HARVARD UNIVERSITY 


al 


LIBRARY 


OF THE 


Museum of Comparative Zoology 


BULLETIN 


OF THE 


MUSEUM OF COMPARATIVE ZOOLOGY 


HARVARD COLLEGE, IN CAMBRIDGE 


VOL. XXII. 


CAMBRIDGE, MASS., U.S. A. 
1891-92. 


Reprinted with the permission of the original publisher 
KRAUS REPRINT CORPORATION 
New York 
1967 


Printed in U.S.A. 


CONTENTS. 


No. 1.— Contributions from the Zoölogical Laboratory. XXVIII. Observa- 
tions on Budding in Paludicella and some other Bryozoa. By C. B. 
DAVENPORT. (12 Plates.) December, 1891 


No. 2. — Contributions from the Zoölogical Laboratory. XXIX. The Gas- 
trulation of Aurelia flavidula, Pér. & Les. By Frank Smirnu. (2 Plates.) 
December, 1891 


No. 3. — Contributions from the Zoölogical Laboratory. XXX. Amitosis in 
the Embryonal Envelopes of the Scorpion. By H. P. Jonnson. (8 Plates.) 


January, 1892 


No. 4. — A Fourth Supplement to the Fifth Volume of the Terrestrial Air- 
Breathing Mollusks of the United States and Adjacent Territories. By 
W. G. Binney. (4 Plates.) January, 1892 


PAGE 


163 


No. 1.— Observations on Budding in Paludicella and some other 
Bryozoa. By ©. B. Davenrorr.! 


CONTENTS. 
PAGE PAGE 
A. SPECIAL PART. III. Budding in Marine Gymno- 
lemata . «s . 40 
1, Architecture of the ‘Stock . 40 
2. Origin and Development of 
the Individual. . . . 53 
3. Regeneration of the Polypide 64 


I. Introduction ie Bike ab 

IL. Budding in Paludicella A eK 
1. Architecture of the Stock 
2. Histology of the seinen 


Region . . 4 4 
3. Origin of the Polyplde. in ANS g= nA pS ok ane 66 
ssue ylactolemata . 
the Terminal Bud. . . . 7 oie y 
4. Origin and Development of B. GENERAL CONSIDERATIONS. 
> Lateral Branches . . ç 
m p) I. Laws of Budding . oen 71 


ao 


» Development of the Body 
WEHR 12 
6. Development of the Polyps 18 
7. Origin of the Muscles . . 27 

8. Formation of the Neck and 
Atrial Opening . . . . 81 

9. Development of the Com- 


II. Relation of the Obserrations on 
Budding in Bryozoa to the 
serm Layer Theory . . . 88 

II. On some Characteristics of 
Gemmiparous Tissue . . 98 

IV. Relationships of Endoprocta 


canes and Botapnoctä, s.i ace a0 
munication Plate . . . 82 

10. Role of the Mesodermal Summar al ee ee OG 

Vacuolated Cells . . . 84/Literaturecitd . ..... . 109 


A. SPECIAL PART. 


I. Introduction. 


Tun somewhat heterogencous studies here brought together have been 
prosecuted at different times and in different places, as opportunity for 
getting light on the problem of non-sexual reproduction as exhibited in 
the group of Bryozoa has presented itself. 

While studies on the fresh water species were pursued chiefly here at 
Cambridge, those on marine Bryozoa were made while occupying one of 
the tables of the Museum at the United States Fish Commission Labora- 


1 Contributions from the Zoölogical Laboratory of the Museum of Comparative 
Zoölogy, under the direction of E. L. Mark, No. XXVIII. 
VOL XXIL — NO. l. 1 


2 BULLETIN OF THE 


tory at Wood’s Holl, Mass., during the summer of 1889, and while at 
Mr Agassiz’s Newport Laboratory during the summer of 1890. To my 
instructor, Dr. E. L. Mark, for many valuable suggestions during the 
progress of my work and the writing of this paper, to Mr. Alexander 
Agassiz, for the kind hospitality accorded me at his Newport Laboratory, 
and to Hon. Marshall McDonald, United States Commissioner of Fish 
and Fisheries, and Dr. H. V. Wilson, Assistant at Wood’s Holl, for fa- 
vors shown me while at the Wood’s Holl Laboratory, I make grateful 
acknowledgment of my indebtedness. 

A word as to localities. The marine Bryozoa were found especially 
abundant at Newport on floating eel-grass in the cove and on the piles 
of the wharf. The embryos of Cristatella and Plumatella were found in 
colonies which literally covered the bottom of some parts of the south 
or shady side of Trinity Lake, Pound Ridge, New York. They occur 
especially in densely shaded and fairly deep water near the shore. 

The Gymnolemata present many difficulties to finer technique. They 
possess a chitinous covering, often very thick, and frequently, in addition, 
a calcarous skeleton. When the latter is present, picro-nitric acid mixed 
with sea water is a fairly good fixing reagent ; when it is absent, hot cor- 
rosive sublimate was most serviceable, The objects must be transferred 
through the grades of alcohol with extreme caution, to prevent the col- 
lapse of the ectocyst. I used the chloroform-paraffin method of em- 
bedding in order to make transfers more gradual at this stage. Some 
difficulty was experienced in staining such small objects on the slide, since 
the tissues are very loosely associated ; and on the other hand in toto 
staining is unsatisfactory in some cases, owing to impenetrability of the 
ectocyst. Often it-was necessary to open the body cavity of each indi- 
vidual by means of a sharp knife or needle. The best results were 
obtained with alcoholic dyes like Kleinenberg’s hematoxylin and 
Mayer’s cochineal; although Ehrlich’s hematoxylin was often used 


with success. 


II. Budding in Paludicella. 
J. ARCHITECTURE or THE STOCK. 

Paludicella, as is well known, occurs in quiet streams and forms 
stocks on the under surfaces of stones and other objects. Seen with the 
naked eye these stocks appear as a fine lacework, composed of constantly 
branching lines of individuals, Some of the stocks which I have meas- 


ured are over 25 mm. in length along their greatest diameter. 


MUSEUM OF COMPARATIVE ZOÖLOGY. 9 


When the stock is studied more carefully, it is seen that the individ- 
uals which compose it are arranged one in front of the other, forming 
lines. (Figs. 1, 2, 2%) We may distinguish (1) a single primary branch, 
which formsa continuous line from the oldest individual, which has been 
derived directly from the egg, to the terminal one; and (2) secondary 
branches, which arise from the individuals of the primary branch and at 
right angles to theiraxes. Typically, a secondary branch arises from both 
the right and left sides of each adult member of the primary branch, but 
in some cases the secondary branch of only one side appears to be formed. 
The secondary branches are composed, like the primary, of a continuous 
line of individuals placed end to end. These in turn give rise to ter- 
tiary branches, which run out at right angles to the right and to the 
left of the secondary ones, and hence parallel to the primary branches. 
Quaternary branches may occur in like manner, but I have never seen 
branches of a higher order than the fourth. All of these branches may 
lie in one plane, but frequently some of the lateral buds are so placed 
that they give origin to secondary branches which rise above the plane of 
the object upon which the stock lies. A study of Figure 1 and the cor- 
responding diagram, Figure 2, reveals some additional facts. The two 
lateral buds of an individual do not arise at the same time, and there is 
a tendency for the first, and therefore oldest and most developed, sec- 
ondary branches to arise alternately on opposite sides of the ‘primary 
branch. This last rule has many exceptions, however. 

The long axis of the individual coincides with that of its branca ; the 
sagittal plane lies in that axis, and at right angles to the substratum. 
The atrial opening is near the distal end of the individual in the sagit- 
tal plane, and is turned away from the substratum. The anal aspect of 
the polypide is placed nearer the tip of the branch, — hence distad ; 
the mouth, on the contrary, proximad. 

A very casual observation shows that not all branches nor all individ- 
uals are of the same size. The shortest and therefore youngest branches 
are placed most distally, and are seen as small buds. The terminal indi- 
viduals of the branches are also evidently less well developed than the 
more proximal ones. The adult individuals measure from 1.5 to 2.0 mm. 
in length and from 0.30 to 0.35 mm. in width. The younger individ- 
uals differ from the older in form also. The outline of the adult branch, 
looked at from the side, and disregarding the atrial opening, is formed by 
a series of beautiful sigmoid curves (Fig.9). The concave and convex 
points of the upper and lower sides of an individual are not placed exactly 
opposite each other, and the lower (abatrial) side approximates more 


BULLETIN OF THE 


4 


nearly to a straight line, The point at which the upper and lower 
curves most nearly approach each other is where the separation of two 
individuals takes place; that at which they are farthest apart is the 
middle of the zoccium, occupied by the polypide and sexual organs. 
The outlines of the young zoweia are straighter, and their breadth is 
considerably less than that of the adult. 

From what we have already seen, the method of growth of the stock is 
perfectly evident: it is by the formation of new median buds at the tips 
of existing branches, and of new branches from lateral buds. In order 
to understand the origin of the individuals of the primary branches, to 
which subject we will first turn our attention, we must study the tips 


of the branches. 


9. HıstoLogy or tim Buppine REGION. 


Figures 7-9 will serve to show more in detail the method of formation 
of new terminal individuals. We find in these cases one polypide already 
pretty well developed and attached to the body wall by means of the kamp- 
toderm at about the point at which the pyramidal muscles (mu. pyr.) 
are seen to be forming. That portion of the animal which extends from 
about the region of formation of the muscles to a point a little proxi- 
mad of the tip represents the region which will go to form the new in- 
dividual. The tip itself, for reasons which will presently appear, is not 
to be included in the terminal individual. The tip of the branch is to 
be regarded as homologous with the margin of the corm in corm-building 
genera of Gymnolemata. Figures 7-9 (gn.) also show the position of 
the bud which is to produce the polypide. By consulting first Figure 9, 
in which the polypide bud is apparent, the significance of the swellings 
of the body wall in Figures 8 and 7 becomes clear. 

Figure 14 (Plate II.) represents a stage in the development of the 
polypide bud, somewhat later than that shownin Figure 9, and this may 
serve us as a starting point in our study of the origin of a new individual, 
and, first of all, of the new polypide. The whole of Figure 14, from the 
tip down to the neck of the older polypide (cev. pyd.), may be divided, for 
convenience, into three zones: first, that distad of the young bud, which 
may be called the tip of the branch (Fig. 14, a to B) ; secondly, the region 
of the bud itself, which may be called the gemmiparous zone (B toy); aud 
thirdly, the region between this last zone and the neck of the older poly- 
pide, which, for want of a better name, may be called the proaimal zone 
(y to 8). In the formation of a new polypide between a and £, that 
region will in turn become divisible into the three zones just named, 


MUSEUM OF COMPARATIVE ZOÖLOGY. 5 


exactly as the region a to § represented the Zip of the branch when the 
older polypide, whose neck is shown at cev. pyd., was of the age that the 
younger bud is now. It will be necessary first of all to study carefully 
each of these three regions before treating of their origin and fate. 

The tip of the branch consists of the two layers of cells which are found 
in other parts of the body wall, — the ectoderm and the mesoderm, as the 
coolomic epithelium may, for brevity’s sake, be called. The cells of the 
ectoderm at the extreme tip (Plate I. Fig. 6) are greatly elongated, form- 
ing a columnar epithelium. There are about 25 or 30 of the larger cells. 
They have a length of 28 p to 32 a, and a diameter of about 4u. They 
possess an ovoid nucleus averaging 5.7 a by 2.6 a, which lies in the middle 
of the cell but slightly nearer the coolomic epithelium than the cuticula. 
It possesses a large nucleolus over la in diameter, which often appears 
stellate owing to the threads of plasma surrounding and proceeding from 
it and forminga nuclear network. As the figure shows, the plasma of 
the cell is filled with large, apparently deeply stained granules, some of 
the largest being over 0.6 u in diameter. The coarser granules lie chiefly 
in the immediate vicinity of the nucleus, but are also found arranged in 
long lines at right angles to the surface throughout the greater part of the 
cell, becoming finer the farther they lic from the nucleus. A fine network 
can sometimes be made out between the large granules, but this appear- 
ance is more evident at the peripheral portion of the cell, where there are 
no large granules. At the outer and inner ends of the cells one finds large 
vacuoles, the largest of which are of about the same size as the nucleus ; 
these become smaller the nearer they lie to the nucleus. In many cases 
the larger vacuoles are each seen to be partly filled by a body which stains 
slightly, and, as focusing determines, is more highly refractive than the 
plasma. Similar highly refracting, slightly staining granules are found 
Owing to the 


in, and in fact often composing, the smaller ‘ vacuoles.” 


fact that the deeply staining granules lie near the nuclei, and that the 
vacuolated and finely granular plasma lies more remote, there is a very 
marked deeply staining band occupying the middle of the ectodermal 
layer, and having about four tenths the thickness of the whole layer. 

At the outer ends of the cells, and doubtless secreted by them, there is 
a cuticula about 1 thick, Its inner surface is sharply marked off from 
the underlying plasma; its onter surface is less sharp, and there are 
usually very minute particles of dirt attached to it (not represented in the 
figure). The whole cuticula forms in section a continuous band of 
substance, which stains deeply in Ehrlich’s hematoxylin (but not at all 
in alum cochineal), and covers nearly the whole tip. Looked at from 


6 BULLETIN OF THE 


the surface after staining in hematoxylin, it appears uniformly dark. 
The mesoderm of the tip is highly modified, and a description of it 
will be more instructive after I shall have described the normal coelomic 
epithelium, as I shall do later. 

Passing from the extreme tip towards 8 (Fig. 14), one finds the ecto- 
dermal cells gradually changing in form, size, and structure, and becoming 
slightly broader, and very much shorter. Their nuclei lie near the inner 
ends of the cells, possess a thick “ nuclear membrane,” and are more 
nearly spherical than those of the columnar cells, but of about the same 
size. They each possess one very large, centrally placed nucleolus, whose 
diameter equals and sometimes exceeds one third that of the nucleus, and 
whose outline is often somewhat stellate. Outside of the nucleus in the 
cell body there are fewer and fewer vacuoles as we pass from the tip, but 
the plasma is still coarsely granular, and here, as before, these stained 
granules surround the nucleus. It is now the regions between cells 
rather than those at the inner and outer ends which remain unstained, 
so that the cells are separated from one another by light spaces. 

The mesodermal layer becomes somewhat thinner than at the tip, that 
is to say, its cells are flattened. The nuclei are elongated in the axis of 
the branch, and average about 4p by 2.2 p. They possess one spheri- 
cal nucleolus, whose diameter is about two thirds of the minor axis of 
the nucleus. Small, clear vacuoles often with highly refractive spherical 
bodies are abundant in the cell protoplasm, which stains as a whole less 
deeply than does the ectoderm. Such highly vacuolated elements will 
be called reticulated cells. 

If we study the gemmiparous zone at a stage considerably earlier than 
that shown in Figure 14, in fact at a stage in which a polypide is about to 
arise, we find an appearance of the layers represented by Plate I. Fig. 3. 
In such a region the ectoderm consists of cuboid cells about 7p high by 
6.5 abroad. The nuclei are large, nearly spherical, and vary in size from 
3.5t06.0p. The largest nuclei are those in the region from which a bud 
is about to arise (ex.). One in this region (to the right of ex.) is 6.5 u 
by 6.0 p in diameter, with a nearly spherical, eccentrically placed nucleo- 
lus of about 3.0 p in diameter. This nucleus is the largest which I have 
found in the whole tissue of Paludicella, and the same is true of the nucle- 
olus. From the examination of many regions from which buds are about 
to arise, I can assert that such regions always, in Paludicella, possess large 
nuclei and large deeply staining nucleoli, I shall have occasion to de- 
scribe similar conditions elsewhere, and to point out the probable signifi- 

sance of these facts. The cell body possesses a highly granular, deeply 


MUSEUM OF COMPARATIVE ZOOLOGY. T 
staining plasma; the inner ends of the cells, however, do not stain so 
deeply as the middle or peripheral portions. 

The cuticula (omitted from Fig. 3, see Fig. 5) is usually somewhat 
different in appearance from that at the extreme tip. In section we can 
distinguish two layers : an outer, thicker, deeply staining layer, which is 
not continuous but appears broken into larger or smaller bits; and an 
inner, thin, non-stainable and highly refractive portion, from which the 
first layer is often slightly separated. This second layer is closely applied 
to the underlying cells, which doubtless secrete it. Looked at from the 
surface (Fig. 10, æ.) the deeply stainable layer is seen to be broken into 
irregular polygonal pieces ranging from 2 a to 17 p in diameter and sepa- 
rated from one another by spaces ranging from 0 to 6 p. 

The mesoderm forms a loose epithelium, whose average width is less 
than that of the ectoderm (Fig. 3, ms’drm.). As a whole, moreover, it 
stains less deeply. In a portion of the gemmiparous zone, which lies 
about 180° from the budding region, the mesoderm has become so delicate 
a layer, if it exists at all, as not to be easily distinguishable. In the vicin- 
ity of the bud its cells have irregular outlines and extend out into the 
cœlom as though possessed of the power of amoeboid movement. The 
nuclei are spherical or ovoid, smaller than those of the ectoderm, and on 
the whole have smaller nucleoli. The cell body is highly vacuolated.. 
The vacuoles are not large and clear in outline, but whole regions of the 
cell body seem to be reduced to a non-stainable condition, and in some of 
these regions a fine network may still be observed. 

The proximal zone (Fig. 14, y to 8) is distinguished, soon after the 
first rudiment of the bud appears, by the diminished thickness of the 
ectoderm. The cells have become transformed from a columnar to a 
pavement epithelium. The nuclei are smaller, the nucleoli less prom- 
inent, and the cell body stains much less deeply. The cuticula is of 
two kinds, as before, but with this difference: the deeply staining outer 
part is less conspicuous, and the pieces are smaller and more widely sep- 
arated. Looked at from the surface, we find an appearance like Figure 
10, c., in which the dark bodies represent the deeply staining cuticula. 
These pieces are much smaller than those of the gemmiparous zone, 
ranging from 0.6 a to 9.5 a in diameter, and separated from each other 
by spaces ranging from 0 to 13 p. 


3. ORIGIN or THE POLYPIDE IN THE TERMINAL Bun. 


Observation having shown that budding in Paludicella follows definite 
laws, we ought to be able to discover the place and time at which buds 


8 BULLETIN OF THE 
will arise ; and it is necessary to do this ın order to study the origin of 
the gemmiparous cells, and the changes which they undergo preparatory 
to an actual involution. 

The study ot tips of branches shows that the necks of the polypides of 
any branch all lie in one plane, and that this plane also includes the 
youngest poly pide ; also that the youngest polypides always arise distad of 
the next older. Knowing these facts, our observations may be confined to 
a short line running from the neck of the youngest apparent buds to the 
tips of the branches studied. The time at which to search for incipient 
buds and the place in the line where they will be found is illustrated by 
Figure 7 (Plate I.). The youngest developed bud is one the axes of whose 
tentacles are approximately parallel to the axis of the branch, and whose 
brain cavity, gn., is not yet constricted off from that of the oesophagus. 
The place of origin is near the tip, immediately beyond the point at which 
the ectoderm changes rapidly from a columnar to a pavement epithelium. 

Figure 3 is from a section across the branch in the region of an incip- 
ient bud. I have already described the conditions of the cells of this 
region. Those near ex. are larger than the surrounding ones, and show 
signs of cell division both in the ectoderm and mesoderm. In both cases 
shown in the figure, the direction of division is such as will tend to 
increase the superficial area of the layer in which it occurs. The ecto- 
derm seems to be the most important layer of the two in the process of 
invagination which is about to take place. I think one is led to this 
conclusion if one considers a folding of an epithelium to be due to an 
increase ir, the area of the epithelium within a certain circumference 
without a correspondingly great increase in the circumference itself. 
Such a conception implies, first of all, mutual pressure of the cells of 
the invaginating epithelium. The cells of the mesodermal layer do not 
seem to be under mutual pressure; in some cases they are barely in 
contact. The cells of the ectoderm are evidently closely applied, and 
probably, therefore, under mutual pressure. 

The one case of cell division which is occurring in the ectoderm is at 
the inner end of the cell. In fact, the centre of the nuclear plate is much 
nearer the deep end than are the centres of the adjacent nuclei. The 
effect of this division is to increase the area on the inner surface of the 
ectoderm more than that on the outer, as appears from a study of the 
sections shown in Figures 4 and 5. In Figure 4 certain cells lie already 
below the niveau of the surrounding ones, very much as though they had 
moved downward on account of this being the direction of least resist- 
ance. A later stage of this process is shown in Figure 5. Here the 


MUSEUM OF COMPARATIVE ZOÖLOGY. 9 


nuclei are already arranged in a deep saucer-shaped layer. The transi- 
tion to the U-shaped arrangement of Figure 37 (Plate IV.), in which 
the invagination of the inner layer of the bud is completed, is not a diffi- 
cult one to understand. It is to. be observed, however, that the folding 
is of such character that it can hardly be termed a typical invagination. 
Comparing Figures 4, 5, and 37, it appears rather to be of a type some- 
what intermediate between typical invagination and typical ingression. 
The cavity of the bud first arises through a rearrangement and reshap- 
ing of the cells of the inner layer of the bud. At this stage the nuclei 
of the invaginated region stain very deeply, and have large nucleoli. 
Figure 21 (Plate III.) shows the condition of the bud at this stage as 
seen in longitudinal section. The proliferation which gave rise to the 
rudiment of the bud is shown, by a comparison of Figures 37 and 21, 
not to have been confined to one point, but to have occurred along a line, 
so that the resulting bud is boat-shaped, and not cup-shaped. The whole 
mass is therefore bilaterally symmetrical. Even at this early stage one 
can distinguish a difference in the form of the bud at the anal and oral 
ends. At the oral end (Or.) the bud passes more abruptly into the body 
wall than at the anal end. Later, this feature becomes more marked. 
This is an indication of a fact for which later stages will bring better 
evidence; that the formation of the bud proceeds from the oral towards 
the anal end, and that the increased length of the bud that one finds in 
the stage represented by Figure 22 is due to growth at the anal end. 


4. ORIGIN AND DEVELOPMENT or THE LATERAL BRANCHES. 


The first lateral branch appears as a prominent protrusion of the lat- 
eral walls of an individual of the primary branch when the ganglion of 
that individual has already nearly closed, and when the bud of the next 
younger individual has attained a stage somewhat later than that shown 
in Figure 37. The zone in which the lateral buds arise lies about mid- 
way between the neck of the median polypide and the tips of its tenta- 
cles at this stage. The place of appearance in this zone is approximately 
90° to the right or left of the neck of the polypide of the median indi- 
vidual, In one case measured, however, that shown in Figure 20 (Plate 
II.), the centres of the two lateral buds seemed to be unequally distant 
from the neck of the polypide, and each over 90° from it (approximately 
100° and 110° respectively. (Compare page 3.) 

A cross section of the branch through the region in which the lateral 
bud is arising shows that the condition of the body wall at the bud is 
quite different from that of the rest of its extent, Figure 19 represents 


10 BULLETIN OF THE 


a longitudinal section of a portion of the body wall passing through the 
non-budding region. The wall seems to consist of one layer only of cells, 
and a fine, non-stainable cuticula. This layer of cells is the ectoderm, 
for it can be traced directly into the outer layer of the tip. The meso- 
dermal layer is not represented in the region from which the figure 
was drawn, but I believe it is not entirely absent from this part of the 
individual, for occasionally extremely flattened cells, spindle-shaped in 
section, may be seen lying inside of the ectodermal layer, quite sharply 
marked off from it by a distinct line, Further evidence of the existence 
of two layers is found in the fact that one occasionally sees in the flat- 
tened body wall two nuclei lying together, one nearer the cœlom than 
the other. The cells of the ectoderm are seen to be very much flattened 
(average 2.5 p), and their nuclei are widely separated (35 p). The nuclei 
are oval, and rather smaller than those near the tip. They possess ¢ 
single, rather large nucleolus, which does not stain intensely. The cell 
protoplasm stains very little. The cuticula is about 0.5 p thick. 

If we study the body wall in the budding region, when the latter is 
first indicated on the surface by a marked protrusion of the outline of 
the zooecium (Plate II. Fig. 15), we find that this protrusion is due to 
an elongation of cells. There are about twenty-two cells in this section, 
which are more or less thickened, Since the section figured passes 
through the centre of the circular thickening, and is about one sixteenth 
the diameter of the circle in thickness, it follows that there are over 250 
cells of the ectoderm which have already at this stage become somewhat 
enlarged previous to evagination. The highest of these cells are the 
central ones, of which the largest is 22 » high. The largest nuclei are 
4 p by 6.3 m, which approximates the size of those in the gemmiparous 
region (page 6). They are placed nearer the coclomic epithelium than 
the exterior, are nearly spherical, and each possesses one large nucleolus 
and a quite apparent network with deeply stainable nodal points. The 
cell body is stained as a whole rather deeply by Ehrlich’s haematoxylin, 
but particularly around the nuclei. The outer parts of the central cells, 
however, are stained very little, and the deep ends of some of the late ral 
cuboid cells not at all. The network of plasma contains only fine 
granules, and these seem to lie in rows parallel to the long axis of the 
cell. The structure of the outer-layer cells, at a somewhat earlier stage, 
is shown in Figure 18, under a higher magnification, The network is 
very apparent in these large spherical nuclei, and the plasma of the cell 
is geen to contain coarse granules, which lie near the nuclei and stain 


deeply. 


MUSEUM OF COMPARATIVE ZOÖLOGY. IE 


While the cuticula of Figure 18 is seen to be that of the normal body 
wall in this region, that shown in Figures 15 and 16 appears under the 
microscope after staining in hematoxylin to be of two distinct kinds : 
(1) that outside of the central region, which is highly refractive and not 
at all stained; and (2) that which lies immediately over the central elon- 
gated cells of the bud, which is also highly refractive but stains deeply. 
In fact, the central cuticula resembles in every way that already described 
for the tip of the branch, and shown in Plate I. Fig. 6. Moreover, it 
has other points of resemblance to the latter. It does not stain at all 
in alum cochineal; the outer boundary of the branch is often uneven 
at this place (Fig. 16) ; and particles of dirt are often found adhering to 
it, while the rest of the cuticula is comparatively free. The difference in 
staining properties of the central and lateral cuticulas indicates that the 
former undergoes with age a change in its chemical properties ; the irreg- 
ular outer boundary and the adhesion of dirt particles seem to indicate 
that the newly formed cuticula is viscid. The mesoderm of the stage ol 
Figure 15 consists of a single loose layer of subspherical cells of the two 
kinds already noticed, reticulated and non-reticulated. The series of 
Figures 18, 15, and 16 shows the behavior of columnar cells in the forma- 
tion of a typical outfolding as distinguished from the slipping in of cells 
to form the polypide (Figs. 3, 4, and 5). 

In stages later than that of Figure 16, the tip of the branch becomes 
further removed from the body wall of the median branch. The cells at 
the tip always retain their elongated columnar condition. A polypide 
is soon formed on the upper part of the body wall immediately behind 
the tip, exactly as in the case of the median branch. A septum is 
very early formed, cutting off the lateral from the median individual, 
and the lateral secondary branch becomes the median primary one of 
new individuals (Plate VI. Fig. 58). 

We have already traced out the origin of the polypide of the median 
branch from the mass of cuboidal cells near the tip; it remains to de- 
termine whether the cells which give rise to the lateral branch can be 
traced directly back to the cuboidal cells of the tip, or whether they have 
arisen from the flattened epithelium of the general body wall and sec- 
ondarily acquired their plump “ embryonic” character. 

Figure 18 ( Plate IT.), to the cellular conditions of which I have already 
referred, shows an early stage of the lateral branch, and Figure 20, gm. bss 
shows on a smaller scale the different cellular conditions in the body wall 
in the region of two lateral buds which are yet far from showing ex- 


ternal signs of evagination. The cells are cuboid and much higher than 
5 5 g 


12 BULLETIN OF THE 


those of the adjacent body wall. Have they been so ever since they were 
derived from the tip, or have they secondarily become so ? I believe 
that these cells have never been flattened pavement epithelial cells, for 
the following reason. All ectodermal cells of the body wall near the tip 
are cuboidal; if these cells have only secondarily acquired this form, 
they must have passed through a stage in which they were flattened epi- 
thelium. Now, if these cells could be distinguished by greater thickness 
from the cells of the surrounding body wall, at a time at which the lat- 
ter cells had only just begun to emerge from the cuboidal condition to 
become differentiated into the pavement epithelium of the body wall, it 
would follow that, even though they had secondarily increased in size as 
a result of an impulse preparatory to evagination, and even though they 
would have been at a stage only a very little earlier indistinguishable 
from the other cells of the body wall, yet they would never have passed 
in this case through a flattened condition, because at a stage only a very 
little earlier the whole body wall was composed of cuboid cells. 

The conditions which I have set as the criterion of our problem are 
fairly realized in Figure 17, which represents a portion of the body wall 
of a median branch which extends from the gemmiparous region above to 
the thickened body wall of the nascent lateral bud below (gm. L). It will 
be seen, by a comparison of the body wall of this region with that shown 
in Figure 19, which is taken from the same individual farther from the 
tip, that even the most differentiated part of the body wall of Figure 17 
is in a relatively indifferent condition as compared with the pavement 
epithelium of the ectoderm of Figure 19, in which the mesoderm, indeed, 
has become so thin and insignificant as scarcely to be visible. We may, 
therefore, maintain that the ectodermic cells of the body wall have only 
just begun to lose their cuboidal condition to become pavement epi- 
thelium, and therefore conclude, in accordance with the argument just 
presented, that the cells of the lateral bud (gm. L.) have never passed 
through a stage in which they were flattened epithelium. It is evi- 
dent, also, that the Anlage of the second lateral bud is also derived 
from near the tip, because, as in Figure 20, we find two lateral regions of 


cuboidal cells. 


5. Deverorment or THE Bopy Watt. 


It is, of course, almost impossible to gain direct evidence upon the 
’ ’ 
place of origin and method of development of the body wall, and one is 
therefore forced to the collection and weighing of circumstantial evi- 
oD © 
dence. Braem (90, pp. 127, 128, 131) believes that the body wall (the 
2 ’ d 


MUSEUM OF COMPARATIVE ZOOLOGY. 13 


cystid, in Nitsche’s sense) has a double origin in Paludicella: “ Ein 
Theil des Cystids zwar vor dem Polypid, ein anderer aber erst später 
angelegt und zwar aus der polypoiden Knospe selbst entwickelt wird.” 
The part developed from the bud of the polypide is the elliptical region 
of the body wall, whose main axis lies in the sagittal plane and which 
has the neck of the polypide at the distal focus and the attached ends 
of the retractor muscles and the parietovaginal (or pyramidal) muscles 
lying in the proximal circumference, — the greatest part of the ellipse 
thus lying oral of the atrial opening. The evidence for this conclusion 
Braem finds in the following facts, which my own observations confirm : 
“The great retractor first appears in the angle between the oral part of 
the polypide bud and the cystid wall [cf. Figs. 23, 24, el. mu. ret.]; then 
its cells gradually become elongated, and as its point of origin retreats 
farther and farther from the polypide, it finally appears as a bundle 
which joins a point lying between the mouth of the polypide [neck of 
the polypide] and the inferior septum with the pharynx, and, as I be- 
lieve, also with the cardial part of the stomach” (p. 125). Compare 
the muscles at the left end of Plate I. Fig. 8. Further on he says: 
“Die Parietovaginalmuskeln [pyramidal muscles] erscheinen an der 
Knospe zuerst in Form zweier seitlichen Leisten, in welchen die einzel- 
nen jugendlichen Fasern senkrecht zur Längsaxe verlaufen. Indem sich 
alsdann lateral von der Knospe das Cystid durch Neubildungen erweitert, 
werden die Fasern verlängert und die beiden Bündel treten in Flügelform 
deutlicher zur Reebten und Linken der Mündung hervor. Ihr Ursprung 
an der Cystidwand rückt nun von der Mündung immer weiter ab und 
gelangt schliesslich auf die gegenüberliegende Seite, wo er anal und late- 
ral seinen definitiven Platz findet.” Compare Plate I. Figs. 7-9, mu. pyr., 
and Plate VI. Fig. 63, mu. pyr.. From these observations Braem (90, 
pp. 127, 128) concludes: “ So scheint es sicher, dass auch hier ein grosser 
Theil des definitiven Cystids, das ja zum anderen Theil schon vor der 
polypiden Knospenanlage entwickelt war, aus dem Material dieser letz- 
teren hervorgeht. Das folgt namentlich aus der Art und Weise, wie 
sich die Muskeln bilden. .. . Auch hier würde, wie bei den Phylac- 
toleemen, oral vor der Knospe nach dem Retractor hin, ein grösseres 
Gebiet der Leibeswand der Knospenanlage entstammen, als seitwärts 
und hinten.” 

An analysis of the facts has led me to conclusions differing somewhat 
from those of Braem ; namely, that all or nearly all of the cells of the body 
wall (eystid) are derived from the tip of the branch or from the immedi- 
ate descendants of cells so derived. The number of cells contributed to 


14 BULLETIN OF THE 


the formation of the body wall by the neck of the polypide is much 
smaller than Braem has suggested, and probably insignificant in amount. 
The retreat of the points of origin of the retractor and parietovaginal (py- 
ramidal) muscles may be in part accounted for by the normal growth in 
area of the body wall, and in part by the actual movement of the point 
of origin with reference to the cells of the body wall. These conclusions 
rest upon the following circumstantial evidence, 

Owing to the small number of cells in the body wall at the tip, and 
the comparatively slow growth of the cystid, karyokinetic figures are 
much less frequent than in the polypide. Quite a long search has there- 
fore not afforded cases enough to enable me to draw any perfectly satis- 
factory conclusions as to just where, and where only, growth was taking 
place. I have, however, seen nuclear division occurring in the elongated 
cells of the extreme tip, rather more abundantly in the cuboidal cells 
between the extreme tip and the gemmiparous zone, and most abun- 
dantly in the gemmiparous zone, but here evidently having to do with 
the origin of the polypide, muscle cells, etc. Proximal to the gemmipa- 
rous zone, I have noted few cases of nuclear division excepting about 
the neck of the polypide. It seems probable that the cells of the tip 
of the branch are not to be regarded as forming a differentiated organ 
whose elements rarely divide, but as quite capable of adding new cells to 
the body wall. On the other hand, there is by no means a Scheitel in the 
botanical sense, but the cells added to the body wall continue for a time 
to divide vigorously, and finally give rise to the polypide, to the Anlage 
of the lateral branches, and to the body wall. The cells belonging to 
the proper cystid then cease to divide rapidly. 

I have already shown how the cells of the tip secrete a cuticula, which 
becomes gradually replaced by a second cuticula secreted beneath it as 
the body wall attains its adult dimensions. It appears as though the 
first cuticula were secreted by the cells of the tip only. This being so, 
since the area ‘of the body wall increases, this first cuticula must either 
stretch to cover the enlarged area, or else it will fail to cover it and 
appear as isolated patches upon the body wall, and these isolated patches 
will become more and more widely separated as the area of the body 
wall increases. This latter condition seems to be the one realized in this 
case. The presence of the old cuticula is easy to demonstrate, since it 
stains deeply in hematoxylin ; and it may be easily distinguished from 
that formeil later, for with the same reagent this stains not at all. Figures 
6, 11, 12, and 13 show different appearances of the cuticula at different 


parts of the body wall. At the extreme tip (Fig. 6) there is a continu- 


MUSEUM OF COMPARATIVE ZOÖLOGY. 15 


ous deeply stained band of cuticula, In Figure 11 it no longer appears 
quite homogeneous, but is darker at some places than at others. The 
ectoderm is here composed of cuboidal cells. At a later stage of devel- 
opment the ectodermal cells have become very much flattened. A. thin, 
unstainable, more deeply lying cuticula has already begun to form, and 
the outer deeply stainable cuticula is seen to be broken up into bits. 
Figure 13 is from the adult body wall. The ectoderm is flattened. The 
inner cuticula has attained a great thickness, and the outer cuticula is 
represented by only a few deeply staining patches. One attains a simi- 
lar result by studying the surface of a stained individual. Figure 10 
shows the condition of the outer cuticula at intervals along the same 
branch from the gemmiparous region a to a nearly adult region, d. The 
bits of cuticula become more and more widely separated and smaller, as 
I have already described in detail on page 7. Here, then, we have not 
merely an interesting case of replacement of one cuticula by another to 
meet the needs of the enlarged body wall by a method which has no par- 
allel, so far as I know, in any other group of animals, but for the specific 
purposes of our problem a criterion of growth of the body wall quite as 
satisfactory as karyokinesis, and much easier of application. 

Let us apply this criterion in our attempt to answer the question, Is 
that portion of the body wall lying between the neck of the polypide 
and the points of origin of the pyramidal muscles (Plate VI. Fig. 63, 
b-a, b-c) derived wholly from the neck, or is it merely the result of 
interstitial growth of that part of the original cystid which was pre- 
formed in the neck region? If the first condition is true, we should 
expect to find no indications of the outer cuticula secreted by the tip of 
the branch; if the second, we should expect to find the outer cuticula 
broken into bits, and underlaid by the inner lately formed cuticula. 
Figure 63 shows clearly the deeply stained outer cuticula here sep- 
arated into bits, and, to my mind, thereby proves that this part of the 
eystid has had an origin similar to that, of the rest of the body wall. 
Moreover, a comparison of the portion of the section figured with the 
remainder (and this comparison has been made on many sections from 
several individuals) shows that the parts of the cuticula about the neck 
are indeed rather smaller and farther removed from each other than at 
the opposite side of the branch; but the difference in this respect is not 
very marked, and may well only signify that there is a more rapid 
growth of the body wall in the vicinity of the neck of the polypide than 
at the opposite side. 

But how then do the points of origin of the pyramidal muscles come 


16 BULLETIN OF THE 


gradually to move away from the neck of the polypide at which they 
arose, in order finally to lie so that the muscle fibres are nearly parallel ? 
If the points of origin remain fixed with reference to the surrounding 
cells, they can hardly come to lie absolutely closer together, but only 
relatively so by growth of the body wall between these points and the 
neck. If, however, we find that in older individuals the points of origin 
are not only relatively but absolutely closer together, we are driven to 
the conclusion that these points move relatively to the surrounding cells. 
To decide whether the points of origin come to lie closer together behind 
the neck absolutely or only relatively, I measured cross sections of four 
individuals through the region of the neck in which the muscle fibres 
showed evident differences of length, and therefore of age. I may 
preface a table of these measurements with the statement that the mus- 
cles first appear plainly differentiated at a stage when the polypide is 
well formed (Fig. 7, mu. pyr.), and that the growth of the body wall in 
circumference is not very considerable after this time. The numbers 
indicate measurements in micra : — 


Distance on periphery between origins of mus- N°1- No.2 No.8. No.4, 


ClasyAtial Aldo urn «25 cea iia th. s ELOU 154 187 260 
Distance on periphery between origins of mus- 

cles, abatrlal Bildes sei) ay ey ie), oe OOF 286 264 220 

Total length of periphery . . . .. . 447 440 451 480 


The distance on the “atrial side ” signifies the distance measured over 
a, b, c, Figure 63 (Plate VI.). The length of the remainder of the sec- 
tion is the distance on the “ abatrial side.” 

From these measurements it appears that the “ origins” ofsthe py- 
ramidal muscles approach each other absolutely, —a condition which 
Braem’s hypothesis cannot explain, and which can be reasonably inter- 
preted, it seems to me, only by assuming, however unique and difficult 
of conception such a condition may be, that the points of origin move 
relatively to the surrounding cells of the body wall. (Compare also 
the movement of parietal muscles referred to on page 29.1) 

It is not necessary to assume that the increase in extent of the body 
wall after the polypide is first formed is due to the addition of cells from 


1 Professor Mark has called my attention to a discussion of the movement of 
the fixation-point of a muscle in Mollusks by Tullberg (’82, pp. 26, 27, 44). This 
author says that he has undertaken no special investigation of the method of 
migration, but concludes that this motion must result from the absorption of the 
inner muscle fibres as new ones are formed on the outside, I do not find any 
evidence of such a process in Paludicella. 


