 BE. A. MINCHIN AND J. D. THOMSON. 


The next day (1: viii), rats 369a, 370a and 37la were removed from the 
bell-jars B, C, and D, and the fleas on them carefully recovered and put 
back into their respective bell-jars. Rat 368b was left in bell-jar A. In 
the sequel rats 368b, 369a, and 370a did not become infected; rat 371a 
showed infection in due course. 

Two days later (3: viii), clean rats 369b, 370b and 371b were placed in 
bell-jars B, C, and D, and left in till they should become infected. Rat 


369b did not become infected ; rats 370b and 371b became infected in 
due course. 


The results of the experiment may be summarised in the 
following manner. We start with four batches (A, B, C, D), 
each of thirty fleas, which had been exposed to infection on 
the same rat for one day (22: vii to 23: vii). 

(1) Batch A (bell-jar A) : 

Put on rat 368 from 25: vii to 26: vii (about three days 
after exposure to infection); result negative. 

Put on rat 368a from 28 : vii to 29 : vii (about six days after 
exposure to infection) ; result negative. 

Put on rat 368b, 31 : vii (about nine days after exposure to 
infection), and left onthe rat ; result negative. 

This batch therefore did not become infective at all. 

(2) Batch B (bell-jar B) : 

Starved for five days (23: vii to 28: vii), then put on rat 
369 from 28 : vii to 29 : vii (about six days after exposure to: 
infection) ; result negative. 

~ Put on rat 369a from 31: vil to 1: vii (about nine days 
after exposure to infection) ; result negative. 

Left in with rat 369b on 3: viii; result negative. 

This batch therefore did not become infective at all. 

(8) Batch C (bell-jar C) : 

Starved for five days, then put on rat 370 from 28 : vii to 
29 : vii (about six days after exposure to infection) ; result 
negative. 

Put on rat 370a from 31 : vii to 1: viii (about nine days after 
exposure to infection) ; result negative. 

Left in with rat 370b on 3 : viii; result positive (the 
examinations of the rat indicate that infection took place 
between 4: viii and 7: viii). 


THE RAT-TRYPANOSOME, TRYPANOSOMA LEWISI. 673 


This batch therefore became infevtive. 

(4) Batch D (bell-jar D) : 

Starved for five days, then put on rat 371 from 28 : vii to 
29: vii (about six days after exposure to infection) ; result 
positive. 

Put on rat 37la from 31 : vii to 1: viii (about nine days 
after exposure to infection) ; result positive. 

Left in with rat 37la on 3 : viii; result positive. 

This batch evidently became strongly infective. 

Remarks.—From the above summary it is seen that 
batch A (not starved) and batch B (starved) failed to become 
infective, while batches C and D (both starved) became 
infective. 

Batch © did not produce its first infection before 3 : vill; 
that is to say not until eleven or twelve days, at least, after 
exposure of the fleas to infection. It is not legitimate, how- 
ever, to conclude from this that the developmental cycle of 
the trypanosome was retarded, since it has been shown above 
that infective fleas often fail to infect. The fleas may very 
well have been infective when placed in contact with rats 370 
and 370a, but they were in contact with these rats for only 
about twenty-four hours. The most probable explanation for 
the two failures to infect is that only a small number of fleas 
in this batch were infective. 

Batch D produced its first infection between 28: vii and 
29 : vii, and since it was exposed to infection between 22: vil 
and 23 : vii it follows from these figures that the incubation- 
period—that is to say, the length of time taken by the 
developmental cycle of the trypanosomes—must have been 
between five and seven days. We are justified therefore in 
concluding that the trypanosomes in this batch went through 
a cycle of perfectly normal duration. The further fact that 
this batch never failed to produce infection during the time 
the experiment was carried on, indicates that the trypanosomes 
went through their cycle and established themselves success- 
fully in a relatively large number of the fleas. 

To conclude: Batches C and D show that starvation of the 


674 E. A. MINCHIN AND J. D. THOMSON. 


fleas during the incubation-period does not inhibit the develop- 
ment of the trypanosomes; and batch D shows further that 
the development is not necessarily retarded by starvation. 


(xix) Starvation of the Flea following immediately 
on an Infective Feed favours the Establish- 
ment of the Haptomonad Phase in the Rectum, 
while Starvation begun after the Incubation- 
Period in the Flea is over favours Migration 
to the Post-Pyloric End of the Intestine and 
the Establishment of the Haptomonad Phase 


there. 


Experiments 49 and 50 were carried out with the object of 
ascertaining what effect, ifany, varying food-conditions might 
have on the incidence and location of the established hapto- 
monad phase in the flea’s gut. 


Experiment 49.—21: iii: °14—A number of fleas collected from 
the non-infected breeding-cage two days previously were put into a 
bell-jar with a well-infected rat at eight a.m., and were recovered again 
at twelve noon. They were then divided into two batches. Batch A, 
consisting of fifteen fleas, was put into a flask with moist sand at 
the bottom. Batch B (about forty fleas) was put into a bell-jar with 
a clean rat (rat 477). 

26 : iii : °14.—Five fleas of batch A and four of batch B were dissected 
and examined. 

Of batch A four were positive, one was negative. Of the four positive 
three showed developing forms of the trypanosomes in both the 
stomach and the rectum. Two of the three showed large numbers in 
both stomach and rectum, and in one of the stomachs intracellular 
forms were found. The fourth positive showed a haptomonad infection 
in the rectum. 

Of batch B only one of the four dissected showed trypanosomes, and 
these were found free in the stomach-slide. 

27: iii: "14.—Hight fleas of batch A and eight of batch B were 
dissected and examined. 

Of batch A seven were positive; one was negative. Of the seven 
positive one showed long active forms in the stomach and haptomonads 
in the rectum, while six showed developing forms in the rectum only— 
three scanty and three in fair numbers attached mostly round the 


THE RAT-TRYPANOSOME, TRYPANOSOMA LEWISI. 675 


rectal surface of the projecting intestine, but in other parts as well. 
Of batch B only one was found infected, and it showed haptomonads 
attached about the middle region of the rectum. 

The remaining fleas of batch B were now divided into two batches— 
batch Al and batch Bi. Batch A1, consisting of fifteen fleas, was put 
into a flask with moist sand at the bottom, and batch B1 was left in the 
bell-jar with rat 477. 

4: iv :’°14.—Rat 477 was removed from the bell-jar and clean rat 478 
was put in its place with batch B1. 

Four fleas of batch Al and four of batch Bl were dissected and 
examined. 

Of batch Al three were positive and one was negative. Of the three 
positive two showed trypanosomes in abundance in the post-pyloric 
region of the intestine and nowhere else. In these the trypanosomes 
were long and slender, and some were club-shaped; while in a third, 
which showed one or two in the rectum also, there was a swarming in- 
fection of haptomonads, as well as long, slender and club-shaped forms 
in the post-pyloric region. The fourth flea was negative. 

Of batch B1 all were negative. 

The remaining fleas of batch Al were now allowed to feed on a clean 
rat for a short time, 

9: iv: 14.—Five fleas of batch Al and four of batch Bl were dis- 
sected and examined. 

Of batch Al only one flea was positive, and it showed a fair number 
of slender trypanosomes in the post-pyloric region and nowhere else. 
Of batch BI all were negative. 

11 : iv: ‘14.—The remaining fleas of batch Al were allowed to feed on 
a clean rat for a short time. 

16: iv : “14.—Five fleas of batch Al and five of batch B1 were dis- 
sected and examined. 

Of batch Al two were positive and three were negative. The two 
positives showed large numbers of trypanosomes, some free, long and 
active, some club-shaped, and others were small, round and pear-shaped, 
in clumps and attached so as to form a lining to the wall of the gut. 
All were inthe post-pyloric region and nowhere else. Of batch Bl 
all five were negative. 

Rat 477 became infected in due course. Rat 478 never became 
infected. This agrees with results of the examinations. 


Experiment 50.—20: v : 14.—A number of fleas collected from the 
non-infected breeding-cage two days previously were put into a bell-jar 
with a well-infected rat late in the evening, were left overnight, and 
were recovered next morning. They were then divided into two 
batches. Batch A was kept in a flask with moist sand at the bottom. 


676 E. A. MINCHIN AND J. D. THOMSON. 


Batch B was put into a freshly-prepared bell-jar with a clean rat (rat 
496). 


26:v: °14.—Five fleas of batch A and five of batch B were dissected 
and examined. 