MUSEUM OF COMPARATIVE ZOÖLOGY. I7 


the neck. The change in form of the ectodermic cells from a columnar 
to a pavement epithelium must alone cause a great increase in the ex- 
tent of that layer. Some measurements that I have made seem to me 
to prove that the area of the body wall does increase greatly, even out- 
side the region whose growth Braem attributed to the addition of cells 
from the neck of the polypide. Thus, in one case, the distance from the 
distal end of the polypide bud, which becomes the neck of the adult, 
to the point of origin of the young retractor muscle was 0.17 mm.; from 
the same point to the septum separating the young individual from the 
next older was 0.27 mm. In the next older individual, from the neck to 
the origin of the retractor muscle was 0.72 mm. ; from the neck to the 
septum was 2.0 mm. Thus assuming that the older individual passed 
through a stage exactly equivalent to that in which we find the younger, 
the distance from the neck to the origin of the retractors has increased 
0.55 mm., and from the origin of the retractors to the septum 1.18 mm, 
The first distance is that in which Braem has assumed the body wall to 
grow by additions from the neck of the polypide, and this assumption was 
apparently made to account for the increase in extent of this region ; but 
the area between ,the origin of the retractors and the septum, which is 
outside the region to which additions such as Braem contemplates could 
have been made, has grown in this case very considerably more in extent. 
This case is not a typical one, however, for we rarely find the distance 
from the origin of the retractor to the septum to be so great. In gen- 
eral, from observation of a number of cases, I should say that in the 
adult the distance between the neck of the polypide and the origin of the 
retractors, is to the distance between the latter and the septum about as 
5:4, and that therefore the growth of the first region is slightly greater 
than that of the second, From the fact, however, that the cells around 
the neck of the polypide for a long time retain a somewhat embryonic 
character, and may quite frequently be seen in division, this was to have 
been expected. The conclusion which I draw from this last series of 
conditions is, then, that it is unnecessary to suppose the addition of cells 
from the neck of the polypide to account for the fact that the origin of 
the retractors is carried backward from the polypide. Normal growth of 
the body wall, such as occurs elsewhere, is quite sufficient to account 


for it. 
To recapitulate. That portion of the cystid lying in the vicinity of 


the neck can hardly be derived from the neck alone, for the cells still 

show adhering to them the cuticula which they derived from the tip of 

the branch. It is not necessary, in order to account for the movement 
‘ 


VOL, XXII. — NO. 1. 2 


18 BULLETIN OF THE 


of the origins of the muscles away from the neck, to suppose that the 
eircumcervical region is derived in that way; for (1) the origins of the 
pyramidal muscles actively migrate away from the neck to a certain 
extent, and (2) the normal growth of the body wall is sufficient to ac- 
count for the carrying backward of the origin of the retractors. 

From the facts already gained it seems clear that the ectocyst (cuticula) 
is first formed at the tip, and then, to meet the wants of the growing 
colony, this is replaced later by a cuticula of different chemical compo- 
sition, which becomes thicker as the body wall grows older. At a late 
stage we find a separation of the thick cuticula itself into two layers, of 
which the outer one is much the more highly refractive. (Plate II. 
Fig. 13; Plate III. Figs. 26, 29.) 


6. DEVELOPMENT or THE POLYPIDE 


We have already (pages 8, 9, Figs. 5, 14, 37) seen how the foundations 
of the polypide are laid by the ingression of cells of the outer layer of 
the body wall pushing before them the mesoderm, and how, finally, those 
cells arrange themselves in a boat-shaped mass to form the inner layer 
of the bud (Plate III. Fig. 21), which possesses no actual cavity, and 
is constantly separated from the external world by the ectoderm which 
remains behind to form the neck of the polypide. Even when a cavity 
is formed later, it does not communicate with the exterior until the 
permanent atrial opening has arisen. The earliest differentiation in the 
bud is, as mentioned by Allmann (’56, p. 36), the formation of a cavity 
which is to become that of the atrium. This cavity is first formed 
at an early stage as an extremely slight fissure in the midst of the inner 
layer. Figure 22 shows a longitudinal section of this stage. Cell di- 
vision is taking place throughout the whole mass, but especially at the 
neck of the polypide, cev. pyd. The position of the cavity is represented 
by the central non-nucleated space, and this gives rise, as the later his- 
tory of development shows, to the atrium and the pharynx. 

Figure 23 represents a stage which is doubtless of short duration, for 
I have found it only twice. The bud is much more developed at the 


1 Such a two-layered condition of the cuticula was long ago described by Reichert 
(70, pp. 265, 266) for Zoöbotryon. He distinguished “ eine äussere, festere, stärker 
lichtbrechende und sprödere Schicht und die innere weichere.” Realizing that the 
“ectocyst” or cuticula undergoes many changes in form, — formation of lateral 
buds, of septe or communication plates, and increase in size of the stolon, — he 
suggested, without having observed the process, that probably during these 
changes the more rigid outer layer disappeared and was replaced by the inner 
softer one. 


MUSEUM OF COMPARATIVE ZOÖLOGY. 19 


anal end than in the last stage, and there is a second cavity below the 
atrium, from which it is separated by a line of nuclei. This is plainly 
an early stage in the formation of the alimentary tract, which thus first 
appears at the anal side of the bud, as in Phylactolemata, and is progres- 
sively formed towards the oral end. An appearance similar to the one 
figured would be given by a slightly oblique section of a later stage; but 
this section is strictly sagittal, and no trace of the lumen appears in adja- 
cent sections, I have found a similar condition in a series of longitu- 
dinal sections at right angles to the sagittal plane of the bud (Plate IV. 
Figs. 39 and 40). Figure 39 shows that the atrio-pharyngeal cavity is 
first developed at the anal end, and in Figure 40, which is three sections 
(about 15 u) below Figure 39, the anal end only of the alimentary tract 
is formed. It is worthy of notice that the cells of the mesodermic layer 
of the bud are often greatly vacuolated at this stage, as in Figures 39 and 
40, vac. Braem (90, p. 126) says of this stage: “ Die der Resorption 
dienenden Darmabschnitte, Magen und Enddarm, werden gemeinsam an- 
gelegt, indem auf jeder Seite der Knospe eine Längsfalte die Wandungen 
nach innen und gegen einander zu einbiegt, worauf die benachbarten 
Theile des inneren Blattes verschmelzen und so durch eine Art Abschnür- 
ung das primäre Knospenlumen in den vorderen Atrialraum und die 
hintere Darmhöhle getrennt wird.” While I thoroughly agree with this 
statement, the additional fact of the formation of the tract progressing 
from the anal towards the oral end is interesting, in that it shows that the 
process of formation of the organ in Paludicella is fundamentally similar 
to, although differing slightly in detail from, that of the Phylactolamata. 
Figure 24 shows in sagittal section a still later stage in the development 
of the alimentary tract. A cross section of this stage is seen in Figure 
30 (Plate IV.), in which the separation of atrial and gastric cavities is 
demonstrated. The inner layer of the bud is here seen to be separated 
from the ectoderm by a distinct line, and, to a certain extent, even by 
the mesoderm. The distal (oral) part of the cavity of the alimentary 
tract next becomes considerably enlarged to form the stomach (Fig. 25). 
The outer layer of the bud, ms’drm., penetrates between the stomach 
and the atrium, and a depression is formed at the bottom of the atrial 
chamber which will give rise to the @sophagus, Even at this stage the 
@sophagus is not in communication with the stomach, but their cavities 
are separated by two layers of cells of the inner layer of the bud. These 
two layers become those of the cardiac valve (Plate IV. Fig. 36, vlw. cr.). 
By a further comparison of Figures 25 and 36 it will be noticed that, 
whereas in the earlier stage, as in Endoprocta, there is no cocum to the 


20 BULLETIN OF THE 


alimentary tract, in the later stage the cocum has already begun to form, 
as in Phylactolemata, by an outpocketing of nearly the whole of the 
lower wall of the stomach. (Compare also Plate I. Figs. 7, 8, and 9.) 

Very soon after the establishment of the alimentary tract, and between 
the stages shown in Figures 24 and 25 in sagittal section, there begin 
to appear organs which have a very considerable phylogenetic signifi- 
cance; namely, the lophophoric ridges, ring canal, and tentacles. 

The lophophorie ridge is a fold which surrounds the mouth, and from 
which at intervals tentacles arise. The ridge, however, arises before the 
tentacles. The general position of the ridge, as well as its method of ori- 
gin, may be learned from an inspection of a series of sections of the age 
of those shown in Figures 31-34. In a section lying near the oral 
end of the bud (Fig. 33), one finds two spaces, — a lower, which is that 
of the stomach, and an upper, the oosophagus and atrium. This upper 
space is broader above than below, and the cell layer which lines it is 
thick below, but above, or nearer to the body wall of the budding 
individual, it is thinner. The transition from one condition to the 
other is quite abrupt, and is marked by a salient curve (loph.). Ina 
section near the anal end of the bud (Fig. 31), it will be seen that here 
too the inner layer is thick below and thin above. The characters men- 
tioned are still more strikingly shown in the median section, Figure 32. 
That the differences in thickness of different parts of the inner layer are 
recently acquired modifications of an earlier simpler condition is indi- 
cated by comparing Figure 32 with Figure 30, which is from a younger 
bud. The series of points (loph.) of transition from thick to thin epi- 
thelium forms on the reconstructed polypide a curved line, convex above, 
This line is the ridge of the young lophophore (compare Fig. 25, loph.). 
I have said that the lophophoric ridge arises before the tentacles, The 
evidence for this assertion is found in a series like that referred to above, 
where, although the ridge exists along the entire side of the atrium, 
one finds nascent tentacles in the middle region only (Figure 32, left 
hand). 

As Figure 25, of a later stage than Figures 31-33, shows, the lopho- 
phore curves downwards rapidly at the anal end, so that it here lies at 
right angles to the axis of the rectum, but does not extend at all beyond 
the anus. Orally, there is in the median plane only the slightest trace 
of the lophophoric ridge. By the formation of this ridge in the wall on 
each side of the atrial chamber, the original atrio-pharyngeal cavity has 
become separated into two regions. The space lying within or below the 
ridge forms the pharynx and the intertentacular space ; that lying with- 


MUSEUM OF COMPARATIVE ZOOLOGY. 21 


out and above, the atrium of the adult. (Plate III. Fig. 25; Plate IV. 
Fig. 32, atr.) Since the lophophore curves rapidly downward to the 
anus and does not extend behind it, the act of cutting off the lower part 
of the atrio-pharyngeal cavity from the upper (atrium proper) does not 
continue behind the anus, which therefore opens directly into a part of 
the atrium. This part has the form of a compressed funnel, and is 
bounded behind and laterally by the kamptoderm, and orally by the 
hinder ends of the lophophoric ridges, and also, since the latter do not 
meet in the median plane, by the pharyngeal cavity. Thus it has come 
about that the anus, which at first opened into the common atrio-pha- 
ryngeal cavity of the bud, has now, in the separation of the two regions, 
come to lie near their point of division posteriorly, but to open distinctly 
into the atrial cavity. The more pronounced separation of the part of 
the atrial cavity into which the anus directly opens from the remainder 
of the atrium takes place much later, and will be described further on. 
In Figure 33, the ring canal (can. cre.) is seen to be already formed. At 
this stage it is found on one side only, the left, if one looks at the poly- 
pide from the tip of the branch. It occurs in only four sections (each 5 a 
thick), being found on the next section behind Figure 33, and on two sec- 
tions nearer the oral end. At its oral extremity, it terminates blindly 
as a thickening of the outer layer of the bud; at its anal end, one sees 
cells of the outer layer extending out partly over the canal, but failing 
to enclose it; in the next section the mesoderm is undisturbed, In sim- 
ilar sections of an older polypide (corresponding in age approximately to 
Plate IV. Fig. 35), the canal is found on both sides, and near to the oral 
end, but at about the middle of the series (cf. Fig. 35) it is found to 
open again into the body cavity. I therefore conclude that the ring 
canal makes its first appearance at the base of the lophophore in a 
region just oral of the middle of the polypide. Exactly how it arises, 
whethor by a growing together of the lips of a shallow furrow formed 
from the mesodermal layer, or by the formation of a pocket, which, elon- 
gating, penetrates between the inner and outer layers of the polypide at 
the base of the nascent lophophore, I have not been able to determine. 
Two facts induce me to believe that the later formation of the canal 
oralwards results from the penetration of a sac-like mass of mesodermal 
cells between the two layers of the polypide at the base of the nascent 
lophophore. One usually finds, (1) as in Figure 33, can. crc., a double 
mesodermal wall between the lumen of the canal and the cœlom, and 
one layer between the former and the inner layer of the bud; and 
(2) at the oral blind end of the ring canal a number of loose cells 


22 BULLETIN OF THE 
(occasionally dividing) representing the blind end of the pocket and 
lying between the inner and outer layers, both of which are intact. 

Braem (’90, p. 50) describes the formation of the ring canal in Phylac- 
tolemata as taking place in the manner just suggested for Paludicella. 
His studies were made, he says, preferably on statoblast animals. Nitsche 
(75, p. 358) concluded that in Phylactolemata the ring canal was first 
a furrow, whose lips fused, and my own study (’90, p. 129) has led me to 
the same conclusion. Since reading Braem’s account I have looked over 
some of my own sections of Cristatella again. Certainly the process is 
not so clear in the buds of the adult colony as in the statoblast embryo 
which Braem figures. Nevertheless the series of sections (90, Plate IV. 
Figs. 33-38) given as evidence of my statement still seem to me capable 
only of the conclusion I drew from them. Perhaps the processes may 
be different in detail in the two cases; certainly the two explanations 
are not fundamentally dissimilar. 

The ring canal being established in the oral part of the polypide, it 
grows forward, as I have said, and, secondarily, the canals of both 
sides meet in the median oral line and their lumina become confluent 
(Plate VI. Fig. 52, can. erc.). From what has already been said, it is 
clear that the lateral parts of the ring canal are not now continuous 
with each other behind. They become so only after the formation of 
the tentacles. 

The tentacles arise upon the lophophoric ridge at a stage a little later 
than that represented in Plate IV. Figure 32. At the stage represented 
by Figure 35, however, the tentacles have begun to form, as indicated by 
the fact that in the series from which this figure was taken the fold into 
the upper part of the atrium appears now deep, now shallow, according 
as the section passes through the length. of a young tentacle, or only 
through the lophophoric ridge between the tentacles. The position of 
the section (Fig. 35) is about the middle of the series, corresponding 
to Figure 32. 

By a comparison of Figure 35 with Figure 32 in respect to the tenta- 
cles, it will be apparent, first of all, that the lophophoric ridge itself has 
been heightened and that this heightening has been effected, not by a 
deepening of the fold existing in Figure 32, the lips of the fold remain- 
ing quiescent, but by a movement downwards of the outer lip (*) of the 
groove which is to form the ring canal. The movement is of course ac- 
companied by an increase in the length of the kamptoderm, kmp. drm. 
This growth of the lophophoric ridge naturally does not result in making 
the tentacles project farther above the ridge. Their elongation must 


MUSEUM OF COMPARATIVE ZOOLOGY. 23 


take place quite independently of the former’s. The lophophoric ridges 
have now become elongated folds lying upon the right and left of the 
polypide, which at this stage has a very compressed appearance (Plate IV. 
Fig. 41). The folds occupy the position of the ridges, and therefore do 
not lie throughout their whole extent in one plane, but oralwards are 
nearly parallel to the body wall (Plate III. Fig. 25), analwards trend 
nearly at right angles to it. It results from this fact, that one cannot 
see the anal tentacles when looking at the polypide from the side of 
the body wall to which it is attached. Figure 41 (Plate IV.) shows also 
that no tentacles have yet made their appearance at the oral ends of the 
two lophophoric ridges. The tentacles are here seen to be arising in 
two long rows, and so that those of one row are placed opposite the in- 
tertentacular spaces of the other. There are six tentacles in each row. 
The rows are not continuous with each other oralwards or analwards. 

The separation of the atrial and oral cavities, begun by the first 
formation of the lophophore, is, now that the tentacles have arisen, much 
more pronounced, Other changes now occur in this region, which pro- 
duce an extensive modification in the form of the polypide. 

One of the first of these changes is the close approximation and 
finally fusion of the anal extremities of the lophophoric ridges oralward 
of the anus. A stage in this is shown in Figures 43 and 44 (Plate V.), 
which are sections in the position of the lines 43, 44, of Figure 25 
(Plate II.), but through a slightly older polypide than that represented 
by Figure 25. The section shown in Figure 43 passes across the rec- 
tum, grazes the outer lip of the ring groove of the anal tentacles, and 
finally cuts, nearly longitudinally, one of the middle tentacles of the 
row. The two lophophores are not yet completely fused in front of the 
rectum. In Figure 44 (compare Plate III. Fig. 25, 44) this break in 
the continuity of the lophophore is more prominent. 

By the completion of the union of the lophophores in front of the 
anus, the rectum is quite cut off from communication with the inter- 
tacular space. It now opens only into the thin-walled, funnel-shaped 
depression of the atrial cavity. 

Pari passu with this operation the stomach and rectum are being 
more completely separated from the pharyngeal cavity by the penetra- 
tion of a double layer of mesoderm between these regions from each side, 
and a fusion of the corresponding layers of the two sides. Finally, the 

1 Compare Plate IX. Figure 77, which is a superficial view of the young lopho- 


phore from Flustrella, in which the process is similar to that in Paludicella, only 
the down curving of the anal tentacles occurs later than in the latter case. 


24 BULLETIN OF THE 


walls of the stomach and pharynx become separated from each other by 
a part of the coelomic cavity, as in Plate IV. Figure 36. This process 
of separation of the alimentary tract proceeds analwards, and finally the 
rectum is far removed from the cosophagus. 

The anus thus comes to lie farther outside of the anal tentacles. 
Finally, the ring canal, which is formed progressively farther and farther 
analwards, follows the fusion of the anal ends of the lophophores, and 
thus completes the canal behind the wsophagus. (Plate IV. Fig. 36 ; 
Plate VI. Fig. 53, can. ere.) 

The anal part of the ring canal is doubtless not merely a groove, 
but a tube; but the ring canal is not closed at this, and probably not 
at any stage throughout its entire extent, for in Plate VI. Figure 52, 
two sections below Figure 53, an opening is shown to exist on each 
side (at can. crc.), putting the cavities of the ring canals and the 
coelom into communication with each other. These openings lie at 
the sides of and slightly above the ganglion (yn., Fig. 52); a position 
exactly comparable with that of the openings in the ring canal of Phy- 
lactoleemata, which leads from the cœlom into the lophophoric arms on 
the one hand, and into the circumoral part of the ring canal on the 
other. 

3y a comparison of Figure 41 (Plate TV.) with the sections shown in 
Figures 60-62, it will be seen that the row of tentacles has undergone 
a change of form: from being laterally compressed, it has become cir- 
cular. This change of form has not resulted from an increase in the 
number of the tentacles, for at the stage of Figure 41 there are six ten- 
tacles on each side already formed (the sixth not visible), and there 
are in front of the mouth spaces already reserved for the two additional 
tentacles. There are also, probably, two nascent tentacles at the anus, 
although these are little developed, making a total of 16. In Figure 61 
there are only 15 tentacles; moreover, the actual diameter of the ten- 
tacular corona in the sagittal plane is less than at the earlier stage 
of Figure 41. This change of form is perfectly normal, all young 
polypides having tentacles arranged in two parallel rows, and adult 
polypides having a circular lophophore. 

These changes in the form of the tentacular corona are correlated 
with important changes in the direction of the axes of other organs. 
These changes may be understood by comparison of Figures 25 and 36, 
together with the assistance of Figures 7-9, all of which are oriented 
in the same manner. In Figure 25 the points fixed by the cardiac valve 
(vlu. er.) and anus (an.) lie in a line which is approximately parallel to 


MUSEUM OF COMPARATIVE ZOÖLOGY. 25 


the body wall. In Figure 36 the line passing through the same points 
makes an evident, but not very large, angle with the body wall. This 
line has undergone, then, a slight change of position only. The axir 
of the anal tentacles lies in both cases nearly parallel to the body wall, 
and so does the neural wall of the pharynx. The oral tentacles, on the 
contrary, whose axes in the earlier stage are directed perpendicularly 
to the body wall, lie in the later stage with their axes parallel to the 
wall; and the base of the lophophore, which in the earlier stage trended 
at its oral end parallel, at its anal perpendicular to the body wall, in 
the later lies throughout its whole extent in one plane perpendicular 
to the body wall. The axes of the oral tentacles have rotated through 
an angle of nearly 90° relatively to most of the other organs of the poly- 
pide. The canse of this rotation must evidently be sought in unequal 
growth in different parts of the polypide. A comparison of the length 
of the kamptoderm on the anal side in Figures 25 and 7 indicates 
that it has grown more in length than on the oral side. This excessive 
growth would tend to rotate the line vlv. er. — an. to a position perpen- 
dicular to the body wall. Since this rotation has not occurred to so 
great an extent ası,was to have been expected, we must look for a com- 
pensating growth on the oral side of the polypide, between wv. er. and 
the neck of the polypide, which shall be nearly equal to the excessive 
growth of the anal kamptoderm, and which must be outside of the oral 
kamptodem. These conditions of location are fulfilled only by the 
oral wall of the @sophagus, and it is by change of position and growth 
of this wall that the extension of the anal kamptoderm is nearly com- 
pensated for on the oral side of the polypide. By this growth in the 
wall of the csophagus the oral part of the ring canal has been brought 
to lie over the anal part, the sagittal diameter of the tentacular corona 
has been reduced, and the compressed lophophore has been transformed 
into a circular one. 

Concerning the number of tentacles, Dumortier et van Beneden (50, 
p. 46) observe that in the adult there are ordinarily 16, although 
individuals with 18 tentacles occur not infrequently, an observation 
which Kraepelin (’87, pp. 98, 99) confirms. In addition to these num- 
bers, I have found 15 and 17. The growth of the odd tentacle is quite 
interesting. The sections reproduced in Figures 60-62 (Plate VI.) will 
serve to illustrate a condition which I have quite frequently found in 
a polypide with 17 tentacles. In this particular series there are only 
15 tentacles. The successive sections abundantly demonstrate that 
the odd tentacle (*) is anal in position, and that it is younger than any 


26 BULLETIN OF THE 


of the others; thus in Figure 61 its tip is cut, in Figure 60 there are 
only 14 tentacles visible, and these are found in the two following sec- 
tions. Since there are six sections which pass through the tentacles, 
and the odd tentacle is found in only three of these, it follows that it 
is only about one half as long as the others. 

The nervous system arises, as in Phylactolemata, by a depression in 
the floor of the common atrio-pharyngeal cavity, in the region which later 
becomes the anal surface of the pharynx. As in Phylactolemata, we 
first see a shallow pit (Fig. 25, gn.). This appears to become deeper, 
sinking downward and somewhat toward the cardiac valve (Fig. 78, git). 
Finally it becomes constricted off from the wall of the esophagus, and 
then appears as a cellular mass closely attached to it and surrounded 
exteriorly only by mesoderm. (Plate I. Fig. 8; Plate VI. Figs. 52, 
53, gn.) Even before the closure of the ganglionic pocket is completed, 
the formation of the circumcesophageal nerve, first described by Krae- 
pelin (’87, pp. 62, 63) in the adult, begins. 

Figure 52 (Plate VI.) shows a transverse section of a young polypide 
in which the ganglion is solid, and not provided with a large cavity as 
in Phylactolemata. There is a small cavity in the upper part of the 
ganglion, and this is not yet wholly closed from the osophagus. The 
ganglion is continuous with a pair of hornlike processes (n.) which 
partly enclose the csophagus, and at a later stage do so wholly 
(Plate IV. Fig. 36, ”/.) The cells of these horns are found dividing in 
unusual abundance. The horns lie next to the digestive epithelium, 
and between it and the mesodermal lining of the ring canal. From the 
method of growth, and from the sharp line of separation between the 
tips of the horns and the surrounding tissue, there can be little doubt 
that the circumoosophageal nerve of Paludicella, like the lophophorie 
nerves of Phylactolemata, arises as an outgrowth of the brain. 

Serial sections show that the ganglion suddenly diminishes in size 
immediately below the point at which the cireumoral nerves arise, but 
one can trace a layer of cells continuous with the brain downwards for 
ten or fifteen micra farther, to near the cardiac valve. At this point 
one can still see nuclei of a third layer lying between the digestive epi- 
thelium of the valve and the mesoderm. It seems to me, therefore, 
that this may be regarded as a gastric nerve, which seems to originate 
by a single root and later to give rise to two nerves, one of which lies 
on either side of the cardiac valve. 


MUSEUM OF COMPARATIVE ZOOLOGY, 27 


7. ORIGIN or THE MUSOLES. 


a. Retractor. — After its first formation the bud becomes elongated 
in the direction of the axis of the branch. The derivation of this elon- 
gated stage from the much shorter earlier one might be effected in one 
of two ways: either, first, by the ingression of cells from the ectoderm 
at points successively more and more remote from the point of primary 
invagination, the additions to the length of the bud being made by a 
continuation backwards of that process by which the first foundations 
were laid; or secondly, by cell proliferation at the point of first invagi- 
nation pushing the oral end of the buds farther and farther from the 
neck of the polypide. 

I think there can be little doubt that the second is the method 
by which the bud becomes elongated; and for the following reasons, 
(1) The oral end of the bud, on the supposition of continued invagina- 
tion of the body wall, should become very gradually of less diameter, and 
transverse sections at that end should exhibit the ingression (potential 
invagination) of cells which were observed in the earliest stage; but 
as a matter of fact the oral end is abrupt (Plate III. Fig. 22, 23, Or.), 
and no stages of ingression are to be found there. (2) On the first 
assumption, the inner layer of the bud should be at all points in equally 
close relation to the ectoderm of the body wall; on the second, the 
inner layer should be closely connected with the ectoderm at the neck 
of the polypide (Plate III. Fig. 22, cev. pyd.), but elsewhere it should 
be sharply separated from it. As a matter of fact, a sharp line can 
be distinguished, in a sagittal section, separating the inner layer of the 
bud from the overlying ectoderm at all points except at the neck (anal 
part) of the polypide (Plate IIT. Figs. 22-25). Moreover, cross sections 
of the anal part of the bud show the inner layer passing directly into 
the ectoderm, and oralward the outer layer of the bud tends to pene- 
trate more and more between the ectoderm and the inner layer. 
Therefore I conclude that the inner layer of the bud is constantly 
augmented by cell proliferation in its mass, and especially at the neck 
of the polypide, and this explanation also accounts for the active cell 
proliferation observed at the neck in Plate III. Figure 22, cev. pyd. 

Since the polypide later becomes attached to the body wall by the 
comparatively narrow “neck” only (Figs. 7, 9scev. pyd.), a Separation of 
the oral part from the body wall has to take place. This process begins 
at the oral end. In its earliest stages it is indicated by the sharp sep- 
aration of the inner bud-layer from the overlying ectoderm, and the 


BULLETIN OF THE 


28 


partial penetration of the mesoderm on each side into the space between 
these two layers (Plate IV. Fig. 30, ms’drm.). At a later stage the 
mesoderm may be seen as a single cell layer lying between the ectoderm 
and the inner layer of the bud midway between the oral and anal ends 
(Plate IV. Fig. 32, ms’drm.), and as a double cell layer at the oral end of 
the bud (Fig. 34, ms’drm.). It is from these cells at the oral end of the 
bud that the retractor muscles are to arise (Plate III. Figs. 23-25, cl. 
mu. ret.). As the oral end of the kamptoderm and csophagus to which 
their inner ends are attached moves away from the ectoderm, and as the 
area of the latter itself increases, the two ends of the cells move farther 
and farther apart, and the young muscle cells become drawn out into 
spindle-shaped muscle fibres. (Plate III. Fig. 25, cl. mu. ret.; Plate IV. 
Fig. 36, mu. ret.) The retractor thus arises unpaired and remains so 
at its origin, but nearer its insertion in the ring canal and osophagus 
one can distinguish a division into right and left masses. The adult 
muscle fibres consist of two parts at least, the inner contractile portion 
and an outer less modified protoplasmic portion, which can be traced 
over the whole of the first part, but is most evident around the nucleus, 
where it has a granular appearance. 

b. Pyramidalis. — At about the stage of Figure 25 (Plate IIT.) one 
finds, on cross sections of the branch which pass through thé neck of the 
polypide, that the mesoderm of the body wall on each side of the neck 
is greatly thickened, and that its closely packed cells, which lie three 
or four deep, have become somewhat elongated. Cell division is quite 
common in the ectoderm of this region, and by it the area of the circum- 
cervical region is increased and the two ends of the muscle fibres are 
carried farther apart, one end remaining attached to the neck of the 
polypide and the other moving towards the abatrial surface. I have 
given reasons above (page 16) for believing that the abatrial ends of the 
muscles are not carried towards the abatrial side passively, and solely by 
the growth of the body wall, but that the ends move relatively to the 
cells of the body wall. A somewhat late stage in the development of 
the pyramidalis is shown in Figure 63 (Plate VI.). Nearly the whole 
of the mesoderm of the body wall has here been transformed into 
muscle cells. The insertion of the muscles is in the mesoderm of the 
neck of the polypide. (Plate VI. Fig. 63; Plate V. Fig. 45.) 

c. Parietal muscles first make their appearance at about the stage 
of the terminal individual of Plate II. Figure 14, immediately below 
the bud and to the right and left, i. e. so that the muscles, which 
usually arise paired, have their long axes parallel to the sagittal plane 


MUSEUM OF COMPARATIVE ZOÖLOGY. 29 


and perpendicular to the long axis of the branch. They arise from cells 
of the mesoderm, most of which in this region are filled with vacuoles, 
and often project into the colom, But in my opinion the muscle cells 
do not themselves arise from such vacuolated cells, for at even an earlier 
stage (corresponding to Figure 21, Plate III.) one can distinguish thick- 
ened patches of elongated cells in the mesoderm which are undoubtedly 
the young muscle cells; but they do not show the slightest traces of 
being vacuolated, and in fact are sharply distinguished from the adjacent 
cells by their uniformly granular appearance and their deeper coloration. 

3raem (’90, pp. 124, 125) has already stated that the parictal muscles 
arise in pairs, and come to traverse the cœlom, not remaining in the 
body wall. The truth of this statement I can confirm in the case of 
the parietal muscles first formed, which lie near the future septum. 
Plate V. Fig. 42 shows the origin of the muscle fibres on both sides 
of the branch. They have already migrated into the ewlom. As Braem 
plainly states, the component parts of this pair of muscles, developed 
from the mesoderm, migrate towards «ach other and finally fuse into 
one unpaired mass, as we see in Plate III. Figure 26. It is perfectly 
evident, in this case at least, that both ends of two muscles originating 
far apart migrate in some manner towards each other so that the cor- 
responding ends come to lie close together. Such a migration cannot 
be accounted for merely by growth of the body wall. The ends of the 
muscle fibres must move relatively to the body wall. 

When the muscles have reached their permanent positions in a 
diameter of the branch, we find their ends attached to the cuticula. 
As the muscle fibres stain deeply in hematoxylin, they can be distinctly 
traced through the vacuolated and poorly stained cells of the body wall 
(Plate IIL Fig. 26). Figure 29 shows a bit of the wall mechanically 
separated from the cuticula, the end of the muscle fibre remaining in 
place. Fine lines can be distinguished in the contractile, deeply stain- 
ing portion of the fibre. The surface by which attachment is effected 
appears very slightly erenulated on longitudinal sections of the muscle 
fibre. I could not distinguish any structural peculiarity on the part 
of tho cuticula to which the muscle was attached, — nothing to indicate 
how attachment is effected. 

Freese (88, pp. 15, 22, Fig. 11) has described a similar method of 
attachment of the muscles to the cuticula for Membranipora.! 

! My friend, Dr. G. H. Parker, tells me that a similar method of attachment of mus- 
cle fibres to the cuticula occurs in Crustacea. According to Tullberg (’82, pp. 27, 44, 
45), the adductor muscle fibres are in Mollusks attached to the cells of the ectoderm, 
The same condition as in Mollusks seems to exist in Annelids (Hisig, ’87, pp. 25, 86) 


30 BULLETIN OF THE 


At a later stage smaller bundles of muscles arise successively toward 
the neck. These muscles are free from the body wall at their middle 
region. They do not usually pass through the coolom in a diameter of 
the branch, however, but rarely subtend as chords an are of more than 
120°. As Braem supposed, such muscles, although arising later than 
the most proximal pair, originate in a similar manner to them (Plate 
VI. Fig. 55). The mesoderm is very thin at the region at which they 
are first seen, and they are quickly discerned by their larger nuclei 
and prominent cell body. At a later stage they have grown much 
longer, and become freed from the body wall at their middle part. 

As is well known, there are two funiculi in Paludicella, called by 
Allman respectively anterior (nearer the atrial opening) and posterior. 
The origin of the funiculi of Paludicella was observed by Dumortier 
et van Beneden as long ago as 1850. They say (p. 54), “La couche 
muqueuse une fois formée s'étend rapidement dans l’interieur et touche 
bientöt par son extrémité inférieure les parois opposées de la loge. 
Les cellules muqueuses dont le tout est encore composé contractent de 
Padhérence dans cet endroit, et c’est ce qui donne naissance au muscle 
rétracteur de Vestomac [= funiculi].” Allman (56, p. 36, Plate XI. 
Figs. 7-9) also describes and figures very clearly and correctly this pro- 
cess, and Braem (’90, p. 127) has recently confirmed their observations. 

It is perhaps unnecessary to redescribe the more evident part of 
this process, the contact of the polypide with the abatrial wall of the 
branch. The mesoderm of the bud comes into contact with that of 
the body wall, the cells of each of the two layers become attached to the 
other, and by the withdrawal of the polypide the attachment persists 
at two points forming a long drawn out string of tissue. Figures 36° 
and 38 (Plate IV.) are contributions to a knowledge of the finer details 
of this process. Apparently the upper funiculus is developed earlier 
than the lower, as I have always found it longer at about this stage. 
The lower funiculus at present consists of only the two mesodermal 
layers of body wall and polypide intimately united. The funiculus 
itself consists of a cord several cells thick ; but I believe these oer- 
tainly to be derived from the mesoderm only. Very early some of 
these cells show an appearance of highly refractive and deeply staining 
fibres, which I interpret as muscular differentiation (Plate IV. Fig. 38, 
fun. su.), 80 that the funiculi must be regarded as partly muscular 
in function. As in Phylactolemata, these fibres lie near the axis of the 
funiculus. Braem (90, pp. 66, 67,) has demonstrated that the museular 
fibres of the funiculus of Plumatella pass directly into the muscularis 


9 


MUSEUM OF COMPARATIVE ZOÖLOGY. ol 


of the body wall. It is interesting to find them persisting in the fu- 
niculus of Paludicella, beneath the mesodermal covering, although there 
is apparently no muscularis developed in the body wall of this region. 


8. Tun Formation or THE Neck AND ATRIAL OPENING. 


This is the last act in the history of the polypide that I shall con- 
sider. The body wall around the neck of the polypide continues to 
possess a less differentiated character than the remaining portion for 
some time after the oral tentacles have undergone their revolution. 
One still sees the cells of this region dividing, and the body wall is 
gradually protruded at this point above the general level. (Plate II. 
Fig. 14, cev. pyd.) The neck of the polypide to which the kamptoderm 
is attached consists, at a somewhat earlier stage than that just referred 
to, of a disk of greatly elongated columnar cells in the centre of which 
there is a distinct notch caused by the presence of shorter cells at that 
point. (Plate VI. Fig. 63 6.) At the inner ends of the columnar cells 
of the neck lies a flat epithelium quite sharply marked off from the 
latter, but which is nevertheless undoubtedly derived from the same 
source as the columnar cells and the inner layer of the bud. This flat 
layer is directly continuous with the inner layer of the kamptoderm. 
At a later stage, the columnar cells of the ectoderm become elongated 
still more, and lose their staining capabilities at their outer ends. Still 
later one sees them arranged in the form of a cup whose cavity is sep- 
arated from the outside world only by a cuticula which becomes slightly 
invaginated at this point. The cells are soon found with their long 
axes perpendicular to the edge of the cavity they line. 

There is one point that I have not been able to determine; namely, 
how the new cuticula, which is certainly formed at the ends of the cells 
which lie next to the cavity, becomes continuous with the old cuticula 
of the non-invaginated body wall, as it is in Figure 50 (Plate V.). The 
presence on the new unstainable cuticula of the remains of the stainable 
one, whose origin I have already discussed at length, may serve as a 
guide to the limits of the old eutieula. The new cuticula is being secreted 
by cells lying deep in the inner end of the neck, and apparently in one 
rod-like mass. Unfortunately, I lack stages between this figure and 
Figure 45 (Plate V.), which shows the neck of a nearly or quite adult 
polypide cut lengthwise. The solid cuticular rod has now become a hol- 
low cylinder, whose inner (deep) edge is embedded in the deep-lying cells 
of the neck, Moreover, one finds superficial to the cuticula of the gen- 
eral body wall a second cuticular cylinder, which is free at its outer end, 


32 BULLETIN OF THE 


but at its inner end fuses with the surrounding cylinder of cuticula. This 
inner cylinder, which is probably formed, as Kraepelin (87, p. 40) sug- 
gested, by splitting of the delicate cuticula at the base of the marginal 
thickening (Randwulst), has been compared by Kraepeliti to the “ collare 
setosum ” of Ctenostomes. The Randwulst itself I believe to be the equiv- 
alent of the Diaphragma of Nitsche, as I shall try to show later. 

At the deep end of the neck (Fig. 45), the inner layer of the bud is 
seen to be continuous with the ectoderm, The region of transition may 
be called the atrial opening, of. atr. Surrounding the atrial opening 
is a fold in the ectoderm, and between the layers of this fold is a thin, 
non-stainable homogeneous layer, slightly more refractive than the sur- 
rounding protoplasm. This membrane extends also a short way into 
the kamptoderm, and here lies between its two cell layers. Embedded 
in this homogencous membrane in the fold, one can distinguish still 
more highly refractive bodies, spht. On account of their form and 
high refractivity, I believe these to be muscle fibres cut across. The 
homogeneous membrane has also the same general appearance and 
relation to the muscularis as the so-called supporting membrane of 
Nitsche, and it is the only representative of that structure that I 
have found in Paludicella. 


9, DEVELOPMENT OF THE COMMUNICATION PLATE. 


In their description of Paludicella, Dumortier et van Beneden (’50, 
p. 40) say : “Il se compose de plusieurs loges ou cellules placées bout a 
bout . . . en sorte qw’il n’y a aucune communication entre les differents 
animaux.” Also Allman (56, pp. 114, 115) refers to the presence of a 
perfectly formed septum separating the cavities of adjacent “ cells.” To 
Kraepelin (87, p. 38) belongs the credit of having first carefully studied 
this structure in the adult by means of sections. He came to the con- 
clusion from the appearances which he figures (cf. my Plate V. Fig. 49), 
that there are small canals passing through the nearly homogeneous 
central mags, and therefore “dass wir in dem ganzen Apparat eine Vor- 
richtung zu erblicken haben, durch welche Nährstofllösungen des einen 
Tieres mittels siebartig wirkender Cautclen in die Körperhöhle des 
Nachbarindividuums übergeführt werden.” 