Of batch A three were positive and two were negative. Of the three 
positive two showed haptomonads in the rectum, one being a swarming 
infection, and the third showed many long, active forms in the post- 
pyloric region only. 

Of batch B all were negative. 

28:v:°14.—Seven fleas of batch A and seven of batch B were 
dissected and examined. 

Of batch A five were positive and two were negative. Of the five 
positive four showed developing forms in the rectum only, and of these 
two were swarming infections. The fifth showed haptomonad infection 
of the rectum, and also free active forms in the stomach. 

Of batch B only one of the seven was positive, and it showed a 
scanty infection of the rectum only. 

The remaining fleas of batch B were now divided into two batches— 
batch Al and batch Bl. Batch Al was put into a flask with moist 
sand in the bottom. Batch B1 was put into a bell-jar with a clean rat 
(rat 497). 

vi: °14.—Fourteen fleas of batch Al and fourteen of batch Bl 
were dissected and examined. 

Of batch Al four were positive and ten were negative. Of the four 
positives one showed haptomonads on the rectal surface of the project- 

ing intestine and three showed trypanosomes (one Seu aS in the 
post-pyloric region and nowhere else. 

Of batch B1 five were positive and nine were negative. Of the five 

positives one which had its stomach full of red blood showed a scanty 

infection in the rectum only. The other four showed infection in the 
post-pyloric regions only. Of these two (females) had small ova and 
their stomachs were empty. The remaining two had a fair quantity of 
bro wnish-coloured blood-débris in their stomachs. Rats 496 and 497 
became infected in due course. 


Summary. 


A.—Fleas starved from immediately after the infective feed, dissected 
and examined five and six days after the infective feed. 


THE RAT-TRYPANOSOME, TRYPANOSOMA LEWISI. 


677 


| 


| post-pyloric only, 1 


~ Number of! 
No. of ex- Number | a: : . 
. fleas exa- | ; | Site of infection. Remarks. 
periment. | “ined. | infected. | | 
| 
j — rt | 
49 13 1 Stomach and Intracellular forms 
rectum, 4; in one stomach | 
rectum only, 7 | 3 scanty, 4 swarming 
i—infections mostly in 
| upper part of rectum | 
| 50 12 | 8 Stomachand | | 
| 
rectum, 1; 
rectum only, 6; 3 scanty and 5 swarm- 
| ing 
| 


B.—Fleas that were put into bell-jar immediately after the infective 


feed, along with clean rat, on which they could feed at any time. 
sected and examined five and six days after the infective feed. 


Dis- 


= Number of} ~ | 
| No. of ex- |- Number | . 
| periment. Pens oe Saerarede | Site of infection, 
49 12 «| 2 Stomach only, 1 
y 
Rectum only, 1 
50 12 1 | Rectum only, 1 


Remarks. 


~Haptomonad_infec- | 
tion in middle 
| region. 


| 


Al.—Fleas in which starvation was begun six days after the infective 


feed. During the six days the fleas had been fed on clean rats. 


Those 


of Experiment 49 were dissected and examined fourteen, nineteen, and 
twenty-six days after the infective feed. Those of Experiment 50 were 
dissected and examined thirteen days after the infective feed. 


Remarks. 


nad infection in 2. 


Pile carpet infection 
upper part; swarm- 
ing haptomonad in-- 


| 
es _ |Number of} 1 
eee fleas exa- anes Site of infection, 
mined, 
| 
, 49 14 6 — Post-pyloric only,5;| Swarming haptomo- 
post-pyloric and 
rectum, 1 
| 50 14 4 Rectum only, 1; 
post-pyloric only, 
3 
fection in 1. 


G73.) E. A. MINCHIN AND J. D. THOMSON. 


B1.—Fleas kept with clean rats in bell-jar ever since the infective 
feed. Dissected and examined as in Al. 


7 Number of} » 
No. of ex- Number . ; ‘ i 
F fleas exa- | . Site of infection. Remarks, | 
periment. | “shined. infected. 
49 13 0 — — | 
50 14. 5) Rectum only, 1;| Stomach full of red | 

| 


post-pyloric only,| blood; !stomach | 
4 empty and ova) 
small in 2; stomach 

contained fair quan- . 
tity of blood-débris | 
in 2. | 


These results seem to throw light on the important function 
that the nectomonad forms described as occurring in the 
established rectal-phase may have in maintaining the infec- 
tion in the flea. The optimum food-conditions for the 
establishment of the haptomonad stage seem to lie some- 
where between abundance and poverty, and between partial 
and complete digestion of the blood-supply. More extended 


1 Although the fleas of batches B and B1 in both experiments had 
the chance of feeding on clean rats at any time from immediately after 
the infective feed onwards, all may not have equally availed themselves 
of the opportunity. Itis certain, in fact, from the condition of the ova 
and of the stomachs of two females of batch Bl that showed post- 
pyloric infection, that they, for some reason not ascertained, had 
starved in the midst of plenty; and these should really be transferred 
to batch Al. The remaining two fleas of batch B1 that showed post- 
pyloric infection had evidently fed, but not quite recently. The other 
infected flea of batch B1 had its stomach distended with red blood, and 
in it the infection was in the rectum only. 

In order to test the infectivity of batches Al and B1 in Experiment 50, 
twelve fleas of each batch were put on clean rats, two fleas to each clean 
rat. The result was that two of the six rats belonging to batch Al became 
infected, while none of the six belonging to batch Bl became infected, 
This may indicate that a period of starvation heightens the infectivity 
of infected fleas, perhaps by inducing increased production of the final 
propagative forms of the cycle; but further experiments would be 
required to justify such a deduction. 


THE RAT-TRYPANOSOME, TRYPANOSOMA LEWISI. 679 


and varied observations are required, but so far as these 
experiments go they show that the incidence, location, and 
continued existence of the haptomonad stage in the flea’s gut 
depend to a large extent on the food-supply. When, under 
conditions of partial starvation, a sufficient supply of nourish- 
ment cannot be obtained in the rectum, the haptomonad stage, 
if established there, would die out, and the flea would lose its 
infection were it not that the nectomonads produced in the 
rectum migrate forwards and re-establish this stage nearer to 
the food-supply. In like manner it may be assumed that 
when the food-supply in the post-pyloric end of the 
intestine becomes continuously too rich and abundant, the 
nectomonads produced there migrate backwards to the rectum 
and so the balance is maintained and the infection in the flea 
is kept up. 
LIstER INSTITUTE, 
June, 1914. 


BIBLIOGRAPHICAL REFERENCES. 


Baldrey, F.S. H.—‘ Versuche und Beobachtungen iiber die Entwicklung 
von Trypanosoma lewisi in der Rattenlaus Hematopinus 
spinulosus,” ‘Arch. Protistenkunde,’ xv, 1909, pp. 826-332, 2 
text-figs. 

Breinl, A., and Hindle, E.—‘ Observations on the Life History of 
Trypanosoma lewisi in the Rat Louse (Hematopinus 
spinulosus),” ‘Ann. Trop. Med. Parasitol.,’ iii, 1909, pp. 553- 
564, pls. i and ii, 

Brumpt, E.—‘‘Evolution de Trypanosoma lewisi, etc., chez les 
Puces et les Punaises,” ‘Bull. Soc. Pathol. Exot.,’ vi, 1913, pp. 
167-171. 


Chatton, E., and Delanoe, P.—* Leptomonas Pattoni (Swingle) et 
Tr. Lewisi (Kent) chez l’adulte et la larve de Ceratophyllus 
fasciatus,’ ‘C. R. Soc. Biol. Paris,’ Ixxiii, 1912, pp. 291-293, 

17 text-figs. 

and Léger, M.—* Sur un mode particulier d’agglutination et de 
cytolyse simulant un enkystement chez les Leptomonas des 
drosophiles.” ibid., Ixxii, 1912, pp. 171-173, 15 text-figs. 


680 E. A. MINCHIN AND J. D. THOMSON. 


| Gonder, R.—‘‘ Untersuchungen iiber arzneifeste Mikroorganismen: I, 
Trypanosoma lewisi,”’ ‘C. B. Bakt. Parasitenkunde’ (I, Abt. 
Orig.), lxi, 1911, pp. 102-113. 

| Léger, L., and Duboseq, O.—* Les Grégarines et l’epithélium intestinal 
chez les Trachéates,” ‘ Arch. Parasitol.,’ vi, 1902, pp. 377-473, pls. 
li-vii, 8 text-figs. 

- MacNeal, W. J.—* The Life-History of Trypanosoma lewisi and 
Trypanosoma brucei,” ‘ Journ. Infect. Diseases,’ 1, 1904, pp. 
517-543, pls. xi-xvi. 

Manteufel.—‘Studien iiber die Trypanosomiasis der Ratten, etc.,” 
‘Arbeiten k. Gesundheitsamte,’ xxxiii, 1909, pp. 46-83. 
~~ Minchin, E. A.—‘The Structure of Trypanosoma lewisi in Rela- 
tion to Microscopical Technique,” *Quart. Journ. Micr. Sci.,’ 53, 
1909, pp. 755-808, pls. xxi-xxiil. 


“\»— “On some Parasites observed in the Rat-flea (Ceratophyllus 
fasciatus),” ‘Hertwig’s Festschrift,’ i, 1910, pp. 289-302, pl. xxiii. 
Ww and Thomson, J. D.—‘ The Transmission of Trypanosoma 


lewisi by the Rat-flea (Ceratophyllus fasciatus),” ‘ Proc. 

Roy. Soe.’ (B), Ixxxii, 1910, pp. 273-285. 

“The Transmission of Trypanosoma lewisi by the 

Rat-flea (Ceratophyllus fasciatus),” ‘Brit. Med. Journ.,’ i, 

1911, pp. 1809, 1310. 

“On the Occurrence of an Intracellular Stage in the 
Development of Trypanosoma lewisi in the Rat-flea,” ibid., 
ii, 1911, pp. 361-364, 

\onicoll, W.. and Minchin, E. A.— Two Species of Cysticercoids from 
the Rat-flea (Ceratophyllus fasciatus),” ‘Proc. Zool. Soce.,’ 
1911, pp. 9-13, 2 text-figs. 

Noller, W.—‘ Die Ubertragungsweise der Rattentrypanosomen durch 

Fléhe,” ‘Arch. f. Protistenkunde,’ xxv, 1912, pp. 386-424, 5 text- 
figs. 


SS 


Nuttall, G. H. F—‘The Transmission of Trypanosoma lewisi by 
Fleas and Lice,” ‘ Parasitology,’ i, 1908, pp. 296-801. 

Prowazek, S. v.—‘ Studien iiber Saugetiertrypanosomen,” ‘Arbeiten k. 
Gesundheitsamte,’ xxii, 1905, pp. 351-395, pls. i-vi. 

Rabinowitsch, L., and Kempner, W.— Beitrag zur Kenntnis der Blut- 
parasiten, speciell der Rattentrypanosomen,” ‘ Zeitschr. f. Hyg. 


Robertson, M.—* Transmission of Flagellates living in the Blood of 
certain Freshwater Fishes,” ‘Phil. Trans.’ (B), ccii, 1911, pp. 29- 
50, pls. i and ii, 4 text-figs. 


THE RAT-TRYPANOSOME, TRYPANOSOMA LEWISI. 681 


Robertson, M.—‘ Notes on certain Aspects of the Development of 
Trypanosoma gambiense in Glossina palpalis,” ‘Proc. 
Roy. Soe.’ (B), lxxxv, 1912, pp. 241-248. 

Rodenwaldt, E—‘Trypanosoma lewisi in Hematopinus 
spinulosus,” ‘C. B. Bakt. Parasitenkunde’ (Abt. I, Orig.), li, 
1909, pp. 30-42, pls. i-il. 

Strickland, C.—* The Mechanism of Transmission of Trypanosoma 
lewisi from Rat to Rat by the Rat-flea,” ‘ Brit. Med. Journ.,’ 
i, 1911. p. 1049. 


— “Agrippina bona nov. gen. et nov. sp. representing a new 
family of Gregarines.” ‘ Parasitology,’ v, 1912, pp. 99-108, pl. iv, 
oo text-figs. 

and Swellengrebel, N. H.—* The Development of Trypano- 
soma lewisi in the Rat-flea (Ceratophyllus fasciatus),” 

‘Proc. Cambridge Phil. Soe.,’ xv, 1910, pp. 531-533. 

“Notes on Trypanosoma lewisi and its Relation 
to certain Arthropoda,” ‘ Parasitology, iii, 1910, pp. 456-454, 
1 text-fig. 

Swellengrebel, N. H., and Strickland, C—‘The Development of 
Trypanosoma lewisi outside the Vertebrate Host,” ibid., ii, 
1910, pp. 360-369, 21 diagrams. 

“Some Remarks on Dr. Swingle’s paper, ‘The Trans- 
mission of Trypanosoma lewisi by Rat-fleas ete,” ibid., iv, 
1911, pp. 105-108. 

Swingle, L. D.—‘* The Transmission of Trypanosoma lewisi by 
Rat-fleas (Ceratophyllus sp. and Pulex sp.) with Short 
Descriptions of three new Herpetomonads,” ‘ Journ. of Infect. 
Dis., viii, 1911, pp. 125-146, pls. i-iv. 

Thomson, J. D.—*‘ Cultivation of the Trypanosome found in the Blood 
of the Gold-fish,” ‘Journ. Hygiene,’ viil, 1908, pp. 75-82, pl. iii. 

Wenyon, C. M.—* Experiments on the Transmission of Trypano- 
soma lewisi by means of ‘Fleas,’ ‘Journ. London School 
Trop. Med.,’ ii, 1913, pp. 119-123. 


682 E. A. MINCHIN AND J. D. THOMSON. 


DESCRIPTION OF PLATES 36 to 45, 


Illustrating Mr. E. A. Minchin and Dr. J. D. Thomson’s 
paper on “The Rat-Trypanosome, Trypanosoma 
lewisi, in its Relation to the Rat-Flea, Cerato- 
phyllus fasciatus.” 


(In all cases where the age of a trypanosome is stated, it is to be 
understood as reckoned from the first infective feed of the flea, or, in 
cases where the fleas were left on the infected rat for some days, from 
the time when the fleas were first given the chance of feeding on it.) 


PLATE 36. 


[Various stages of the stomach-phase in film-preparations fixed with 
osmic vapour and stained with Giemsa’s stain; drawn with the camera 
lucida at a magnification of 3000. | 


Figs. 1-12.—F ree (extracellular) trypanosomes. 

Figs. 1 and 2.—Six hours; fig. 1, practically unmodified; fig. 2, 
slightly modified in structure. 

Figs. 3 and 4.—Twelve hours. 

Figs. 5 and 6.—Highteen hours. 

Figs. 7-10.—Twenty-four hours. 

Figs. 11 and 12.—Forms of crithidial structure from an abnormal 
flea (see p. 519); fig. 11, a free specimen; fig. 12, a couple adhering 
together. 

Figs. 13-17.—Recurved forms. 

Figs. 13, 14.—Twelve hours ; both from the same preparation, but 14 
s flattened out by having dried up before fixation, while 13 is normal. 

Figs. 15-17.—T wenty-four hours. 

Figs. 18-20.—Forms in which the multiplication of the nuclei have 
begun without the typanosomes having assumed the recurved form (or 
perhaps forms which have straightened themselves out again second- 
arily); twenty-four hours, all from the same preparation. In fig. 19 
the trypanosome is enclosed in the remains of the host-cell. 

Figs. 21-23.—Rolled-up forms with 2n and 2 NN each; from the 
same preparation as the preceding. 


THE RAT-TRYPANOSOME, TRYPANOSOMA LEWISI. 683 


Figs. 94 99—Rolled-up forms with » and N still single, in 27 begin- 
ning to divide. Figs. 24 and 25 are from the same preparation as the 
preceding ; figs. 26-29 are also twenty-four hours; the specimen in 
fig. 26 is evidently enlarged artificially by having dried before fixation. 

Figs. 30-37,—Multiplication of nuclei to 2 or 3 nn and 2 NN. 

Fig. 38.—Small sphere with 4 nn and 4 NN. 

Fig. 39.—Stage with 3 nn and2 NN. N.B.—Figs. 35-39 are all from 
the same preparation, showing the intense staining and consequent 
opacity of the body characteristic of the intracellular stages; only the 
principal flagella can be made out clearly, the daughter-flagella being 
scarcely, or not at all, visible. Twenty-four hours. 

Fig. 40.—Sphere with 3 nn and 3 NN, from the same preparation as 
fig. 26; evidently enlarged by flattening due to drying. Twenty-four 
hours. 