The descriptions of Kraepelin concerning the structure of the “ Roset- 
tenplate” are confirmed by my own observations, and seem to justify his 
conclusions concerning its function, The development of the organ has 
not, however, been carefully observed heretofore. Korotneff (74, Plate 
XII. Figs. 1 and 2) gives figures to show this process, but I have never 


MUSEUM OF COMPARATIVE ZOOLOGY, 33 
seen any such circular groove surrounding the branch as he figures. In 
all cases the two layers of the body wall form a circular fold, in which, 
however, there is never, even at the earliest stages, a space between the 
ectodermal layers, nor any infolding of the cuticula as Korotneff (75, p. 
369), according to Hoyer’s rather incomplete abstract, maintains (Plate 
V. Fig. 47). When the circular fold has advanced until only a small pore 
remains, by which the cavities of the older and younger individuals are 
kept in communication, the mesodermal cells at the angle of the fold 
begin to undergo a metamorphosis both in form and histological charac- 
ter. In the first place they become much elongated and extremely 
attenuated, passing from one surface of the septum to the other, and 
forming the lips of the pore. In the second place their plasma becomes 
first deeply stainable, and later, in addition, homogeneous and highly 
refractive. These metamorphosed cells form what may be called the 
teeth of the plate. They are derived wholly from mesoderm. 

The cells in the upper mesodermal layer next increase rapidly in 
number and size, and the number of teeth is also augmented (Plate V. 
Fig. 48). The metamorphosis of the cells extends still farther away 
from the communication pore, and involves the lower mesodermal layer ; 
but, apparently, each cell of the latter is metamorphosed only to a 
slight depth within its cell wall (Fig. 51), whereas in each of the upper 
cells the ends which project into the communication pore are modified 
through and through (Fig. 46). At a later stage (Fig. 49) the meta- 
morphosed part of the cell seems quite sharply cut off from the active 
part, and the slits between the metamorphosed teeth are considerably 
reduced. Nevertheless, I believe a transfer of fluids may still occur 
between them, for even in the adult communication plate one can trace 
continuous lumina when the cells are by accident torn off from the 
“teeth” which they have produced. It is important to note that the nu- 
clei are not destroyed in the cell metamorphosis. Some lie above, others 
below the pore, and become deeply stainable. The ectodermal layers of 
the communication plate secrete a cuticula between them, This is thin- 
ner than that of the body wall, and does not extend, of course, to the 
centre of the communication plate, but ends in a thickened ring, whose 
diameter is about one tenth the diameter of the plate, or, absolutely, 
about 9.4 pt 


1 Reichert (70, p. 267) first carefully described the Rosettenplate of Cteno- 
stomes in Zoöbotryon, and the organ in Paludicella must be regarded as homologous 
with it, The central circular hole in the cuticula of Zoöbotryon is from 7 to 10 u 
in diameter, and from one ninth to one seventh that of the entire plate. Similar 


9 


VOL, XXII — NO. 1. 2 


BULLETIN OF THE 


10, RÔLE or THE MESODERMAL VACUOLATED CELLS, 


Allman (56, p. 36) observed that at the time a lateral branch was 
well formed, and before the origin of tbe polypide, the internal outline of 
the body wall was uneven, and he figures (Plate XI. Fig. 4) very large 
cells lying on the inside of the body wall. Korotneff (74, Taf. XI. 
Figs. 1-3, ’75, pp. 369, 370) progressed a step farther, and recognized 
a distinction between large, coarsely granular cells projecting into the 
cavity of the bud, especially near the tip, and the surrounding epithelial 
cells. Braem (90, p. 126), finally, has described them more accurately. 
He finds cells filled with numerous granules in the youngest branches of 
the colony. Immediately around the bud, such cells are less abundant ; 
probably, he says, because their granules have been absorbed in the 
process of formation of the polypide. He compares the granules with 
the yolk spherules of the statoblast cells, and believes that they are to 
be regarded as food matter. 


My observations and conc.usions, achieved independently of B 


raem’s, 
fully confirm his. I have succeeded, moreover, in obtaining some addi- 
tional evidence as to the function of these cells, a subject to which I 
have paid some attention. 

First as to the distribution of the cells, and their frequency in different 
regions. We can best get an approximate idea of this by counting the 
number of the reticulated cells in each section of a series which in- 
volves a young polypide and the regions immediately above and below 
it. It is not possible to do this with perfect accuracy, because there is 
no sharp line of distinction between reticulated and non-reticulated cells ; 
but I have made the count without prejudice, and I believe as fairly as 
possible. When the bud of the polypide has reached about the stage 
shown in Plate III. Figure 28, the number of reticulated cells seems to 
have nearly reached a maximum. In the series from which this figure 
was taken there was an average of 4.8 reticulated cells to the section in 
the ten sections distal of the bud. There was an average of 11.2 reticu- 
lated cells to the section for the twenty sections which passed through the 
bud, and 11.2 for the eleven sections proximal of the bud in the region 


perforated organs have been described by Smitt (’67, p. 426), Nitsche (71, pp. 
429-422), and Vigelius (’84, p. 26) for Flustra, by Freese (’88, p. 7, 18, 14) for 
Membranipora, by Ostroumoff (’864, p. 13) for Lepralia, by Claperede (’70, p. 160) 
for Bugula and Serupocellaria, by Ehlers (’76, p. 14) for Hypophorella, and by 
Joliet (77, p. 222) for Bowerbankia. Nitsche alone (’71, p. 455) has had anything 
to say upon their origin, and this apparently not the result of direct observation. 


MUSEUM OF COMPARATIVE ZOÖLOGY. 3D 


at which muscle fibres were arising. A similar series through a slightly 
older bud gives for the same regions respectively 5, 14, and 13 cells per 
section. In series through older buds, a rapid decline in the number of 
these cells occurs so that at the stage of Figure 30 (Plate IV.) there is an 
average of only about 3.1 cells per section through the bud, and about 
2.2 immediately below. These reticulated cells are not very numerous 
in the region of the bud at the time this is about to arise, as a look at 
the sections Figures 3 and 4 shows. One finds reticulated cells in the 
mesoderm at the tip, and most abundantly at a rather early stage in 
the development of the bud. The number of these cells diminishes as 
one leaves the young individual to pass into the next older of the same 
branch. In the adult such cells are rather rare; so rare, in fact, that 
Kraepelin (’87), who studied with care the body wall of the adult indi- 
vidual, makes no mention of them. Nevertheless they do occur in the 
cells which are to go into the lateral branch (Plate II. Fig. 15), as well 
as elsewhere on the body wall. The place in which one finds the reticu- 
lated cells most abundant, however, is in the young lateral branches near 
the time when the polypide bud is about to arise. Here every cell of the 
mesoderm is greatly enlarged, and filled with the vacuoles (Plate VI. 
Fig. 58). These are very apparent upon a surface view of the branches. 
Reticulated cells occur not only in the mesodermic cells of the body 
wall, but also in those of the polypide bud, which were, indeed, only 
lately a part of the mural mesoderm (Plate III. Fig. 28, Plate VI. Fig. 
56). Thus, in general terms, we may say that the reticulated cells of 
the mesoderm are chiefly confined to regions in which there are young 
buds developing; and since these arise at intervals only, there is a 
periodicity in their appearance,—a time of maximum development 
followed by one of decline, then one of reproduction of such cells in 
the ends of branches culminating in another maximum, and so on. 
Turning our attention now more particularly to the structure of these 
reticulated cells at the period of their best development, we find (Plate 
VI. Figs. 56, 57, 59) that they possess a large nucleus lying at the 
deep end of the cell and containing a relatively large nucleolus, and that 
this is surrounded by a granular protoplasm with included vacuoles. It 
is very common to find the nuclei in various stages of division, and thus 
it is frequently seen as a mass of chromatic substance without any nu- 
clear membrane or nucleochylema. The vacuoles, which in the more reg- 
ular cells lie in a semicircle nearly peripheral (the nucleus being at the 
centre), are highly variable in number, some of the cells containing as 
many as 20 to 30. They often appear as perfectly clear homogeneous 


36 BULLETIN OF THE 


spaces, but more frequently at this stage contain a spherical body, which 
frequently fills the entire vacuole and is more refractive than the sur- 
rounding plasma (Fig. 59). Not unfrequently one sees a less refractive, 
clear space, surrounding the highly refractive body (Iig. 57). 

The description just given corresponds to the condition seen in a ter- 
minal branch whose polypide has attained the development of that shown 
in Figure 28 (Plate 111.). At the time immediately preceding the ori- 
gin of the bud, the cuboidal cells of the mesoderm show traces of vac- 
uolation, but their form and size have suffered no appreciable disturbance. 
This vacuolation of cells proceeds hand in hand with the development of 
the bud, and one first notices the homogeneous, highly refractive bodies 
in the vacuoles when the bud is well established. At about the time the 
alimentary tract has become formed, the reticulated cells begin to show 
sigus of degeneration. The highly refractive bodies have disappeared, 
and the skeleton of the cell which remains becomes very irregular. As 
already stated, the number of reticulated cells also decreases, until, at 
about the time of “rotation” of the polypide, there are few reticulated 
cells in the mesoderm, but these few are filled with vacuoles and their 
highly refractive bodies. 

The conditions of the mesodermal cells at the tip are slightly different 
from those found elsewhere. Usually, instead of many small vacuoles, 
one finds only one or two which fill almost the entire cell, — sometimes 
perfectly homogeneous in structure, sometimes containing small highly 
refractive granules. 

These appearances I believe to be explicable only upon the assumption 
that the mesodermal cells are capable, at the time at which the young poly- 
pide is arising, of imbibing the fluids of the body cavity and storing them 
up for the purpose of supplying the rapidly growing cells of the bud with 
nutrition. It is desirable to show reasons for believing, first, that the 
contents of these cells are nutritive matter ; secondly, that this has been 
taken up from the body cavity; and, thirdly, that it is supplied to the 
bud for its nutrition. 

It must be admitted that the strongest argument for the belief that 
these are absorbing cells is derived from a comparison of the appearances 
which we find in these cells with those described for Protozoa, and by 
Metschnikoff (83, Taf. I. Figs, 18-35) for mesodermal trophic cells. 
At the same time, it must be acknowledged that similar cells are found 
in other cases where the function is believed to be not ingestive, but 
excretory, as in the chlorogogen cells of Annelids, as shown by Kiiken- 
thal (85), Eisig (87, pp. 751-762), and others, and indeed even in the 


MUSEUM OF COMPARATIVE ZOOLOGY. OF 


cells of coelomic epithelium. Eisig (’87, p. 752) has already clearly ex- 
pressed how, in view of the many cases of high excretory activity of 
peritoneal and blood cells demonstrated by him, “ kiinftighin bei der 
Beurtheilung gewisser Zelleneinschliisse erst genau festzustellen sein wird, 
ob man est mit von aussen aufgenommenen (gefressenen), oder aber mit 
von der Zelle ausgeschiedenen Producten zu thun habe.” 

A criterion for judging this matter may be found, in the first place, I 
believe, in this: that the products of exeretion increase with the activi- 
ties of the cells, and are thrown out, usually in the shape of concrements, 
either from the cell or with the cell into the cwlom ; whereas bodies 
taken in from without for digestion decrease with the activities of the 
region. In the second place, vacuoles are less characteristic of excretory 
tissue than of imbibitory. But vacuoles are the important feature of the 
reticulated cells in Paludicella, and the highly refractive bodies are less 
constant phenomena. As for the latter, they are not found in the later 
stages, nor in the earliest. Moreover, these bodies differ from excretion 
concrements in this, that they are always transparent, often almost indis- 
cernible in the vacuole, except by their higher refractiveness, and there 
is no sharp demarcation between cases of vacuoles filled by such bodies 
and those the contents of which are less highly refractive. ‘The degree 
of refractiveness is variable, at one end of the series grading off into the 
undifferentiated fluid of the vacuole. What significance is to be assigned 
to these highly refractive bodies in the vacuoles? There are two reasons 
why I do not believe that they represent solid food particles devoured 
as such by the mesodermal cells. First, I do not find such highly 
refractive bodies lying loose in the body cavity before the stage at which 
they first appear in the cells; and, secondly, one can find all gradations 
between less highly refractive vacuoles and highly refractive ones (which 
I have assumed to be entirely filled by one highly refractive body), and 
between the latter and vacuoles containing a small body surrounded by 
a broad, clear area. I believe, therefore, that the vacuoles are rather 
cavities filled with chemically different nutritive fluids, which are acted 
upon differently by the reagent. 

I have assumed that the contents of the vacuoles represent material 
taken up from the body cavity, because it seemed most reasonable to 
look there for the source of their supply. ‘The ectoderm is covered on 
its outer surface by an apparently continuous cuticula, so that food 
cannot be gained from the outside world directly. It is, moreover, not 
unreasonable to suppose that some of the products of digestion elabo- 
rated by the adult polypides of the colony pass through the wall of the 


38 BULLETIN OF THE 


alimentary tract in solution, and thus into the body cavity, from which 
they may be taken up by the mesodermal cells at the growing part 
of the body wall. Nor is there anything unreasonable in insisting that 
the body cavity functions, in these animals without blood-vessels, as a 
heemo-lymph system, for in many animals with incomplete vessels, such 
as Arthropods, Hirudinea, ete., it evidently does so to a certain degree. 
Moreover the constant motion of the fluids of the body cavity of Bryozoa 
points to the same thing. It is conceivable that the food in the digest- 
ive cells might be distributed throaghout the body wall without passing 
into the body cavity, since all parts of the body wall are continuous 
with the digestive epithelia of the polypides of the colony. Two consid- 
erations make it improbable that the cells of the tip gain their nutri- 
tion in this manner from the digestive cells of the youngest functional 
polypide: first, the considerable distance of the rapidly growing, and 
hence rapidly consuming tip, from the youngest functional polypide ; 
and, secondly, the fact that the tip is separated from that polypide by 
one or two septa, whose central cells are highly metamorphosed, and 
apparently cuticularized, thus serving to break the continuity of the 
cell wall. An objection to the assumption that the mesodermal cells of 
the tip derive their nourishment from the products of digestion which 
have been elaborated by the alimentary tract of the youngest polypides 
and passed into the body cavity, might be based on the fact that the 
communication plates are always fully formed between the bud and 
the next older polypide before the older polypide has become functional. 
If the communication plate were a closed septum, this would be a fatal 
objection. But it is not closed to fluids carrying food in solution, The 
very persistence of an opening indicates that it has a function, and favors 
the hypothesis here presented. 

Positive evidence for the conclusion that the reticulated mesodermal 
cells take up food material from the body cavity is derived from the 
fact that these cells often show evidences of being amosboid. Thus they 
are sometimes found with pseudopodia-like prolongations of the cell body 
(Figs. 54 and 59). <A large percentage of all reticulated cells of this 
stage show similar appearances. Although they here seem to keep their 
places in the mesodermal epithelium, their movements being confined 
to their free surfaces, the cells derived from the homologous Jayer in 
marine Bryozoa are migratory. Therefore these may be considered as 
morphological equivalents of migratory cells, which have come to remain 
in or have never departed from the mesodermal layer, although possess- 
ing some of the characters of these notoriously trophic elements. 


MUSEUM OF COMPARATIVE ZOÖLOGY, 39 


That the nutritive matter in the cwlomic cells is supplied to the 
young bud is what we should expect, since the cells of the bud, being 
most actively engaged in growth, will require most nutriment. The 
actively dividing cells of the outer layer of the bud are thick and cuboid, 
and are rarely so highly vacuolated as the more passive ones of the body 
wall; yet occasionally one finds one or two huge cells in this layer full 
of vacuoles, which contain highly refractive bodies. In most cases these 
cells send out processes into the colom, and in a few instances I have 
seen them united with similar processes from cells on distant parts 
of the body wall. This remarkable phenomenon, shown in Figure 54 
(Plate VI.), may possibly signify that cells of the cœlomic epithelium at 
times directly communicate with those of the outer layer of the bud to 
supply it with nourishment. Nutrition of the bud is also probably 
effected through the presence of large reticulated cells at the angle 
between the bud and the body wall, A condition like that shown in 
Figure 56, cl. ret., is very common, 

Every author from Dumortier et van Beneden to Braem, who has 
studied the origin of the polypide in Paludicella, has mentioned the 
presence of highly refractive bodies in the alimentary tract at the time of 
its formation. “These are very striking in some living specimens, and 
in whole animals after killing. I have found that this highly refractive 
substance in the bud is exceedingly variable in amount and position, 
and that sometimes it is apparently absent. When present, it usually 
occupies the lumen of the forming alimentary canal ; but, as sections 
show, it is often located in large vacuoles in the future digestive cells 
of the alimentary tract. It seems highly probable that, as Braem sug- 
gests, this is nutritive substance, and it has doubtless come from the 
body cavity through the agency not only of the outer layer of the bud, 
but also of other parts of the coolomic epithelium. 

I am inclined to interpret the phenomenon of cells filled with nutri- 
tive material as an adaptation to the peculiar conditions of Paludicella, 
in which the individuals are early separated from one another, except 
for the communication plate, through which at best fluids can pass only 
slowly, and in which a rapid growth of the body wall to produce the 
poly pide 3 place periodically. The mesodermal cells rapidly absorb 
the nutritive fluids of the body cavity and store them in their substance 
before the formation of the communication plate, and give them out 
again during the period of the polypide’s most rapid growth chiefly to 
this part of the individual. This hypothesis has been mainly derived 
from considering the fact of the great development of the reticulated 


40 BULLETIN OF THE 


cells in the lateral bud and the very early completion of its commu- 
nication plate, the immediate needs of the polypide, which arises only 
after the formation of the plate, being met by this supply of stored 
nutriment.! 

But why is the septum (communication plate) formed so early, if it 
is desirable for the species that the growing tip should be well nourished 
by the fluids of the body cavity? Here again I must resort to pure 
hypothesis. I assume that the early formation of the septum is a pro- 
vision for the protection of the stock against a rapid influx of the sur- 
rounding water in case the brauch is broken, One can understand how, 
if the body wall and growing regions depend upon the fluids of the body 
cavity for nutrition, an open communication of this cavity with the out- 
side world would be a serious obstacle to regeneration of the body wall in 
the lost part, or the growth of the stock in any other direction. There is 
a fact which ought to be mentioned in this connection, as bearing on this 
hypothesis of the function of the septa. One frequently finds that in 
stocks which have been handled with reasonable care the median branches 
are broken off at either end, and in almost every colony one or more lat- 
eral branches are missing from the parent branch. Apparently, then, 
the lateral branches are unusually subject to destruction, and we find 
the septs developed at a much earlier period between them and the an- 
cestral branch than between individuals of the median branch. Compare 
Plate IT. Figure 14, in which the communication plate has not yet begun 
to form, with Plate VI. Figure 58. 


III. Budding in Marine Gymnolsmata. 


1, ARCHITECTURE OF THE STOCK. 


T have already described the process of stock-building in Paludicella, 
and have attempted to show that it follows a certain law. I desire now 
to present a few observations upon the architecture of certain stocks of 
marine Gymnolsemata, which will aid in arriving at some general con- 
clusions later on. Other observers have worked out the architectural 


laws of single species or groups, and I shall refer to their studies either 


1 Similar conditions to those in Palndicella exist in some marine Bryozoa, and in 
one of these cases, Bowerbankia, I find them fulfilled by a similar arrangement. The 
young buds of the stolon which give rise to the “nutritive zodids” are, at an early 
stage, loaded with food granules, As in Paludicella, so in Bowerbankia the 


communication plates are formed early. 


MUSEUM OF COMPARATIVE ZOÖLOGY. 41 


in connection with the species which I have used in common with them, 
or in the general part of this paper, in considering the process of budding 
in Bryozoa as a whole. 

I will begin my description with Bugula turrita* of Verrill, which I 
gathered in the summer of 1889 at Wood’s Holl, where it occurs abun- 
dantly on the piles of the wharf. The stock is bushy, and, when its 
polypides are active, of an orange color. In its simplest form the 
stock consists of a central axis, which is somewhat zigzag, and gives off 
lateral branches like the trunk of a tree. The lateral branches are in- 
serted on the trunk in a spiral line. Each lateral branch is fan-shaped 
(Plate VII. Fig. 64), the part corresponding to the handle of the fan 
being the point of attachment, and the fans are smaller the nearer they 
are to the tip of the trunk. The attachment of the branch to the 
trunk is effected by one primary individual. Each fan-shaped branch 
extends from its point of attachment obliquely upward and outward, 
and, although it is slightly concave on its upper inner surface, the 
concavity is not sufficient to prevent its being spread out upon the slide 
for study without materially disturbing the interrelation of the indi- 
viduals in the stock. 

I have studied several branches flattened in this way (one of 400 
individuals), and have made camera drawings of them. Since the 
results in the different cases are substantially in agreement, I have con- 
cluded that they are significant. One of these camera drawings is shown 
in the figure just referred to. 

To designate individuals in che stock, I have adopted a simple no- 
menclature. The forty-four terminal individuals are numbered from 
l to 44. The successive generations (if I may be allowed to use this 
word in a loose way) are indicated by the Roman numerals from I. to 
XII. Any one individual is indicated by placing the numbers of the 
radial line or lines to which it belongs first, and following this by 
the Roman numeral of the generation to which it velongs. Thus, 
27-30 IV. is an individual near the base of the twig 27-30 and of 
generation IV. Figure 64° (Plate VII.) is a diagram showing the 


1 This species is very similar in general habit to B. avicularia, Linnaeus, and to 
B. turbinata, Alder (Hincks, ’80, pp. 75-80). It differs from the first named spe- 
cies by possessing only one spine, on the outer upper edge, as described by Leidy 
(’55, p. 142), instead of having three, — two outer upper and one inner and upper. 
It differs from Hincks’s diagnosis of the second in having only two “cells” in each 
branch, instead of 3-6 in the upper portions. The form of the avicularium would 
seem to ally it more closely to B. avicularia. 


42 BULLETIN OF THE 


arrangement of the individuals in Figure 64. The radial lines repre- 
sent the rows of individuals; the concentric lines separate adjacent 
individuals of the same radial row. The same nomenclature is used 
as in Figure 64, 

In studying Figures 64 and 64°, one of the first facts which attracts 
our attention is that (1) the individuals of the twigs are in pairs, and 
the adjacent individuals of the two rows “ break joints.” In general, one 
finds that the individuals of the same twig are of the same length ; but 
since the two rows of any twig ultimately rest upom one, either the prox- 
imal two individuals of these rows must be of unequal length, or else 
they must arise on different parts of the individual which supports 
them. Both of these cases occur. Sometimes one individual (26 IX.) 
has nearly twice the length of the other (25 IX.), and in other cases 
(9, 10 VL, 11, 12 VI.) the more proximal of the two individuals 
(9, 10 VI.) arises so far proximally on the side of the supporting indi- 
vidual 9-12 V. as to have a total length quite equal to that of the more 
distal (11, 12 VI.). Owing to their different positions upon the indi- 
vidual 9-12 V., these two individuals may be designated as lateral 
(9, 10 VI.) and terminal (11, 22, VI.). The terminal individuals con- 
tinue the ancestral row; the lateral individuals are the first of lateral 
branches. 

This distinction is an actual, and by no means a meaningless one. 
The constant difference in position of the two individuals which rest 
upon one shows conclusively that this branching cannot be regarded as 
dichotomous,.and I may say parenthetically that I shall try to show 
in the general part of this paper that true dichotomy is not com- 
mon in Bryozoan stocks, if indeed it exist at all. Now, since in 
the rows of individuals in which there is no lateral budding the 
distal lies directly terminal to the proximal individual, that individual 
which fulfils this condition at the region of bifurcation of the twig 
must be regarded as continuing the ancestral branch; and that 
individual, conversely, which arises from the side of the single prox- 
imal individual must be regarded as the lateral one. Thus we have 
the stock composed of ancestral and lateral branches as represented 
in Figure 64% 

(2) When two lateral branches are given off from two ancestral ones 
which have had a common origin (and are consequently themselves re- 
spectively ancestral and lateral branches), they are given of’ towards 
each other, This is equally true whether the two lateral branches in 
question arise in the same generation (32 X., 33 X.) or in different 


MUSEUM OF COMPARATIVE ZOOLOGY. 43 


generations (24 X.,25 IX.). This may be expressed by saying branches 
are given off on the side towards the axils. 

By consulting Figure 64" and tracing out the finely dotted lines which 
connect the second, third, etc. axils of all branches counting from the 
proximal end of the fan, it will be seen that (3) lateral buds tend to arise 
on two closely related branches in the same generation. There are several 
slight deviations from this rule. The less closely related the branches, 
the less marked the tendency, although it is still discernible. (Of. 
branches 9-16, 23-30.) 

This rule does not hold, however, so well on the margins as in the 
middle region of the fan, for here another and a superior rule seems to 
obtain. This is that (4) lateral budding occurs more frequently at the 
margins of “ fans” than elsewhere. Thus in Figure 64* there is at the 
margins, on the average, 1 case of lateral budding to 4.3 cases of median 
budding. Elsewhere the average is as 1 to 6.5. In larger fans the 
difference is even more pronounced, This is true not only for the 
“fans,” but also, to a less degree, for the two “ subfans” which arise re- 
spectively from the two individuals of generation II. (but 17, 18 is very 
anomalous in this respect). In general, any rule deduced for the mar- 
gin of the fans holds true also for subfans to any degree of subdivision ; 
but the less perfectly, the higher the degree. 

By consulting again the diagram, it will be seen that the branches 
have attained different lengths. Thus 9, 10, 29, and 30 contain repre- 
sentatives of generation XII., while the terminal individual of branch 1 
is of generation X., and those of branches 35-44 are of generation XI. 
So the curve which connects the tips of the branches (see dot-and- 
dash line, Fig. 64°) would rise from 1 to 9-10 as a maximum, and fall 
again till it reached the margin of the first subfan; then rise again, 
reaching a second maximum in the middle at 29-30, and finally fall 
again to the other margin. In general, then, (5) the marginal branches 
are shortest, the intermediate ones longest, i. e. give rise to the greatest 
number of generations. 

Although the marginal individuals of say generation II., IV., or V. 
do not support branches with so many generations as the intermediate 
ones, yet they are not therefore necessarily less prolific in individuals, 
because the number of branches arising distally of such individuals is 
greater according to rule 4 than the number arising distally of the 
intermediate ones. Thus, if we count the number of individuals borne 
on each of the eight individuals of the fourth (IV.) generation of 


Figure 64, we find in the given case :— 


44. BULLETIN OF THE 


ter | 1. 1-8 IV. (an outer individ.) gives rise to 8 branches and 34 individ, 
2. 9-12 IV. (an inner Cine) ee ed Ter at 
md 8. 13-16 IV. (an inner K adres ee re Katkin aai 
l4. 17,18 IV. (asubouter “ ET TO Kr e > 
Tara, | 5. 19-22 IV. (arshouter “ jJ «€ er! bi REN 
6. 23-26 IV. (ar ‚ner Sie S ei! ki ier en 
Oren § 7. 27-80 IV. (an inner BR EH Eee i megg st 
(8, 81-44 IV. (an outer Mishel ef ld rs 700 4 


According to the rule that inner branches are slightly prolific, we should 
expect cases numbered 4 and 5 in the above table to contain the fewest 
branches and individuals; in accordance with the rule that marginal 
branches even of subfans are more prolific, we should expect them, on 
the contrary, to contain more branches and individuals than cases num- 
bered 3, 6, ete. The result is usually a condition intermediate between 
that of the middle and outer branches, such as is partially realized in 
case number 5, Case number 4 seems to present an unusual condition, 
which may be correlated with the fact of its close approximation to 
number 5. (See Fig. 64, 17-20.) From the consideration of this and 
other cases, I think this conclusion may fairly be drawn: (6) Of the 
Jour proximal individuals from which a fan arises, the outer two will bear 
the greater number of individuals, the inner two the lesser. 

Since from rule 2 median individuals (ancestral branches) occupy the 
margins of fans (or subfans of any degree) and the lateral branches are 
intermediate, it follows, as a corollary to rule 5, that, in general, the an- 
cestral branches are the shorter, the lateral branches the longer; and, as 
a corollary to rule 6, that from any axil the ancestral branch will of the 
two give rise to the greater number of individuals; the lateral branch, 
conversely, to the less, other conditions being equal. 

We have deduced the laws of lateral budding on different parts of the 
circumference. We find also that there is a regular variation in the 
frequency of lateral budding, dependent upon the distance of the region 
from the primary individual of the fan. This rule, like any other, is not 
invariable, whatever the other conditions may be; but it is more or less 
dependent upon them. A small and regular fan having seven genera- 
tions gives this result. 

No. of Generation, Number of Individ. Increase per Cent. 


1 


MUSEUM OF COMPARATIVE ZOOLOGY. 45 

In this table the first column gives the number of the generation, the 
second the whole number of individuals in the generation, and the third 
column the increase per cent of individuals in each succeeding genera- 
tion over the last. In this specimen the increase underwent a very 
regular diminution. 

With larger colonies so great a regularity as that just shown is hardly 
to be expected, nor is it found. The following table is based on 
Figure 64, and is like the preceding; but in addition the percentage 
increases have been averaged — i. e. the means of successive increases 
taken in pairs have been given—to eliminate what may be called 


accidental variations, 


Gene | N | > | Genera- |Number of) Increase 
hei Fa ee Average. tion. Individ. | per Cent. Average. 
Ren | | 
1 1 29 
VIII. 28 27 . 
II 2 100 23 
100 IX. 33 18 
III. 4 100 18 
100 Sg 39 18 
IV. 8 100 16 
88 XI. 44 13 
Y 14 75 
49 XII. 20 ! 
VI. 17 22 Incomp lete. 
26 || XIII 4 |) 
VII 22 30 


Hence we conclude, Zhere is a diminution in the rate of increase of in- 
dividuals in the “ fan” as it grows older. 

In searching for an explanation of this phenomenon, I first drew 
a line from the centre of the primary individual of the fan to the 
periphery, and divided it into four equal parts. I then described 
ares with the primary individual as a centre, and with radii equal to 
4, 2, 3, and 4 of this line respectively. Counting the number of in- 
dividuals cut by these ares respectively, and dividing those numbers 
by the length of .the corresponding ares, I found that there is almost 
exactly the same number of individuals per unit of are for each of the 
four ares, (Rule 7.) The previous conclusion, that there is a dim- 
inution in the rate of increase of individuals in the fan as it grows 
older, may then be considered as a corollary to this rule, as it ob- 
viously follows from it. 


46 BULLETIN OF THE 


Bugula flabellata, J. V. Thompson.— I have studied this species for 
the purpose of confirming the results obtained in B. turrita, and have 
found the architecture of the two species alike in all essentials. 

The entire colony of B. flabellata (Plate VII. Fig. 66) may be com- 
pared to a single “fan” of B. turrita, only there are usually many more 
individuals in the former, and of course there is no central stem to which 
it is attached ; but the fan is fastened directly by its rhizoids to the 
object which supports it. 

Usually about four rows of individuals are united, instead of two as 
in B. turrita, — a condition which can be easily derived from the latter 
by imagining adjacent branches to become fused together. Here as 
there adjacent individuals break joints. Here as there lateral branches 
are given off towards the axils. 

Rule 3 is not true for B. flabellata. This is entirely annulled by the 
establishment of a new rule, which depends upon the new conditions 
found in this species ; namely, that more than two rows cling together, 
and that consequently one or more rows of individuals are enclosed be- 
tween outer marginal rows. In any such twig composed of more than 
two rows (Rule 3a) lateral branches are given off only from the marginal 
rows. (See Fig. 66, 49-54 XVII.) It might possibly result, then, 
that certain of the middle rows of the twig should never give rise to 
lateral branches. But I do not believe that this ever occurs in very 
long rows, for by the splitting up of the twigs the middle rows sooner or 
later become marginal (so 46-51 XV.). In one stock that I have 
drawn, consisting of 17 to 21 generations, every middle row occurring 
as such up to the 13th generation had become at the periphery a 
marginal row. 

As in B. turrita, so in B. flabellata lateral budding occurs most 
frequently at the margins of fans, —in a fan of about 800 individuals in 
the ratio of 1:10 for the margin, and 1:14 for the remainder of the 
fan. By a comparison of these figures with those given on page 43 for 
B. turrita, it will also appear that lateral budding is less frequent here 
relatively to terminal budding than in B. turrita. 

The fifth rule deduced for B. turrita holds equally well here. In one 
case the curve of the tips of the rows rises from the margin of the fan at 


1 The species which I have studied is identified by Verrill (’78, pp. 711, 889) 
under this nathe, and my specimens also agree fairly with Hincks’s (’80, pp. 80-82) 
diagnosis. The two pairs of spines, one longer than the other, could be distinctly 
seen. Hincks says, “ The rows of cells. . . are never, I believe, fewer than four, 
and range as high as seven.” But his Figure 66 shows three rows only in some 
places. 


MUSEUM OF COMPARATIVE ZOÖLOGY. 47 


generation XXIV., reaches 3 maxima of XXVI., XXVII., and XXVIII. 
respectively, and falls again at the other margin to generation XXIII. 
In the subfan from which Figure 66 was taken, the curve begins at the 
outer margin with generation XVII., rises to generation XXII. at two 
points, and falls again to XX. at the inner margin of the subfan. 

Of the four proximal individuals in any fan here, as in Bugula turrita, 
the outermost, ancestral give rise to the greater number of individuals. 
In one case, for instance, the marginal individuals lie at the base of 31 
rows with 184 individuals, while the inner ones support only 7 branches 
with 65 individuals. Similar results were obtained from other stocks. 

With the middle of the primary individual as a centre, I passed an 
are of a circle through the extremities of the branches of a large camera 
drawing of a fan of B. flabellata, divided the radius into eighths, and 
passed arcs through these points. The number of individuals cut by the 
different arcs was then counted and tabulated; the arc with the longest 
radius cut through 87 individuals. By measuring the length of the ares, 
the number which should be cut by each arc on the assumption that the 
number of individuals per unit of arc is constant for all radii was deter- 
mined. This was then compared with the actual number found, with 
the following results: — 


ST | Me ete | TAM) Re a ne 4 N 
1 3 2} 5 87 40 
2 7 6 6 56 56 
3 13 13 7 68 70 
4 22 25 8 87 [87] 


In this instance, then, the 7th rule deduced for B. turrita evidently 
holds true for B. flabellata. 

While at Mr. Agassiz’s laboratory at Newport, during the summer of 
1890, I had frequent opportunity to examine other stocks of Bryozoa, 
which occur there very abundantly. I will take four species as typical 
examples of the groups they represent, and treat of the architecture of 
their colonies. 

Lepralia Pallasiana, Busk.'—It is not at all easy to determine 


1 I do not feel perfectly certain that the specimen shown in Figure 71 (Plate 
VIIL.) belongs to this species, because the characters of the young stocks differ some- 


48 BULLETIN OF THE 
from a young stock what has been the order of succession of individuals. 
One has to view the object from both sides, make a careful examination 
of the walls of the zooecia and of the relation of the polypides to one 
another, and, when he has done his best to determine what are the facts, 
he must feel that his conclusions are after all more or less subjective. 
By a careful study of the colony shown in Figure 71, I have constructed 
the diagram shown in Figure 71°. 

The stock of Lepralia is a creeping one, and all of its rows of individ- 
uals are in juxtaposition. This juxtaposition is continued into the adult 
stage. Even the young stock begins to show evidence of a quincunx 
arrangement of individuals. This is less evident in the youngest indi- 
viduals than in the older part of the stock, and is most evident in old 
colonies. That there is not here a true dichotomous division of rows of 
individuals, resulting in the annihilation of the ancestral row and tho 
establishment of two new ones, is evident from a glance at the youngest 
generation in rows 11, 12, or, better, 2, 3, in which the relation of ter- 
minal (11, 3) and lateral (12, 2) individuals is very different. The for- 
mer continue the ancestral line, the latter establish new rows. Lepralia 
differs from Bugula in this: that two lateral branches may be given off 
from the ancestral row in the same generation, as at B, C, and a, a 
(enclosed in circles), Figure 71". 

In contradistinction to the conditions in Bugula, when only one branch 
arises, it is not given off towards the axil, but away from it. 

The synchronism of the budding process noticed in B. turrita is hardly 
distinguishable in the adult stock of this species; in the young, however, 
it is quite marked, and gives to the whole a very symmetrical form. The 
cleavage of eggs does not proceed by more regular steps. Of the three 
individuals a, O, a (in circles), which follow B, each has given rise to 
three others, a median and two lateral. From each of the three individ- 
nals derived from the two individuals a, a (in circles) has arisen a 
lateral branch. Rule 3 is therefore well marked in the young stock of 
Lepralia. 

Rule 4, concerning the greater frequency of lateral budding at the 
margin, is also exemplified in Lepralia. The ratio of cases of lateral 
to median budding being 1:1 on the margin (rows 1-6 and 15-19) 
and 1: 2.8 in the middle (rows 7-14.) 

In Bugula, as will be recalled, it was concluded that the marginal 


what from those of older ones. Yet it is an Escharine closely allied to Lepralia, and 
Ihave seen in some cases the broad-based spine on the proximal border referred to 
by Verrill as being found in L. Pallasiana. 


MUSEUM OF COMPARATIVE ZOOLOGY. 49 


branches were possessed of fewer generations than the intermediate ones. 

Since by Rule 2 the lateral branches were given off towards the axils, | 
and the ancestral branches therefore always remained marginal, it re- | 
sulted that the ancestral branches were the shorter, the lateral branches | 
the longer. But in Lepralia lateral branches are turned away from the 

axils, and here we find the conditions concerning the relative number of 

generations in marginal and intermediate rows correspondingly reversed. 

Thus, in Figure 71°, the terminal individual D of row 10, a median but 

ancestral row, belongs to generation IV. while the lateral branches 6 

and 15 have five generations of polypides. Thus it is true here, as in 

Bugula, that the ancestral branches are the shorter, the lateral branches 

the longer (page 44), 

That the outer individuals a, a, of rows 6 and 15, have given rise to 
more individuals than the inner C, is clear without further comment. 
Finally, since the individuals retain a nearly constant width, the neces- 
sity of the rule established for Bugula, — viz. that there is almost ex- 
actly the same number of individuals per unit of, are for all radii, — and 
of its corollary, — that the increase of individuals in sucessive gener- 
ations undergoes a regular diminution, — is apparent. 

Plustrella hispida, Fabricius.’ — This stock is a very dense corm-like 
one. The primary individual becomes surrounded on all sides by the 
younger zooecia. It is very evident from an inspection of the position 
of this primary polypide with relation to the periphery, that growth 
occurs most rapidly on each side and in front of the primary polypide. 
In making any diagram of such a stock, it is not very difficult to decide 
upon the origin of the more peripheral individuals of the stock, but it is 
wellnigh impossible to say with any certainty what are the relations 
of the individuals of the second generation to those of the first. Bar- 
rois (77, pp. 227-229) has, however, determined this for this species, 
and my diagram (Fig. 67) is based in part upon his observations. I 
do not desire to insist that the diagram represents the exact method 
of growth of the stock. It is an attempt to represent it, founded princi- 
pally on careful study of Figure 69. The quincunx arrangement of in- 
dividuals is already apparent in the young stock (Fig. 69) ; it becomes 


1 Hincks (’80, pp. 504-506) makes the existence of a larval bivalve shell a char- 
acteristic of this genus, and therefore I assign to it a very common Aleyonidium-like 
form which was extremely abundant on Fucus at Newport. F. hispida is the only 
species of this genus. I found the bivalve shell still adhering to the primary indi- 
vidual of a young colony (Plate VIII. Fig. 69, 0.). In Verrill’s (78, p. 708) catalogue 
this species is referred to under the name “ Alcyonidium hispidum, Smitt.” 

VOL. XXII — NO. 1. 4 


50 BULLETIN OF THE 


more evident in the adult, and when new individuals arise distad to any 
two, one of the new ones is median (ancestral branch), the other lateral. 
(So terminal individual of rows 11 and 10; 22, 23; 43, 44; etc.) In 
the diagram, however, I have not always indicated which is the median 
and which the lateral branch, for in the older parts of the colony, owing 
to a shoving of individuals, it is not easy to distinguish them. 

Lateral branches appear usually to be given off towards the axis. 
Here, as in Bugula, the lateral branches tend to be longer; the ances- 
tral, shorter. 

Jt is evident from the diagram that lateral budding is most frequent 
at the margins of the corm, i. e. that part lying posterio-dextral or poste- 
rio-sinistral of the primary individual, and that the descendants of the 
two lateral individuals of the four belonging to generation II. are more 
numerous than those derived from the middle two. Finally, it is evi- 
dent that the number of individuals per unit of are will be the same for 
ares of all radii, and therefore the rate of increase of individuals will 
diminish through successive generations. 

In Orisia eburnea, Linn.,! we find the same laws illustrated. The 
architecture of the genus has been carefully treated of by Smitt (65°, pp. 
115-142) as forming the basis of classification. Barrois (77, pp. 76-85) 
has described in a masterly way the formation of the young stock of 
tubuliporid Cyclostomata, and the relationships of the different types 
of budding in this group. Harmer (91, pp. 145-173) has recently dis- 
cussed the architecture of the stock in British species, adopting Smitt’s 
graphic method of showing it. I have found his paper of great value 
for my purpose. 

This species grows as a shrub-like stock upon floating eel-grass, etc. 
I was wrong in saying, in my Preliminary (’91, p. 282), that Crisia has 
its branches united in pairs. The comparison of this species made by 
Barrois (77, p. 82) with the “geniculata form ” is conclusive evidence, 
to my mind, that the apparent double row is in reality a single one, and 
that such a branch as 18, Figure 65, is to be represented by a single 
line in the diagram Figure 65%. We find here terminal and lateral 
branches ; no true dichotomy. Branches are given off on the side away 
from the axils, as in Lepralia, not as in Bugula. (But branch 11 is an 
exception to the rule.) They are given off, as Harmer C91, p. 131) 
has shown, alternately to the right and left. 

1 This is the only species of Crisia given by Verrill, and, since my species 
is very common, it must be the one to which he refers. Moreover, it agrees 
fairly well with Harmer’s diagnosis (’91, p. 131). 


MUSEUM OF COMPARATIVE ZOOLOGY, 51 


There is something of a tendency for lateral branches to be given off 
in the same generation from closely related branches. Thus (Fig. 65°) 
from the primary individual, o, of the stock, two individuals, a median 
and a lateral one, arise. Each gives rise in its first generation to two 
individuals, a median and a lateral. Of these four individuals each 
gives rise at the end of three generations of median buds to two buds, 
a median and a lateral. Comparing 2 and 10, first descendants of the 
two branches arising from the second individual of 8, we find that each 
gives rise to lateral branches from their first individual and from their 
fourth. Comparing 14 and 19, first descendants of the two branches 
arising from the first individual of 8, we find each giving rise to lateral 
branches from their first individuals, The law breaks down, however, 
when an attempt is made to carry it to extremes. 

The fourth rule is not always so pronounced in Crisia eburnea as 
elsewhere, although lateral budding seems to be slightly more frequent 
at the margin. 

The extreme marginal branches usually attain far fewer generations 
than the more intermediate ones; thus, in Figure 65%, branch 20 ends 
in the 7th generation and branch 13 in the 7th also, while the more 
intermediate branches 15 and 18 attain 12 and 14 generations respect- 
ively. So, too, while the outer branches 6 and 1 contain respectively 
10 and 11 generations, the inner branches reach 12 and 14. 

It is very noticeable that the outer branches give rise to more indi- 
viduals than the intermediate ones. Figure 65* will serve to illustrate 
this also. Here the outer branch 4, the intermediate 8, and the outer 
15 possess, together with the branches arising from them, 33, 28, and 40 
individuals respectively. Harmer (’91, p. 168) finds this true for his 
Crisia ramosa, for he says, “ It is frequently remarked that the longest 
and most branched parts of the colony are lateral branches, and not 
parts of the main stems.” 

There is, in the long run, a decrement in the rate of increase of indi- 
viduals in successively older generations, yet it is not so regular a one 
as that which we found to exist in Bugula. Thus, in the seven gen- 
erations which even the shortest branches shown in Figure 65* had 
attained, the average increase of the number of individuals in the 
second, third, and fourth generations over the number in the preceding 
is 67% ; in the fifth, sixth, and seventh, 44%. The generations beyond 
the seventh are not complete; they would have contained more indi- 
viduals at a later period, when the branches which have now attained 
only seven generations had grown. Thus the number of individuals 


| 
i 
| 
| 
| 
| 
| 


52 BULLETIN OF THE 


in successive generations ,beyond the seventh increases more and more 
slowly, and finally decreases to zero. Thus the average rate of increase 
of individuals in the generations 7 to 10 over those in the preceding is 
only 16%. 

One finds here, as elsewhere, that the number of individuals cut by 
any unit of arc, the primary individual being taken as a centre, remains 
practically constant, whatever the radius of the arc. 

In studying the creeping stocks of Cheilostomes (Plate VITI. Fig. 71), 
young corms have been chosen because they exhibit fewer irregularities 
of formation than old ones. Such irregularities are chiefly due to some 
unevenness of the surface on which the corms lie, but sometimes 
apparently to a crowding of individuals. Old rows of individuals are 
occasionally entirely cut off and end in the middle of the stock; some- 
times two rows running side by side, perhaps derived from a common 
ancestor, suddenly merge into one again. In one case, Escharella varia- 
bilis, Verrill, I have seen three rows thus merge into one at the margin, 
suggesting the existence of a samknopp (common bud) in the sense of 
Smitt (65, pp. 5-16). Ostroumoff (86%, pp. 338, 339) has observed 
a case in Lepralia Pallasiana. He says: “Dans quelques cas, qu’on 
peut considérer comme des anomalies, il arrive parfois que deux bour- 
geons, provenant de loges différentes, viennent a se fusionner.” It seems 
to me, therefore, that while Nitsche (’71, pp. 445, 446), who opposed 
with such vehemence and success the idea of Smitt that zocecia arise from 
an undivided marginal zone of cells, was quite right in affirming (71, 
p. 447) that even the smallest marginal zowcia are sharply marked off 
from the adjacent ones, yet he overlooked the possibility that under 
certain circumstances the lateral walls might fail to develop, and thus 
one zoocium might arise in the place of two, or even three. 

I have not read Smitt’s Swedish paper, but I do not find anything 
in the translation given by Nitsche to warrant the latter’s conclusion 
(71, p. 446) that Smitt believed the “ Gesammtknospe ” to be “ formed 
from the sum total of the mature peripheral zoccia.” If I understand 
Smitt, he conceived the samknopp not to be derived from the most 
peripheral mature zoœcia, but to be self-proliferating, and to give rise 
to the rows of zomeia, not to arise from them. It is the “bud of 
the colony,” not the sum of the buds of the peripheral individuals of the 
stock. In this I would agree with him exactly. Although usually one 
finds the marginal gemmiparous tissue forming the lateral walls at the 
extreme edge of the corm, and thus apparently separated into wholly 
distinct adjacent gemmiparous masses; under certain conditions, the 


MUSEUM OF COMPARATIVE ZOOLOGY. 58 


lateral wall may not be formed between two or more rows, which will 
then merge into one. 


2, ORIGIN AND DEVELOPMENT OF THE INDIVIDUAL. 


My studies on this subject, which were undertaken for the purpose of 
showing the unity of the type of budding throughout Ectoprocta, have 
been very fragmentary. 

Figure 72 (Plate IX.) has been introduced for the sake of orientation. 
It represents a longitudinal vertical section through the peripheral part 
of a stock of Lepralia Pallasiana. The body wall is thicker at the mar- 
gin (marg.), and gradually becomes thinner as one passes backward, A 
septum (sep.) has already arisen cutting off the youngest zoocium from 
the more proximal one, which contains a young polypide; proximal to 
this is another septum, and the distal end of a third zooecium. 

Nitsche (71, pp. 445-456) has already well described the process of 
forming the zowciwm in Flustra membranacea. In fact, he has studied 
the organogeny more thoroughly in many respects than I have. Nitsche 
CTL, p. 452) showed that the wall of the advancing margin of the colony 
was composed of two layers of cells, — an outer, “ Cylinderepithelschicht,” 
which secretes a cuticula, and an inner, “ Spindelzellschicht mit anliegen- 
den Körnerhaufen.” As the body wall formed directly from these cell 
layers is left behind by the advance of the margin, it becomes continually 
thinner. “Die Cylinderepithelzellen der Wandung platten sich weiter 
nach dem proximalen Ende zu ein wenig ab, besonders die der Unterseite 
verkürzen sich, die einzelnen Zellen rücken auseinander, die Zellgren- 
zen werden undeutlicher, die Kerne jedoch bleiben deutlich erkennbar.” 
Vigelius (84, p. 76) could not find the inner cell layer in Flustra, even 
at the youngest stages, and consequently he believed that only one ex- 
isted at the margin, and that this went to form the “ Parenchym- 
gewebe” of the adult. Ostroumoff (86°, p. 336) seems inclined to doubt 
the existence of any mesodermal layer at the distal portion of the bud- 
ding zoœcium in Cheilostomes, and Seeliger (90, p. 580) has failed to 
find in Bugula “eine zusammenhängende dem Ectoderm dicht anlie- 
gende Schicht von mesodermalen Spindelzellen.” Both Ostroumoff and 
Seeliger, however, believe in the existence of isolated mesodermal ele- 


meats at the budding end. 

According to my own observations, there is usually only one continu- 
ous layer at the budding margin of the stock. Thus, in Flustrella 
(Plate IX. Fig. 79) one can usually distinguish a continuous ectoderm, 
but the mesoderm (ms’drm.) is represented by scattered cells only. At 


54 BULLETIN OF THE 


the margin in Lepralia (Fig. 73) one finds a thick ectodermal layer, 
composed of columnar cells, but the mesoderm consists of an irregular 
thick mass of cells, some of which appear to be amoeboid. They how- 
ever show no signs of having been derived from the outer layer. The 
condition of the budding margin of Escharella resembles that of Lepralia. 
In older parts of the body wall, where the ectoderm is reduced to an ex- 
tremely thin layer, only scattered mesodermal cells appear, and these are 
amceboid or mesenchymatoid. 

On the other hand, one finds in the body wall, around the nascent 
neck of the polypide (Plate X. Fig. 88), even to a late stage, both ecto- 
derm and mesoderm well formed as layers. The ectoderm is a columnar 
epithelium ; the mesoderm is flatter, and often its cells are not sharply 
delimited from one another. It is thus perfectly evident, to my mind, 
that the mesoderm has in general lost its original epithelial character 
in the marine Bryozoa, although it has retained it in Phylactolamata. 
Whenever it does exist in the former group as an epithelium, it is at the 
budding regions (neck of polypide, and Figures 74, 75, 78, 79, ex.). 

Origin of the Polypide, — There are very few problems in modern 
morphology, I fancy, the history of whose investigation shows a less 
satisfactory aspect than that of the origin of the polypide in Gymnola- 
mata, It is hardly to be wondered, however, that investigators have 
sought for another interpretation of the process than the most obvious 
one, because that seemed to oppose many long cherished and wellnigh 
universally held dogmas. While the first recognition of the animal 
nature of marine Bryozoa, which we owe to the studies of Bernard de 
Jussieu in 1742 and John Ellis in 1755, brought with it a knowledge of 
their colonial nature, yet it was not until much later that the most 
characteristic part of this process — the formation of the polypide — 
was clearly observed. Grant (’27, p. 115) and Farre (’37, pp. 400, 409, 
415) first described the process by which is formed this complex of or- 
gans, and settled once for all the controversy which had sprung up as to 
whether these animals were truly stock-builders. Under the influence 
on the one hand of the endosare theory of Joliet (’77), and on the other 
hand of the view promulgated by Hatschek (’77), that similar organs 
in larva and polypide are equivalent as far as regards their origin from 
the germ layers, the more important papers ! between ’77 and ’90 main- 
tained either that the polypide arose independently of the body wall, 

1 Excepting those of Barrois, who, from the study of the favorable material 


presented by metamorphosing larv®, has persistently maintained the correct 
interpretation, 


MUSEUM OF COMPARATIVE ZOOLOGY, 59 


and secondarily acquired connection with it, or that it had a double 
origin. 

To Nitsche (’71, pp. 456-463) belongs the credit of having first described 
the histological changes in the origin and development of the polypide 
of marine Bryozoa, particularly with reference to the part which the 
germ layers play in that process. He says (71, p. 456): “Die Anlage 
des Polypids erscheint zunächst als eine Wucherung der Zellschicht der 
Endocyste in der Mitte der Hinterwand der Knospe, und zwar in dem 
Winkel, den die Hinterwand mit der oberen Wand macht. Bald ordnen 
sich die Bestandtheile des regelloşen Zellhaufens in zwei deutlich geson- 
derte Schichten, und wir sehen nun einen rundlichen Körper, beste- 
hend aus einer äusseren einschichtigen Zellschicht, welche sich scharf 
absetzt gegen die das Innere des Körpers bildenden Zellen.” 

This stood until a year ago as the most satisfactory description of this 
process in the adult stock. The appearance within the last year of the 
two papers of Prouho (°90) and Seeliger (90) marks a distinct epoch in 
the advance of our knowledge concerning the origin of the polypide in 
Gymnolsemata. The paper of Prouho treats of the process in the case of 
the primary polypide of the metamorphosing larva of Flustrella, that of 
Seeliger in the case of the young (practically adult) stock of Bugula. 
According to both authors, the polypide arises from the body wall by an 
invagination of it, and its two layers are from the first distinct and 
separate, and go to form the two layers of the adult polypide, and the 
whole of those two layers. The outer layer of the body wall gives rise 
to the outer layer of the tentacles and the lining of the alimentary tract, 
and the inner layer of the body wall gives rise to the mesodermal lining 
of the polypide. Prouho alone is cognizant. of the method of origin of 
the ganglion, and in addition there are several points of difference be- 
tween these two authors concerning the development of other organs, to 
which I shall refer in the proper place. Thus the latest studies have 
confirmed the assertions of Nitsche, that the polypide arises from a 
single centre of proliferation of the body wall; they have made an ad- 
vance in this, that they have shown that the two layers of the bud do 
not become secondarily differentiated from a single cell mass, but are 
respectively derived from the two cell layers of the body wall. My own 
studies have led me to the same conclusion on this point. 

Figure 75 (Plate IX.) is a vertical radial section through the margin 
of an adult Flustrella stock. The ectoderm is relatively thick at the sole 


(sol.) and margin, and very greatly thickened at the point marked qm. 
Here two layers, sharply separated, are apparent. The cells of the outer 


56 BULLETIN OF THE 
layer are columnar and full of granular protoplasm, the mesodermal cells 
cuboid. The body wall has clearly begun to invaginate in this region. 
Figure 79 is a similar section, and shows a later stage in this process. 
The lumen of the bud is apparent, and has been formed by invagination, 
not, as in Paludicella or Phylactolemata, by ingression. The two layers 
of the bud are apparent; they have been derived from those of the 
body wall. 

Figure 73 (Plate IX.) shows a stage in the development of the poly- 
pide which is intermediate between that of Figures 75 and 79, but from 
another suborder, Cheilostomata. ‘The mesoderm has here a mesenchym- 
atous character, and is loosely attached to the inner layer of the bud ; it 
is not always sharply marked off from it by boundaries, but is quite dis- 
tinct in its reaction with staining reagents. This bud has evidently 
arisen by invagination of the body wall. Seeliger (90, p. 581) also finds 
that there is an actual invagination of the ectoderm in Bugula, the open- 
ing to which he calls “ blastopore.” 

From what has been already shown, it is evident that in Flustrella, as 
well as in Cheilostomata, the first appearance of the young polypide is 
near the margin of the stock, not near the proximal part of the young 
zoœcium. This will also be apparent at 6 and 9, Figure 71 (Plate VIIL), 
where the accumulation of nuclei immediately behind the margin indi- 
cates the neck of the polypide, —the point at which the bud arose. To 
be sure, at quite an early stage, but very much later than that of Figure 
73, the polypides are found near the proximal wall of the zoweium, but 
a delicate funnel-shaped sheath of tissue runs from the polypide to the 
distal part of the zocecium, where the polypide is attached to the body 
wall. 

After invagination the pocket closes at its attached end by a growing 
together of its lips (Figs. 79, 78). Thus the body wall becomes contin- 
uous again over the lumen of the bud, and this union is first broken 
when the fully formed polypide is ready to evaginate itself. Seeliger 
(90, p. 582, Taf. XXVI. Figs. 8, 10) has described and figured a similar 
condition in Bugula. 

The young bud now becomes elongated (Fig. 80), the walls of the bud 
sometimes becoming closely approximated. A little later it begins to 
pass backwards relatively to the distal wall of the zooecium. A trans- 
verse section through the young polypide and the neck of the colony 
shows that the connection has become a less intimate one (Fig. 81, cev. 
pyd.). The tissue by which the connection is still effected is that from 
which the kamptoderm will be formed, It is apparently the existence 


MUSEUM OF COMPARATIVE ZOÖLOGY. 57 


of this stage, in which the kamptoderm is long drawn out and easily 
overlooked in optical as well as actual sections, that led to the belief that 
polypide buds may arise independently of the body wall and only sec- 
ondarily become connected with it. 

At about this time the lumen of the alimentary tract begins to be 
separated from that of the atrium, Thus, in the series from which Fig- 
ure 81 was taken the more oralward lying sections show that the cavities 
of the lower and the upper parts of the bud, which at the anal end are 
broadly confluent, have here become separated by a constriction. A 
sagittal section of a somewhat later stage is shown in Figure 76, which 
is from Flustrella. Here we find the alimentary tract represented by a 
space in the lower part of the bud, broader at its anal than at its oral 
end and separated from the upper cavity — the common atrio-pharyngeal 
cavity, œ. + atr,— by a line of nuclei which represents the line of ap- 
proximation of the inner layers of the two sides of the bud. The bud is 
attached to the body wall at its marginal (anal) end, and is free from it 
oralwards. (Compare with Paludicella, Plate III. Fig. 24.) It seems 
to me highly probable from these and other series of sections that the 
alimentary tract is separated from the rest of the lumen of the bud, not 
by an approximation of the inner layers of the bud along the whole ex- 
tent of the future alimentary tract at once, but that the rectal part is 
first formed and constitutes a large cavity, at first broadly open to the 
atrium above, and that the gastric portion is formed somewhat later by 
a progressive enlargement of the lower cavity of the bud, which now 
becomes constricted off from the atrium and cesophagus above. This 
process is like that found in Paludicella (page 19), which forms a sort 
of transition to that of Phylactolemata, described by Braem (’90, pp. 
45, 46) and myself (90, p. 112). 

Prouho (90, p. 448, Fig. 6) shows that the rectum at first appears 
as a blind sac open to the atrium at its posterior end, although later 
this opening is greatly reduced. Hence in the Flustrella larva also the 
space from which the lumen of the future rectum is to arise is formed 
before that of the stomach, although this part of the alimentary tract 
is the last to be cut off from the atrium, Seeliger (’90, p. 585) says 
concerning the formation of the alimentary tract in Bugula : “ Der ganze 
Basaltheil des Polypids sich in der Mittelpartie durch zwei immer tiefer 
werdende Furchen von dem vorderen abschnürt, während er an zwei 
Stellen, einer oberen und einer unteren, mit ihm in Verbindung bleibt. 
Die obere Verbindung entspricht dem Anus, die untere dem Mund.” The 
author here seems to imply that the whole alimentary tract is formed at 


58 BULLETIN OF THE 


one time; but as he has not attended particularly to this point, this can 
hardly be said to militate against my view. 

There is, however, in my opinion, a more important error in Seeliger’s 
description of the origin of the alimentary tract, — an error into which 
Nitsche (’71, p. 457) also fell. As in Phylactolemata and Paludicella, 
so also in marine Bryozoa in general, so far as I have studied them, the 
posterior and anterior parts of the alimentary tract are formed indepen- 
dently, and their cavities coalesce only secondarily. The constriction 
which separates the lumen of the bud into a cavity nearer, “ vorder,” and 
one more remote from the body wall, “ basal,” does not separate off the 
whole alimentary tract from the atrium. Neither does that constriction 
result in the formation of a space opening into the cavity nearer the 
body wall, “ Vordertheil,” at an upper [distal] point (anus) and lower 
[proximal] point (mouth). Thus if one examines a complete series of 
sections through a polypide even of so late a stage as Figure 92 
(Plate X.), one finds that, while there is an open connection between the 
anal end of the alimentary tract and the atrium, the oral end is at all 
points sharply separated from the cavity above by a double-layered wall 
of cells, as is shown in Figure 92, between œ. and ga. Such a condition, 
moreover, has been found by Barrois (’86, pp. 73-76) in the primary 
polypide of Lepralia, and by Prouho, as just stated, in the primary 
polypide of Flustrella. 

Origin and Development of the Ring Canal and Tentacles. — Nitsche 
(71, p. 430) first described in Flustra a ring canal surrounding the 
mouth-opening and lying at the base of the tentacles, but did not refer 
to the origin of it. Seeliger (90, p. 588) describes it in a young pol- 
ypide of Bugula, as derived from the mesodermal layer. 

My own sections also show that it arises on each side of the cesopha- 
gus as a groove lined by mesoderm (Plate X. Fig. 92, right). This 
canal, which is shown cut along its course in Plate IX. Fig. 82, can. 
cre., is not wholly separated from the body cavity, but communicates 
with it below the brain. This communication occurs in the section 
below that shown in Figure 82, near the point can. erc. This ring 
canal at an earlier stage is shown in Figure 87. It has not yet been 
formed backwards nearly so far as the brain; anteriorly the section has 
traversed the tentacles under which it runs. The canal is also shown 
cut across in Figure 86 at the base of a tentacle, with whose lumen its 
cavity is directly continuous. 

The formation of the tentacles is closely connected with that of the 
ring canal, from the upper wall of which they arise. Since the upper 


MUSEUM OF COMPARATIVE ZOOLOGY. 59 


wall of the ring canal is two-layered, the tentacles are two-layered also. 
The outer layer of the tentacle is thus derived from the inner layer of 
the bud; the inner layer, on the contrary, from the outer layer of the 
bud. It would be hardly necessary to make this statement, which 
agrees both with early and the most recent observations, had not Bar- 
rois (’86, p. 75, Fig. 48) referred to and figured the tentacles as having 
been formed from the inner layer of the bud only. 

My observations fully confirm Seeliger’s (90, p. 587) description of 
the manner of growth of the tentacles ; that is, that the outer edge of 
the ring canal, together with its tentacles, moves downward and outward 
along the sides of the polypide, turning the axis of the tentacle from a 
nearly horizontal to a vertical position, and increasing the area of the 
kamptoderm. Thus in Figure 92 this process has progressed farther on 
the left side than it has on the right. 

Nitsche (71, p. 458) lays some stress upon the statement that the ten- 
tacles are not at first few in number, gradually becoming more numer- 
ous ; on the contrary, he says, “Ich sah stets, beim ersten Auftreten von 
Tentakelanlagen, 16, 17, oder 18 Stück gleichtzeitig erscheinen.” See- 
liger (’90, p. 584) agrees with Nitsche in this respect ; but Prouho (’90, 
p. 449) finds the conditions different in Flustrella. Here the tentacles 
“ne se développent pas simultanément sur tout son pourtour, mais ap- 
paraissent d’abord de chaque côté du plan de symétrie, puis se multi- 
plient vers V’arriére.” As I have shown, 14 of the 17 tentacles arise 
nearly simultaneously in Paludicella, for here there are few of them ; 
and this is the case also in Escharella variabilis with its 17 tentacles. 

As the tentacles of both Flustra and Bugula are few in number,’ the 
statements may easily be considered to be correct for these genera. The 
tentacles of Flustrella hispida are much more numerous (30-35), and 
Prouho’s statement may well be true for bis form. In fact, my own ob- 
servations on this species are fully in accord with those of Prouho. 
Figure 77 (Plate IX.) represents a young polypide of a Flustrella corm, 
viewed from the roof as an opaque object. Six tentacles were visible on 
each side of the bud, but the oral and anal parts of the corona were yet 
incomplete. The remaining nine or ten pairs of tentacles subsequently 
arise oralward and analward of these rudiments. 

Much disagreement has prevailed concerning the number of layers in- 
volved in the kamptoderm of marine Gymnolemata, in both the adult 
and the developmental stages. As in so many other cases, we owe to 

1 Bugula avicularia has 14 or 15 tentacles, and Flustra (Membranipora) mem- 
branacea 20, according to Hincks (’80, pp. 76 and 140). 


60 BULLETIN OF THE 


Nitsche (’71, pp. 431, 432) our first intimate knowledge of this organ. 
He believed it to consist in the adult of Flustra of a single cell layer, 
in which are imbedded (or applied?) longitudinal and circular muscle 
fibres. He believed the kamptoderm to be formed gemmigenetically only 
by the outer cell layer, the derivative of the mesoderm. Repiachoff (’75, 
pp. 138, 139) observed in Tendra (Membranipora) “die Doppelschichtig- 
keit der Tentakelscheide nicht nur bei den jungen Knospe sondern auch 
bei den ganz ausgewachsenen, offenbar schon längst functionirenden, in 
ihrem mittleren Theile ganz braunen ‘ Polypiden,’” and later (’76, p. 152) 
a similar two-layered condition of the kamptoderm (Tentakelscheide) 
in Membranipora and Lepralia. Ehlers (76, p. 37) finds a single layer of 
cells in the kamptoderm of the adult Hypophorella (Ctenostome), which 
he believes is continuous with the endocyst of the body wall, and thus 
is ectodermal, He finds neither longitudinal nor circular muscle fibres, 
Haddon (’83, p. 517) believes the kamptoderm to be derived from both 
the inner and outer layer of the polypide bud. Vigelius (’84, pp. 33, 82) 
deScribes it as arising from the mesoderm only (Parenchymegewebe), and 
as being essentially one-layered, both longitudinal and circular muscles 
lying in this layer. Barrois (86, p. 74) derives the kamptoderm from 
the mesodermal layer only. Ostroumoff (’86*, p. 15) believes the kamp- 
toderm to be two-layered and provided with muscles ; it is in his opinion 
derived from both layers of the bud. Freese (’88, pp. 18, 19) studied 
only the adult of Membranipora. He admits the presence of muscle 
fibres, but believes the kamptoderm one-layered. Pergens (’89, p. 507) 
states only that in the Cheilostomes studied by him the tissue of the 
kamptoderm is composed “aus abgeplatteten Zellen, zwischen welchen 
Längs- und Ringmuskelfasern eingebettet sind.” Prouho (’90, p. 451) 
states that in the primary polypide of Flustrella this organ is early 
differentiated, “ et les deux couches de rudiment prennent part à sa for- 
mation.” Finally, Seeliger (90, p. 587): “Es kann danach keinem 
Zweifel unterliegen, dass die Tentakelscheide ektodermalen Ursprungs 
ist... . . Das Mesoderm erscheint auf allen gelungenen Schnitten von 
der Tentakelscheide scharf abgesetzt.” 

It is my belief that throughout the group of marine Gymnolemata, as 
in Paludicella and Phylactolsemata, the kamptoderm is derived from both 
of the two layers of the polypide bud, is provided with a strong system 
of longitudinal and a slight one of circular muscles, and contains in the 
adult two layers, or at Teast modified representatives of two layers. I 
have arrived at this conclusion from a careful study by sections of the 
following genera : Bugula, Lepralia, Escharella, Flustrella, Bowerbankia, 


MUSEUM OF COMPARATIVE ZOOLOGY. 61 


and Crisia. The existence of two layers was easily demonstrated in all 
cases in the young polypide by cross sections of the “neck.” The two 
layers are of nearly the same thickness, and distinctly separated from 
each other. The presence of two layers in the adult is more difficult to 
determine, but it was always indicated by the occasional presence of 
two nuclei lying side by side, and especially at the attachment to the 
diaphragm. The presence of muscles was demonstrated in all cases 
(except Bowerbankia, where my few sections did not show the proper 
region) upon tangential sections of the sheath, I may add, that the 
existence of muscles is wellnigh conclusive @ priort evidence of the 
existence of the mesodermal layer, since nowhere else in Bryozoa, so far 
as I know, do muscles arise from any other layer. Prouho’s evidence in 
support of his position is perfectly satisfactory to my mind, certainly 
more so than the negative evidence of Seeliger in support of able hi 
further support of my statements I may refer to the condition of the 
kamptoderm (kmp’drm.) in Figures 92 and 83, Plate X. 

Nervous System. — Since Dumortier discovered, in 1835, a ganglion in 
Lophopus, there has been seen in marine as well as fresh water Bryozoa 
a body which has been considered, with greater or less certainty, to con- 
stitute the central nervous system. Overlooked by Farre, it was, I be- 
lieve, first described for marine Gymnolemata in 1845 by van Beneden, 
co-worker with Dumortier, for Laguneula (Farrella). Nevertheless, up 
to the present the evidence of its being a ganglion homologous with that 
of Phylactolamata has not been satisfactory. The homology can be estab- 
lished only by determining its similar origin with the brain of Phylacto- 
lemata ; its function can be best established by showing the existence of 
ganglionic cells and fibres. I hope to have advanced our knowledge in 
both of these directions. 

At about the time that the œsophagus and stomach have become con- 
fluent, one notices a papilla-like elevation of the floor of the atrio-pha- 
ryngeal cavity. This has been noticed by Korotneff (74) in Paludicella, 
and by Nitsche (71, p. 459) and Seeliger (’90, p. 586) in Cheilostomes, 
It has been called by them “ Epistome,” and compared with that of Endo- 
procta or Phylactolemata. In my own opinion, it is merely a structure 
brought into prominence by the sinking down of the floor behind it to 
form the ganglion (Plate X. Fig. 86, gn.). This depression has been 
scen by Barrois (’86, pp. 74, 75) and Prouho C90, p. 450), and rightly 
interpreted by them as probably destined to give rise to the central 
nervous system. That this is the correct interpretation is shown by 
later stages from different species, as Figures 89 and 83, in which we see 


62 BULLETIN OF THE 


the ganglion gradually assuming the position it has in the adult, on the 
anal side of the pharynx at the base of the anal tentacles. 

A section across the pharynx in such a stage as Figure 83 is shown 
in Figure 87. A comparison with Figure ‘51 (Plate V.) of my Crista- 
tella paper (Davenport, ’90) will show a great similarity of conditions at 
about the same age, and can leave no doubt concerning the homology of 
the regions marked in both cases lu. gm. ; or compare Taf. VIII. Fig. 100, 
nh., of Braem’s (90) magnificent work. A section through a later stage 
is shown in Figure 82. The brain has already sent out circumoesopha- 
geal nerves, as in Paludicella. The central part of the ganglion does 
not stain; one sees only a granular mass, sometimes with signs of short 
fibres. In the cornua (n’) one occasionally sees very large clear nuclei 
with a single nucleolus, lying in the midst of a cell mass which is 
spindle-shaped and stains more deeply than adjacent cells. These remind 
one strongly of bipolar ganglionic cells, but fibres could not be traced 
far from their pointed ends. Series of sections of Flustrella parallel to 
Figure 82 show, as one passes below the level of the ganglion, a con- 
tinuous band of cells extending down from it towards the cardiac valve 
and between the cell layer lining the esophagus and the surrounding 
mesoderm. One is reminded of the exactly similar conditions in Palu- 
dicella (page 26), and of the “linienartige Zeichnung” seen by Nitsche 
(71, p. 431) and Vigelius (84, p. 42) in the same place in Flustra. 
These facts go to indicate the existence of a gastric nerve. 

At about the time at which the ganglion arises, the cavities of the 
stomach and the wsophagus become confluent (Fig. 86 @.). At this 
stage (somewhat earlier than Figure 86) the alimentary tract consists 
of a U-shaped tube of nearly uniform calibre, and without any indica- 
tion of the cecum. The tentacles lie in two parallel rows in the middle 
of the bud, the corona being incomplete both in front and behind, but 
less so oralwards than towards the anus (Fig. 77, atr.). In fact, while 
new tentacles are formed later towards the oral median line, they never 
appear behind the line atr. This hinder region has another fate. Its 
wall increases very greatly in area, diminishes correspondingly in thick- 
ness, and forms a large part of the kamptoderm lying behind the post- 
oral tentacle in Figure 86. With this growth of the kamptoderm the 
anus is carried backwards, and farther and farther from the posterior 
ends of the rows of tentacles, immediately behind which it formerly lay. 

As the kamptoderm grows in area, the polypide comes to lie in the 
proximal part of the zowcium, Pari passu with this process occurs the 
rotation of the oral tentacles, as in Paludicella. The oral tentacles which 


MUSEUM OF COMPARATIVE ZOOLOGY. 63 


at first lie perpendicular to the roof of the colony (Fig. 86) gradually 
come to lie parallel with it (Figs. 89 and 83). The wsophagus loses its 
elongated, laterally compressed form, and becomes circular, and the gan- 
glion lies just below the mouth-opening. Not until now, in fact, can one 
speak ofa mouth. It was not at all formed synchronously with the anus. 
To illustrate this process I have taken three different genera represent- 
ing different stages. Similar stages could have been obtained from each 
genus. By using three genera, the similarities as well as the dissimi- 
larities of the process are indicated. Among other things, the larger 
size of the polypide and shorter kamptoderm of the Ctenostome Flus- 
trella (Fig. 89) is noticeable. 

Lastly, the coecum is formed as a wholly secondary differentiation of 
the alimentary tract. This arises in some species relatively earlier than 
in others; thus it is better developed in Figure 86 than in the later 
stage of Figure 83. 

The lining cells of the alimentary tract now rapidly undergo the dif- 
ferentiations characteristic of the different regions. The most extreme 
modification takes place in the pharynw. In Cheilostomes the cells of 
this region gradually become vacuolated, until finally very little stain- 
able protoplasma remains, The nucleus lies at the deep end of the cells. 
A very peculiar modification of the cell walls takes place, in that they 
become plainly perforated by holes through which the adjacent cells 
are in communication (Fig. 85). It is in a region similar to this that 
the cells become cuticularized in Bowerbankia to form the so-called 
gizzard. The pharyngeo-wsophageal region is also provided with a very 
powerful musculature of circular muscles (mu, Figs. 85, 86). 

Concerning the origin of the muscles 1 have made very few studies. 
The parieto-vaginal muscles seem to arise, as in Paludicella, from 
around the neck of the polypide, and the retractors from the oral end 
of the polypide bud (mu. ret., Fig. 89). 

The neck of the polypide sinks below the general level of the body 
wall by an infolding of the latter, as described for Paludicella, and the 
mass of columnar cells which passes down with it forms, I am confident, 
the diaphragma of Nitsche (71, p. 432), which is thus exactly com- 
parable with the mass of cells around the atrial opening of Paludicella 
in Figure 45, of. atr. (Plate V.). According to this view, then, the dia- 
phragma is not placed at about the middle of the kamptoderm, but at 
its proximal end, and all that lies between it and the outer body wall — 
the non-evaginable portion — has been formed in the elongated neck, 
exactly as the non-evaginable portion is formed in Phylactolemata (see 


64 BULLETIN OF THE 


Davenport, ’90, Plate IX. Fig. 77, Plate XI. Fig. 98) and Paludicella 
(Plate V. Figs. 50 and 45). 

As my purpose is not so much to present a complete organogeny of 
3ryozoa as to show the method of origin of the bud and the fate of the 
layers, I have had to desist from carrying on my studies further in the 
organography, and have left many interesting and important questions 
unsolved ; such, for instance, as the development and structure of avi- 
cularia, the presence of an excretory system, and the degenerative pro- 
cesses which occur with regularity in the polypides. 


3, REGENERATION OF THE POLYPIDE. 


I have been led to study the regeneration of the polypide because 
Ostroumoff seems to believe that in regenerating buds the digestive epi- 
thelium of the stomach is derived from an extraneous source, — the 
brown body. Thus he says (’86", p. 340) the brown body appears as a 
ceocal appendage of the young digestive tube. “est sur ce dernier 
[tube digestif] qu’on trouve un groupe de cellules affectant la forme 
d’un bonnet et se réunissant très tôt à Vangle proximal du rudiment 
ectodermique. A mesure que les cellules du bonnet, ainsi que la masse 
brune, sont employées & la formation de la portion moyenne du tube 
digestif, ces dernières se débarrassent de leur contenu,” etc. 

The external phenomena of regeneration are well known. In the 
Membranipora stock, for instance, one sees polypides being produced at 
the margin, and one finds them older and older as one passes backwards, 
until finally they are seen to be wholly degenerate, and to be replaced by 
young polypides. Thus, in passing backward along a single row of indi- 
viduals in a Membranipora stock about 18 mm. long, I have seen this 
process of regeneration recurring four times. In Aleyonidiun, too, one 
finds an apparently regularly recurring degeneration and regeneration 
of polypides. In the mat-like Cheilostomata the regenerating polypide 
(Plate VIII. Fig. 71, pyd. rgn.) is always found at one place, — namely, 
on the operculum, — that is, proximal of the opercular opening.t In 
Flustrella it is found in a similar position on the dorsal body wall, proxi- 
mal of the cuticularized introverted portion. My studies have been 
chiefly made on the Cheilostomata. Figure 91 (Plate X.) represents an 
early stage in the formation of a regenerating polypide. Here, as in the 
marginal polypides, there is a typical invagination involving the two 

1 Haddon (’83, pp. 522, 523) has found the regenerating polypide arising from 
the same place in Flustra membranacea and in Eucratea, and Ostroumoff (786s, 
p 389) in Cheilostomes in general, 


MUSEUM OF COMPARATIVE ZOOLOGY. 65 


layers of the body wall (i., ex.). Owing to the reagent, the body wall is 
shrunken from its contact with the operculum (op.). 

If one inquires what has been the histological conditions of this region 
antecedent to this stage, one must look to younger adjacent and mar- 
ginal zowcia, since they reproduce these conditions. I will again call 
attention to Figure 88, which represents a cross section of the body wall 
through the region of attachment of the kamptoderm of a young pol- 
ypide of about the stage of Figure 83. This, then, represents the neck 
of the polypide, and it is from about this region that the operculum and 
finally the regenerating polypides will arise. The cells are columnar, 
and stain deeply about the nuclei, and both cell layers are well devel- 
oped. Elsewhere in this same individual the body wall is composed of 
smaller, flatter cells, and two layers are not easily distinguished. The 
region of the future operculum possesses at an early stage some of the 
largest, most columnar cells of the body wall. The cells of this region 
do not, however, retain their peculiarly large size throughout life, but 
in the adult we find the same region occupied by a flat epithelium, 
nearly as thin as the epithelium shown in Figure 90. Meanwhile 
the epithelium of the rest of the body wall has become still more 
attenuated. The difference between the body wail of the operculum 
and that of adjacent regions is best shown by the greater abun- 
dance of nuclei under the opercular region when the stained stock is 
looked at in toto from the roof (Plate VIII. Fig. 71). The regions of 
the future opercula are seen, in young zocecia (Fig. 71, 4, 6), to be 
patches of densely packed nuclei. The opercula of older zowcia show a 
slight preponderance of nuclei, and thus indicate more numerous cells. 
It is from such a region, then, that the young regenerating polypide 
arises, 

As in the case of the marginal polypides, so here, the lips of the 
invagination pocket close and become fused to form the neck of the 
polypide (Plate X. Fig. 84). The later stages of the development of 
the regenerating polypides seem to be the same as those of the marginal 
buds, Figures 74 and 89 are, indeed, regenerating polypides. I cannot 
find any evidence that the alimentary tract, or any part of it, is formed 
in regenerating buds by a method differing in any essential particular 
from that in marginal buds, 

It is well known, however, that the degenerated polypide which forms 


’ in the old zoocium eventually disappears. Haddon 


a “brown body’ 
(83, p. 519) maintains that in the developing regenerated polypide 


“the walls of the stomach, or, more strictly, that portion of the stomach 
VOL, XXII — NO. 1. 


66 BULLETIN OF THE 
which forms the gastric coecum, grow round and envelop the brown body, 
so that the brown body passes as a whole into the alimentary tract of the 
young Flustra.” It seems to me that the burden of proof of such a 
remarkable occurrence lies with him who asserts its existence, and cer- 
tainly sufficient evidence is not presented by Haddon. 