Fig. 41—A large sphere containing 13 nn and 13 NN, and two 
others ; the stain is over-extracted and the flagella are not seen. These 
specimens are from the same stomach as figs. 88-94 on pl. 37, in all of 
which flagella are plainly seen. Thirty-six hours. 

Fig. 42.—Two very large tailed spheres from the same preparation as 
figs. 18-25, 30, 33, 34: the flagella cannot be made out. Twenty-four 
hours. 

Figs. 43-46.—Small rolled-up forms, possibly in process of degenera- 
tion, from the same preparation as the last. 


PLATE 37. 


[Various stages of the stomach-phase from films preserved in sub- 
limate mixtures—sublimate-acetic, Schaudinn’s fluid, or Maier’s fluid— 
and stained with iron-hematoxylin, drawn with the camera lucida at 
a magnification of 3000. | 

Figs. 47-68.— Free (extracellular) trypanosomes, except 58 and 59. 

Figs. 47, 48.—Six hours; fig. 47, quite unmodified; fig. 47, slightly 
modified. 

Fig. 49.—Highteen hours. 

Figs. 50-55.—T wenty-four hours. 

Figs. 56-59.—Three days, all from the same preparation. Figs. 58 
and 59 are evidently early stages of the intracellular multiplication. 

Fig. 60.—Four days. 

Figs. 61-68.—From an abnormal flea taken from the infected breed- 
ing cage (see p. 519). Every possible grade of transition occurs in the 
preparation from quite ordinary forms, such as fig. 61, to large erithidial 


684 BE. A. MINCHIN AND J. D. THOMSON. 


forms, such as fig. 66, and of the latter some occur in couples adhering 
together, as in figs. 67 and 68. 

Figs. 69-73.—Recurved forms, twenty-four hours; in the specimen 
sbown in fig. 72 the flagellum appears to have become torn away from 
the body. 

Fig. 74.—First division of n in a rolled-up form. 

Fig. 75.—Recurved form with 2 nn and 2 NN; forty-eight hours. 

Figs. 76.—A rolled-up form in which the flagellum has become un- 
twisted from the body; from the same preparation as the last. 

Figs. 77, 78.—Two rolled-up forms from the same preparation, 
twenty-four hours. 

Figs. 79-81.—Twenty-four hours, all from the same preparation ; 
fig. 79, n divided, daughter-flagellum arismg, N beginning to divide ; 
figs. 80 and 81, similar stages. 

Figs. 82-86.—Twenty-four hours, all from the same preparation ; 
figs. 82, 83, each 2 nn, 1 N; fig. 85, 2 nn, 2 NN; fig. 84, 3 nn, 2 NN; 
fig. 86, tailed sphere with 8 nn, 8 NN. 

Fig. 87.—Recurved form, twenty-four hours. 


Figs. 88-94.—Various stages, all from the same preparation and from 
the same flea as fig. 41 on Pl. 36. Thirty-six hours. 


Fig. 95.—Small sphere with 3 nn and 3 NN; twenty-four hours. 

Fig. 96.—Large, nearly ripe sphere containing about ten trypano- 
somes, shown also in the photograph on Pl. 44, fig. 318; forty-eight 
hours. 


PLATE 38. 


[Epithelial cells of the stomach and stages of the stomach-phase 
from sections of the flea’s stomach, stained with Giemsa’s stain (with 
the exception of figs. 260a, 26la, 262a, 263a, and 264, for which see 
description of Pl. 42). Figs. 99-103 were fixed with Flemming’s fluid, 
but all the rest were fixed with Maier’s fluid. Drawn with the camera 
lucida at a magnification of 2000. | 


Figs. 97, 98.—Epithelial cells; fig. 97 shows a cell in the columnar 
form ; fig. 98 a flattened cell with, just beneath the border, a layer of 
granules staining quite differently from the blood-débris externally. 

Figs. 99-103.—Various details of the cells after Flemming-fixation ; 
fig. 99, red-staining grains in the upper end of the cell; fig. 100, red- 
staining grains mixed with osmic-blackened (fatty) grains ; fig. 101, cell 
comparatively free from granules ; fig. 102, pseudosphere in the upper 
end of a cell; fig. 103, two ‘ yellow bodies”’ in the upper end of a cell. 


THE RAT-TRYPANOSOME, TRYPANOSOMA LEWISI. 685 


Fig. 104.—Portion of the epithelium close to a erypt of regeneration, 
showing a recurved trypanosome attached to an epithelial cell and a 
trypanosome of ordinary type free in the débris close by. Twenty-four 
hours. 

Fig. 105.—Section shaving the side of an epithelial cell obliquely, 


showing a trypanosome of ordinary appearance attached to it. Twenty- 
four hours. 


Fig. 106.—Stout form of trypanosome free in the débris. Twenty- 
four hours. 

Fig. 107.—Recurved trypanosome within a cell. Twenty-four hours. 

Fig. 108.—Tailed sphere within a cell, also two smaller spheres and 
a portion of a trypanosome cut across; the nucleus of the cell does not 
come into the section. ‘Twenty-four hours. 

Fig. 109.—Portion. of the epithelium showing a very large tailed 
sphere, three smaller ones and a nucleus (N); the cells are much 
exhausted. Twenty-four hours. 

Fig. 110.—Section of a cell (not passing through the nucleus) showing 
a large tailed sphere and fragments of three smaller ones. Twenty- 
four hours. 

Fig. 111.—A large tailed sphere and a slice of a smaller one, in a cell 
near the nucleus (NV). Twenty-four hours. 

Fig. 112.—Numerous multiplication-stages in a partly broken-down 
cell of flattened type, and a cell nucleus (N). Twenty-four hours. 


Fig. 113.—Six spheres in a broken-down cell cast off into the blood- 
débris. Twenty-four hours. 


Figs. 114-116.—Three figures drawn from the same slide. Fig. 114, 
exhausted cell filled with trypanosomes; fig. 115, a large sphere ripe 
for breaking up into a mass of trypanosomes as in the last; fig. 116, 


large sphere. The preparation was over-extracted and does not show 
the flagella. Twenty-four hours. 


PLATE 39. 


[Stages of the stomach-phase, all from the same series of sections 
through nine stomachs preserved thirty-six hours after the infective 
feed in Flemming’s fluid and stained with iron-hematoxylin and Licht- 
grin-picric. Drawn with the camera lucida at a magnification of 2000. | 

Figs. 117-129.—Extracellular trypanosomes. 


Fig. 117.—Slightly club-shaped form, near the epithelium but not 
attached. 


Figs. 118, 119.—Recurved forms, not attached. 
Figs. 120-123.—Attached recurved forms, 


686 E. A. MINCHIN AND J. D. THOMSON. 


Fig. 124.—A small bunch of trypanosomes attached in the interspace 
between two cells. Structure of the trypanosome difficult to make out 
clearly. 

Fig. 125.—Trypanosomes in or attached to the débris of a necrosed 
cell. Two of them are recurved forms, the third possibly crithidial in type. 

Fig. 126.—A slightly hypertrophied epithelial cell containing three 
spheres of different sizes, with portions of the cells on either side of it. 
N, the nuclei of the cells; l.m., longitudinal muscles of the stomach in 
transverse section ; ¢.m., one of the circular muscles of the stomach. 

Fig. 127.—From the next section of the series, showing the same cell 
in which appears one of the same three spheres lodged in a distinct 
vacuole. 

Fig. 128.—Large hypertrophied cell with five nuclei in process of 
being thrown off from the epithelium ; it contains two spheres (sph.) 
and numerous yellow bodies. The same cell is shown in the photograph 
in Pl. I, fig. 3138. 

Fig. 129.—Flattened cell showing a sphere lodged in a vacuole beside 
the nucleus (NV). 

Fig. 130.—Cell containing four spheres close to the nucleus (NV). 

Fig. 131.—Young epithelial cell close to a crypt containing an early 
multiplication-stage just external to the nucleus (NV). 

Fig. 132.—Degenerated and cast-off cell, containing a sphere, the 
flagellum of which is sticking out from the remains of the cell. 

Fig. 133.— Recurved trypanosome within a cell. 

Fig. 1384.—Normal columnar epithelial cell containing spheres both 
above and below the nucleus (NV). 

Fig. 135.—Degenerated and cast-off epithelial cell, full of fatty 
deposits and yellow bodies, containing two spheres. 


PLATE 40. 