To settle this question in my own mind, I cut a series of thin sections 
through a part of a stock of Escharella (which in budding shows a prac- 
tical identity with Flustra), in which all stages of regenerating polyp- 
ides were to be found. From complete series, at critical ages, I utterly 
failed to find any indication of the inclusion in toto of the brown mass 
by the polypide. But I found the alimentary tract of the polypides 
usually applied to the brown body (pyd. dgn.), as shown in Figure 92. 
At this stage the degenerated mass is surrounded by spindle-shaped cells, 
and just within these by a homogeneous or lamellated sheath. At later 
stages the elements of the degenerated mass were seen to be more loosely 
associated. The cells of the alimentary tract at the same time appear 
highly granular, and a granular coagulum often partly fills the alimentary 
tract. Before the new polypide is ready to expand itself, the brown body 
as such has often wholly disappeared. Just as my sections, leave no 
chance for the brown body to be included en masse by the alimentary 
tract, so too do they yield no evidence of the addition to the latter of 
new cells from this degenerate mass, as Ostroumoff, in the sentence 
quoted above, implies. 

The interesting facts of degeneration in Bryozoa deserve a more careful 
study than I have been able to give them. We are quite ignorant of the 
physiological significance of the regularly recurring degeneration and 
regeneration in certain Bryozoan colonies. Ostroumoff (786%, p. 339) has 
offered an interesting hypothesis, to the effect that the degeneration of 
the polypides, the remains of which are taken into the stomach of the 
regenerated polypide and the undigested portion of which is cast out with 
the fæces, is a method of excretion, made necessary to these animals from 
lack of urinary tubules. 


IV. Origin of the Gemmiparous Tissue in Phylactolemata. 


After having found that in Paludicella and the marine Bryozoa, as in 
Phylactolsemata, the growth of the colony takes place at the margin or 
tips, and that it is here primarily that buds originate, and after having 
thus found that throughout the group all of the organs of the polypide 
are derived from two layers, of which the inner gives rise to organs so 


MUSEUM OF COMPARATIVE ZOOLOGY. 67 


dissimilar in origin as the central nervous system and the alimentary 
tract usually are, it becomes a matter of no little importance to solve 
the two problems, what is the origin of these growing regions, and what 
that of the two layers. Through the works of Barrois (86), Ostroumoff 
(’87), Vigelius (88), and especially Prouho (90), on the metamorphosis 
of the larva and formation of the first polypide of Gymnolemata we 
are fairly well acquainted with the facts in this group; but a careful 
study has not heretofore been made of the Phylactolaemata with reference 
to the points mentioned above. Korotneff (’89) and Jullien (90) have 
published quite extensive papers on the ontogeny of Phylactolemata, 
which describe too incompletely the stages which should reveal the 
required facts. 

In order to throw a little light on these questions, I undertook the 
study of the embryology of two species of Phylactolemata. But before 
beginning the account of what I have found, it is necessary to remind 
the reader of some facts concerning the origin of the polypides in the 
adult colonies. For our knowledge of these we are chiefly indebted to 
Braem (’90, pp. 18-32) ; it has also been my privilege to confirm many 
of them. 

The details of the budding process are slightly different in Plumatella 
and Cristatella. In the latter genus the body wall becomes highly mod- 
ified as it grows older by the formation of secreted masses which nearly 
fill most of the ectodermal cells. In Plumatella, on the contrary, the 
ectodermal cells retain, for the most part, a more primitive, unmodified 
condition. Here, moreover, by a rapid growth at the neck of the pol- 
ypides, the individuals are carried to considerable distances from one 
another, whereas in Cristatella there is a less rapid growth resulting in a 
compact stock. 

In Plumatella, the whole of the embryonic tissue from which any bud 
arises does not go to the formation of a polypide, but a part of it re- 
mains as the neck of the polypide, and gives rise by cell proliferation to 
the body wall and the Anlage of a new bud. Thus the Anlage of each 
bud is part of that of a preceding bud. The question remains yet un- 
solved, Whence came the Anlage of the first polypide? Since the em- 
bryonic tissue of the inner layer of the bud, which seems to take the 
most active part in the formation of the bud, gives rise to both the lining 
of the alimentary tract and the wall of the brain, it becomes an ex- 
ceedingly interesting question, From what germ layer is this inner bud 
layer derived ? 

In Cristatella, as in Plumatella, not all of the embryonic tissue from 


| 
| 
| 
| 


68 BULLETIN OF THE 


which any bud arises goes to form that bud; but some of it is, appar- 
ently, passed along under the highly metamorphosed cells of the ecto- 
derm, again to divide itself, one part going to form a new polypide, the 
other to form the Anlagen of new buds. In Cristatella, this embryonic 
mass of cells of the inner layer of the bud seems to be to a considerable 
extent independent of the highly metamorphosed ectoderm, and to form 
at places a sort of third layer, lying below the true ectoderm and above 
the muscularis with the coelomic epithelium. Here, too, while it is 
easy to see buds arise from preceding buds in the adult colony, we 
cannot consider our question answered until we have discovered the 
origin of the cells from which, as from a stolon, the Anlagen of polypides 
successively arise. 

I desire to say that I have avoided giving a full account of the on- 
togeny of these species, both because it is not directly required for the 
solution of the problems in hand, and because we are promised studies 
in this field by Braem. 

The eggs of Phylactolamata arise, as has long been affirmed, from the 
cœlomic epithelium of the body wall. The evidence of this is conclusive, 
for one often gees in a single section various stages in the development 
of the eggs. (Plate XI. Fig. 93, ov.) It is also to be observed that 
they do not arise indiscriminately from any region of the body wall, but 
always close to the neck of a polypide. Sooner or later these eggs, 
surrounded it may be by a few follicular cells, are enclosed in an oœcium, 
and here undergo their development up to the stage of a young stock, 
possessing perhaps a dozen immature polypides. In the figures on 
Plates XI. and XII. the oweium (o«@.) has been usually drawn, but in 
Figures 100 and 104 it has been omitted. As a result of cleavage, a 
blastula is formed, and from one pole of this — the pole nearest to the 
neck of the ocecium — cells are given off which move into the blastocoel 
(Figs. 94, 98) and finally come to line the cavity. It is important to 
observe that in the earliest stage of this process found there were four 
inner cells, of which two are represented in the section (Fig. 94, ms drm. 
+ewdrm.). Thus the two layers of the adult body wall are established. 
Up to this stage the conditions are practically the same in Cristatella 
and Plumatella. From now on, they are somewhat different in the two 
genera, 

The first difference to be noticed is in the oweium itself. In Crista- 
tella the cells composing this rapidly become a pavement epithelium 
(Fig, 97); in Plumatella, on the contrary, the cells of the ocecium 
remain columnar (Fig. 99). The neck of the oweium also differs in the 


MUSEUM OF COMPARATIVE ZOOLOGY. 69 


two cases. In Cristatella it is long, thick, and filled with a dense mass of 
large cells (Figs. 95, cev. oœ., and 101 *, 102 *). In Plumatella (Fig. 99) 
it is very short. 

The second difference concerns the embryo itself, and is connected 
with the formation of the first polypide. In Plumatella (Fig. 99) the 
first indication of the formation of the first polypide occurs at or 
very near the neck of the oweium, or, since the ingression of cells 
into the blastocol took place at the pole of the blastula nearest the 
neck, we may say near to the pole at which ingression occurred, 
The cells of the outer layer (i.) are elongated and contain large ellip- 
soidal nuclei which are often pressed close together. All of the cells 
of the larva stain more deeply at this pole than elsewhere, and those of 
the inner layer rather more deeply than those of the outer. The nu- 
clei are also very large, those of the outer layer being possibly more 
prominent than those of the inner ; but the difference is not so marked as 
in the drawing, where too the nucleoli of the inner layer are represented 
relatively too small. Even at this stage one finds in another section 
of the same embryo the beginning of a second polypide, whose position 
is indicated at *. This second polypide is indicated merely by a con- 
siderably thickened inner larval layer, and a very slightly thickened 
outer one. The two polypides are thus seen to be wholly independent 
of each other. The first invagination further advanced is seen in cross 
section of the whole larva in Figure 96. The entire outer layer would 
seem at first sight to be involved in this invagination ; but even in this 
figure there are seen one or two nuclei which lie under the oœcium at 
the place of invagination. I believe that they will not be involved in 
it, for at a very little later stage (Fig. 104) one finds a layer of cells 
lying over the invaginated bud, which I believe are destined to form the 
ectoderm of the body wall at this place. 

Later stages in the development of the larva in this species are 
not shown. The bud follows, I am confident, the same steps that are 
pursued by the bud in the adult colony. A placenta-like connection 
of the larva with the, ocecium, which was first described by Korotneff 
(87, p. 194), begins at about this stage, and continues until two well 
formed polypides are present. This “gürtelformige Placenta” begins 
to form in about the middle of the young embryo, and the elongated 
cell of the outer layer of the larva, in contact with the oœcium shown 
on the left of Figure 99 below the *, is, I believe, the first indication 
of it. The oocium and larva both continue to increase in size, and 
the walls of the former become thinner with their increase in area. 


70 BULLETIN OF THE 


The attachment of the owcium to the body wall of the mother stock 
always remains small, as in Figure 99, and the embryo, in my experi- 
ence, does not come in contact with it. 

The formation of the first polypide in Oristatella is preceded by 
another process. Just as in the adult colony the inner layer of the 
polypide does not arise by invagination of the ectoderm, but from the 
stolonic cells lying at the base of the ectoderm (see Davenport, ’90, 
pp- 108, 109, Figs. 4 and 15), so too in the embryo. The first process 
then must be the formation of the stolonic cells. Figure 101 shows at 
the point marked sto. (which is at the pole of the embryo whence the 
inner-layer cells originated) that certain of the cells of the ectoderm ap- 
pear to be arching over a disk, containing about six cells in section, and 
thus coming in contact with the cylinder of cells (*) which projects 
from the neck of the occium. By a continuation of this process, the 
central disk of cells gradually comes to lie below the general level 
of the ectoderm, and to be cut off from contact with the neck of the 
ooecium (Fig. 97, sto.). The position of the stolonic mass with refer- 
ence to the neck of the polypide in this last figure must be considered 
abnormal ; it is at any rate exceptional, as it lies at one side of the neck 
of the ocecium, which does not, therefore, appear in this section. The 
next later stage which I have found is shown in Figure 102. The sto“ 
lonic mass seen lying beneath the ectoderm in Figure 97 has here 
already given rise to a young polypide (%., ex.), and its area is increas- 
ing in all directions by cell division (sto.). The beginning of a sec- 
ond polypide is indicated on the right at sto. The ectoderm is seen 
lying above this stolonic mass, and closely applied to the neck of the 
oocium (*). 

Neither at this nor at any subsequent stage have I been able to 
detect in Cristatella any “gürtelformige Placenta” such as exists in 
Plumatella. I am therefore of opinion that the process of nutrition, 
which is effected in Plumatella from the ocecium through its placenta, 
is effected in the Cristatella larva by its attachment to the neck of the 
oocium. I am pleased to see that Jullien (’90, pp. 13, 14) has also 
reached this conclusion in a paper which he has had the kindness to send 
me. At a later stage, the embryo, or young colony, seems to become 
detached from its intimate association with the neck of the oœcium, as 
we see in Figures 95 and 103. 

Figure 103 represents a stage in which there are two well developed 
buds, both shown in the section. There is, in addition, on another section, 
one less developed. The stolon is seen passing oralward of these twe 


MUSEUM OF COMPARATIVE ZOOLOGY. an 


primary polypides, or rather the primary and secondary one. Moreover, 
as the series of sections shows, the stolon does not exist merely in this 
section, but it isa disk which is cut here in one of its diameters. A sepa- 
ration of the stolonic mass has occurred between the two oldest polypides, 
so that the ectoderm is here in contact with the coolomic epithelium, 
just as is the case between buds in the adult stock. As the colony in- 
creases, the inner and outer margins of the stolonic tissue continue to 
extend farther outward, and this tissue forms at first a broad ring of 
ever increasing diameter. Later, as the area of the stock increases, the 
ring becomes broken, so that, instead of growing along an infinite number 
of radii, its growth is confined to a few, as in the adult colony, 

I will defer a discussion of the significance of these facts to the gen- 


eral part of this paper. 


B. GENERAL CONSIDERATIONS. 


I. Laws of Budding. 


Carefully conducted studies on stock building have generally revealed, 
just as these on Bryozoa have shown, a law in budding. This law in 
budding results in the formation of a stock the interrelation of whose 
individuals is a determinate one. I now propose to offer an hypothesis 
to account for the existence of these laws, and then to show how facts 
of budding in Bryozoa and other groups can be explained by means 
of it. 

And first of all I must acknowledge that this hypothesis, although 
perhaps here first formulated, really depends upon observations and de- 
ductions made long ago on this group, first by Hatschek, who from 1877 
has maintained that individuals do not arise independently of one an- 
other, and secondly and mostly to Braem, who in ’88 (pp. 505, 506) 
declared of Phylactolamata “dass in dem Stock keine Knospe entsteht, 
die nicht auf das embryonale, d. h. den specifischen Leistungen der 
Körperwand noch nicht angepasste Zellmaterial einer älteren Knospen- 
anlage zuriickgienge und dass somit in der ersten Knospe des keimen- 
den Statoblasten sämmtliche Knospen des künftigen Stockes implicite 
enthalten sind.” Not less is the following hypothesis indebted to the 
ideas of Roux and Fraisse, and to Nussbaum, who has said (’87, p. 
293): “ Ein lebendes Wesen ist somit als Ganzes oder in seinen Theilen 
soweit individualisirt und vergänglich, als die Gewebebildung und die 
Theilung der Arbeit vorgeschritten ist; das Ueberdauern der Einzelexis- 


2 BULLETIN OF THE 


tenz, die Theilbarkeit auf geschlechtlichem oder ungeschlechtlichem 
Wege, spontan oder künstlich bedingt, ist an das Vorhandensein undif- 
ferenzirter Zellen gebunden und ist um so grösser, je weiter im Organis- 
mus diese Zellen verbreitet sind”; and, finally, to the idea which is 
implied in the conclusions of Nussbaum (’80, pp. 106-113) and Weis- 
mann, that germplasma does not find its origin in the parent individuals, 
but is merely borne by them in its unbroken passage from generation 
to generation. 

This hypothesis is simply that there is in every stock of Bryozoa a mass 
of indifferent cell material which is derived directly from indifferent cells 
of the larva or embryo, and whose function is to form the organs of the va- 
rious individuals, including the polypides. This indifferent cell material 
lies in the body wall, principally at the growing tip or margin of the 
stock. By its growth and differentiation it gives rise to the body wall, 
muscles, etc., and at intervals it leaves behind, as a portion detacned 
from itself, a mass of indifferent cells, which is capable of forming a polyp- 
ide, or of becoming a new centre of growth, or of both, Which of these 
possibilities will be fulfilled, where and when these masses of indifferent 
cells will be left behind, depends upon the necessities of the species, and 
the variations in these respects give rise to the peculiar characters of 
the different stocks, 

This hypothesis differs from that of Braem in that the pre-existence 
of a Knospenanlage assumed by Braem is, according to my view, a non- 
essential feature in the formation of the colony; the pre-existence of an 
indifferent cell mass, which does not itself constitute buds, but may give 
rise to masses which can, is the only essential feature. 

As a first application of this hypothesis I refer the reader to the con- 
ditions of stock formation in Paludicella, already described. We find at 
the tip of the colony a mass of large proliferating cells, which I regard 
as histologically undifferentiated. These cells give rise to the body wall, 
— the cystid, — and at intervals leave behind three masses of cells, which 
I regard, from the fact that they retain their cuboid condition, as well 
as from their ultimate fate, as indifferent or embryonic. The median 
mass of each of these gives rise to a polypide, and to one only. The 
lateral masses form centres of growth similar to the one from which they 
were derived. 

In order to reproduce the arrangement of individuals in the stock re- 
sulting from this manner of budding, we may make use of some graphic 
method of representation, as Smitt (765%, pp. 139, 140) did long ago, and 
as Allman (70), Semper (’77, pp. 67-78), Chun (’88, pp. 1167-1180), 


1 MUSEUM OF COMPARATIVE ZOOLOGY. 73 


Braem (’90, pp. 33 and 44), Ehlers (790, p. 9), and others, have since 
done. I shall represent the mass of indifferent cells by an asterisk, 
and individuals (according to Chun’s nomenclature) by the use of the 
large and small letters of the Roman alphabet, and, finally, by Greek 
letters. The typical stock of Paludicella might then be graphically 
represented thus (cf. Plate I. Figs. 2 and 2°) :— 


* 
* CO * 
x“ OR 
xabax 
* * * 
*b * KR xx 
* * *¥BaaaB* 
xaaak KEIN 
* * a * * * 
(1) 3 D 2) B A 
* xa% Ko 
* xaaax KK %* 
k*k * x Baaaßx 
xbx* Xk k% 
* X% 
xabax 
* Cx 
* 


Here the letters indicate polypides or their Anlagen, and the asterisks 
indifferent tissue. The individuals represented by capital letters may be 
called primary individuals; they may be said to belong to the primary 
series, and to have been derived from the primary indifferent mass. 
The individuals represented by small Roman letters will then be secon- 
dary individuals, belonging to the secondary series and arising from sec- 
ondary masses; etc. It is to be observed that this indifferent tissue is 
here found only at the tips of branches or Anlage of such. No asterisks 
are found adjacent to, the adult polypides A, B, C, etc., which have 
given rise to lateral branches, and these have therefore no power of pro- 
ducing new parts of the colony. The asterisks must not be regarded as 
having been descended from the letters which they adjoin, but from the 
terminal asterisks only; that is to say, in Paludicella embryonic tissue 
has originated from terminal embryonic tissue, and not from indifferent 
tissue left remaining alongside of the polypides. 

Jonditions differing in an interesting manner from these were found 
by Braem (’90, pp. 18-32) and myself (Davenport, '90, pp. 103-106) in 
Phylactoleemata. In Plumatella Braem has shown in the clearest man- 
ner how some of the embryonic tissue around a polypide at the proximal 


74 BULLETIN OF THE 


end of a nascent branch is carried away to the oral side of the “mother 
polypide,” and lays the foundations of another polypide. In like man- 
ner the embryonic tissue around the “mother polypide ” may give rise 
to one or several additional embryonic masses. He has also (pp. 29- 
32) shown in the most convineing way that each mass, particularly in 
the case of secondary buds, consists of two parts, of which one goes to 
form the polypide ; the other contributes to the further growth of the 
common cystid and the formation of new embryonic masses. Since 
here every embryonic mass is in intimate relation with a polypide, and 
since the polypides arise nearly in one plane, only secondarily moving 
out from it, the relation of individuals may be expressed by a formula 
occupying a single line. Braem has thus expressed it: — 


ofall I nl 
(2) Del Be BBA 

According to the system adopted for Paludicella, this may be given 
thus : — 


(8) *a xa xb *A a *B %0% 
or, more developed, thus : — 
(4) *a, xa *B xa xa xb xc «A xa xa xb xB xax O ¥D % 


in both of which the right hand asterisk (%) takes the place of the A at 
the right of Braem’s diagram. These symbols denote that we have a 
mass of indifferent tissue connected with each polypide, or the Anlage of 
such ; and this indifferent mass, as well as the adjacent polypide, was 
derived from some other indifferent mass. Thus the masses connected 
with A, B, C, D are to be regarded as having been cut off from the em- 
bryonic mass at the extreme right ; and each of these secondarily gives 
rise to the polypide buds a, b, etc., and their embryonic tissue. Thus 
we have to do with centrifugal budding only. 

In Cristatella the conditions are essentially similar to those in Pluma- 
tella, the chief difference being that usually only two polypides with 
their embryonic masses arise from each polypide. ‘This condition may 
be represented by the formula : — 


(5) *a, *a *B xa xa xb [*]A a *a-%b [*]B [$] 


in which the embryonic masses originally attached to A, B, etc., are 
bracketed to indicate that they are normally no longer active in giving 


MUSEUM OF COMPARATIVE ZOÖLOGY. 75 


rise to new polypides. As a matter of fact, the secondary rows often 
make. a greater or less angle with the primary ones, and as a result 
lateral branches are formed. Taking this character into account, the 


Cristatella formula might be written : — 


* 
* b 
*a, — a | 
| [*] 
x8 — xa — A 
| 
(6) #b — WB — [8 
| 
ka mug 
* 


This representation indicates the fact that the first formed buds (A, a, «, 
etc.) are lateral ones; the second, median (Davenport, ’90, p. 106). 
Intermediate stages between the condition in Plumatella, in which an 
indefinite number of polypides and gemmiparous masses can be budded 
off from pre-existing gemmiparous masses, and the condition in Crista- 
tella, in which only two such arise, occur apparently in some species of 
Plumatella, in which, as Braem (’90, p: 31) has shown, few polypides 
are produced from any gemmiparous mass, and all but two of these gen- 
erally do not develop. In the young corms of Cristatella, on the other 
hand, more than two polypides may thus arise. 

Other Ctenostomata show a regularity in the budding process similar 
to that of Paludicella, and exhibit instructive variations upon it. 

Victorella, an interesting Ctenostome occurring in slightly brackish 
water, and first described by Kent (70) in 1870, possesses, according to 
the pregnant observations of Kraepelin (87, pp. 75, 76, 154-157), a 
stolon-like tube, from which at intervals polypide-bearing “ cylindrical 
cells” arise. Kraepelin (°87, pp. 155-159) has shown it to be in the 
highest degree probable that the protrusion of the body wall in the neck 
region of the polypide of Paludicella is the homologue of the “cylindrical 
cells ” of Victorella, and that the remainder of the zoocia of Paludicella 
is homologous with the “stolon” of Vietorella. While in Victorella the 
cylindrical cell is developed to such an extent that the retracted poly- 
pide is still included within it, and the stolon remains of small calibre, 
in Paludicella, owing to its shortening, the retracted polypide must seek 
refuge in the stolon, whose diameter is consequently increased to receive 
it. Evidence for this is found in the stolon-like nature of the youngest 
zoœcia of a hatching winter bud of Paludicella Ehrenbergii, and in the 
elongated cylindrical cell of the adult Palndicella Mülleri, Kraepelin, 


76 BULLETIN OF THE 


which must be considered a form intermediate between P. Ehrenbergii 
and Victorella. 

The architecture of the Victorella and Paludicella stocks is, then, sim- 
ilar, in that they both consist of a row of individuals successively formed 
at a stolonic tip. The resemblance is heightened by the fact that, as in 
Paludicella, so also in Victorella, a pair of lateral buds is given off from 
each zooecium to form lateral branches (Kraepelin, ’87, p. 157). As in 
Paludicella, so also in Victorella, communication plates, Mosettenplatten, 
arise early to separate the zoocia from each other. But Victorella 
differs from Paludicella in this, that while in the latter the neck of the 
polypide does not become the centre of origin of new buds, in the former 
it does, just as is the case in Plumatella (Kraepelin, ’87, Plate III. 
Fig. 75) ; that is to say, there are laid down from the tip of the branch 
three masses of bud-producing tissue, besides that which goes to form 
the polypides of the primary branch. The graphic representation of 
this species will therefore be more complicated than that of Paludicella, 
and has this form : — 


* * * 
%* + * * * 
xd» Bs ara *aq x Be 
* * * * * * 
% * * 
axs Ar AHA a 
f * * * * 
* 
()#Dx Cea * Bx b% ax ax Ax cx bx ax ax Br ax 
* 
* * * * * 
»Ax ax DHA a 
* A yd 
* * * * * * 
xb» Bs ar axa Be 
* * * * * 
* * * 


* 
*C% 
%* 
Jompare with (1), page 73, and (4), page 74. 
From around each individual of the series A, B, C, etc., which has 
been derived from the tissue of the stolon tip, there arise- series of 


MUSEUM OF COMPARATIVE ZOOLOGY. Tl 


lateral and of median buds. From around each of the lateral buds, in 
like manner, both lateral and median buds of a higher order arise. But 
from each of the median buds only median buds arise. These median 
buds, are not, however, all of the same kind. The one first produced (a x, 
of Formula 7, # of Form. 8) differs from all formed after it (b *, c *, d x, 
etc.) in this, that it bears no polypide, but forms the tip of a stolon 
from which both median and lateral buds arise (8, a, near extreme right of 
Form. 8). From the second and all succeeding median buds (a b, c, ete., 
Form. 8), there arise only median buds of a still higher order, Of the 
latter, the first, as before, produces no polypide, but becomes the tip of a 
stolon giving rise to both median and lateral buds ; the others give rise 
to only median buds of a still higher order, and so on. Our former for- 
mula assumed that all median buds were alike, and all incapable of giving 
rise to lateral individuals. Their dissimilarity introduces a complication, 
so that the species must be represented by some such formula as 


this; — 


* 
e 
* 
* * 
"rar b xa * 
12 * * 
a 
* * * 
* a, * wt * ay * 
* * * * 
Br ax ar Hat ra ra * Be 
* * * 1. 
* * a, * * ay * 
xb * B * 
* * 
(8) ie 
* * 
„ar a ea* a 
* * * * 
* a 3% 
raw * * a, * * 
2 3 5 * My * 
* * * e 
Dak Cease Bx be ae thar Ax ce Dxttak Br ax at Se Be tax ay «3 
* * * a 
Mg at T * ay * 
*A* * ay, * * 
* a w% 
etc. etc. 


1 Tt must be borne in mind that such a graphic representation as this, while it 
agrees with the descriptions and figures of Kraepelin and Hincks (’80, Plate 79) so 


78 BULLETIN OF THE 


in which the heavy asterisks represent the budding tips of the stock, 
which give rise to new individuals (tips of the stolons), and a, ß,, etc. 
indicate individuals of the fourth order. The lighter asterisks indicate, 
as before, points of proliferation from which new buds may arise. 

It seems highly probable that Victorella finds near allies in Mimosella 
and other genera of the Stolonifera, 

In Hypophorella expansa, according to Ehlers (’76, pp. 5-9) and 
Joyeux-Laffuie (’88, pp. 137-139), the stolon is composed (as in Vic- 
torella), of a number of internodes, each separated from the other by 
communication plates, and bearing on the distal end typically a feeding 
zoöid (Nährthier) and a lateral stolon. It seems to me that the jointed 
condition of the stolon is reasonably accounted for in the same way as 
that of Victorella, by supposing that each internode, together with its 
zocecium, is comparable with the whole indvidual of Paludicella. The 
“feeding zodids” of Hypophorella will then be comparable with the 
Cylinderzelle of Victorella. Two facts are opposed to this view : first, 
the polypide is not formed primarily in the stolon, coming only secon- 
darily to lie in the Cylinderzelle as in Victorella ; and, secondly, there is a 
Rosettenplatte in Hypophorella between the feeding zoöid and the stolon, 
while none exists in Victorella. But upon this assumption one can best 
account for the fact that the stolon is composed of as many joints as 
there are feeding zoöids, —a condition which appears to occur in only a 
few other genera, and these closely allied to Victorella. Thus, in Cylin- 
dreecium pusillum and C. dilatatum of Hincks we have two species which 
may be considered to represent two possible intermediate stages between 
Victorella and Hypophorella, not only on account of the jointed stolon, 
but also on account of the enlarged distal end of the joint, which is emi- 
nently characteristic of the allies of Victorella. The first objection, that 
the polypide is not developed in the stolon, but first arises in the well 
formed zocecium of the feeding zoöid, might result from the increased 
importance of the zocecium over the Cylinderzelle. The formation of the 
plate between the zomwcium and the stolon might be accounted for by 
the physiological need of such an organ resulting from the increased im- 
portance of the zocecium, (cf. p. 40). Such plates exist, in fact, between 
the primary median individuals, and those secondary median ones in 
Victorella which are budded from the Cylinderzelle. This hypothesis 


far as they go, may not fit the conditions in all parts of the colony. Moreover, it 
is to a certain degree idealized, i. e. subjective, for even in the figure of Kraepelin 
(87, Fig. 75) one of the individuals of the series a, b, c, etc. has given rise to no 
stolon as its first bud. 


MUSEUM OF COMPARATIVE ZOOLOGY. 79 


is further supported by the fact that, as a stolon may arise from the 
Cylinderzelle of Victorella, so in Hypophorella such a condition is not 
uncommon, although hardly typical. In accordance with this hypothesis 
the formula for Hypophorella might be given thus :— 


* 
c* 
* bar 
(9) b * * 
* aa* aaß 
* a* * * %* 
%D € B A 


Ehlers (76, pp. 127, 128), in founding the group of Stolonifera, clas- 
sified the different methods of arrangement of the individuals in the 


colony as follows :— 


I. Many polypides (Nährthiere) on the single joints of the stolon 
(Stengelgliedern). 
1. On the entire length of the joints. 
(a.) Arranged in two rows. 
(b.) Arranged in a spiral. 
(c.) Arranged in one row. 
2. At the ends of the joints. 
(a.) In rows. 
(b.) Massed. 
II. Only one polypide Nährthier on a joint of the stolon. 
1. Polypide lateral, near it one or many stolonic joints (Hypophorella). 


2. Polypide terminal. 


In the present state of our knowledge, it is very difficult to say how 
the types of budding shown in those Stolonifera which possess more than 
one Nährthier on a joint of the stolon are related to, or are to be connected 
with, the types of Paludicella, Victorella, Hypophorella, or other genera 
possessing only one Nährthier to a joint. This could doubtless be deter- 
mined, however, by studying the early stages in the development of the 
stocks. Taking them as they are, however, we find a very simple condi- 
tion in the stocks of Class I., in which the Nährthiere are arranged in a 
single row, as in Vesicularia spinosa (cf. Hincks, ’80, Plate 73, Figs. 3-7). 
The tip of the stolon consists, as I have myself observed in allied spe- 
cies, of somewhat cubical cells of variable thickness, and it is from this 
tip that the Anlagen of the individuals arise. Lateral branches occasion- 


80 BULLETIN OF THE 
ally replace a Nährthier, and the latter seems never to produce secondary 
individuals, The formula of the stock might be written : — 


(10) eR —-E—-|—-D—C— 8B — 


In Bowerbankia pustulosa we have two rows of individuals produced 
side by side from near the end of the stolon, This condition would be 
represented by 
(11) wD +O B #A 

aw *D xC *B xA 
provided the individuals of this primary series possess the power of giv- 
ing rise occasionally to secondary buds, as seems certainly to be the case 
in some members of this genus which I have seen. The spiral arrange- 
ment of some colonies is striking ; it is of evident advantage to the stock, 
but its cause in these cases is wholly unknown. 

In every one of these cases, and, in fact, in all of those figured by 
Hincks, which belong to the Stolonifera, there is no trace of dichotomy. 
Throughout we have to do with linear series, which give rise to lateral 
branches. 

Turning now from the Stolonifera to the other grand division of 
Ctenostomata, the Aleyomiide, we find the same prevalence of a law in 
budding. In its typical expression it may be written as follows : — 


* 
»* bt 
%* * (*) * 
“ga * „aaa * 
* * * 
(12) mC (*) B (*) A 
* ‘i ig 
* aA * xa aù a * 
* » (*) * 
«b * 
* 


Although secondary median individuals are not habitually formed, 
yet, owing to the capacity of regeneration possessed by individuals 
A, B, C, etc., an asterisk is affixed in parentheses to show the probable 
persistence of embryonic tissue. Of the lateral series one or both may 


fail to be developed. 


MUSEUM OF COMPARATIVE ZOOLOGY, sl 


It might be difficult to determine whether in this group we have to do 
with dichotomy, did not the tips of the margins at times reveal the fact 
that there is no division of the ancestral series, but that a new one is 
added at the side of an ancestral one (Plate VIII. Fig. 69), where of 
the marginal individuals 4 is clearly median (ancestral) and 3 is lateral, 
13 median and 12 lateral, etc. (see page 49). 

The members of the group of Cyclostomata seem to be closely related, 
and the method of budding is so similar throughout the group that it 
seems fair to interpret the more compact Tubuliporide from the Crisiadee, 
In Crista, as we have seen, individuals are placed in rows, from which at 
intervals lateral rows are given off to the right or left. One may say 
that typically these are given off from each individual to both the right 
and the left, although in some cases, as in Figure 65°, lateral branches 
are typically given off alternately to the right and left, and are often 
aborted. Perhaps the most general formula of all for Cyclostomes should 
be that of two lateral branches from each individual, one or both of 
which may remain undeveloped. Such a formula I believe to be also 
the typical one for Bugula and its allies, and for the Alustrina and 
Escharina. It would be written thus:— 


* 
CH 
+» (*) * 
IRo (Doe 
* * * 
* Db * 
* x (x) * * 
# * (*) * *Beaaa* Be 
*a* eave. ii * * * * 
* * * 
(18) 3 D (*) C (*) B (x) A 
* * * 
* a* *aaar* * * * * 
* # (*) * * B (*)a a a(*) B * 
* * (*) x * 
wbs 
* * * 
ta ba* 
* (x) * 
* 8: % 
* 


in which the parenthesized asterisks indicate the presence of regenera- 
tive tissue. This is identical with (12) änd similar to (1). 

Braem (’90, pp. 130-133) has already called attention to the differ- 
ence between Phylactolamata and Gymnolemata in the orientation of the 


VOL. XXII. — NO. 1. 6 


82 BULLETIN OF THE 


polypide. In Phylactolemata the oral aspect of the polypide is turned 
towards the margin of the corm or the tip of the branching stock ; in 
Gymnolsemata, on the contrary, the anal aspect is turned in that direc- 
tion. This difference is a very striking and constant one. It is corre- 
lated with another difference in the law of budding of the stock, which 
will become evident upon comparing Formulas (4) and (5) on page 74, 
of Phylactolemata, with Formulas (1) on page 73 and (7) to (13). In 
all of these the margin or tip of the stock is at the left, the centre at the 
right. In the formule of Phylactolemata the budding is centrifugal, 
new individuals being produced from the embryonic masses towards the 
margin; in the formule of Gymnolemata budding is centripetal, new 
individuals being produced from the embryonic masses towards the 
centre. In both Phylactolemata and Gymnolemata the anal aspect is 
turned towards the gemmiferous region. 

Braem calls attention to one other difference, namely, that, in the case 
of the retracted polypide, in Paludicella the rectum lies next the at- 
tached surface of the stock; in Phylactolemata, the œsophagus. A 
mechanical cause of this is suggested when this statement is put in other 
words : the polypide in its retracted position is stored in both Phylac- 
tolemata and Gymnolemata proximad of the atrial opening ; i. e. away 
from the tip or margin, and towards the centre of the stock. May not 
this be explained, in part at least, as an adaptation to room} 

I will here add four examples of regular budding taken from other 
groups of animals, to illustrate the general applicability of this method 
of representation. The first of these is that of the Siphonophore Halı- 
stemma whose formula has been worked out by Chun (’88, p. 1169), and 
expanded and illustrated by Korschelt und Heider in their recent text- 
book: (p. 39). It runs as follows : — 


(14) DO ba Oo 4d. ¢ b BL do brava y 8 aA 


According to my interpretation of the case, this formula might be 
written (15) :— 


%#D sc xb xa xC 3d #c xb xa xB ae xe xd xc xD xa xa ey #B ra % A, 


in which the # behind B has been derived from the embryonic mass at 
A, that behind C from B., etc. The %’s represent embryonic masses 
from which a, b, c, etc. are derived. 

If we assume that the terminal individual (A) has not been derived 
from the primary embryonic mass, at the extreme left, but has had its 


MUSEUM OF COMPARATIVE ZOOLOGY, 83 


origin in the embryo, the formula would have to be written somewhat 
differently ; namely, thus (16) :— 
3% Cat ce be ax Bat dx cx br ax Aex de ck bk ak ax yk Be ax [A] 


In a species of Pennaria, common on our coast, which is probably 
Pennaria tiarella, McCready,’ I have noticed the presence of a similar 
law of budding. The whole stock lies in one plane, the lateral branches 
arising alternately from the right and left of a central stock, like the barbs 
of a feather. These lateral branches give rise to a series of secondary 
ones, which are all placed on the same (axial) side of the branch. Each 
branch, of whatever degree, originates as a bud bearing a polyp. From 
the elongating stalk of this terminal polyp, buds arise, — the beginnings 
of branches of the next higher order. The stock may be represented by 


the following formula ; — 


A 
* 
| 
de — 
| 
cr — 
ge d 
* 
be —— 
a 
B * 
c * | | 
* * a eT 
| 
(17) #»D--- - - - - - - -- 
| 
ax — — 
| 
be — 
| 
c* 
* 
B 


Expressed in a linear series, this formula may be written :— 
(18) w Dx Ce Bece be ax Ax dec beat ax Be ax 


which is identical in form with the second formula (16) given for Hali- 
stemma. 


1 This species is figured by Leidy (’55, Plate 10, Figs. 1-5) and Verrill (’78, Plate 
XXXVII. Fig. 277). An allied species, P. gibbosa, is figured by Louis Agassiz 
in the “ Contributions” (Vol. III., Plate XV. Fig. 1). In describing P. Carolinii, 
Weismann (Entstehung der Sexualzellen, p. 122) says that the lateral hydranths 
do not possess the capacity of giving rise to new lateral hydranth-buds (of a higher 
order). But, as indicated above, P. tiarella seems to do this regularly. Leidy’s 
and Verrill’s figures show the same thing. 


34 BULLETIN OF THE 

Lastly, this formula may be applied to certain cases of fission, as in 
fresh water Annelids, As is well known, the fissiparous process is pre- 
ceded by the formation of the so-called budding zones (Knospungszone). 
These arise in Ctenodrilus, according to Kennel (’82, pp. 403, 404), be- 
tween two dissepiments in the middle of a metamere, and new ones are 
continually formed behind the others as the animal grows in length 
by cell proliferation at the tail end. The budding zones are, according 
to Kennel, regions composed of embryonic cells. I think it probable 
that this embryonic tissue has been derived from the embryonic tissue 
of the anal end of the animal. There are as many budding zones pro- 
duced as there are new metameres added by the anal growth, and since 
the budding zones are intrasegmental, each zoöid consists of four parts ; 
viz, (naming them from anterior to posterior end) of the posterior half 
of the preceding budding zone, of the posterior half of the metamere in 
which the budding zone arose, of the anterior part of the next follow- 
ing metamere, and, finally, of the anterior part of the following budding 
zone, Zoöids then are made up of parts of two adjacent metameres, 
and the middle of each zoöid is intersegmental. The zodid has progressed 
little beyond the state of possessing two (half) metameres at the time it 
becomes free. New metameres must become formed by caudal growth. 
The animal is, then, according to my conception of the significance of the 
process, derived chiefly from these budding zones. Evidently, the law 
of production of new individuals (or new budding zones) is a simple 
one, and may be written, in accordance with my nomenclature, 


(19) ae E (+) * D (x) * C (*) æ B (*)* A 


in which A, B, C, ete. represent successive individuals (adjacent halves 
of two metameres), and the asterisks, as before, embryonic tissue, ‘The 
two adjacent asterisks together represent the budding zone, of which 
the posterior half (parenthesized) proves itself the least active. 

The conditions given by Semper (’77, pp. 69, 77) for Chætogaster 
(and Nais) are much more complicated, but may be expressed by 
the use of a formula constructed upon the same plan. Choetognster 
differs from Ctenodrilus in this: that young budding zones, and event- 
ually young individuals, are produced between older ones, instead 
of always at the anal end; and the new zoöids often acquire several 
metameres before becoming free. It seems to me probable that, as in 
Ctenodrilus, the budding zones are derived ultimately from the anal 
zone ; but~ here, in contradistinction to Ctenodrilns, new budding zones 
may secondarily arise from other budding zones produced eartier, thus 


MUSEUM OF COMPARATIVE ZOOLOGY. 85 


giving rise to the phenomena of young individuals interpolated between 
older ones. Representing, then, individuals by the budding zones from 
which they have arisen, we may convert the following formula of Semper 
into one based on our own nomenclature : 


E D H B G Cc F A 


un nen mae ee [en ~~ 


az, -3+0-0 4+0-1+0-0 4+1-1+0-0 4+0-2+0-0 4+1 
in which the succession of generations of zodids is 
Seg Raat Brgy 8, OF lt 


In the above formula A, B, ©, etc. represent zoöids; the numerals 
below the letters, the number of metameres of which each is composed ; 
0, an incomplete metamere about to be derived from a budding zone ; 
az, the anal zone. Written in accordance with my conception of the 


facts, this formula would read: 


(20) $ D(x) *C(*) #a(x) «Bir * b (x) xa(x) xa (x) *A, 


i 
which somewhat resembles Formula (15) of Halistemma, and signifies 
that two embryonic masses are left behind by the anal zone, of which 
the one anterior to the zoöids proper (represented by letters) goes merely 
to form the head parts, and is represented parenthesized. The second 
is caudad the zoöid, and may form a secondary “ anal zone” giving rise 
to new zoöids. From one zoöid two or more anal zones may take their 
origin. Thus, from the embryonic mass caudad of A there have arisen 
that caudad of b, which has given rise to b (x) * a (x), and that caudad 
of a, which has given rise to a! (x).1 

The most general formula given on page 81 undergoes many modi- 
fications in the different groups, but in the midst of these modifications 
certain laws are to be discerned, to some of which I have already called 
attention. I will now proceed to a discussion of the significance of 
these laws. 

The guincunx arrangement of individuals, which is so noticeable in 


1 From a study of surface views of many specimens of Autolytus collected 
at Newport during June, 1891,-I am convinced that the sexual individuals are 
produced by proliferation of cells in the metamere XTIT. or XIV. of “parent 
form,” —the last which remains behind after breaking off of sexual form. Rep- 
resenting the proliferating metamere by (%*), we may write the budding formula 
of Autolytus thus: 

(208) $% E(*) D(x) C(*) Be) Alk) 
in which the parenthesized asterisks indicate the proliferating, but not gemmif- 
erating anal metameres of the sexual form. (Cf. A. Agassiz, ’63, pp. 397-400.) 


86 BULLETIN OF THE 


the phytoid stocks of Bugula, and in creeping corms like Lepralia or 
Cristatella, may be explained as affording additional strength on the one 
hand, and as a device for saving space on the other. 

The absence of true dichotomy, which I have sought to show charac- 
terizes the budding of Bryozoa, is interesting as seeming to indicate 
the fundamental similarity of the process of budding in Paludicella to 
that found elsewhere. The tip of the branch does not divide equally 
in the first nor in the other instances, but constantly maintains its 
precedence, giving off parts of itself to form lateral branches. These 
parts may grow out at right angles to the primary branch, as in Palu- 
dicella, but generally they grow forward nearly parallel to it, as in 
most marine Gymnolsemata. 

In Bugula (Plate VII. Fig. 64°) branches are always given off 
toward the axils, and therefore an ancestral branch gives off all lateral 
branches from one side and the successive orders of branches are given 
off alternately to the right and left. In Crisia, on the contrary, branches 
are given off abawially, and they are given off not from one side only, 
but alternately to the right and left. In both cases the two facts are 
mutually dependent. The first case gives rise to a stock in which the 
branching tends as greatly as possible towards compactness and the for- 
mation of a closely built stock ; the second case gives rise to a diffuse 
and loosely built stock (cf. Figs. 64, 65, and 64", 65"). In the sec- 
ond case there is a maximum space’ to each individual ; in the first, 
a maximum economy of space. 

The rule that lateral buds on two closely related branches tend to 
arise in the same generation is one that, as has been shown, is more 
or less apparent in some cases, but is easily obscured by other rules, 
May not the tendency be due to the same causes that produce the 
synchronism of division in related cells of a cleaving egg! 

That lateral buds should occur in Bugula flabellata on the outermost 
rows only is not surprising when we reflect that there is abundant room 
on the margin, whereas the inner individuals are hemmed in from lat- 
eral expansion by the pressure of adjacent rows. This is very marked in 
certain repent colonies, as, for instance, occasionally in Membranipora 
(Plate VIII. Fig. 70). Here the intermediate branches 6, 7, 8, and 9 
have produced no lateral buds for many generations, while almost every 
individual of the marginal rows has given rise to a lateral branch. It is 
merely a result of the same cause, it seems to me, that lateral budding 
occurs more frequently in Bugula turrita at the margins of fans than 
elsewhere. Here there is room to spread. 


MUSEUM OF COMPARATIVE ZOOLOGY. 87 


The rule (5) that ancestral rows contain fewer generations of indi- 
viduals than lateral ones may perhaps receive a partial explanation from 
the further fact (rule 6) that of the two rows starting from any axil the 
ancestral branch will give rise to a greater total number of individuals 
than the lateral one will in the same time. We should expect a less 
rapid forward growth if the lateral growth is extremely vigorous. One 
might also say that the intermediate rows had grown abnormally in 
length, since that is the direction in which there is most room. 

The reason why the ancestral branches in Bugula give rise to the 
greater total number of individuals is, to my mind, because they are 
marginal, In Crisia it is the lateral branches which are the most 
prolific, and for the same reason. 

The existence of the 7th rule in mat-like species is a mechanical 
necessity ; in the phytoid species, like Bugula and Crisia, it must be ac- 
counted for on another ground ; namely, on the relations of food supply 
to demand, — on the deterrent effects of overcrowding. And this, to my 
mind, is the key to the significance of the 4th, 5th, 6th, and 7th rules. 
The form of the stock is determined by the same law which has deter- 
mined the form of the individuals, — the struggle for existence and the 
survival of the fittest, —the fittest in the present case being those which 
are most advantageously placed with reference to food supply. Abundant 
food supply has made possible the rapid production of lateral individuals 
at the margin, and less abundant food supply has retarded such produc- 
tion in the middle. Therefore has lateral budding oceurred more rapidly 
at the margin; therefore has the number of individuals produced at the 
margin been greatest ; therefore have the median rows grown in length 
only with great rapidity ; therefore has the distance between adjacent 
rows of individuals in phytoid stocks remained constant. 

Many observations on different groups of animals agree in demonstrat- 
ing a relation between rapidity of the budding or fission process and food 
supply. Thus Zoja (90, pp. 25-27) has shown for Hydra, and Zacha- 
rias (86, p. 274) and von Wagner (’90, p. 360) for Turbellarians, that 
abundant food supply results in an acceleration of the processes of 
non-sexual reproduction, and Braem (90, p. 24) has shown that bud- 
ding in Cristatella proceeds less actively during the late fall. This 
diminution in activity has been attributed by Braem to diminished 
temperature ; but we know also that this period is one of scareity of the 
small fresh water organisms upon which the fresh water Bryozoa live 
(cf. Parker, ’90, pp. 597-600), and this fact also must be considered as 


having an important influence in this case. 


88 BULLETIN OF THE 


II. Relation of the Observations on Budding in Bryozoa 
to the Germ Layer Theory. 


No question in Bryozoan morphology has been more thoroughly dis- 
cussed than that of the part played by the germ layers in the production 
of the polypide, and upon none has there been less agreement. Nitsche 
first boldly opened the question, and concluded that we have in this pro- 
cess a fatal objection to the idea of the homology of the germ layers, in 
so far as their homology depends upon a similarity of fate throughout the 
Metazoa. A single layer, the invaginated ectoderm, gives rise to the outer 
covering of the tentacles, to the pharynx, and to the brain, — structures 
elsewhere considered as ectodermal, — and also to the lining of the ali- 
mentary tract, elsewhere universally accounted entodermal. In view of 
these facts, “sind die Keimblätter,” concludes Nitsche (75, p- 398), “ kei- 
neswegs mit einer besonderen histologischen Prädisposition ausgestattete 
Zellschichten, sondern lediglich die flächenhaft ausgebreiteten Elemente, 
aus denen die den Metazoenkörper zusammensetzenden, ineinander 
geschachtelten Röhren sich bilden.” Prouho, although recognizing the 
facts to be as stated by Nitsche, has not discussed the theoretical bearing 
of the question. Seeliger (’89", p. 204) finds in the budding process of 
Endoprocta a shortening and confusion of the embryonic process. “ Wie 
die gesammte Knospenentwicklung verkiirzt ist, erscheinen auch die 
beiden Processe der Kinstiilpung durch welche im Embryo zuerst Ento- 
dermkanal, dann Atrium sich bilden, in einen zusammengezogen.” In 
another place (Seeliger, 790, p. 595) the budding process is considered as 
an “immer sich erneuerende Gastrulationsvorgang.” Braem (’90, p. 116) 
regards the inner layer of the bud as entoderm, and the process of its 
formation as one of gastrulation. In a preliminary notice published last 
February (Davenport, 791, p. 279) I suggested that the embryonic tissue 
from which the inner layer of the polypide arises is to be regarded as 
“neither ectoderm nor entoderm, but as still indifferent, and capable of 
giving rise to either.” A few weeks ago I saw for the first time the paper 
of Oka (90), in which he offers (p. 145) a priori a similar suggestion 
concerning the significance of the embryonic tissue from which the inner 
layer of the polypides arises. I am pleased to find that our ideas, thus 
independently arrived at, are so fully in agreement. My idea of the re- 
lation of the germ layers to the layers of the polypide bud chiefly grew 
out of my studies on the embryology of Phylactolsemata as described in 
earlier pages. 

As there are two layers to the bud, the question of the part taken by 


MUSEUM OF COMPARATIVE ZOOLOGY. 89 


the germ layers in the polypide bud may be subdivided into two: What 
is the significance of the outer layer of the bud? and, What is the sig- 
nificance of the inner ? 

The outer layer of the bud is derived from the calomic epithelium. 
The views of those who have studied-the formation of this inner layer 
of the cystid in Phylactolamata may be classed in two categories : 
(1) those in which it is regarded as entoderm, and the process of its 
formation as gastrulation ; and (2) those in which it is regarded as 
mesoderm. To the former class belong the views of Reinhard (’80, 
p. 208), Korotneff (’89, p. 403), and Jullien (90, p. 19) ; to the latter, 
those of Kraepelin (86, p. 601) and Braem (90, p. 116), and in this 
class the views of Barrois (’86, p. 68) and Haddon (°83, p. 543), founded 
on a priori considerations, must be placed. 

Tt seems to me that, since, as Barrois has demonstrated, there is a great 
similarity between the Phylactolæmatous and Gymnolæmatous larvæ, 
and especially since the former show evident signs of degeneration, we 
are bound to study the phenomena they exhibit in the light of our 
knowledge of the ontogeny of Gymnolæmata. 

But first it is necessary to give reasons for believing that the larva of 
Phylactolæmata is to be regarded as homologous with that of Gymnolæ- 
mata; and to do this I will first name the points of similarity in the two 
larvæ, and then try to show that the differences which exist are not suf- 
ficient to invalidate the attempt to establish a homology. And, first of 
all, it may be said that, since the adult Phylactolæmata and Gymnolæ- 
mata are strikingly similar to each other, and since no one doubts their 
close relationship, we should expect a priori that their larvæ would be 
homologous, especially since the larvæ of Qymnolæmata are admitted to 
belong to the trochosphere type, of whose ancient origin there can be 
little doubt. In the second place, the very existence of a larval stage 
in Phylactolemata is indicative of its inheritance from an earlier condi- 
tion, for two reasons: (a) because in general fresh-water life tends to 
eliminate larval stages from species which have inherited them from ma- 
rine ancestors, and tends little to form them de novo (Hydra, fresh-water 
Turbellarians, Rotifera, Oligocheta, Hirudinea, Astacus, and fresh-water 
Mollusca); and (b) because, specifically, the early stages of development 
of Phylactolæmata are passed within a uterus-like sac, from which the 
embryo is released only when a colony is already well established. In 
the third place, the Phylactolematous larva possesses, in common with 
all Gymnolematous larvæ, the following characteristics. The primary 
poly pides arise in both at a pole, and this pole is in both a prominent disk, 


90 BULLETIN OF THE 


surrounded by a circular fold, — the so-called mantle fold, — from which 
it is separated by a circular groove, — the so-called mantle cavity. This 
organ has a similar origin and fate in the two groups, as shown by Barrois. 

The following points of difference, however, must be recognized. 
First, the absence of a definite ciliated ring, cowronne (Barrois), of an 
internal sac, and of a pyriform organ. But, as Barrois (’86, p. 67) has 
shown, these are absent, or at least (Ostroumoff, ’87, pp. 182, 183) little 
developed, in Cyclostomatous Bryozoa. The ciliated ring and pyriform 
organ are doubtless organs connected with a free locomotive larval life, 
which is greatly abbreviated in Phylactolemata. A second difference 
exists in the fact that, while most Gymnolematous larvae possess either, 
rarely, (1) a functional alimentary tract, or (2) a mass of loose tissue 
lying inside of the ectoderm, the Phylactolemata possess (3) a central 
space lined by an epithelium placed next to the ectoderm. However 
great the difference between the first and third conditions mentioned 
above, it is to a large extent bridged over by the widespread existence of 
the second. In some Cyclostomes, moreover, a similar condition to that 
in Phylactolemata seems to exist. Compare Metschnikoff (’82, p. 310, 
Taf. XX. Fig. 62). Lastly, the origin of two primary polypides, instead 
of one, at the aboral pole, upon which Barrois has laid some stress, can- 
not be considered a very strong objection to the homology, because in 
reality the two polypides do not arise at the same time even in Pluma- 
tella, and in Cristatella this difference is still more pronounced. In fact, 
it is not the formation of two polypides which requires explanation, but 
that of a young stock before hatching. 

There remains, therefore, to my mind, no serious objection to regard- 
ing the larve of Phylactolemata and Gymnolemata as having been 
derived from some common ancestral larva, possessing, of course, more 
points of resemblance to the Gymnolematous than to the Phylactolema- 
tous type; and therefore it is perfectly justifiable to interpret the latter 
by aid of the former. 

Admitting the larva to be homologous, we should expect the process 
of gastrulation to be comparable throughout Ectoprocta. As a matter 
of fact, we do find a great similarity in the »arliest stages. Thus, the 
first indication of the inner layer is the ingression of four cells at one 
pole, which by multiplication give rise to a layer of cells lying inside of 
the ectoderm.’ It is to the comparative study of the fate of this inner 


1 This has been shown for Membranipora (Tendra) by Repiachoff (’78, pp. 
416-420); for Alcyonidium polyoum by Harmer (’87, pp. 445, 446); for Bugula 
by Vigelius (’86, p. 519) ; and for Cristatella in the present paper (page 68). 


MUSEUM OF COMPARATIVE ZOÖLOGY. 91 


layer in the different Ectoproct larvæ that we must look for an explana- 
tion of the layer in the specific case of the Phylactolemata. 

For the purposes of this study, it is desirable to begin with species in 
which there has been a minimum amount of degeneration. Such are 
Membranipora (Cyphonautes), Alcyonidium, and Flustrella, to which 
we must now turn our attention. 

The studies of Repiachoff on Membranipora lead up to a stage in which 
the entoderm lies as a solid mass inside the ectoderm, and is separated 
from it at all points. Neither the origin of the mesoderm nor the forma- 
tion of stomodeum or proctodeum was observed at this time. As for 
the fully formed Cyphonautes, it is certain, as I can confirm from personal 
observation, that there is a well developed functional alimentary tract, 
and that it is provided with a well developed muscular system, including 
cross-striped muscle fibres. There is, therefore, every reason for believing 
that typical entoderm and mesoderm have been formed in it, 

In Alcyonidium (polyoum), Harmer (87, p. 445) has shown that after 
gastrulation a great mass of cells occupies the former blastocel. This, 
in the author’s opinion, represents entoderm and mesoderm. The young 
larva possesses a mouth, ‚an esophagus, and a large stomach, but never 
an anus. No evidence is presented that the oral pole corresponds with 
the pole of ingression. 

Flustrella, which is nearly related to the last species, possesses in 
its young larval stages a pocket, which Prouho (’90, pp. 424-426) has 
shown to represent the anterior part of the alimentary tract, directly 
comparable with that of Aleyonidium polyoum, but less developed. Mus 
cle fibres and an epithelial lining of the entoderm and ectoderm exist to 
indicate the presence of mesodermal tissue. 

These three genera, Membranipora, Aleyonidium, and Flustrella, are 
the only Ectoprocta in whose larvae the presence of an alimentary tract 
has as yet been demonstrated. 

In Bugula, a very careful study of which was made by Vigelius (86 
and ’88), one finds after gastrulation and cell multiplication a mass of 
cells filling the whole interior of the larval body, at first appearing as 
an epithelium surrounding a central space, but later without arrangement 


and often showing signs of degenerescence. No definite separate meso- 
derm could be found, and at no time was any trace of'an alimentary tract 
to be seen. Vigelius calls the mass derived from the four entodermal 
cells Füllgewebe, and he believes it to correspond morphologically to 
both “hypoblast and mesoblast.” Tt is to be noted, however, as a 
point of considerable importance, that in his figures of the metamorphos- 


92 BULLETIN OF THE 


ing larva Vigelius (88, Taf. XIX. Fig. 6) represents this tissne as hav- 
ing almost entirely disappeared ; that which remains giving rise to the 
mesodermal lining— the outer layer of the bud — of the developing 
polypide. 

There can be no doubt that the so-called oral pole of the Bugula larva 
corresponds to the mouth-bearing pole of Alcyonidium, but does it cor- 
respond to the pole of ingression of entoderm? This question has not 
been answered by Vigelius. The existence of homopolar stages like that 
represented in his (86) Figure 25, Taf. XXVI., makes it very difficult 
to establish this doubtful point. 

The formation of the inner layer of Cyclostomes has been studied by 
Barrois (’82, p. 141). He says: “Des les premiers stades les spheres 
vitellines glissent les unes sur les autres de manière à former une espèce 
de gastrula par épibolie et l’on ne tarde pas à rencontrer des stades d'un 
volume extrêmement exigu et déjà composés d’une couche exodermique 
et d’une masse endodermique libre dans son intérieur, La masse endo- 
dermique s’atrophie rapidement et lon arrive à une petite blastula qui 
succède non pas à un stade composé de cellules radiaires dans lequel 
se forme une cavité centrale, mais qui est issu, au contraire, d’une vraie 
gastrula née par épibolie dans les premiers stades de la segmentation 
et dans laquelle la masse endodermique est déjà disparue.” I have 
quoted Barrois thus at length, since his description will show forcibly at 
least one thing, that the fate of the cells which by ingression had entered 
the blastoccel is quite different from that of those in Bugula, where a 
great Füllgewebe is formed. Ostroumoff (87, p. 183), however, has 
shown that the inner layer of the Cyclostome larva does not disappear, 
but comes to line the ectoderm as a very thin layer. In the adult larva, 
however, we find the contents of the ectodermal sac “ filled with me- 
senchymatous cells, which are commingled with yolk granules and glob- 
ules of albumen.” It is these cells that produce the very considerable 
mesodermal layer of the first polypide, which arises in the metamorphosis 
of the larva. Here, as elsewhere, an apparently homopolar stage inter- 
venes between gastrulation and the formation of larval organs, making 
orientation difficult. 

Thus, passing from Cyphonautes, through Alcyonidium and Flustrella, 
Bugula, and finally Cyclostomes, we have a series in all of which the 
inner germ layer is derived from one pole by ingression or by ‘epiboly,” 
and in which there is a gradual reduction of the functional entoderm until 
it seems, in Cyclostomes, to be lost, and a gradual transformation of the 
mesoderm from a cell mass nearly filling the larva, and producing muscles 


MUSEUM OF COMPARATIVE ZOOLOGY. 93 


all and alimentary tract, to a single thin cell 


and a lining to the body w 
or to mesenchymatous cells extending 


layer lying next to the ectoderm, 
through the coelom. 

This same series may be said, also, to be one in which there is a 
gradual decline in the complexity of larval organs. These find their 
maximum development in the bivalve Cyphonautes and Flustrella, and 
the complicated and beautiful Aleyonidium larva. They find their mini- 
mum development in the Cyclostomes, whose larvae, instead of a girdle of 
flagella, possess merely an undifferentiated clothing of cilia, are reduced 
to a cylindrical or ellipsoidal form, lack the pyriform organ of other spe- 
udiment of the internal sac. 


cies, and in some cases possess only the r 
he degraded end of the 


If we were to imagine still another term at t 
series, it would be a form in which the four inner-layer cells that arise 
at one pole of the larva should give rise to little or abso- 
lutely no entoderm, in which the mesoderm should come to form an 
inner lining to the ectoderm, and in which the internal sac should be 
It is just these conditions which are fulfilled by the 


by ingression 


entirely absent. 
Phylactolematous larva. 

Of all these changes, the loss of the entoderm is the most striking. 
What can be said in explanation of it? I would suggest this hypoth- 
esis: that the entoderm of the Bryozoan larva has become rudimentary 
through loss of the alimentary function. 

In direct support of this hypothesis I have little experimental evidence 
to offer. One observation, however, which I made last summer, seems 
to favor this conclusion strongly. This is that larval life is of consider- 
able duration in Cyphonautes, which possesses a functional alimentary 
tract, but is very brief in Bugula, in which no alimentary tract arises. 
As is well known, Cyphonautes occurs in enormous numbers in the 
“tow ” at certain seasons of the year, and this is alone evidence of a con- 
siderable length of life. I have t: ken Cyphonautes thus obtained from 
the tow and have kept them for three or four days, at the end of which 
time they died, or had settled to the bottom of the glass vessel to un- 
dergo their metamorphosis. In fact, from several hundred Cyphonantes 
which I collected, not more than half a dozen completed their full meta- 
morphosis, the others apparently succumbing to unfavorable conditions. 


1 Just as the manuscript of this paper is going to the printer, after long delay 


caused by an accident necessitating the re-engraving of the plates, I find that Dr. 


Prouho read last summer (’90), before the Association Francaise pour l’Avancement 
de la Science, a preliminary communication on the development of Cyphonautes. 
This is published in the printed report of the proceedings of that association. The 


94 BULLETIN OF THE 


The Bugula larve, on the contrary, I have never found in the tow, but 
they swarm out from stocks gathered in the morning and placed in a 
glass vessel; and I can confirm Nitsche’s (69, p. 9) observation that they 
settle and begin their metamorphosis within “a few hours” after hatch- 
ing. One rarely or never finds these larve succumbing to the unfavor- 
able conditions of the aquarium before metamorphosing. From these 
observations I conclude that the Bugula larva has a very much shorter 
life than Cyphonautes. Now, since the larva, owing to its shortened life, 
has no need of functional entoderm, and since entoderm can be of use to 
the larva only, no part of it going over into the tissues of the primary 
polypide of the stock (except as food material), functional entoderm is 
not developed. In other genera, its rudiments have become less and 
less important in the ontogeny, and, finally, in Phylactoleamata are 
wholly lost. 

That the entoderm should reach its last stage of degeneration in 
Phylactolemata is easily understood when we consider that the larval 
period is passed in a closed oœcium, from the wall or neck of which it 
receives nourishment as a parasite docs. Moreover, by the delay in the 
period of hatching, as well as by precocious development of polypides, one 
at least of the latter is usually functional in the just hatched stock, for 
there is sometimes found at least one polypide in the newly hatched 
larva, which is partly extruded, and therefore capable of feeding, and 
thus of supplying the whole stock with nutriment. Of what advantage 
to a species could be the development of a functional larval entoderm, 
which should go to form no part of its adult tissue, provided the larva 
was contained in a uterus during its early stages, and was provided with 
the adult digestive organs in a functional condition before leaving the 
uterus 4 

Those who maintain that the inner layer is to be regarded as entoderm, 
and are still unwilling to place the Bryozoa among the Colenterata, must 
account for the absence of mesoderm. Korotneff (’89, p. 400) finds de- 
generating cells in the blastocosl before this is wholly obliterated by the 
extension of the inner layer. These he seems to regard as degenerate 
mesoderm. According to his view, then, the entoderm gives rise to 
the muscularis, — for this arises from the inner larval layer, according 


author does not there state whether stomodw#um and proctodæum are formed on 
the blastoporie side of the larva. He accounts for the existence of an alimentary 
tract in Cyphonautes by the fact that it undergoes its development disconnected 
with the parent, while almost all other Bryozoa pass their early stages in the 
parent or some protecting zoöid (ocecium, ovisac, ovicell). 


MUSEUM OF COMPARATIVE ZOOLOGY. 95 


to Braem’s (90, Taf. VII. Fig. 89 mb.) observations, which I can abun- 
dantly confirm, — and to the coelomic epithelium of the adult stock, In 
the few series of sections of the proper stage which I possess, I have not 
found with certainty the degenerating cells of which Korotneff speaks ; 
but even if they regularly occur, I should be inclined to regard them as 
the degenerated entoderm, the mesoderm persisting to give rise to the 
muscular tissue and the ccolomie epithelium. 

From a consideration of these facts, — that the larvee are homol- 
ogous and the process of gastrulation is comparable throughout the 
Ectoprocta, that in the least modified larvae both functional entoderm 
and mesoderm are produced by that gastrulation, that one of these 
two germ layers has become rudimentary in Phylactoleemata, that 
it is highly probable that the entoderm has disappeared from loss of 
function, and that the layer which persists gives rise to the muscula- 
ture, sexual cells, and coelomic epithelium, — I conclude that the inner 
layer of the Phylactolematous larva, and therefore the outer layer of the 
bud, is mesoderm, 

If we accept the point of view of Kleinenberg (86, pp. 1-19) and ad- 
mit the existence in general of only two layers, ectoderm and entoderm, 
a clearer conception of the modification undergone by the Phylactolema- 
tous larva may be gained. We may divide the entoderm arising in 
Bryozoa into two parts; viz. (1) that which gives rise to the lining 
of the midgut, as in Cyphonautes, and (2) all the rest of the inner 
layer. Now, since no midgut is formed in the Phylactoleematous larva, 
part (1) of the entoderm has ceased to be differentiated ; all which 
remains, then, is part (2); but this is equivalent to “ mesoderm” in 
the sense in which I have employed it, and therefore T am justified 
in saying that “mesoderm” only is produced. 

The question has now to be answered, What is the significance of the 
inner layer of the bud? Two different answers have been given to this ques- 
tion. It has been maintained, on the one hand, that it is to be regarded 
as ectoderm; on the other, as entoderm. There are serious difficulties 
in the way of accepting the first view, — so serious, in fact, that few 
authors have maintained it, although at first glance it seems to be re- 
quired by the facts. Although we have not yet sufficient grounds for 
declaring that organs formed by budding. must be built up from the 
same germ layers as corresponding larval ones, —although we may ad- 
mit that gemmigenesis recapitulates phylogeny and corresponds with 
ontogeny only in an imperfect and confused way, — still, from the expe- 
rience gained by tracing the development of hundreds of animals from 


96 BULLETIN OF THE 


the most widely separated groups of the animal kingdom, the idea that 
a functional alimentary tract is ever wholly derived from differentiated 
ectoderm will not be accepted by most embryologists without conclusive 
evidence. 

The second view is that the formation of the inner layer of the bud 
is a process of gastrulation, giving rise to entoderm, and that the so- 
called “ gastrulation ” of the sexual ontogeny of Phylactolaemata is to be 
regarded as a precocious ingression of mesoderm only. 

Two considerations are opposed to this view. In Membranipora there 
is a gastrulation which gives rise to the entoderm and mesoderm of the 
larva; and since the gastrulation of Phylactolemata is similar, these 
clements must be potentially present here also. The “gastrulation” 
in Bryozoa is a normal one; if there is any entoderm in the body wall 
giving rise to the inner layer of the bud, it must have been ento- 
derm which failed to become invaginated. But what, in the second 
place, is to be gained by assuming that the inner layer of the bud is 
formed from entoderm? Here is as great a difficulty as before, since the 
nervous system originates from this layer. It has been maintained in 
many cases that the nervous system arises from mesoderm, and Seeliger 
(89, p. 602) believes that it is formed from that layer in the non-sexual 
reproduction of some Tunicates; but I know of no good evidence of its 
origin in any of the Triploblastica from entoderm. 

Before going on to state my conception of the significance of the 
inner polypide layer, I desire to call attention to the conditions in the 
region at which it is first formed. I have shown above (page 69) that 
the primary polypide or polypides arise from the pole of ingression in 
Phylactolemata, and that therefore in this group the aboral pole (in the 
sense of Barrois) corresponds to the pole of ingression. As I under- 
stand Barrois, he means by oral pole merely the pole which in Cypho- 
nautes, for instance, bears the nõouth, — the pole also by which the 
larva attaches itself. Braem (90, p. 123, foot-note), however, interprets 
“oral side” in Barrois’s sense to mean in the last instance the place at 
which gastrulation takes place. Perhaps Barrois does somewhere state 
such to be the significance of his term (I have not found the place), 
but in that case I can only say that, to my mind, he has not produced 
sufficient evidence to prove that the oral pole of the larva of Gymno- 
læmata is the same as the pole of ingression in the gastrula ; nor, in my 
opinion, has any other investigator done so. Nearly all species studied 
have a stage early in their development when their poles are very sim- 
ilar, and orientation certainly would be exceedingly difficult. One of the 


MUSEUM OF COMPARATIVE ZOOLOGY. 97 


best figured series in which to trace the homology of poles is that shown 
by Repiachoff ('80, Taf IIT.) for Bowerbankia. So far as the figures 
go, one would conclude that Figure 10 A and its predecessors were 
oriented in the opposite direction to Figure 11 and its successors, 
which would result in placing the pole of ingression (Fig. 9) at the 
aboral pole of the larva, — the pole which here, as in all other Gymno- 
lemata, and, I believe, in Phylactolemata also, gives rise to the primary 
polypide. I have given above additional evidence for this conclusion, 
in my argument to prove the homology of the larvee and larval organs in 
Phylactolemata and Gymnolemata. 

The polypides arise in Phylactolemata at the pole of ingression, which 
is probably homologous with the aboral pole of Gymnolemata. The pole 
of ingression, or the region of the lips of the blastopore, must be regarded 
as being a region of less pronounced differentiation than the rest of the 
Its cells cannot be said to be either ectodermal or ento- 


gastrula. 
It is an interesting fact, that it is just these indifferent cells 


dermal. 
— not yet either ectoderm or entoderm — that give rise to the inner 
layer of the polypide, from which organs usually considered ectodermal 
as well as those considered entodermal arise. 

My conclusion, then, the objections to which I fully realize, may be 
stated in the following words: The inner layer of the polypide bud is 
composed of cells derived from the rim of the blastopore. Such cells are to be 
regarded as still indifferent, and as first becoming differentiated into ecto- 
derm and entoderm in the formation of the young polypide. 

Just when and where, on this hypothesis, the differentiation into 
ectoderm and entoderm occurs, is an important question ; but unfor- 
tunately I cannot answer it decisively. Tt may be pointed out, however, 
that it has now been shown for most Ectoprocta that the lining of the 
middle part of the alimentary tract is formed independently of the 
œsophagus, and by an actual or potential outpocketing of the primitive 
simple sac of the bud. In Endoprocta there is a similar outpocketing, 
which, however, arises in connection with the esophagus, and is formed 
independently of the rectum. 

This is perbaps the proper place to ‚all attention to the fact that the 
mesodermal outer layer of the bud has a very embryonic character at 
the budding region. This is indicated by the fact, that in Phylactolemata 
(in which group alone I have studied the subject) eges always arise 
from that part of the cœlomic epithelium which lies in the budding 
region (cf, Plate XI. Fig. 93). In Pyrosoma, also, according to the 
researches of Seeliger (789, pp. 598-602) the mesoderm of the budding 


VOL. XXII — NO. 1. 


98 BULLETIN OF THE 


region, the stolon, gives rise to eggs. The same condition seems to exist 
in other Tunicates. 


III. On some Characteristics of Gemmiparous Tissue. 


In the preceding part of this paper the words “embryonic tissue,” 
“undifferentiated tissue,” have often recurred, and they are terms in 
wide usage in modern zodlogy. I do not know of any attempt to define 
further the real character of this tissue, nor to give its more detailed 
characteristics, other than that usually employed in the term plasma- 
reich, or “rich in plasma.” The persistence of yolk granules is, as 
Nussbaum (’80, pp. 2-14) and Goette (75, pp. 31, 32, 831) have shown 
in the case of amphibian embryos, indicative of the embryonic condition 
of cells, when these have been derived from an egg filled with yolk. 

It is very far from my purpose to go into a detailed discussion of the 
significance of embryonic tissue, for which I am not yet fitted ; neverthe- 
less, I wish to call attention to the minuter characters of gemmiparous 
tissue as I have found it in Phylactolemata and Paludicella. I have 
described it in some detail in preceding pages. 

First, then, gemmiparous tissue seems to stain more deeply than non- 
gemmiparous tissue in the same section. This character has been re- 
peatedly observed before by others, and Braem calls attention to it 
several times. T have already described how I found, by the use of 
high powers, that much of this depth of stain was due to the unusually 
large number of deeply staining granules scattered through the cell, but 
chiefly gathered about the nucleus (Figs. 6, 17, 18, ete.). So marked 
is the greater depth of the stain around-the nuclei, that, with a power so 
low that the nuclei are hardly distinguishable, their position is indicated 
by a deeply staining band. 

Secondly, gemmiparous tissue, as I have found it in the cases referred 
to, is distinguished by the possession of large cells, nuclei, and nucleoli, 
I had already noticed this fact in my studies on budding in Cristatella, 
and [ find that Braem has figured the nuclei in the budding region as 
larger than the average (cf. Braem, ’90, Taf. VII. Figs. 86, 88-90). 
My own figures show this repeatedly (Plate I. Figs. 3, 4, 5, 6, Plate II. 
Figs. 15, 17, Plate XI. Fig. 99, ete.). I have also noticed this to a 
certain extent in the marine Bryozoa, but, since the cells of the latter 
are smaller, and as I did not succeed in obtaining from them sections so 
satisfactorily stained, the results are not so reliable. In attempting to 
obtain an explanation of this phenomenon one involuntarily recalls to 


MUSEUM OF COMPARATIVE ZOOLOGY. 99 


mind the condition in young egg cells, where the nucleus attains a rel- 
atively enormous size. This great size of the nucleus in young egg cells 
is explained by Korschelt (’89, p. 92) as due to its participation in the 
trophic activity of the cell: “Sein grösster Umfang fällt in die Zeit des 
energischen Wachsthums der Eizelle.” So in the gemmiparous regions 
the large size of the nuclei must be considered as connected with the 
growth of the cells. 

But if the growth of the cells is accompanied by a rapid ingestion of 
food material (which the larger nucleus implies), some evidence of that 
fact should be observed in the cells themselves in the presence of food 
granules. Such food material in rapidly growing ovarian egg cells lies 
near the nucleus, Stuhlmann (’87, pp. 13, 14) describes such a condi- 
tion in the ovary of Zoarces. “Neben dem Keimbläschen, jedoch ein 
klein wenig von seiner Membran entfernt, bilden sich an verschiedenen 
Stellen jetzt eigentiimliche Verdichtungen des Protoplasmas, die sich 
ein wenig stärker mit Saffranin färben als das Zellplasma.”” Such a 
thickening of the protoplasma is represented in the figures as minute 
granules, Korschelt (’89, pp. 123-125) mentions several other such 
instances. 

It has seemed to me possible to interpret the stainable granules lying 
near to the nucleus in gemmiparous tissue as such food material,’ par- 
ticularly since we know that food material döes exist in the ccelomic 
epithelium lying next to the cells which are about to divide rapidly and 
to give rise to the inner layer of the polypide. That food is being taken 
in by the inner layer cells from the calomic epithelium is indicated by 
the fact that the nuclei of the former cells lie near the latter epithelium 
(cf. Figs. 15, 17, 18, 28, 56, ete.) ; for, as Korschelt has shown, the nu- 
cleus tends to move towards the centre of activity of the cell. That these 


1 Granules similar to these appear to exist in the protoplasm of all cells. It 
is their extraordinary abundance in the geminiparous tissue upon which I lay 
stress. They have been variously interpreted by different authors. Bütschli (’88, 
pp. 1469-1472) describes various kinds of stainable granules in Ciliata which are 
food products, and the general character of which accords with that of the gran- 
ules referred to above. ‘Excretion granules” of Ciliata do not stain, according 
to this author, which is an indication that the bodies in gemmiparous tissue are not 
such. Iam particularly struck by the fact that the food products of Protozoa are 
chiefly found in parasitic forms, — Gregarinid® and parasitic Ciliata. These take 
up food in solution from their hosts exactly as the cells of the body wall of Bryo- 
zoa do from the body cavity. Altmann (’90) has recently interpreted similar 
deeply staining granules in other cells, as “die Elementarorganismen.” I can 
see no reason, on Altmann’s theory, for the peculiar distribution of the granules 
that I have found. 


100 BULLETIN OF THE 


granules observed in the cells are food material is indicated by their 
abundance in cells lying next to the reticulated cells of the calomic 
epithelium (Figs. 6, 28, 56). 

My conclusion, then, is this: Gemmiparous tissue is a rapidly as- 
similating tissue, possessing large nuclei because actively assimilating, and 
staining deeply because full of food material. 

While for Nussbaum, as already quoted (page 71), “indifferent cells” 
are essential to the reproduction of individuals by non-sexual as well as 
by sexual methods, Seeliger (90, p. 596) has concluded that “ die Vor- 
gänge bei der Knospung der Bryozoen uns zeigen, wie histologisch sehr 
bestimmt differenzirte Gewebe einen ganz embryonalen Charakter wie- 
dergewinnen können. Mehr noch als bei der normalen Knospung am 
freien Stockende ist dieses Vermögen bei der Regeneration der Polypide 
der Ektoprokten oder der Köpfchen der Pedicellinen ausgebildet. In 
diesen Fällen schen wir ein plasmaarmes, äusserest feines Plattenepithel, 
das über sich eine mächtige Cuticula ausgeschieden hat, sich in kubische 
und cylindrische plasmareiche Zellen zurückverwandeln und durch eine 
Einstülpung ein neues Polypid bilden, in welchem schliesslich die man- 
nigfachsten Gewebsformen vertreten sind.” 

It seems to me that many facts in the budding of Bryozoa are strongly 
in favor of Nussbaum’s hypothesis, On this assumption, we can best 
understand why in Cristatella there is not an invagination of the ecto- 
derm, and why instead a stolon is formed in the embryo, which passes 
along at the base of the ectoderm and at intervals gives rise to the 
inner layer of the body wall. I believe it is becanse the outer layer 
of the body becomes so rapidly differentiated by the secretion of the 


1 Other observers describe gemmiparous tissue as being either rich in food or 
deeply staining. Seeliger (’85, p. 588) speaks thus of the mesodermal gemmiparous 
tissue in Salpa: “Die einzelnen Zellen sind grossblasig, enthalten einen runden 
Kern und führen Oel- und Fettsubstanzen die als Reservematerial beim Aufbau 
des embryonalen Leibes weiterhin in Verwendung gelangen.” Von Wagner (’90, 
p. 377) says of the indifferent cells which are being transformed into the new 
pharynx of dividing Microstoma : “ Dieselben nehmen an Grösse zu, . . . indem 
gleichzeitig ihre Protoplasmaleiber feinkörnig granulirt und für Farbstoffe imbibi- 
tionsfähiger werden.” 

In some sections of gemmules of Esperella fibrexilis, H. V. Wilson, of which 
Dr. Wilson has very kindly sent me several slides, I find the outer layer of young 
gemmules, in which the inner layer has been newly formed, stained very deeply. 
Observed with a Zeiss Apochr. 4.0 mm., Ocs. 8 and 12, the cell contents are seen to 
be evidently of two kinds, — light and deeply stained. The latter appearance is 
due, in part at least, to small dark granules, which can be discerned without much 
difficulty. 


MUSEUM OF COMPARATIVE ZOÖLOGY. 101 


gelatinous balls in its cells as to be incapacitated for the work of build- 
ing organs. In Plumatella the outer layer of the body wall, which is 
derived, as Braem has shown, from the neck of the older polypide, re- 
tains for a long time its embryonic condition, so that its deeper cells can 
and do go to form the inner layer of the polypide bud. 