(Epithelial cells of the flea from sections of stomachs, all, except 
fig. 147, fixed with Flemming’s fluid and stained with iron-hematoxylin 
followed by Lichtgriin-picric. Figs. 136-140 are drawn at a magnifica- 
tion of 1000; all the rest at 2000. } 

Fig. 136.—Four young cells showing the partially developed border 
and the different positions of the nucleus in the cell. 

Fig. 137.—A cell showing the more granular condition. 

Figs. 138, 139.—Two stages of the fatty degeneration of the cell. In 
138 the nucleus, though obscured, is still faintly visible ; in 189 the cell 
has become an opaque black mass. 


THE RAT-TRYPANOSOME, TRYPANOSOMA LEWISI. 687 


Fig. 140.—T wo contiguous cells, of which the one to the left shows 
the beginning of the “ yellow necrosis.” 

Figs. 141, 142—The upper ends of cells containing “ yellow bodies ” ; 
in fig. 141 several such bodies, in fig. 142 one large one. 

Figs. 143, 144.—To show the deposition of osmic-blackened fatty 
grains in cells beginning to degenerate; fig. 145 in a flattened cell, 
fig. 144 in a columnar cell. 

Figs. 145, 146.—Cells containing pseudospheres (ps.) in their upper 
portions. 

Fig. 147.—To show the deposition of deeply-staining grains below 
the border of the epithelial cell. External to the cell are seen the 
coarse black grains of the blood-débris. Maier’s fluid, iron-hema- 
toxylin, Lichtgriin-picric. 

Fig. 148.—Outer end of a cell containing a sphere (sph.) which is 
seen to contain a darkly-staining mass (chromatin ? or fatty deposit ?). 


PLATE 41. 


[Various stages of the rectal-phase. Figs. 201, 202 are from sections 
fixed with Flemming’s fluid, stained with Giemsa’s stain and drawn at 
a magnification of 2000; all the other figures are from smear-prepara- 
tions fixed with osmic vapour, stained with Giemsa’s stain, and drawn 
at a magnification of 5000. | 


Figs. 149-152.—Early “tadpole” forms from the rectum, twenty- 
four hours; fig. 152 is perhaps an early stage of division. 

Figs. 153, 154.—Early stages from the rectum, twenty-four hours. 

Fig. 155.—Rectum, thirty-six hours. 

Fig. 156.—Rectum, twenty-four hours. 

Figs. 157, 158.—Rectum, thirty-six hours. 

Figs. 159-161.—From the rectum of another flea, thirty-six hours. 

Fig. 162.—Rectum, sixty hours. 

Figs. 163, 164.—rectum, three days. 

Figs. 165-170.—From a swarming rectal infection, three days. 

Figs. 171-176.—From a swarming rectal infection, about three and a 
half days old. 

Figs. 177, 178.—F rom the rectum, forty-eight hours. 

Figs. 179-182.—From the rectum about three and a half days. 


Figs. 183-187.—Early stages from the rectum, four days. In the 
clump shown in fig. 183 two trypanosomes of quite ordinary type are 
seen; these are probably forms which have come in with a later feed 
and have attached themselves secondarily to the clump. 

vot. 60, PART 4.—NEW SERIES. A7 


688 E, A. MINCHIN AND J. D. THOMSON. 


Figs. 188-190.—From the rectum, five days. 

Fig. 191—From the rectum of another flea, five days. 

Figs. 192-200.—From the recta of fleas taken from the infective 
breeding-cage, age uncertain. Fig. {192, early form from the rectum 
figs. 193-196, from stomach-preparations (post-pyloric?); fig. 197, 
rectum ; fig. 198, rectum; fig. 199, stomach (post-pyloric) ; fig. 200, 
rectum. 

Fig. 201.—F ree clump from a section of a rectum. 


Fig. 202.—From a section of the intestine close behind the pylorus, 
showing a number of haptomonads attached to the cuticle, and between 
them, projecting up above the level of the haptomonads and distin- 
guished from them by their lighter stain and terminal mn are numerous 
examples of the final trypanosome-forms (T). 


PLATE 42. 


[Various stages of the rectal-phase. Figs. 273-277 are from sections 
of recta fixed with Flemming’s fluid, stained with iron-hematoxylin, 
and drawn at a magnification of 2000; all the other figures are from 
smear-preparations fixed with sublimate mixtures, stained with iron- 
hematoxylin and drawn at a magnification of 3000.] 


Figs. 203, 204.—Forms from the intestine in process of migration to 
the rectum, twenty-four hours (see p. 568). 

Figs. 205-212.—Early “tadpole” forms from the rectum, twenty- 
four hours. 

Figs. 215, 214.—Early division-forms, rectum, forty-eight hours. 

Figs. 215, 216.—Early forms, one dividing, rectum, three days. 

Fig. 216 a.—Couple produced by an early division ? 

Figs. 217-220.—F rom a well-infected rectum, about three and a half 
days. 

Figs. 221-240.—Various forms from a swarming rectal infection of a 
flea that had been on an infected rat for five days; figs. 221-225, hapto- 
monads; figs. 226-230, division-forms; figs. 231-234, growth of flagel- 
lum figs. 235, 236, nectomonads ; figs. 237, 238, transitional forms ; 
figs. 239, 240, slender and stout trypanosome-forms. 

Figs. 241, 259.—Various forms from a swarming rectal infection, 
eight days; figs. 241-245, rounded haptomonads without free flagella ; 
figs. 246, 250, 251, flagella growing out from rounded haptomonads ; 
figs. 247, 248, 249, pear-shaped haptomonads without free flagella ; 
fig. 252, clump of pear-shaped haptomonads, some with well-developed 


THE RAT-TRYPANOSOME, TRYPANOSOMA LEWISI. 689 


flagella ; fig. 253, division-stage; fig. 254, nectomonad ; figs. 255-258 
transitional forms; figs. 259, final trypanosome-form. 

Figs. 260-264.—Various forms from a swarming rectal infection, 
taken from the infected breeding-cage, age unknown; figs. 260 a-263a, 
the same specimens as 260-263, restained in Twort’s stain ; fig. 264, a 
nectomonad stained with Twort’s stain. (N.B.—Figs. 260 a-263a and 
264 have been transferred to Pl. 38.) 

Figs. 265-272.—Various forms from a swarming rectal infection, age 
unknown; fig. 265, early form ? fig. 266, division-form ; fig. 267, rounded 
haptomonad, without flagellum, beginning to divide; figs. 268, 269, 
rounded haptomonads with free flagellum beginning to grow out ; 
fiz, 270, transitional form ; figs. 271, 272, final trypanosome-forms. 

Fig. 273.—Clump of crithidial forms attached in the intestine just 
behind the pylorus, in sections of a stomach of a flea preserved thirty- 
six hours after the infective feed. (x 2000.) 

Figs. 274-276.—From sections through the rectum of a flea preserved 
eight days after the infective feed. Fig. 274, free clump ; fig. 275, clamp 
attached to the cuticle of the rectum ; fig. 276, forms attached singly to 
the cuticle. (x 2000.) 

Fig. 277.—Haptomonads attached to the cuticle of the rectum, from 
sections through the rectum of a flea, preserved eleven days after the 
infective feed. (x 2000.) (This drawing is from the same section as that 
photographed in Pl. 44, fig. 317; the patch drawn lies just to the right 
hand of the middle of the three pointers that start from ¢ in fig. 317.) 


Figs. 278-288.—Leptomonas pattoni, various forms from the 
rectum of the flea. Figs. 278-284 a are from preparations fixed with 
Maier’s fluid and stained with iron-hematoxylin; figs. 285-288 are from 
preparations fixed with osmic vapour and stained with Giemsa. In the 
specimens drawn in figs. 281, 282, note small bodies marked # which 
appear to be endogenous buds (“ infective granules’’); in fig. 283 a 
similar body is shown free, and figs. 284, 284 a appear to be stages in 
the development of the bud into the leptomonad. Fig. 285 is evidently 
a nectomonad form, and fig. 286a stage in the development of such a 
form; they differ from the nectomonads of T. lewisi in the great 
prolongation of the body behind NV. 


PLATE 48. 


[Degenerative forms, fixed with osmic vapour and stained with 
Giemsa, drawn with the camera lucida to a magnification of 3000. | 


Figs. 289, 290.—Six hours, rectum. 


Figs. 291-296,—Twelve hours, rectum. 


690 E. A. MINCHIN AND J. D. THOMSON. 


Figs. 297, 298.—Recurved forms from the rectum, twelve hours. 
9 


Figs. 299-301.—From the rectum of another flea, twelve hours. 

Figs, 302-304.—Rectum, thirty-six hours. 

Fig. 305.—Stomach, forty-eight hours. 

Fig. 506.— Rectum, twenty-four hours. 

Fig. 307.—Rectum, sixty hours. 

Fig. 308.—Clump of degenerative forms, rectum, eighteen hours. 

Figs. 509, 310.—Agglomerating trypanosomes from a flea fed for the 
second time on an infected rat the day previously. 


PLATE 44. 


Fig 311.—Clump of degenerative forms (compare fig. 308). Photo. 
(x 1500.) 

Fig. 312.—Clump of developmental crithidial forms; at 7’ is seen a 
final trypanosome-form attached to the clump. From the same pre- 
paration as figs. 241-259. Photo. (x 2000.) 

Fig. 313.—Infected patch of stomach-epithelium in which the cells 
are breaking down and being thrown off. The Jarge multinucleate cell 
marked C is the cell, part of which is drawn in PI. 39, fig. 128. Photo. 
(x 600.) 

Fig. 314.—Section through an epithelial crypt of regeneration. Close 
beside it is seen a black, degenerated epithelial cell. Photo. (x 800.) 

Fig. 315.—Section of an infected patch of the stomach-epithelium 
showing the cells being thrown off; the cells contain coarse black 
(fatty) granules and stages of intracellular multiplication, which are 
not clearly seen. Photo. (x 600.) 

Fig. 316.—Section of the stomach-epithelium showing an epithelial 
crypt towards the middle ; to the left of the crypt the epithelium is old 
and degenerate and full of blackened fat; to the right of the crypt is 
new, clear epithelium budded off from it. Compare Text-fig. 2. Photo. 
x 300. 

Fig. 317.—Section of the wall of a rectum showing a swarming 
crithidial infection of the “ pile-carpet”’ type. Fig. 277 is drawn from 
this section, from the part between the middle and the right-hand 
pointers which start from ¢ to indicate the serried ranks of the crithidial 
forms. Photo. (x 500.) 

Figs. 318.—A sphere (the same that is drawn in fig. 96) and near it 


one of the long free trypanosomes of the stomach-type, from a smear, 
Photo, (x 1000.) 


THE RAT-TRYPANOSOME, TRYPANOSOMA LEWISI. 691 


Fig. 319.—Section passing through the opening of the pylorus into 
the intestine. Photo. (x 300.) 


PLATE 45. 


[General diagram of the entire life-cycle of Trypanosoma lewisi 
in the flea, combined from the observations and experiments set forth 
in this memoir in order to give a summarised idea of the complete 
course of events. The stomach-phase is represented on the upper side 
of the diagram, the rectal phase on the lower side and to the left; in the 
right-hand lower corner is shown the secondary infection of the pylorus. 
Magnification 2000. | 


1. Trypanosome as taken up by the flea from the rat. 

2. Trypanosome slightly modified after a few hours in the flea’s 
stomach. 

3-12. Cells of the epithelium of the stomach containing the various 
phases of the intracellular multiplication. 

3. Cell with two trypanosomes attached to it, one of ordinary, the 
other of recurved, type. 

4. Penetration of an epithelial cell by a trypanosome, drawn from 
Ndller’s description of the process. 
Recurved form in the cell. 


Or 


. Rolled-up form in the cell. 

. Early multiplication-form. 

_ Later multiplication ; 8 nn, 8 NN. 
. Large tailed sphere. 


Oo © aI SD 


10. Large ripe sphere without tail. 

11. Daughter-trypanosomes free in the exhausted cell after bursting 
of the sphere. 

12. Daughter-trypanosomes escaping from the cell. After being set 
free, the trypanosomes may do one of two things as shown by the 
arrows. They may each penetrate another epithelial cell and repeat 
the process of multiplication by which they were produced. Or they 
may pass through the pylorus and down the intestine to the rectum to 
give rise there to the rectal (crithidial) phase. 

13-18. To show the two possible ways in which the established rectal 
phase may arise from the stomach-trypanosomes : 

(a) 13, 14, 15, 16. Four successive stages of the contraction of a 
stomach-trypanosome into a pear-shaped form which at 17 divides by 


692 E. A. MINCHIN AND J. D. THOMSON. 


equal or subequal binary fission to produce two equipotential daughter- 
products, similar to 18, which by further-repeated binary fission produce 
the ordinary crithidial phase. . 

(b) 14a, 14b. Two successive stages of the contraction of a stomach- 
trypanosome into a club-shaped form, which at 17a divides by unequal 
binary fission to produce two inequipotential products, the larger 
parent-form which does not develop further and the smaller daughter- 
form, similar to 18, which by repeated binary fission produces the 
ordinary crithidial phase. 

19. Established rectal infection, showing the various forms of the 
rectal phase. h.h. Haptomonads, some of them dividing, some of them 
transitional to the other types. u.n. Nectomonads. ¢r. tr. Forms 
transitional from the crithidial type to T.T., the final trypanosome-form. 

The final trypanosome-forms, as shown by the arrows, pass finally 
out of the flea with the feces. 

The nectomonads can, under certain conditions of food-supply, 
migrate back along the intestine and fix themselves close behind the 
pylorus in order to give rise to— 

20. Secondary infection of the pyloric region of the intestine: the 
letters have the same significance as before. Final trypanosome-forms 
(7.T.) arise and pass down the intestine into the rectum and so out with 
the feces ; nectomonads (nn) also arise and may migrate back again to 
the rectum and re-establish the crithidial phase there ; these migrations 
are indicated by the arrows. 


INDEX FO, VOL: -60; 


NEW SERIES. 


Acrossota liposclera, a new 
genus and species of Aleyonarians, 
by Gilbert C. Bourne, 261 

Ageregatide, sporogony of, by Helen 
Pixell-Goodrich, 159 

Aleyonarians, a new genus of, by 
Bourne, 261 


Beaver, blastocyst and placenta of, 
by Arthur Willey, 175 

Bourne on A crossota liposclera, 
a new genus and species of Alcyo- 
narians, 261 


Centrifuged egg of the Frog, by 
Jenkinson, 61 

Ceratophyllus fasciatus, the 
Rat-flea, and its relation to the 
Rat-trypanosome, 463 

Chromosome complex of Culex 
pipiens, by Monica Taylor, 377 

Conus, anatomy of, by H. O. N. 
Shaw, 1 

Culex pipiens, chromosome com- 
plex of, by Monica Taylor, 377 

Culex pipiens, trypanosome from, 
399 


Dendy on the gametogenesis of 
Grantia compressa, 313 


Egg of the Frog, centrifuged, by 
Jenkinson, 61 


Flea of the Rat and the Rat- 
trypanosome, 463 

Frog, centrifuged egg of, by Jenkin- 
son, 61 


Gametogenesis of Grantia com- 
pressa, by Dendy, 313 

Grantia compressa, gameto- 
genesis of, by Dendy, 313 


Hexmoprotozoa, by Woodcock, 399 
Hamm on Stylops and Stylopisation, 
435 


Jenkinson on the development of the 
centrifuged egg of the Frog, 61 


Minchin, E. A., and J. D. Thomson 
on the Rat-trypanosome, Try- 
panosoma lewisi, inits relation 
to the Rat-flea, Ceratophyllus 
fasciatus, 463 


Nemertines, proboscidean system of, 
by Wynhoff, 273 


Owl, Little, development of its try- 
panosome, 399 


Pixell-Goodrich on the sporogony 
and systematic position of the 
Ageoregatide, 159 

Placenta and blastocyst of 
Beaver, by Willey, 175 


the 


694. 


Rat-trypanosome and its relation to 
the Rat-flea, by E. A. Minchin and 
J. D. Thomson, 463 


Sex, experimental analysis of, 
Part 11, by Geoffrey Smith, 
435 


Shaw on the anatomy of Conus, 
1 

Smith, Geoffrey, on Stylops and 
Stylopisation, 435 

Stylops and Stylopisation, by 
Geoffrey Smith and A. H. Hamm, 
435 ; 


| 


INDEX. 