On Nussbaum’s hypothesis we can best understand why in Paludicella 
the Anlagen of the lateral branches exist from the beginning as cuboidal 
cells, quite different from those of the rest of the body wall; we can 
understand why the cell layers of the margins of the stock, the tips of 
branches, and the ends of stolons from which buds arise, are thicker 
and more rapidly dividing than the rest of the body wall (cf. Figs. 14, 
71, 73, 75); and we can also understand why the regenerating buds 
always arise from the region of the neck of the degenerated polypide, 
— the same region from which that degenerate polypide had arisen by 
budding. 

There is no doubt, however, that at times buds do arise from tissue 
which, as Seeliger says, has lost its cuboidal nature only to regain it. 
From such tissue apparently the polypide of Figure 79 has arisen ; 
from such tissue certainly, as Secliger says, do regenerating polypides 
arise. But is the process by which cuboidal cells become a pavement 
epithelium one of so fundamental differentiation that, in accordance 
with Nussbaum’s doctrine, we should not expect, under favo rable condi- 
tions, to see these cells regain their cuboidal form? No doubt we have 
many other cases in the animal kingdom in which flat epithelial cells 
regain their cuboidal form. Thus, for instance, among the Bryozoa, 
Oka (90, p. 182) has shown how the flat cells of the outer layer of the 
statoblast begin to thicken again at the return of warmth, and at the 
beginning of the active assimilative processes, not only at the pole from 
which the primary polypide is to arise, but also opposite to this. 

Many facts indicate that cells may beeome flattened epithelia, and yet 
not lose their embryonic character. Maas (790, pp. 541-54 4) has re- 
cently shown step by step how the columnar ectoderm of the fresh water 
sponge is forced, on account of the great increase in surface which it is 
called upon quickly to cover, to become broad and flat. It finally gives 
rise to an epithelium so flat that its existence was long overlooked, and 
has been denied by so competent an observer as Goette ; and yet in its 
flattened condition it possesses to a remarkable degree the capacity of 
sending out pseudopodia-like processes, a condition indicative much less 
of a high degree of differentiation or specialization, than of an unspecial- 
ized, primitive or embryonic condition, 


102 BULLETIN OF THE 


I have already stated (page 65) that the region from which the re- 
generating buds of Cheilostomes arise, although one of flattened epithe- 
lium, is one in which many more nuclei persist than elsewhere in the 
adult (cf. again Fig. 71). This fact, coupled with the constancy of 
position of regenerating buds with reference to the degenerated polypide, 
is to my mind evidence against the assertion that buds arise here from 
“histologically very definitely differentiated tissue.” 

As for regeneration in Endoprocta, no one is more competent to speak 
than Seeliger himself. I am the more surprised, therefore, to find that 
in Ascopodaria macropus, which is quite closely allied to the species 
studied by Seeliger (°89), the cells at the part of the stalk immediately 
below the ‘“ head,” from which regenerated buds arise, are, as Elhlers’s 
magnificent Pedicellina work shows, very large and cuboidal (Ehlers, 
90, Taf. II. Figs. 26-33). I think one may conclude that a similar 
condition obtains in some cases in Pedicellina, even judging from Seeli- 
ger’s own drawings, although they are drawn to a scale that is not quite 
large enough to allow of settling this point (Seeliger, ’89, Taf. X. Fig. 
35, a, Fig. 41, etc.). 

If the increase in size of the flattened cells, and their subsequent rapid 
division and invagination to form a bud, are due to their more active nour- 
ishment, it would be difficult to see why certain cells of any region should 
quickly undergo this modification, while the adjacent cells apparently 
as favorably situated with reference to the acquirement of food retain 
their flattened, quiescent condition, if we assumed such favorable situation 
to be the only requisite. Still less satisfactorily would such an assump- 
tion explain the regular position of regenerating buds. It is taking only 
one step farther back, but, to my mind,.a helpful step, to assert that cell 
proliferation in any region which produces invagination depends upon 
the capacity of the cells of that region to become better nourished than 
their fellows. This may evidently be effected by a diminution in the 
feeding capacity of the surrounding cells, or by an increase in this respect 
in the growing cells. 


IV. Relationships of Endoprocta and Ectoprocta. 


I discussed this topie in my earlier paper (Davenport, ’90, pp. 132, 
33). I have only to add, that later studies have confirmed my 
opinion of Nitsche’s correctness in placing these two groups close to- 
gether, and in regarding the Endoprocta as nearer the ancestral types. 
The stages of Figures 25 (Plate ILI.) and 77 (Plate IX.) probably rep- 


MUSEUM OF COMPARATIVE ZOÖLOGY. 108 


resent roughly a phylogenetic stage ancestral to both groups of Bryozoa, 
but most clearly allied to adult Endoprocta. The formation of new ten- 
tacles anteriorly and posteriorly in Figure 77; would reproduce the adult 
Endoproct condition. Two changes lead to the Ectoproct stage: first, 
the closure of the tentacular corona posteriorly in front of the anus (Plate 
V. Fig. 43), and, secondly, the formation of the pharynx or anterior 
part of the oesophagus by the growth of the oral tentacles over the floor 
of the atrium towards the atrial opening. Thus the brain, which lies at 
the floor of the atrium in Ectoprocta, comes to lie on the pharynx. The 
pharynx would, upon this assumption, be a new structure, not found in 
Endoprocta. Such appearances as are exhibited by Figure 77 lead me 
to retract, my former expressed opinion, in which I agreed with Nitsche in 
saying that the earliest condition of the tentacular corona is a U-shaped 
one. Rather, the tentacles are formed first on each side of the atrium, 
and only secondarily grow around the mouth in front, as later they grow 
in between mouth and anus. The U-shaped stage is therefore not the 
primary one, but secondary. 

The close relationship of Endoprocta and Ectoprocta has recently been 
doubted by Cori ('90, p. 16), but his chief argument depends upon the 
dissimilarity of the Endoproct and Ectoproct kidney. Unfortunately, 
our knowledge of the latter is still very imperfect, and we may well hope 
for renewed researches in the subject by this skilful investigator. 

Ehlers (90, pp. 149-154) has recently re-expressed his former (76, p. 
132) utterances concerning the lack of homology between the tentacles 
of Ectoprocta and the “ cirri” of Endoprocta, He finds the homologue 
of the latter in the “ Diaphragma” or “ Kragen ” of Ectoprocta. This is 
the organ which I have believed to be homologous with Kraepelin’s 
“Randwulst” (which may be Anglicized as marginal thickening), — an 
organ occurring in all Ectoprocta. It is nothing but the “neck of the 
polypide,” which has sunk below the general level of the body wall. It 
is always provided with sphincter muscles, and in Ctenostomes forms the 
base of insertion of the cylindrical or comb-like “ collare setosum.” It 
can hardly be that Ehlers refers to this latter structure by the term 
“Kragen,” since this is merely cuticular. In my opinion the “ Dia- 
phragma” of Nitsche cannot be homologized with the cirri of Endo- 
procta, because it is merely a part of the body wall comparable to that 
part from which the “ polypide ” of Endoproctous Bryozoa arises, and 
beneath which the tentacles or “ cirri” arise. This part of the body wall 
is provided with a sphincter in Endoprocta as well as Ectoprocta, and by 
it the atrial cavity may in both cases be closed. 


| 
i 
| 
| 
| 
| 


104 BULLETIN OF THE 


To my mind, the most significant difference between the two groups 
exists in the fact that the outpocketing to form the stomach arises from 
the oral end of the future alimentary tract in Endoprocta, and from the 
anal end in Ectoprocta. One is led to believe that in the ancestral form 
either two nearly equally important outpocketings from both the oral 
and anal sides existed, or that the two existing methods are remnants 
of a method different from either (such as the formation of the whole 
alimentary tract at once), or, finally, that the Endoproct condition repre- 
sents the ancestral one, and that the rectal evagination has secondarily 
become of greater importance in Ectoprocta, and that the oral evagina- 
tion has become less significant. Oka (90, pp. 134, 141) has recently 
asserted that in the polypide buds of the statoblast and adult colony of 
a Pectinatella of Japau (P. gelatinosa) the cesophagus and stomach are 
formed by one evagination, which acquires secondary connection with 
the rectum, This condition reminds one, then, of Endoprocta., I must, 
however, doubt the accuracy of Oka’s conclusions until more satisfactory 
evidence is forthcoming; the more so, since Pectinatella magnifica, 
Leidy, presents a method of budding exactly comparable to that in 
Cristatella and Plumatella, as my own sections show with sufficient 
clearness. 

The homology of the Ectoprocta and Endoprocta implies a homology 
of their larvae, and demands that the life history of the two groups should 
be directly comparable. 

It is well known from the researches of Hatschek (77) on Pedicellina, 
and of Harmer (85) on Loxosoma, that the surface of the larva which 
bears the mouth and anus, i. e. its oral side, corresponds with that of the 
blastopore. How, then, is the oral aspect of the Ectoproct larvee, which 
I have tried to show is opposite to the pole of the blastopore, to be 
homologized with this? 

The month and anus of the Endoproct larva undergo a rotation after 
the larva has settled, so that they come to occupy the pole opposite to 
that at which the blastopore was. This stage of the Endoproct larva 
is comparable to the whole larval stage of Ectoprocta. I believe the 
two stages to be homologous, and that, just as polypides are pre- 
coctously formed upon the Phylactoleematous larva, its larval digestive 
tract having dropped out from the ontogeny, so the mouth and anus of 
Gymnolemata are precociously formed on the pole opposite the blasto- 
pore, the primitive stage during which they existed at the blastoporic 
pole having dropped out of the ontogeny. 

It is well known from the works of Harmer (86) and Seeliger (89), 


MUSEUM OF COMPARATIVE ZOOLOGY, 105 


that in Pedicellina, in which the metamorphosis of the larva has been best 
studied, the stolon arises from the base of the stalk — that is, at the pole 
where mouth and anus were first formed — at the pole of invagination. 
I have shown that this is true for Phylactoleemata, and probably for 
Gymnolamata. 

If the interpretation which I have put on Gymnolamatous ontogeny 
becomes confirmed, the larvæ and the budding areas will be homologous 
throughout all Bryozoa. The following diagrams will explain my idea 
of the relation of the different ontogenetic stages in the two groups. 


ENDOPROCTA. ECTOPROCTA. 


The left hand vertical series represents stages in the development of 
Endoprocta ; the right hand one, stages of Ectoprocta. The blastopore a 
is throughout turned upwards in the figures. Stage I. is in both cases ¢ 
young gastrula. Stage IT. is that of the free-swimming larva of Endo- 
procta, This stage is lost in the ontogeny of Ectoprocta, in which, by 
abbreviation of larval life, the free-swimming stage corresponds to the 
condition of the fixed Endoproct after it has undergone its rotation. 
This stage, or one slightly later, is shown in III. Both larvæ are fixed, 
the Endoproct by the blastoporic, the Ectoproct by the opposite pole. 
The position of the stolon, or of the first polypide of the colony produced 
by non-sexual methods, is represented at gm., near the blastoporic pole. 


BULLETIN OF THE 


Summary. 


The following general scheme of the budding process in Ectoprocta, 
derived from my own and other recent studies, may be now drawn up. 
The references are to pages of this paper. 

All Ectoprocta build stocks or corms. The individuals in these are 
arranged in rows radiating from a centre, — the larva, or statoblast, — 
and are placed one in front of another (Figs. 2, 64", 65%, 67, 71%, etc.). 

New rows or branches are constantly being produced peripherally. 
There is no dichotomy in the branching (page 86), but the ancestral or 
median branch gives rise to one or more lateral branches, which in turn 
become median branches of their part of the stock. 

The body wall and polypides of the median branch, as well as the 
Anlagen of lateral branches, arise from a pre-existing mass of embry- 
onic tissue, the gemmiparous mass (pages 72-82). This may exist cen- 
trally of the forming region, as in Phylactolemata, or peripherally, as in 
Gymnolemata. 

The anal aspect of the polypide is turned towards the gemmiparous 
mass (page 82). 

The outer layer of the body wall in the budding region is one 
of rapidly assimilating and rapidly dividing tissue ; the inner layer of 
the body wall becomes filled with food taken from the body cavity in 
species in which the latter is early cut off by a partition (Paludicella, 
Bowerbankia, Lepralia ?) ; it shows no tendency to do so in species with 
a coonocal (Phylactolsemata, Alcyonidium). 

The first impulse to the formation of the polypide is found in the 
outer layer of the body wall (excepting when this is highly modified, as 
in Cristatella), and many cells seem to be involved in its formation from 
the beginning (pages 8, 56). 

This outer layer of the body wall is embryonic tissue, derived from the 
tip of the stock (margin of the corm) as in Gymnolemata, or from the 
neck of pre-existing polypides, as in Phylactolemata, It is the direct de- 
scendant of the gemmiparous tissue of the larva, which in turn has been 
derived from the region around the blastopore, — in Phylactolæmata cer- 
tainly, in Gymnolemata probably (pages 8, 11, 12, 69). 

The inner layer of the body wall is also embryonic in the budding 
region, as indicated by the fact that ova arise near the neck of the 
polypide, in Phylactolemata at least (page 68). 

The outer mural layer becomes the inner bud layer by invagination, 


with or without the formation of a cavity. In the former case (many 


y d 


MUSEUM OF COMPARATIVE ZOÖLOGY. 107 


iymnolaemata) the mouth of the invagination pocket rapidly closes to 
give rise to the neck of the polypide (page 56). In the latter case, the 
cavity of the bud arises only secondarily by a separation of its walls 
(page 18). 

By a rapid growth of the walls of the bud, its distal part, in which the 
alimentary tract is to arise, is formed. Since this rapid growth occurs 
earlier at the anal side than at the oral, the rectum is formed first, the 
stomach last (pages 19, 57). 

By an approximation of the lateral walls, alimentary tract and atrio- 
pharyngeal cavity become separated. 

The œsophagus arises as a pocket of the atrio-pharyngeal cavity, and 
secondarily unites with the stomach (pages 19, 58). 

The lophophore arises first as two lateral thiekenings of the atrio- 
pharyngeal wall, then as two lateral folds, whose cavity becomes the ring 
canal (pages 20, 58), 

Tentacles appear on the ridge of the lophophoric fold thus established, 
and like it are formed first at the sides of the polypide, then anteriorly 
and posteriorly (pages 22, 59). 

The posterior end of the lophophoric ridge is the last to be formed, 
and, in forming, it cuts off the anal part of the atrium from the inter- 
tentacular cavity (pages 23, 62). 

The compressed intertentacular cavity becomes circular by change in 
position of the oral tentacles (pages 24, 62). 

The ganglion arises as a depression in the floor of the intertentacular 
room, and becomes included in the pharynx, which is differentiated by 
the change in position of the oral tentacles (pages 26, 61). 

Muscles and funiculi arise from the coelomic epithelium of both the 
body wall and the bud (pages 27-31, 63). 

The neck of the polypide may sink to a considerable distance below 
the general level of the body forming the “ Randwulst” of Phylactole- 
mata or “ Diaphragma ” of Gymnolemata (pages 31, 63, 103). 

The atrial opening first arises at a late period by separation of the 


cells of the neck. 


The communication plate arises in Paludicella as a circular fold 
of the layers of the body wall, the mesodermal cells at the centre of 
which become cyticularized. It is not so completely closed as to pre- 
vent communication between the colomata of the two individuals it 
separates. 

The mesodermal cells of Paludicella become stored with food mate- 


108 BULLETIN OF THE 
rial before the formation of the communication plate, and yield it up to 
the rapidly growing bud. 

The regenerated polypides, like the marginal ones, arise in Choilo- 
stomes in a definite position, —on the wall of the operculum from tissue 
left behind to give rise to the polypide, but not wholly used up in its 
formation. They arise wholly from the body wall, come to lie next to 
the “ brown body,” and cause its disintegration. 


The more important theoretical conclusions to which I have arrived 
are :— 

a. There is in every stock or corm of Bryozoa a mass of indifferent 
cell material, which is derived directly from the indifferent cells of the 
larva or embryo, and whose function is to form the organs of the different 
individuals, including the polypides. This mass by constant growth and 
division affords the embryonic material for lateral branches, 

d. The form of the stock and interrelation of individuals is in large 
part controlled by food supply. 

c. The inner layer of the Phylactolamatous larva represents meso- 
derm only: the entoderm has become rudimentary through loss of the 
alimentary function. 

d. The polypides arise in Phylactolamata at the pole of ingression, 
which is probably homologous with the aboral pole of Gymnolemata, 

e. The inner layer of the polypide bud is composed of cells derived 
from the rim of the blastopore, and they are to be regarded as still 
indifferent, and as first becoming differentiated into ectoderm and ento- 
derm in the formation of the young polypide. 


f. Gemmiparous tissue is a rapidly assimilating tissue possessing 
large nuclei because actively assimilating, and staining deeply because 
full of food material. 

g. The Endoproct and Ectoproct larvae are to be compared by assum- 
ing that the act of rotation of the axes occurring in the former has been 
leaped over in the ontogeny, the mouth and anus arising at once on the 


pole opposite the blastopore. 


CAMBRIDGE, Mass., June 1, 1891. 


MUSEUM OF COMPARATIVE ZOÖLOGY. 109 


LITERATURE CITED. 


Agassiz, A. 
>63. On Alternate Generation in Annelids, and the Embryology of Autolytus 


cornutus. Bost. Jour. Nat., Hist., VII., p. 384. 


Allman, G. J. 
’56. A Monograph of the Fresh-water Polyzoa. viii + 119 pp., 11 pls. 
London, Ray Society. 
70, The Genetic Succession of Zoöids in the Hydroida. 
Edinb,, XX VI. 1, p. 97. 
Altmann, R. 
90. Die Elementarorganismen und ihre Beziehungen zu den Zellen. Leip- 


zig, Veit & Comp. 145 pp., 21 Taf. 


Trans. Roy. Soc. 


Barrois, J. 
77. Recherches sur l’ömbryologie des Bryozoaires. Travaux Inst. Zool. de 


Lille, I., 305 pp., 16 pls. 

82. Embryogénie des Bryozoaires. Essai, ete. 
XVIIe Ann., p. 124. [Transl. in Aun. and Mag. Nat. Hist, (6), X., 
p. 265.] 

’86. Mémoire sur la métamorphose de quelques Bryozoaires. Ann. Sci. 
Nat., Zool., (7), 1.1, p.d. 


Braem, F. 
'88. Untersuchungen über die Bryozoen des süssen Wassers. Zool. An- 


zeiger, XI., No. 288, p. 503; No. 289, p. 533. 
90. Untersuchungen über die Bryozoen des süssen Wassers. Bibliotheca 
Zoologica, Heft 6. 184 pp., 15 Taf. 
Biitschli, O. 


'88. Protozoa. 
Abth. III., Infusoria, etc., pp. 1281-1584. 


Jour. Anat. et Physiol, 


Bronn’s Klassen und Ordnungen des Thier-reichs, Bd. T., 


Chun, C. 
83. 


ausgeführte Reise. 


Bericht über eine nach den Canarischen Inseln im Winter 1887/88 
Sitzungsber. Akad. d. Wiss. Berlin, XLIV., p. 1141. 
Claperede, E. 
‚70. Beiträge zur Anatomie und Entwieklungsgeschichte der Seebryozoen. 
Zeitschr. f. wiss. Zool, XXI. 1, p 137 


Cori, C. J. 
‚90. Ueber Niereneanälehen bei Bryozoen. Totos, XT. 18 pp. 


110 BULLETIN OF THE 


Davenport, C. B. 
90.: Cristatella: The Origin and Development of the Individual in the 


Colony. Bull. Mus. Comp. Zoöl. at Harvard College, XX. 4, p. 101. 
'91, Preliminary Notice on Budding in Bryozoa., Proc. Amer. Acad. Arts 
and Sciences, XXV., p. 278. 
Dumortier, B. C., et Beneden, P. J. van. 
'50. Histoire naturelle des polypes composés d’eau douce, ete. Bruxelles, 
M. Hayez. 130 pp., 6 pls. 
Ehlers, E. 
76. Hypophorella expansa. Ein Beitrag u. s. w. Abhandl. d. königl. 
Gesellsch. d. Wiss. zu Göttingen, XXI. 156 pp., 5 Taf. 
90. Zur Kenntniss der Pedicellineen. Abhandl. königl. Gesellsch. d. Wiss. 
zu Göttingen, XXXVI. 200 pp., 5 Taf. 
Eisig, H. 
’87. Monographie der Capitelliden des Golfes von Neapel. Fauna u. Flora 
Golfes v. Neapel, XVI. 906 pp., 37 Taf. 
Farre, A. 
137. Observations on the Minute Structure of some of the Higher Forms of 
Polypi, etc. Philosoph. Trans. Roy. Soc. London, Year 1837, I., p. 387. 
Freese, W. 
’88. Anatomisch-histologische Untersuchung von Membranipora pilosa L. 
u.s. w. Archiv. f. Naturgesch., LIV. 1, p. 1. 
Goette, A. 
75, Die Entwicklungsgeschichte der Unke. Leipzig, Leopold Vose. 964 
pp.» 22 Taf. 
Grant, R. E. 
‚27. Observations on the Structure and Nature of Flustre. Edinburgh New 
Philos. Jour., (2), III., pp. 107, 337. 


Haddon, A. C. 
’83. On Budding in Polyzoa. Quart. Jour. Mior. Sci., XXIIT., p. 516. 


Harmer, S. F. 

‚85. On the Structure and Development of Loxosoma. Quart. Jour. Mier. 
Sci., XXV., p. 261. 

86. On the Life-History of Pedicellina. Quart. Jour. Micr. Sci, XXVII., 
p. 239. 

'87. Sur lembryogénie des Bryozoaires ectoproctes. Arch. Zool. expér. 
et gen. (2), V., p. 443. 

’91. On the British Species of Crisia. Quart. Jour. Mier. Sei, XXXIT. 2, 
p. 127. 


Hatschek, B. 
77, Jimbryonalentwicklung und Knospung der Pedicellina echinata. Zeit- 


schr. f. wiss. Zool., XXIX. 4, p. 502. 


MUSEUM OF COMPARATIVE ZOÖLOGY. tli 


Hincks, T. 
’80. A History of the British Marine Polyzoa. 2 vols, 8°. (11) + exli + 


601 pp., 83 pls. London, John van Voorst. 

Joliet, L. 
"77. Contributions 
Arch. Zool. expér. et gen., VI. 2, p. 193. 


à Phistoire naturelle des Bryozoaires des côtes de France. 


Joyeux-Laffuie, J. 
'88. Description du Delagia cheetopteri (J. J.-L.). 
gen., (2), VL, p. 135. 
[This is the Hypophorella expansa of Ehlers, ’76.] 


Arch. Zool. expér. et 


Jullien, J. 
90. Observations sur la Cristatella mucedo, G. Cuvier. Mém. Soc. Zool. 


de France, III., p. 361. 
Kennel, J. v. 
'82. Ueber Ctenodrilus paradilis, Clap. Ein Beitrag u. s. w. Arb. zool.- 
zoot. Inst. Würzburg, V. 4, p. 373. 
Kent, J. S. 
7), On a New Polyzoan, “ Vietorella pavida,” from the Victoria Docks. 
Quart. Jour. Mier. Sci, X., p. 34. 
Kleinenberg, N. 
’86. Die Entstehung des Annelids aus der Larva von Lopadorhynchus. 
Zeitschr. f. wiss. Zool., XLIV. 1, p. 1. 
Korotneff, A. 
‘74, Towronanie Paludicella. Bull. Roy. Soc. Moscau, X. 2, p. 45. [Russian. ] 
’75. (Abstract of Korotneff, 74, by Hoyer, in Hoffman u. Schwalbe’s 
Jahresber. Anat. u. Phys. f. 1874, III. 2, p. 369 | 
’87. Zur Entwicklung der Alcyonella fungosa. Zool. Anz., X., No. 248, 
p. 193. 
'89. Sur la question du développement des Bryozoaires d’eau douce. Mém. 
Soc. Naturalistes de Kiew, X. 2, p. 393. [Russian.] 
Korschelt, E. 
’89. Beiträge zur Morphologie und Physiologie des Zellkernes. Zool. 
Jahrb. (Spengel), Abth. f. Anat. u. Ontog., Is lepo 
Kraepelin, K. 
’86. Ueber die Phylogenie und Ontogenie des Süsswasserbryozoen. Biol. 
Centralblatt, VI. 19, p. 599. 
‚87. Die Deutschen Süsswasser-Bryozoen. Eine Monographie. I. Anato- 
misch-Systematischer Teil. Abhandl. naturwiss. Verein in Hamburg, X. 
168 pp., 7 Taf. 


Kükenthal, W. \ 
‚85. Ueber die lymphoiden Zellen der Anneliden. Jena. Zeitschr., XVIII, 


p 819. 


BULLETIN OF THE 


112 


Leidy, J. 
55. Contributions towards a Knowledge of the Marine Invertebrate Fauna 


of the Coasts of Rhode Island and New Jersey. Jour. Acad. Nat. Sci. 
Philad., (2), IIL, p. 135. 
Maas, O. 
90. Ueber die Entwicklung des Süsswasserschwammes. Zeitschr. f. wiss. 
Zool., L. 4, p. 527. 
Metschnikoff, E. 
‚82. Vergleichend-embryologische Studien. Zeitschr. f. wiss. Zool., 
XXXVII. 2, p. 286. 
’83. Untersuchungen über die intracelluläre Verdauung bei wirbellosen 
Thieren. Arb. aus d. zool. Inst. Wien, V. 2, p. 141. 
Nitsche, H. 
’69. Beiträge zur Kenntniss der Bryozoen. I. Beobachtungen über die 
Entwicklungsgeschichte einiger chilostomen Bryozoen. Zeitschr. f. wiss, 
Zool, XX. 1, p 
71. Beiträge zur Kenntniss der Bryozoen. ILL. Ueber die Anatomie und 
Entwieklungsgeschichte von Flustra membranacea. LV. Ueber die Mor- 
phologie der Bryozoen. Zeitschr. f. wiss. Zool., XXI. 4, p. 416. 
75, Beiträge zur Kenntniss der Bryozoen. V. Ueber die Knospung der 
Bryozoen. Zeitschr. f. wiss. Zool, XXV. Suppl. 3, p. 343. 
Nussbaum, M. 
’80. Zur Differenzirung des Geschlechts im Thierreich. Arch. f. mikr. 
Anat., XVIIL, p. 1. 
‚87. Ueber die Theilbarkeit der lebendigen Materie. II. Mitth. Arch. f. 
mikr. Anat., XXIX. 2, p. 265. 
Oka, A. 
90. Observations on Fresh-water Polyzoa (Pectinatella gelatinosa, nov. sp.). 
Jour, Coll. Sci., Imper. Univ. Japan, LV. 1, p. 89. 
Ostroumoff, A. 
’86", Contributions à Pétude zoologique et morphologique des Bryozoaires 
du golfe de Sébastopol. Arch. Slaves de Biologie, IL., pp. 8, 184, 329. 
’86». Einiges über die Metamorphose der Süsswasserbryozoen. Zool. Anz., 
IX., No. 232, p. 547. 
’87. Zur Entwicklungsgeschichte der cyclostomen Seebryozoen. Mitth. aus. 
d. zool. Stat. zu Neapel, VII. 2, p. 177. 
Parker, G.H. 
’89. Report upon the Organisms, etc. Rept. Mass. State Board of Health 
on Water Supply and Sewage, I., p. 581. 
Pergens, E. 
’89. Untersuchungen an' Seebryozoen. Zool. Anz., XII., No. 317, p. 504. 
Prouho, H. 
90, Recherches sur la larva de la Flustrella hispida (Gray), structure et 
métamorphose. Arch. Zool. expér. et gen., (2), VIII. 3, p. 409. 


MUSEUM OF COMPARATIVE ZOÖLOGY. ile 


91, Sur le développement de la Membranipora pilosa. Assoc. Frangaise 
pour PAvanc. d. Sci., XIX., 2° part., p. 517. 
Reichert, K. B. 
’70. Vergleichende anatomische Untersuchungen ü 
(Ehrenberg). Abhandl. königl. Akad. Wissensch. zu Berlin, a. d. Jahre 
1869, II., p. 233. 
Reinhard, W. W. 
80. Zur Kenntniss der Süsswasser-Bryozoen. Zool. Anz., IIL, No. 54, 
p- 208. 
Repiachoff, W. 
75. Zur Entwickelungsgeschichte der Tendra zostericola. Zeitschr. f. wiss. 
Zool., XXV. 2, p. 129. 
"75%, Zur Naturgeschichte der chilostomen Seebryozoen. Zeitschr. f. wiss. 
Zool., XXVI. 2, p. 189. 
78. Ueber die ersten embryonalen Entwickelungsvorgänge bei Tendra 
yostericola. Zeitschr. f. wiss. Zool, XXX., Suppl, p. 411. 
’80. Ko Mopipoxorin Muranons- 69 pps, 4 Tab., 8°. Odessa, 1880. 


ber Zoobotryon pellueidus 


Seeliger, O. 
’85. Die Knospung der Salpen. Jena. Zeitschr., XIX. 3, p. 578. 
89. Zur Entwickelungsgeschichte der Pyrosomen. Jena. Zeitschr., 
XXIII., p. 595. 
89%, Die ungeschlechtliche Vermehrung der endoprokten Bryozoen. 
Zeitschr. f. wiss. Zool, XLIX. 1, p. 168. 
90. Bemerkungen zur Knospenentwicklung der Bryozoen. Zeitschr. f. 
wiss. Zool., L. 4, p. 560. 
Semper, C. 
"77. Beiträge zur Biologie der Oligochaeten. Arb. zool.-zoot. Inst. Würz- 


burg, IV. 1, p. 65. 
Smitt, F. A. 

65. Om Hafs-Bryozoernas utveckling och fettkroppar. Ofversigt Kongl 
Vetenskaps-Akad. Forhandl., XXII. 1, p. 5. 

65%, Kritisk förteckning öfver Skandinaviens Hafs-Bryozoer. Öfversigt 
Kongl. Vetenskaps-Akad. Förhandl., XXII. 2, p. 115. 

>67. Kritisk förteckning öfver Skandinaviens Hafs-Bryozoer. Ofversigt 
Kongl. Vetenskaps-Akad. Förhandl., XXIV. 5, p. 279. 


Stuhlmann, F. 
87, Zur Kenntnis des Ovariums der Aalmutter. Abhandl. naturwiss. 


Verein in Hamburg, X. 48 pp. 
Tullberg, T. 
’82. Studien über den Bau und das Wachsthum des Hummerpanzers und 
Kongl. Svenska Vetenskaps-Akad. Handl., XIX. 3. 


der Molluskenschalen. 
57 pp. 12 Taf. 


VOL. XXII. — NO. 1. 8 


114 BULLETIN OF THE MUSEUM OF COMPARATIVE ZOOLOGY. 


Verrill, A. E. 
’73. Report upon the Invertebrate Animals of Vineyard Sound, etc. Rept. 
U. S. Com. Fish and Fisheries, 1871-72, p. 295. 
Vigelius, W. J. 
’84. Die Bryozoen, gesammelt während 3. u. 4. Polarfahrt des “ Willem 
Barents.” Bijdragen tot de Dierkunde, XI. 104 pp., 8 Taf. 
’86. Zur Ontogenie der marinen Bryozoen. Mitth. zool. Stat. zu Neapel, 
VI. 4, p. 499. 
’88. Zur Ontogenie der marinen Bryozoen. Mitth. zool. Stat. zu Neapel, 
VIII. 2, p. 374. 
Wagner, F. v. 
’90. Zur Kenntniss der ungeschlechtlichen Fortpflanzung von Microstoma 
u. s. w. Zool. Jahrb. (Spengel), Abth. f, Anat. u. Ontog., IV., p. 329. 
Zacharias, O. 
’86. Ueber Fortpflanzung durch spontane Quertheilung bei Süsswasser- 
planarien. Zeitschr. f. wiss. Zool, XLIII. 2, p. 271. 
Zoja, R. 
’90. Alcune ricerche morfologiche e fisiologiche sul? Hydra. Bollettino 
Scientifico, XII. 3, p. 65. 
’91. Quelques recherches morphologiques et physiologiques sur l Hydre. 
Arch. ital. de Biol., XV. 1, p. 125. [Résumé of Zoja, ’90.] 


DAVENPORT, — Budding in Bryozoa, 


PLATE I. 


ABBREVIATIONS. 
cev. pyd. Neck of polypide. gn. Ganglion. 
cla. Normal cuticula of adult h Inner layer of bud. 
body wall. kmp’drm. Kamptoderm. 
eta. Cuticula secreted by tip. ms’drm.  Mesoderm. 
ec’drm.  Ectoderm. mu. pyr. Pyramidal muscles. 
ex. Outer layer of bud. a. CEsophagus (pharynx). 
ga. Stomach. rt. Rectum. 
gm. Bud. ta. Tentacle. 
All figures are of Paludicella Ehrenbergü. 
Fig. 1. Stock of Paludicella Ehrenbergii, viewed as an opaque object. X 4.5. 
Fig. 2. Diagram representing the interrelations of individuals in stock shown in 
Figure 1. A-H are individuals of the ancestral (median) branch ; 
a, b,c, etc., lateral branches given off from the ancestral branch to the 
right; a’, b, branches given off to the left; a, B, etc., lateral branches 
of second order given off in the direction of the distal end of the 
ancestral branch; a’, B’, ete., given off in the direction of proximal end; 
a,’, lateral branches of third order — to left. 
Fig. 2°. Diagram of another (smaller) stock. Letters have same significance as 
in foregoing. 
Fig. 3. Cross section of branch near tip, showing the first trace of the bud of the 
polypide at ex., i X 685. 
Fig. 4. Cross section of branch near tip, showing bud of polypide slightly older 
than in Figure 3. X 695. 
Fig. 5. Cross section of slightly collapsed branch near tip, showing ingression 
of cells at ex. to form inner layer of bud. X 635. 
Fig. 6. Longitudinal section of tip of branch to show cell structure. Zeiss, qfy 
oil immersion, Oc. 1. X 1000. 
Figs. 7, 8, 9. Optical sections (nearly in sagittal plane) of three tips of branches 


in successive stages of development, showing relations of young bud, 
gm., to next older polypide. In Figure 8 the branch is slightly 
shrunken, X 87. 


gm 


gm 


Davenport. — Budding in Bryozoa, 


PLATE II. 
ABBREVIATIONS. 
cev. pyd. Neck of polypide. ec’drm. Ectoderm. 
cta. Normal cuticula of body gm. l. Anlage of lateral bud. 
wall. kmp’drm. Kamptoderm. 
cta’ Cuticula secreted by tip. ms’drm. Mesoderm. 
All Figures from preparations of Paludicella Ehrenbergii. 

Fig. 10. Surface view of cuticula near the end of a branch at intervals, a being 
nearest the tip, and d farthest from it. The branch was stained in 
Erlich’s hematoxylin, the color being taken up by superficial cuticula 
only. x 320. 

Figs. 11, 12, 13. Cross sections of the cuticula taken at different distances from 
the tip, to show the stainable and non-stainable cuticule. Figure 11 
is from near the tip, Figure 13 farthest from it. >< 1000. 

Fig. 14. Longitudinal median (sagittal) section through the tip of a branch show- 
ing cells of tip and an early stage in the development of the polypide. 
x 410. 

Fig. 15. Cross section of branch showing origin of lateral bud. x 635. 

Fig. 16. Longitudinal section of body wall of branch through the point at which 
a lateral bud is originating. Polypide of ancestral branch is nearly 
adult. X 685. 

Fig. 17. Longitudinal section of body wall from near the tip through the Anlage 
of a lateral bud. X 410. 

Fig. 18. Cross section of branch showing histological conditions of Anlage of 
lateral bud. The polypide has reached a stage of development cor- 
responding to that of Figure 36, Plate IV. 1000. 

Fig. 19. Longitudinal section through body wall from the same branch as Figure 
17, but farther from the tip. Histological conditions are to be com- 
pared with those of Figure 17, which represents a less differentiated 
condition. 410. 

Fig. 20. Cross section of branch in which the polypide has reached a stage slightly 


younger than that of Figure 36. To show Anlage of two lateral buds 
with their cuboidal undifferentiated cells. X 410. 


Davenport. — Budding in Bryozoa. 


PLATE II 


ABBREVIATIONS. 


An. Anal side of polypide. kmp’drm. Kamptoderm. 

atr. Atrium. loph. Lophophore. 

cev. pyd. Neck of polypide. ms’drm, Mesoderm. 

cl. mu. ret. Young cells of retractor mu. par, Parietal muscle. 
muscle. a. Œsophagus. 

cta. Cuticula. Or. Oral side of polypide. 

ee’dılm. Ectoderm. rt. Rectum. 

ga. Stomach. vlv, er. Cardiac valve. 

gn. Ganglion. 


All figures from preparations of Paludicella Ehrenbergii. 


Figs. 21-25. Longitudinal sections through buds of polypides at successively older 


Fig. 21. 
Fig. 22. 


Fig. 23. 


Fig. 24. 
Fig. 25. 


Fig. 26. 


Fig. 29. 


stages. The tip of the colony, and therefore the anal aspect of the 
polypide, is to the right in all cases. All figures X 410. 

Stage of Figure 87 (Plate IV.). Few nuclei in central region. 

Shows rapid growth of bud, chiefly at neck of polypide. ‘The two inner 
cell layers are about to separate to form the common cavity of atrium 
and cesophagus. 

Beginning of formation of alimentary tract at rectum, rt. The row of 
nuclei separating the atrio-esophageal cavity from the alimentary tract 
is due to the fusion of the two inner layers of the bud along this line. 

Rectum and stomach completed. Retractor muscles begin to form. 

Lophophore and young tentacles have made their appearance, and 
wsophagus and pharynx are separated from atrium. Beginning of 
formation of brain at gn. 

Part of cross section of a branch of stage of Figure 80. Parietal mus- 
cles, mu. par., occupy a diameter of the section, and are attached to 
the cuticula. X 635. 

Young parietal muscle at stage of Figure 28. This is one of the pair 
which in a later stage are found lying together in Figure 26. X 635. 

Cross section of branch showing young polypide, and reticulated vacuo- 
lated cells. x 410. 

Bit of body wall, with euticula separated from underlying ectoderm to 

show ends of parietal muscles. X 690. 


ZUDDING IN 


te 


COMI 


chur’ 


On 


msärp 


cou pel. 


io 


erden 


20 


PUL par, 


Davenport. — Budding in Bryozoa. 


PLATE IV. 

ABBREVIATIONS. 
An. Anal side of polypide. i. Inner layer of bud. 
an, Anus. loph. Lophophore. 
atr. Atrium. ms’drm. Mesoderm. 
can. cre. Ring canal. mu. Muscle fibre in funiculus. 
ec’drm. Ectoderm. n’ Circumæsophageal nerve. 
ex. Outer layer of bud. æ, (Esophagus. 
Jun. inf. Inferior funiculus. Or. Oral side of polypide. 
fun. sup. Superior funiculus. rt. Rectum. 
ga. Stomach. vac. Vacuole. 
gn. Ganglion. vlv, er. Cardiac valve. 


All figures from preparations of Paludicella Ehrenbergii. 


Fig. 30. Cross section of polypide bud of stage of Figure 24, Plate III. The posi- 
tion is indicated by the line 80, Figure 24. X 410. 

Figs. 81-34. Four cross sections of a branch through a young polypide, some- 
what younger than that of Figure 25. Figure 81 is nearer the anal, 
Figure 34 nearer the oral surface. In Figure 34 that part only of the 
section of the polypide which lies near the body wall is represented. 
x 410. 

Fig. 35. Cross section of branch through polypide of age of Figure 25. To show 
origin of tentacles and ring canal. X 410. 

Fig. 36. Sagittal section of young polypide at period of closure of ganglion, gn. 
x 410. 