Taylor, Monica, on the chromosome 
complex of Culex pipiens, 377 
Thomson, J. D., see Minchin. 
Trypanosoma lewisi and 
relation to the Rat-flea, 463 
Trypanosoma noctue in Culex 
pipiens, by Woodcock, 399 


its 


Willey on the blastocyst and placenta 
of the Beaver, 175 

Woodcock on Avian Hemoprotozoa, 
399 

Wynhoff on the proboscidean system 
of Nemertines, 273 


ADLARD AND SON, IMPR., LONDON AND DORKING 


¥ 


bd 2 

rie aia 

| re 8 

_ ion 

= 

? 4 ; 
» 


Trg im So 


Quart. PurnMNicr SéVot, b0 NS HOT, 


oe 
' 
oO 
% 


$V 
V 

V 

Vv ; 


Huth, Lith? London. 
Poms S TEXTILE. 


BHIQyu,. YOUWTN,. WAT NNN CY OL LV ACSA 


Hath, Lith? London. 


SHAW COWS: TWIP As 


PCIe ee eA Nelo) ss \ Atel a) tay le SANOD—MVHS 
“TopmoeT SqIwy YT 
Estes (cj ----4"70 


‘Le 


SAB SM 09S YP Cong prong 


SHAW —CONUS TE 


Quart. Sourn.Mion Seu, Vol, CONS. Lt, 6. 


Huth, Lith” London. 


MmooONWS TULIPA. 


= 


Ul 7 _ 7 —~ ao, oe 2 eo ee ee eee 


= 
a 


Ns 


JENKINSON— CER 


y ST 


Duart. Yourn Mior Sei. Vol, 60. NSE E&Y, 7, 


f2 . 
Oa. 


Huth, Lok? Londo 


UGED FROGS EGG. 


a 
A ete. 


= 

se 

som tades. 
5 

a! 


Deano 


ot 
=> = 


sO ZogI6 
39 989 DON 
CRU ee 


sr) 
ts) 
lore) 


20, 
een 


ose 
Fe BO 


3.00 
B59, 


Quart Journ NMior SHVALONS J46. . 


OAc S357, 9, 
apenotg 
< 


CaS 
“ohana 
pay 


. 


2se east 3 
Sao 2 Omens 
he she 


380. 
29 88% 


S 


o 


2 


, 
ie 
° 


a 


C 
Ly 
Ct) 


0 
do’ 
oes. 


a ©, 
98% 92, 
a0 40°°, 


Huth, Lith London. 


) 7 
' 
4 £ j 
, - ’ 
ae - ¥ 
~ a 
ial 
.4 7 - 
= . 
= f 
a “ 
a 
7 
a ‘ > 4 
a 
* ’ Y . 
* 
: 
7 . 
7 
‘ 
= - 


Quart Gourn Mier Si Vb CONE I, 


sessberer a 2: 


At 


ee 


a 


Yo. 
exe) 


ee 
eibacbe sages: 
%! 


9. 


2899 


aie 


pions 
Oe aao2eee: 
oan 


0: 
oe Sok re ed 
CER aA BoP n aces 
S59; exehlal 6 EOD L6 Ss PS Oo 
LT Tihge a a 20h OO! 


2S “J 
Ries 


Dery 3, 
s 
PRs 
Moret 
Benen 
SifaP 
. 


JENKINSON—CENTRIFUGED FROGS EGG. 


Huth, Lith® London. 


a 


Quart. IournNicorsAVot CONS H10. 


Huth, Lith™ London. 
JENKINSON-CENTRIFUGED FROGS EGG. 


aime 


JENKINSON— CEI 


St 5 


ere. 6) 


41. 


FROGS. BGG. 


Quart. Sourn Mioresi. Vol, GON SM, 1] 


4.3 
Huth, Lith” London 


a 


JENKINSON —GIl 


Luar fourn MiorSiVAbO NS #12. 


Huth, Lith® London. 


feck FROGS EGG. 


H.L.M.GOODRI 


IT, 


FOWUVIVC 


- 


othr U, e 


ondson 


uth, L 


GGREGATIDEA. 


7 


WILLEY—S8&. 


rb, Lourry Mice Sci, ol, 60 ME LU, 14. 


Huth, Lith? London 


'S PLACENTA. 


et 


WILLEY—Bes 


QuartDourndhior Sc: Yt CONS ICM. 


, 00 9erhveho-4ega a4 
0 Go°'O.6 © OFS 9 Caerare ge 


= O8 


Huth. Lith’ London. 


iS PLACENTA. 


fw 


WILLE Y=3aae 


Quart. ourrucMicr isi: UACONS Fb16. 


So 
98 


FDA SO“BIPG © 


Huth. Lith? London. 


~ CK OLN > .... 
' . PAD Oso RAs = 

; & —s—s—s—s ss OobG0 6 OOS 

l : 

re) } : 


[ole 
7m meoleye) CoaocsouuD ole 


RPS OOOO 


RBOoSo 2) 


co 
. . v HO Ol OOo 
4 > = eee ROO 


ne STO OTOIOSS 


’S PLACENTA 


WILLEY-BEA 


Quart-PurrvMicrstilt: CON SIA,17. 


a 


a a 


fol 


OORT 


POETICS) 


TOISIOO OMS TSS 


ee SOOIS>) 


Bese TITIS> 


PAIOOO MIAO cOOaw 


ey 


Gr 


£s 
TortoMTolose 
OBLGHOTLO (OJOS?) 


hath. Lith® I,ondon. 


PLACENTA. 


A p> “er htt: PENT ae ee Utes 


STD, ; NM Zz 
2 Spat = 


ISOS af 


£0) 


QE. 
ee =~ Wipiaieksan 
‘ 5 eqyanialan 


WILLEY= BEA 


c 
Ae 
), 


ti UAOON ES: PLE 


Quart. fwrMicr 


Huth. Lith? London 


PLACENTA. 


Qa scccccee 


Ss 


W Tole. 


te DourmMior SUA CON SGA, 19. 


i 


FEE © LF & LA. 


Huth. Lith? London. 


\VER’S PLACENTA, 


WILLEY — BES 


QyartGourabicrnSci: Ut CONS L4: 20. 


nth. Lith’ London. 


5 PLACENTA. 


+ : ‘ 
Te 
- if 
a 
? 
<< f . 
. ey By 
5 . 
tam 
&. Res : ‘ 
ty 
. . 
" ’ 
. 
° : 
r 
. 
, ” BS - 
e ‘ re =. — wr - 


Quart; SourrMuy Sci, Vol, 60 VS &%21 


. 


Be. Huth hith 


* Landon 


WILLEY—-BEAVER’S PLACENTA. 


© 


BOURNE— 


QDuartofourv Mice iS Vol, 60 NE IY, Z be 


Huth, Lath" London. 


ROSSOTA. 


et ee re, Fa ae ee ne Jn eee ee 


39 
Auth, Lith? Landen. 


Cit tee 2 ee I ee ley eS 


2 
; 
’ 
) 
5 


PAL 


A.Dendy del. 


Sourn Mor Sci. Vol, ONS. 4h, 24. 


Pi 


Huth, Lath? London. 


re 


ite J 


Huth, Lith? London. 


74, yee 


we 


A.Dendy,del, 


Luart foun Mor sa.Vi. CONS. G4. 26. 


Huth, Lith? London, 


we 


t i 
) ah a 4 . aba’ 


-o2 Tm, ‘olmm, 


Scale of figs.3 &6. : Scale of all figs. except 3 &6. 


M. TAYLOR-CHR( 


Luan. Iaurn. Mier Si. Vol. CON SH. 27 


mes OF CULEX, 


1" 


ca 


Quart SournMior SiVAbCONS £4.28. 


M.TAYLOR-—CHROMOSOMES OF CULEX. 


Quart. Fourn Mion’ Sci.Vot.60,NS. F1.29. 


Huth, London. 


(/PANOSOMA. 


Quant. Sourn. MicrSiV’ltA CONS 30, 


Huth, London. 


ue 


TNOCLTUASZ 


DEVELOPMENT IN THE MOSQUITO. 


wooodcoc! 


AYPANOSOMA,. 


Qnant, Foun. Mier SVot. CONS. Pt.3). 


T. FRIN GILLINARUM. 


DEVELOPMENTAL STAGES IN CULTURES. 


re 


om" 


ee 2 ae oe 


— 


a 


b 


. 


’ 
‘ 
. 


au 


GEOFFREX am 


a 
A 


Quant.furn Mien Sei, Vt CONE Lt. 3 


C) 


@) 


a 
as fi ; % i 
j (fy )§ 
fall AS aya) 
\ Ae = g 
LBA be 
iff 7? /o— 
4 \L6 
r, } j 
p Mg 
es 
- YZ 
i g 
Oe Vv 
Y 


i cj 
on 


} \ 
za I 
j { He 2 


o. 


“Hath Lith? London 


mot YLOPS. 


. 


GEOFFREY Sm 


) 
Ces 


LU, 


7 


Micwirei. Ul. 00,.N.5 


QDuart.fowrrn 


SY) 
- 
if 


car anns 


ah} 


{A-. 


Thaker 
ust eS 


0) . & Uf rar, 94S AI NE AD, 23 
WUE OL UME Wat 60 Nio.kd 4, 


1 


St, ‘Lopised é 


20.m7 
Stylopised ? 


omen 2 en OS 


Stylopts ed £ 


Hith Lith? London 


W.G. Browmng del 


GEOFFREY SMITH — STYLOPS 


47 


9) / Gi tf. (7 Vy fay Ne Oo) 2 | 
Hart CLOMYVY, NULIIIN, C MH) wy i a) SALOU, 


eer 


Normal & 


~T A 
Normal o 


GEOFFREY SMITH —STYLOPS. 


ra 
3 
& 
q 
rd 
o 
nm 
o 
ro 
E ; 
= 


t JourneNienSct Vet. 60 NS. La. 36. 


Huth,London _ 


Ww 
ze?) 
{ 
Cs 
ao 
| 
= 
Oo 
< 
= 
oO 
a oe 
ie) 
| 
< 
uJ 
eo 
ra 
uw 
ES 
as 
= 
a 
3 


‘ 

; J 

eu - 
e be 

a L 

- = 
: 7 


-_ . a = . 


iia=as yi ae iy FO: eye hk, ee, 


GE 


vn 
Meant) 3 


TF 


M. Rhodes, ad nat del. 


MINCHIN AND THOMSON —TRYPANOSON 


“2 


LE 


( 
Huth Lith? London. 


MéenSci: ol. 00 NS AH. 


Quart: Lown, 


WIS| IN THE FLEA— STOMAGCH—PHASE. 


) yt 
‘) a 
f 
1 
. 
. 
<= 
« 
/ 
. 
‘ 
: 
me 
» 
; 
g 


'q s i a Vit~ met Perio. i he | 7 


> ( 
red! 


aN 


4 


M.Rhodes adnat.del. 


Quare. Pourn. Mor sci: VAt.CONS $6.36. 


264. 
Huth, Lith? London. 


WISI] IN THE FLEA —STOMACH-—PHASE. 


te" 


ye Fer See 7 es lee Pe ee Ort fer agi hee | “j . 


M.Rhodes,adnat.del. 


MINCHIN AND THOMSON —TRYPANO 


Gis 


Fal 


-EWIS) IN THE FLEA—STOMACH —PHASE. 


Huth Lith? London 


se 


t) 


“ 


Quart, PournrMicmse,Vol, 60 N.S L 40 


‘i tel 
a 


148. 
M.Bhodes,adnat.del. Hoth Jath? London 


MINCHIN AND THOMSON —EPITHELIUM OF THE STOMACH OF THE FLEA. 


* 


Quart, PourncMicnse,Vol. 60 NS, Lt, 40 


M. Hhodes,adnat.del. Hnth Tath? London 


MINCHIN AND THOMSON —EPITHELIUM OF THE STOMACH OF THE FLEA. 


ee: 
M Rhodes, ad nat. del. 


Quart, SournMuorSeaVot. 60 NS. 447. 


¥ 168. Huth,London 
VISI IN THE FLEA —RECTAL PHASE. 


U \ 2842 ; 
288. 2887. 286. 285. AE Be 
M.Bhodes ad nat del. 


MINCHIN AND THOMSON—TRYPANOSOMA LEWIS! & LE 


Quart. fourm, Micr’ Sci: Vot.6C0NS.F6.42. 


ae, 221. 222. 
a> 


228. 


ey 


aol, "260: 


Huth London. 


P1.38. 


For Fig? aoe ei262%263, 264, See 


MONAS PATTON! IN THE FLEA—RECTAL PHASE. 


i) 3 2 ee 


Dart. Souvr Mier Sct.U'A. CON EAU. 


a 


M.Rhodee adnat.del. w Huth, Lith? London. 


MINCHIN AND THOMSON — TRYPANOSOMA LEWIS! IN THE FLEA. 


ee ee ee ee, ee ee —_— ss eas fF 


3ll. | 


Gl 


MINCHIN AND THOMSON — TRY PA 


Quart. fourn Mier Sci VA.b60 NS A44 


33; 


OMA LEWIS! IN THE FLEA. 


MINCHIN AND THOMSON—LIFE—CYCLI 


Quart, Lourm, Mor Sc. Vot.6GO N. CMA. 


Ly Des . 
ve z Ga i Pa 
PENA 
Vasteyy) 2 
q f . Z 


Ld w 
6 45 RE ot v Huth, London. 


TRYPANOSOMA LEWIS! IN THE FLEA. 


[Po Pod 


A Coa 


ms 
4 


Al vy va 


Ba i) te 
byt Vaan 


sf 
hai 
DI | 
f hi : 
} Pid 
‘ ‘ 
; ‘ 
ae 
' a, { 
s id | 
' e 
ry ; vt 
pati: 
UN 
yt RAY on} 
' " 


wie! 


{ tO Nhy F 
Bt , . AD) Dusk i 4 ) 
i | rh ba iI ah | | if 


wy 


’ 


yi Hive, 


—_— ae ee 


Tien) ore 
ay 


UM etd 
- mf Ai , 
1 a s)he 
i! i yea 
st 
aa vi, 


iy 
1 
i re 
% 
i 
rg Fal 
é 
of 
: q 
i 
‘ 
= 
if 
“ 
; i 
iy) 
q 
f , 
i, 


i A 
cd 


BN 
a 


\ ay Ni y Ye | ty 
My AR yi aa (yOu 
i : ha A A hie Ji \ iM , 
: ne i ‘ (shalt ae i a ’ } 6) aL iy 
f We: i i} - Mie i : t i Ly, i 4 ’ 
‘ os +4 ae i has boty Wa A al 
1 i i 2 mat) CN ar am 
ak 1 ue Dd ay tm r bib 4 Mt M vy Ad 4 
i * y ho} play ih} ve Ve, | \ Viney | if / v0 } 
te ke ar . Wik 1) ae oi delt cf 
L or | F af, Noe i | 5 Us ity i ; i \ 
} ' : ‘ if iv ‘ Law : 
. Pe VM Nuh 
Dest’? es re Ne: Rae es ii! 7 
4 { i Th | MAT. ey at) 
he Wale We, it i i Ni rick ibh| We ny 


‘9 
Te an 


yy Re: : 5 Ma PNM ERD MAK Sy AD ed ee 
Tae nk ae MPO RN AY Wit AOR Toei 


Bema ae) 


i f?) ani! bP ien 
Dit pe ant! Minny eye UTE ih a ated | 
an tea, vat 1th) oy u Yr Vere rv 
cP 5) a RA din Baty 
re beg aE 'y) aly OLN hea pad) zi) aa: 
Ain Ween it AT) Pe ala PAD if " 
BA: fi hic AN Oe nn a ra 


v4 : i i) Ril ? 
Vauiie 


2 


? y JOGRD te ran aa 
= 7 Ure La (tye 7 
sony oe BAe nu .tfA i Ligce even Ma 
Ye an, AA aU Noth Na te 
: ; i MNS 


NE 
ge iy 


i 
i 


4 ic@ } 
“ ah cet 0 im : 
f Aven Nr: 
05 ay et ie he 
RVs) | 
i 3 
5 ; tre y ; 
Wy PPS TY " 
i ye i ‘ 
‘ 4 if u i 


: ry = e 
r . , _ 7 » * Poe id - As se i to 


QH Quarterly journal of micro- 
201 scopical science 

Q2 

D's 

v.60 

cop.2 


Zeology- 


BIOLOGICAL & 
MEDICAL SERIALS 


PLEASE DO NOT REMOVE 
CARDS OR SLIPS FROM THIS POCKET 


UNIVERSITY OF TORONTO LIBRARY 


UND 
1E 


a ne 


OeN oe 


Saami eng em eee 


ee iad 


ynhim intr “oe 
Sy iaiea nean pad al natty 


eee 
ere 


lS OS 


naan nat enbnapnyn bere or rete aa? 


eee 
enehatent wtehelee wen ane