Fig. 36°. Bit of same polypide a few sections to one side of plane of Figure 36, 
showing origin of inferior funiculus. X 410. 

Fig. 37. From cross section of branch showing early stage in development of the 
bud. X 410. 

Fig. 38. From a sagittal section of nearly adult polypide, showing the two funiculi 
and their muscles. X 410. 

Figs. 39 and 40. Two neighboring sections parallel to the body wall through a bud 
of the stage of Figure 23. Figure 40 lies three sections below Fig- 
ure 39. Figure 39 shows the atrial cavity, formed as yet only on the 
anal side. Figure 40 shows the beginning of formation of the ali- 
mentary tract at the anal end. Note the vacuolated condition of the 
mesoderm. X 410. 

Fig. 41. Polypide of about the stage of Figure 25 looked at en face. The anal 
tentacles, being turned under, do not appear. To show compressed 
condition of polypide, and alternating position of tentacles, Cf. 

Figure 77, Plate IX. X 820. 


T = BUDDING IN BRYOZOA 


scher 


> g ate 
a a Bo è 


msarm 


£ D D Mpe 2 
z z) AR 
. iw 
P R * {+ 
a. Oy Sun set $ 
` 
> F; S Ct 
is) @ o a 
5 m PH 
f t N. 
á i 
ae 5) insti 
fi f 
e & DY DMD ys D Gy DOM Se ae 
9 y ? E Op ¢ 
vor AT D 


Davenport. — Budding in Bryozoa. 


PLATE V. 
ABBREVIATIONS. 

an. Anus. kmp’drm. Kamptoderm. 
cev. pyd. Neck of the polypide. loph. Lophophore. 
clr. set. Collare setosum. ms’drm. Mesoderm. 
cta. Normal cuticula of adult mu. par. Parietal muscles. 

body wall. mu. pyr. Pyramidal muscles. 
cta! Cuticula secreted by the tip of. atr. Atrial opening. 

of branch. rt Rectum. 
ec’drm. Ectoderm. spht. Sphincter. 


AU figures from preparations of Paludicella Ehrenbergü. 


Fig. 42. Cross section of branch of age of Figure 37, Plate IV., to show origin of 
primary parietal muscles. X 410. 

Figs. 48 and 44. Successive sections through a polypide slightly older than 
of Figure 25, cut perpendicularly to the long axis of the branch. Dur- 
ing this period the lophophore becomes more nearly circular, and its 
aboral ends meet oralwards of the rectum, rt. Figure 44 is nearer the 
tip of the branch. X 410. 

Fig. 45. Axial section of neck and atrial opening of polypide just sufficiently devel- 
oped to be capable of extrusion. Shows the collare setosum in place. 
x 410. 

Fig. 46. Section of communication plate cut across the branch. Two sections 
(10 u) above Figure 51. X 635. 

Figs. 47-49. Three stages in the development of the communication plates. Lon- 
gitudinal sections of the branch. In Figure 47, the polypide has 
reached the stage of Figure 22; in Figure 48, the stage of Figure 23 ; 
and in Figure 49, the stage of Figure 24. X 635. 

Fig. 50. Longitudinal section through neck of young polypide, showing the sink- 
ing of the neck below the general surface of the body, and the method 
of forming the inner cuticula of neck. X 890. 

Fig. 51. Cross section of branch through communication plate. The left side 

` of the section includes the cuticula and the underlying flat ectodermal 
layer. The right side cuts a little lower into the mesodermal cells. 


x 685. 


that 


an. 


el, rtl. 
cta. 


cta’ 


ec'drm. 
ex. 
gm. 
gn. 


Fig. 52. 


Fig. 58. 
Fig. 54. 
Fig. 55. 


Fig. 56. 


Fig. 57. 


Fig. 58. 


Fig. 59. 


Fig. 68. 


Davenport. — Budding in Bryozoa. 


can. cre. 
cev. pyd. 


PLATE VI. 


ABBREVIATIONS. 

Anus. i: Inner layer of bud. 
Ring canal, kmp’drm. Kamptoderm. 
Neck of the polypide. la. comn. Communication plate. 
Reticulated cells. ms’drm. Mesoderm. 
Normal cuticula of adult mu. par. Parietal muscles. 

body wall. mu. pyr. Pyramidal muscles. 
Cuticula secreted by tip. nw Circumosophageal nerve. 
Ectoderm. æ. Güsophagus. 
Outer layer of bud. tts Rectum. 
Bud. vac. Vacuole. 
ganglion. 


All figures from preparations of Paludicella Ehrenbergii. 


Cross section of a branch through a polypide slightly older than that 
shown in Figure 36, The section passes through the brain and whole 
extent of the ring canal, together with its opening into the coclom. 
x 685. 

Next section below Figure 52 of same series ; showing the beginning of the 
circumosophageal nerve ring. X 635. 

Shows connection of mesodermal cells of body wall, ms’drm, with those 
of the outer layer of bud, ex. X 1030. 

Origin of the secondary parietal muscle cells from mesoderm of body 
wall. X 685. 

Histological conditions of the budding regions. The cells have large nu- 
clei, the mesodermal cells are vacuolated and rapidly dividing; the 
cells of the bud are densely granular. Zeiss, 7g oil immersion, Oc. 1. 
x 1070. 

Normal vacuolated cell, full of food particles. X 1030. 

Longitudinal section of young lateral branch, showing highly reticulated 


character of mesoderm, and nearly complete ‚ormation of communi- 


cation plate. X 410. 
Reticulated cell, showing one of the pseudopodia-like processes which 
frequently appear on them, projecting into the calom. x 1030. 


Figs. 60-62. Three successive sections from a series across the tentacles of a pol- 


ypide which has 15 tentacles, and is of about the stage of Figure 86. 
The odd tentacle (*) is shorter than the others, and lies opposite the 
rectum, rt. X 295. 

Cross section of branch through neck of polypide of about the age of Fig- 
ure 36. Shows also the young pyramidal muscles. X 410. 


kmpilren 


gi mein.) 


me 


"ae 


JD 


ea 
eo 


s$ wy É R ~i Amprirm 


Davenport, — Budding in Bryozoa. 


PLATE VII. 
For explanation of notation employed on this plate, see page 41. 


Fig. 64. Outline drawing of one of the lateral “fans” of Bugula turrita, taken 
from the axis of the colony and spread out flat on the slide. 
Kear 12, 

Fig. 649, Diagram showing arrangement of individuals in Figure 64. 

Fig. 65. Outline drawing of one of the lateral branches of a stock of Crisia eburnea, 
spread out flat on the slide. 16. 

Fig. 65%. Diagram showing arrangement of individuals in Figure 65. 

Fig. 66. Part of stock of Bugula flabellata. 10. 


Op. 


Fig. 67. 
Fig. 68. 


Fig. 69. 


Fig. 70. 


Fig. 71. 


DAVENPORT. — Budding in Bryozoa, 


pyd. 


. Plan of Figure TY. 


PLATE VII. 


ABBREVIATIONS. 


Operculum. pyd. rgn. Regenerated polypide. 
dyn. Degenerated polypide. 


Diagram to show interrelation of individuals in the corm, Figure 69. 

A part of a corm of Membranipora pilosa, to show regular arrangement, 
with a single median branch, each of whose individuals gives rise to 
two lateral branches. The * indicates margin of frond on which 
stock was growing. X ca. 8. 

Young corm of Flustrella hispida, to show arrangement of individuals. 
x 10. 

Young corm of Membranipora pilosa, with several median branches, show- 
ing regular arrangement. The marginal ones alone give rise to lateral 
branches. X 10. 

Young corm of Lepralia Pallasiana, showing arrangement of individuals. 
On the left, the nuclei of the cells of the body wall are shown, to 
indicate the inequality of their distribution. On the right, nuclei are 
omitted. At pyd. rgn. a regenerating poly pide is seen, on the opercu- 
lum. x 43. 


Davenport. — Budding in Bryozoa, 


PLATE IX. 


ABBREVIATIONS. 
An. Anal side of polypide. lu. gm. Lumen of bud. 
atr. Atrium. marg. Margin of corm. 
can. cre. Ring canal. ms’drm. Mesoderm. 
cev. pyd. Neck of the polypide. n. Circumosophageal nerve. 
cta. Cuticula. æ. (Esophagus. 
ec’drm. Ectoderm. Or. Oral side of polypide. 
ex, Outer layer of bud. Th Rectum. 
ga. Stomach. sep. Wall of zocecium in the corm. 
gm Bud. sol. Sole of the corm. 
gn. Ganglion. tet, Roof of the corm. 
i. Inner layer of bud. 
Fig. 72. Longitudinal vertical section through the peripheral part of the corm of 
Lepralia Pallasiana, showing the margin of the corm and two zoœcia, 
the olđer of which contains a polypide. X 160. 
Fig. 73. Longitudinal vertical section through the margin of a corm of Lepralia 
Pallasiana, showing the two layers of this region and the origin of the 
polypide. X 410. 
Fig. 74. Young regenerating polypide of Flustrella hispida. The section passes 
through the sagittal plane. X 380. 
Fig. 75. Vertical section through margin of corm of Flustrella hispida, to show 
origin of polypide. x 410. 
Fig. 76. Sagittal section through young polypide of Flustrella hispida, to show 
early stage of development of alimentary tract. X 410. 
Fig. 77. Superficial view of young polypide from upper surface of corm of Flus- 
trella hispida, showing young tentacles and their relation to the anus 
(at atr.). X 820. 
Fig. 78. Bud of polypide of Flustrella hispida at the time of closure of the pore of 
invagination. >< 390. 
Fig. 79. Radial section through margin of corm of /lustrella hispida, showing bud 
of polypide. X 410. 
Fig. 80. Young polypide of Flustrella hispida. : X 380. 
Fig. 81. Bud of Lepralia Pallasiana immediately before the formation of alimen- 
tary tract, showing relation of the rectal pocket (rt.) to the atrio- 
pharyngeal cavity above. X 410. 
Fig. 82. Section through polypide, through lately formed brain and circum- 


wsophageal nerves (n) growing around osophagus (œ.). 410. 


mary 


y IN 
un 


msdm 


ecélrm 


80 


i 
qm 
s ( 


uhr 


mard. 


Davenport, — Budding in Bryozoa. 


PLATE X. 


ABBREVIATIONS. 


An. Anal side of polypide. kmp’drm. Kamptoderm. 

an. Anus. lu, gn. Lumen of the ganglion. 

atr. Atrium. ms’drm. Mesoderm. 

can. cre. Ring canal, mu Musculature of oesophagus. 
cev. pyd. Neck of polypide. mu. ret. Retractor muscle of polypide. 
ce. Cocum. æ. CEsophagus. 

cla. Cuticula. op. Operculum, 

di’sep. Wall of zocecium in the corm. Or. Oral side of polypide. 

ec’drm. Ectoderm. or. Mouth. 

ex. Outer layer of bud. pyd. dgn. Degenerated polypide, “ brown 
fun. Funiculus. body.” 

ga. Stomach, rt, Rectum. 

gn. Ganglion. ta. Tentacle. 

2 Inner layer of bud. 


Fig. 83. Sagittal section through young polypide of Escharella variabilis. X 820. 

Fig. 84. Regenerated polypide of Lepralia Pallasiana on operculum (op.). X 880. 

Fig. 85. Cross section of pharynx of adult polypide of Escharella variabilis, show- 
ing perforated cell walls. X 635. 

Fig. 86. Sagittal section of young polypide of Lepralia Pallasiana, showing forma- 
tion of brain. X 320. 

Fig. 87. Section parallel to sole of a corm of Escharella variabilis at about the stage 
of Figure 86, showing atrium, ganglion, and rectum. X 430. 

Fig. 88. Vertical section through a bit of roof of corm of Escharella variabilis at 
neck of polypide, showing also the region of future operculum and of 
origin of future regenerated buds. Compare with Figure 90. X 410. 

Fig. 89. Sagittal section of young regenerated polypide of Flustrella hispida inter- 
mediate in age between Figures 86 and 83. Shows the origin of the 
ganglion and rotation of the oral tentacles. X 820. 

Fig. 90. Vertical section of a bit of body wall from same individual as Figure 88, 
to show the comparatively less embryonic condition of cells here than 
at neck of polypide. X 410. 

Fig. 91. Operculum of Zepralia Pallasiana cut verpendicularly to surface, show- 
ing originof aregenerating polypide. Body wall somewhat shrunken 
from cuticula. X 410. 

Fig. 92. Section through a regenerated polypide of Escharella variabilis, showing 

relations of alimentary tract to “ brown body” (pyd. dgn.). x 410. 


Rep branen i Runt 2 
ORT - BUDDING IN BRYOZOA Pi 
op 
eS £ 85 
ar SES 
er \ 
fiin 
86 an 
eji i nia 
atr 
a | 
m 
Fi Ah 
67 a 
fe k 2 N a 
“ay € 3 ù lugn 4 
7 0 "é CANEN 3 
, % 6 SR ff z x N 
Ampdtrm At 80) OX ER 
in 5 vi on 
altr 2 ` 5 p 
` ye 5 ga COF 
3 ` ru 
A z SO 
canere ii: Be: 
kmparm Asep 
89 
l = cev. pve 
0 
eae 
óe gn Q, 
Lee A i ; a 
muret Pe 
Br 
ga 
ri 
90, 
peter 
E T EG cee Ry 
a, 
“ss m 
91 


matlrm 


on 


Davenport. — Budding in Bryozoa. 


Fig. 96. 


Fig. 97. 


Fig. 98. 


Fig. 99. 


PLATE XI. 


ABBREVIATIONS. 


cev. oe. Neck of occium. ms’drm. Mesoderm. 
cal, Colom. 008, Ocecium, 
ec’drm. Ectoderm. ov! Oöblasts. 
en’drm. Entoderm. pyd. Polypide. 

ex. Outer layer of bud. sto. Stolon. 

i. Inner layer of bud. tet. Roof of stock. 


lu. gm. Lumen of the bud. 


Fig. 98. A portion of a longitudinal section through a young stock of Plumatella 


polymorpha, about two weeks after hatching from statoblast (killed 
12th May, 1890), showing the body wall just analward of the neck of 
a young polypide (pyd.), at the oral side of which a younger bud has 
already arisen. ‘The inner (mesodermal) layer of the body wall 
shows oöblasts (ov.’) in various stages of development. X 600. 


Fig. 94. Longitudinal section of o@cium of Cristatella showing embryo which is 


giving rise to the colomic epithelium by ingression of cells at its 
proximal pole, —i. e. the pole nearest the neck of the owcium. There 
are in the next section two other cells in the cavity of the blastula, 
one of which appears degenerate in that it contains a huge vacuole, 
and has no distinct nucleus, the chromatic substance lying scattered 
loose near the cell wall. x 600. 


Fig. 95. Longitudinal section through ocecium of Cristatella and its contained em- 


bryo. One polypide bud and the stolon (sto.) are shown here. There 
are two other buds in the embryo further developed than this one, 
lying to one side of it, and on the side of each of these buds is the 
Anlage of another. The stolon is seen to be well developed, lying be- 
tween the ectoderm and mesoderm throughout the region bounded 
by the three older buds, and extending as a zone beyond them, and 
even beyond the Anlage of the youngest polypides. The embryonic 
tissue thus forms a disk about 75 X 150 u in extent. >< 390. 

Transverse section of omeium of Plumatella, showing origin of first pol- 
ypide. Compare with Figure 99, which represents an earlier stage. 
x 390. 

Longitudinal section through omeium and contained embryo of Cristatella. 
The stolon is already cut off from the ectoderm. This stage imme- 
diately follows that of Figure 101, Plate XII. The forming bud is 
that of the first polypide. 890. 

Oblique section through oœcium of Plumatella, showing a later stage in 
development of the inner layer of the larva (cf. Fig. 94). X 600. 
Longitudinal section of oæcium and contained larva of Plumatella, The 
bud shown at i, ex. is the first in the colony. An incipient (second) 
bud is shown five sections to one side in the region indicated by an 


asterisk. X 410. 


9, y ; 
$ 5 nad Y eciirm 
| ery | . 
waa E a 
Bia EA R 
\ 9 ae \ 
a EN \ y 
è a n/A EN ecilrm 
4 ° $ x a í 
C a n ~ . * 4 
s A EA cr \ 
bY > Oy EN vA @ 
* © ae Se ` \ 
0 ) 
mariem y 5 
e i ; 
mire * ooe a \ 
y oni D x 
oy) * a 2 ` Dee é 
* s, p- . à x t N} * 
i a x . x : “ye ; ale 
z * 7 GOA Sie x r ” A CPOO 
ai © OTe OOM ONS ` d 4 \ « } 
ATO Rane) ees Oe . 94 sit» end 
Pin we Š 
msile ” D N OIO SEO e s. d * 
2 te x me v 
or LOCO nia x a YA 
IO er : as 
cr K z \ in agm * eedem 
on ; % poi 
i “an 9 À s s " $ * 
{ NORD S ‘ rool ecilrm 
? A‘ FS 00 x a 
ae Ki 5 Ya ha noe b i k 
D MO QOS Oks i x ay " 
> BBS ri be ae A N hi A i i 
ie x Win an] Gee ` ex 
X FARO 5 
er ärm ge ON 
SER ` ODP. ` DARD L 
97 ee : z -> 
8, trt i en 
= : x RN 2 a m $ 
Ar) ` x Dale s $ F 
< Oe x 
* + “x = 3 
fy à ® we 
fe è x 
L) br mselrm S & io x 
G 2 Dr 
C Ale y Ha Ae i 
; a/e $ we 
i wo 6 2 4 ets 
u PA 8, ë 
è s D a 
‘ Ma o ` J 
\ D EN 
z gos a Yay morn 
i 5 ` 0 
$ 7 4 N 
aoe à G 
: AR AR a N 
s; N N a š 5 ; 
ate 2 s F 
ej y eta 
è . s 
K . S * 4 g 
s ek 
An R as: X . 
x o 
* as > } 
arm AS g > s * ` N 
` 5 N r 
: ‘ oN 
D y ` A i r4 Aaa 
i vi < 
* A OR ` 
i A i : 
i $ f ` 
2 a 
$ ae SG 
* ® $ ` 
s * 


Davenport, — Budding in Bryozoa, 


PLATE XII. 


ABBREVIATIONS. 
ec’drm. Ectoderm. ms’drm. Mesoderm. 
ex, Outer layer of bud. ow. Ocecium. 
i. Inner layer of bud. sto. Stolon. 


lu. gm. Lumen of the bud. 


Fig. 100. Longitudinal section of a larva of Plumatella polymorpha, in which the 


two layers are established; the pole of ingression is directed upward, 
on the plate. 


Fig. 101. Section of upper part of zoæcium of Cristatella mucedo, with its contained 


larva. Showing the formation of the stolon at the pole of ingres- 
sion and the attachment of this pole to the placenta-like neck of the 
oweium (*). X 390. 


Fig. 102. Section through an ocecium of Cristatella, with its contained larva. One 


polypide is already established, and a second is arising. The two 
are the only buds in the larva. On the left of the older bud the 
stolon is seen to be intruding itself between the ectoderm and meso- 
derm of the larva. 390. 

Section through the two oldest polypides of the Cristatella larva, to- 
gether with the stolon. This larva contains one other less developed 
bud at one side of these two. 890. 


Fig. 104. Plumatella polymorpha. Stage of first bud later than that shown in 


Figure 96, exhibiting pore of invagination closed by overgrowth of 
ectoderm. X 3890.. 


NPORT ~ BUDDING IN BRYOZOA. ' PRAE 


{+ 
D 
100 DS 
Hi ` 
VONA 
POA D f 
j 10} ac) j 
S w . & 
Q Bs 
b = 3 
; Ones x à X 
G fs N 
@ Sur, A x z 
? i@ ` . a 
è of % we > > i 
Ä . $ * A : 
ed d : 5 * « 
à ; 
5 eve x ON 
. S x + y © 
t . 4 Mn 
. a AtS "~ 
* CPN WSs 5 < 
“ fi © VOOM ` x 
s OY WV Ye . 
%e r 5 Q 0) NE > Ñ 
be è ° 
s 
” = 
Gi $ ç 
x X fs 
Se è EN \ 
0 * N D 
` Q Q ` A 
o) h À (s ie ) 
Ö) 102 5 8 
. ale 2 N 
D x 
x x X ` > è O 
NO D7 OE cs . en N 
è ` ` AN < 
& + BEN RR © 
SEO) a yee ` Re 
s ` i š NONAS v oor 
. Dj a DVL ve 3 A 
VG SQ) Qyae . x X ODAN 
D x “i BR CRA 
py 
a\* E © 
HO) 7 
e e ‘a » ° nd * 
š my SA 3 
e * ` > 0 SS ae 2 uk 
s y 2 e ` $ 2 ade , 
x 5 S 2 M A ningar ORO Q 
O N . a A S « VAS . 
s : 3 e Yin DN s . ` OTON 
x © 2 . * wis : 5 OR 
è « SE ` £ * ey be 
` “ Na $ he “ty aye `, 


No. 2.— The Gastrulation of Aurelia flavidula, Per. & Les. 
By Frank Smiru.} 


Precepine the appearance of Goette’s (87) publication in 1887 upon 
the development of Aurelia aurita and Cotylorhiza tuberculata, the gas- 
trulation of Aurelia had been regarded, in the light of the studies of 
Kowalewsky, Haeckel, Claus, and others, as the result of invagination or 
at least of a process nearer to invagination than to any other method of 
gastrulation. 

Gvette’s work seemed to show, however, that, instead of an invagina- 
tion, there is an ingression of cells to form the entoderm, and that the 
first result of this ingression is the production of a solid gastrula, or 
sterrogastrula, which is only subsequently hollowed out, and is put into 
communication with the exterior through the formation of a prostoma 
at a still later period. Recently, in a paper dealing especially with the 
development of Cotylorhiza tuberculata, Claus (°90) reaffirms the posi- 
tion taken in his previous paper (’83), in which the gastrulation in 
Aurelia was represented as being simply a modification of invagination. 
In recent papers by Hamann (’90) and McMurrich (’91), Goette’s views 
are adopted, and form part of the basis for statements that, in the devel- 
opment of the Scyphomeduse, invagination, instead of being the rule, is 
the exception. 

This want of agreement among those who have given the subject 
most attention makes the determination of the actual method of gastru- 
lation in Aurelia a matter of considerable interest, and it may be 
assumed that any contribution to the solution of the question will not 
be unwelcome, 

Early in the current year, at the suggestion of Dr. E. L. Mark, I 
undertook to investigate the method of gastrulation in A. flavidula. 

Through the kindness of Mr. B. H. Van Vleck of the Boston Society 
of Natural History, I was enabled to spend two months of the summer 
of 1887 at his seaside Laboratory at Annisquam, Mass., where I then 
collected the material used in the present study. The embryos were 
killed with picro-nitric acid, and preserved in 90 per cent alcohol, in 
which they have been kept during the three intervening years. Of the 

1 Contributions from the Zoölogical Laboratory of the Museum of Comparative 
Zoölogy, under the direction of E. L. Mark, No. XXIX. 

VOL. XXII, — NO, 2, 


116 BULLETIN OF THE 


various staining fluids tried, Erlich’s acid haematoxylin gave decidedly 
the best results for sections. For examination of the whole embryos, 
Grenacher’s alcoholic borax-carmine and Czokor’s alum-cochineal each 
gave good results. The latter stain possesses the peculiarity of stain- 
ing embryos of different ages with corresponding degrees of intensity, 
the youngest stages being stained the least, the degree of intensity in- 
creasing with the age of the embryo up to the planula stage. 

The result of segmentation is a one-layered blastosphere, as in 
A. aurita. Although the diameter of the blastocel, or segmentation 
cavity, presents some individual variations at a given stage of develop- 
ment, it in general corresponds very nearly with that of A. aurita, as 
described by Goette (’87, p. 3). It increases slightly as the process of 
gastrulation advances. The cells of the blastosphere are usually some- 
what shorter at one pole than elsewhere, and it is from this region that 
the entoderm is formed. The nuclei of all the cells are situated very 
near the outer surface of the blastosphere. Small spheroidal bodies con- 
stitute the greater portion of each cell; they are very evenly distributed 
through its substance, except in the vicinity of the nucleus, where they 
are somewhat less abundant. Vacuoles of variable sizes are usually 
found in some of the cells. The nuclear region stains a little more 
deeply than the remaining portion. 

The method of gastrulation in A. flavidula is similar to that in A. 
aurita as described by Claus (’83, pp. 2 and 3), although it resembles 
even more closely a typical invagination. When the process of cleavage 
has resulted in the formation of a blastosphere composed of somewhat 
more than four hundred cells, a depression of limited extent appears in 
the portion of the wall which is composed of the shorter cells, From this 
depressed region is formed the entoderm, which develops as a single con- 
tinuous layer of cells surrounding a small cavity, the coslenteron. At the 
beginning of the process, and throughout its duration, the colenteron is 
in communication with the exterior by means of a narrow passage, the 
blastopore, or blastoporic canal. See also Explanation of Figures (Plate 
I. Figs. 1-4). From these figures it is apparent that only a small por- 
tion of the wall of the blastosphere is concerned in the invagination, 
and to that extent it must be regarded as deviating from the typical 
invagination, where one half of the ‘wall of the blastosphere is infolded 
to form the entoderm. The colenteron is, however, at all stages of gas- 
trulation, an open sac-like cavity, and therefore noticeably different from 
that of A. aurita, of which Claus (83, p. 3), says: “Mit dem weiteren 
Nachrücken der die Mundspalte begrenzenden Zellen in das Innere des 


MUSEUM OF COMPARATIVE ZOOLOGY, 117 


Larvenleibes ändert sich jedoch allmählig das frühere Verhältniss zu 
Gunsten der Entodermfüllung, die noch immer keine wahre Höhle, sondern 
eine schmale lineare, mit der Hauptachse des Leibes zusammenfallende Spalte 
besitzt.” |} With the growth of the entodermal layer, the coelenteron 
enlarges, and the cleavage cavity is diminished, until finally it is 
entirely obliterated and the entoderm everywhere comes into contact 
with the ectoderm (Plate I. Figs. 4-6, Plate II. Fig. 11). 

During the process of gastrulation, and also for a short time after its 
completion, the thickness of the entoderm, which is much less than that 
of the ectoderm, does not increase. Figures 5 and 6 (Plate I.) are 
from ‘sections of two embryos at different stages of development. 
Figure 5 is from an embryo soon after the completion of gastrulation ; 
Figure 6 is from an older stage. Since in each case the section is from 
the middle of its series, it follows that a decided thickening of the ento- 
derm takes place between the stages represented by these Figures. This 
thickening is apparently due to an increase in the number of the cells, 
which are soon unable to find room for themselves except by elongation. 
The entodermal cells are quite different in appearance from those of 
the ectoderm; they are,approximately spherical, and do not have as 
numerous spheroidal yolk bodies as the latter. Their nuclei, however, 
closely resemble those of the ectoderm, and usually lie in the portion of 
the cell nearest the coelenteron. 

As is to be seen from Plate II. Fig. 7,— a section nearly perpendicular 
to the blastoporic canal, — the blastopore in A. flavidula is very small. 
A similar condition has been shown by Claus to exist in A. aurita and 
by Metschnikoff (86, Taf. X. Fig. 14) in Nausithoe marginata. 

The nuclei of the cells composing the wall of the blastosphere are sit- 
uated, as has been stated, near the surface of the sphere. But at about 
the time of the beginning of the invagination, sometimes a little earlier, 
a few of the nuclei are found in the deeper portion of the wall. At first 
there are only one or two such displaced nuclei to be observed in the 
whole embryo, but as development progresses they increase in number. 
A careful examination of sections shows that the cells to which they 
belong do not extend, like the remaining cells of the wall, through its 
whole thickness, but that they are wedged in as it were between the 
bases of the ordinary cells. The latter are much elongated, and from 
mutual pressure are prismatic, whereas the deep cells are spheroidal and 
project in some cases into the segmentation cavity. Since these cells 
are found at various intermediate positions between the outer and inner 

1 The original is rot Italicized. 


118 BULLETIN OF THE 


surfaces of the wall, I infer that they result from a process of migration 
inward, either at the time of cell division or independently of that pro- 
cess. Indeed, there is obviously no other possible source whence these 
cells could come, but the exact process of transfer is not easily determined. 
I believe that this increase in number is at first for a considerable time 
due exclusively to the migration of cells which once shared in forming 
the external boundary of the sphere, but later the division of cells which 
have already migrated into the deeper portion of the ectoderm undoubt- 
edly contributes to this increase. 

We have now to turn our attention to a phenomenon of considerable 
importance, the study of which from preserved material is, however, 
attended with difficulties. I refer to the ingression of cells from the 
wall of the blastosphere into the cleavage cavity, which begins a con- 
siderable time before the invagination commences. The latter does not 
take place until the number of cells forming the wall of the blastosphere 
has exceeded 400, whereas the ingression, as far as can be inferred from 
the cases which I have studied, may occur at any time after the blasto- 
sphere contains about 100 cells up to the period of invagination. The 
phenomenon of ingression in A. flavidula is not of constant occurrence, 
but when it does take place is similar to that represented by Goette 
(87, Taf. I. Figs. 1-5) for the earlier stages of the blastula in A. 
aurita, It consists of a migration into the cleavage cavity of one or two, 
rarely more than three, of the cells of the blastospheric wall. With 
the exception that they assume a spherical form, because relieved from 
pressure, they are at first similar in size, as well as in nuclear and other 
characters, to the cells remaining in the wall. 

The study of ingression upon preserved material is attended with diffi- 
culty, since in any one specimen we have the condition at only one stage 
of development, and cannot say wh certainty what its condition has 
been in past stages, or what it might have been during some subsequent 
period. This can be determined only by studying the conditions exist- 
ing in other embryos killed at other stages, and arranging all in their 
probable natural sequence. In view of this fact, I have sectioned and 
examined several hundred embryos which were killed at different stages of 
development. As far as possible the results obtained from these sections 
have been verified by the study of embryos cleared and mounted whole. 
Although this ingression occurs before invagination, I have deferred the 
discussion of it until now, because invagination is constant in its occur- 
rence, whereas the ingression does not appear to be so; indeed, the 
majority of the specimens have shown no indications of it. 


MUSEUM OF COMPARATIVE ZOOLOGY. 119 


The subsequent history of these cells, as shown by the comparison of 
specimens of succeeding stages of development is both interesting and 
peculiar. I imagine that it is such cells as these to which Claus (90, 
p. 3) refers when he says: “Ich habe den vereinzelt eingetretenen zwei 
bis drei Zellen, weil sie nicht regelmässig in jeder Blastula sich ablösen, 
der am vegetativen Pole einwuchernden Zellenmasse gegenüber keine 
weitere Bedeutung beigemessen, so dasse ich dieselben zwar auf einer 
Abbildung darstellte, im Texte aber nicht besonders erwähnte, und bin 
auch jetzt noch der Ansicht, dass diese auffallend kleinen Zellen wieder 
rückgebildet werden und überhaupt nicht zur Bildung des Entoderms 
beitragen.” In my judgment, a part of the difference of opinion be- 
tween Goette and Claus is due to the fact that there are two kinds of 
cells which find their way into the cleavage cavity. These are the large 
cells described by Goette as beginning to be formed at an early stage of 
the blastula, and much smaller cells, of which I shall have more to say 
hereafter, that make their appearance only at later stages of develop- 
ment. Claus seems to have seen “very small cells,” and to have 
assumed that they were equivalent to the large cells figured by Goette. 
I am unable to say with ‘certainty that the cells seen by Claus are the 
equivalents of those figured by Goette, but Claus assumes that they are, 
and I have the more reason to believe it because the large cells are of 
more frequent occurrence than the small ones. But if this be so, I do 
not understand how Claus could speak of them as “diese auffallend 
kleinen Zellen.” But however that may be, I have reason to believe 
that the supposition of Claus, that they ultimately degenerate, is 
correct. 

Soon after the ingression of a cell its nucleus undergoes changes 
which result in its disappearance as such, for instead of a nucleus there 
can be seen only one or more small, isolated, deeply stained particles, 
which I judge to be scattered portions of the nuclear chromatine 
(Plate II. Figs. 8 and 10). Even these are often wanting. I have 
said that this nuclear change follows soon after the ingression of the 
cell, because out of the numerous instances in which these cells have 
been present there is not one in which the nucleus retains its original 
condition after the cells in the wall of the blastula have given evidence, 
by their diminished size, that they have undergone division since the 
ingression took place. This conclusion is in part based on the assump- 
tion that at the time of ingression the ingressing cells are of about the 
same size as those which remain in the wall of the blastula. The in- 
gressing cells sometimes persist, without any further apparent changes 


120 BULLETIN OF THE 


than the disintegration of the nucleus, until the process of gastrulation 
is completed. Such cases are not as common, however, as others, where 
there is to be found ‘in the cleavage cavity material which appears as 
though it had resulted from the disintegration of similar cells. This 
material has a spongy or vacuolated appearance, and contains faintly 
staining bodies or granules similar to those found in the ectodermic 
cells; it does not possess definitely circumscribed boundaries; on the 
contrary, it fills the cleavage cavity more or less completely, but is not 
of uniform density throughout. The fact that this material is not homo- 
geneous, and that it contains granules, etc., prevents the conclusion that 
it has been produced as a simple secretion into the cleavage cavity, 
although it may have been formed in part by such a process. The fre- 
quent association of this material with ingression cells in the same spe- 
cimen (Plate II. Fig. 8), and the lack of other ways of accounting for 
its presence, lead me to believe that it is produced by the disintegration 
which I have suggested. 

There is another peculiarity of the development which I believe to 
be connected with this process of nuclear disintegration. It is this: 
after having once entered the cleavage cavity the immigrating cells seem 
to lose their power of division, and consequently do not become more 
numerous, while the cells composing the blastospheric wall undergo 
repeated divisions, as is shown by their increased number and dimin- 
ished size. 

The number of these immigrating cells is small, usually only one or 
two, very rarely more than three, so that I hav®not been successful in 
finding the “ Verbindungsglieder” connecting the conditions shown by 
Goette (87, Taf. I.) in his Figures 5 and 6, which Claus (90, p. 4) re- 
garded as essential to the substantiation of Goette’s view of the method 
f gastrulation. 

Reference has been made to the fact that in some cases the ingrowing 
cells persist both during and after the process of invagination. In the 
latter case, they are to be found in the celenteron rather than in the 
cleavage cavity. Figure 11 (Plate II.) is drawn from such a specimen. 
Figures 9 and 10 represent two sections of one individual in which the 
invagination is not completed, and furnish a hint as to the process by 
which the cells pass into the cwlenteron from the cleavage cavity. The 
entoderm being composed of less closely fitting cells than the ectoderm, 
doubtless admits the passage of the large immigrated cells through it 
more readily than the latter would (Plate II. Fig. 9). The immigrated 
cell is of course passive in this process. Since it is prevented by the 


MUSEUM OF COMPARATIVE ZOOLOGY. 121 


firm wall of the ectoderm from escaping, the pressure exerted upon it by 
the enlarging entoderm is probably sufficient to cause it to be forced 
through the entodermic wall into the cwlenterie cavity. From Figure 10 
it is to be seen that one cell has already reached the gastral cavity. In 
speaking of these peculiarly situated cells I have thus far assumed that 
they are such as originally reached the cleavage cavity by an early 
ingression, where, with changed nuclear condition, but apparently with 
no further alteration, they have remained until the time of gastrula- 
tion. That this is their source is evident from the following consid- 
erations. First, the small diameter of the blastoporie canal (Plate II. 
Fig. 7), which is from the same series as Figures 9 and 10, precludes 
the assumption that they might have entered the gastrula cavity from 
without. Secondly, in their large size and general appearance they are 
unlike the cells of either ectoderm or entoderm at any time during 
gastrulation, and so could not have been derived from those sources 
during that process. Thirdly, they do correspond in size and general 
characters, except in their nuclear conditions, with the cells of the 
blastospheric wall as the latter appear at the time when ingression 
takes place. 

It is difficult to state either the cause or the purpose of this immigra- 
tion. That it is not essential to the welfare of the embryo, either by 
affording nourishment to the developing cells of the entoderm, or in 
any other way, is evident from the fact that in a large number of cases 
it does not occur. That it is not an inherited tendency, derived from a 
more primitive method of gastrulation by ingression, is probable from 
the fact that the immigrating cells do not appear to have any share 
whatever in the formation of the entoderm. On the other hand, its 
oceurrence scems to be much too frequent to be considered as acci- 
dental. 

I have stated previously (p. 119) that two very different kinds of cells 
are to be found at times in the cleavage cavity. Besides the large immi- 
grating cells already described at length, I have found in a much smaller 
number of cases very small cells (Plate I. Fig. 2), one or two in num- 
ber, that appear precisely like the deep-lying ectodermal cells already 
described. Because: of their strong resemblance to the latter, their 
exceptional occurrence, and the fact that they do not appear until after 
the beginning of the development of the deep-lying ectodermal layer, 
I incline to the opinion that they are derived from that layer, and that 
their occurrence is entirely accidental. 

At first it appeared to me surprising that two investigators could 


122 BULLETIN OF THE 


reach such different conclusions as those published by Claus (’83 and 
90) and Goette (87), concerning the method of gastrulation in the 
same animal, A. aurita. Since studying this process in A. flavidula, it 
seems less strange. The results obtained from my first sections led me 
to think that the conclusions reached by Goette would be confirmed in 
the case of A. flavidula. Better staining, thinner sections, and more 
accurate orientation have made it certain, however, that the method of 
gastrulation in this species is much more in accord with the description 
given by Claus, and that the process really is one of invagination. 
Certain considerations weaken my confidence in the position defended 
by Goette. A comparison of his Figures 6-9 (’87, Taf. I.) with some 
of my thicker sections, or with those which were made when the gastrula 
was go oriented as not to be cut parallel to the blastoporic canal, makes 
it appear, to me probable that his results are based upon similar inade- 
quate sections, In Figure 8 (Plate IL) there are only about one half 
as many nuclei visible as there are cells, the nuclei of a portion of the 
cells being contained in adjacent sections. In figures of corresponding 
stages of A. aurita as represented by Goette (’87, Taf. I.), nuclei are 
figured in nearly all the cells. I believe this to be evidence that his 
figures were drawn from thick sections. The blastopore, because of its 
very small diameter, is quite easily overlooked in thick sections, and 
especially if the plane of sectioning is somewhat oblique to the longitu- 
dinal axis of the blastopore. Since, as previously stated, the nuclei of 
the entodermal cells are usually situated in the portion of the cell near- 
est the ewlenteron, it is easy to find in thick sections of an in yaginating 
embryo conditions like those represented by Goette in his Figures 6-8, 
My Figure 12 (Plate II.) reproduces a section of the same series as 
that represented in Figure 3 (Plate I). The intervening section (not 
figured) is quite similar to Goette’s Figure 8. An examination of the 
cells bordering the blastoporie canal in Figure 3 will show how sections 
like Figure 12, or such as are a little oblique to the chief axis of the 
embryo have the appearance of containing immigrating cells, Such 
sections also exhibit the flattening in the region of the shorter cells to 
which Goette (’87, p. 4) has called attention in the following words : 
«Schon während der Gastrulation zeigt sich eine Stelle des Keims im 
Bereich seiner kürzeren Zellen etwas abgeplattet.” 

Additional considerations increase the probability of the correctness 
of the view which I have advanced to explain Goette’s error. With 
advancing stages of development, I have found an increase in the num- 

ber of the cells composing the ectodermic wall. This is undoubtedly 


MUSEUM OF COMPARATIVE ZOOLOGY. 128 


subject to slight individual variations, but the number of such cells is 
nevertheless in quite close correlation with the stage of development. An 
examination of Goette’s Figures 6-9 (’87, Taf. I.) reveals such a simi- 
larity in the number and size of the cells composing the ectoderm in each 
of the four supposed stages, that I am driven to the conclusion that 
they represent sections from specimens of a single stage of development, 
which may have been produced by cutting in planes having different 
relations to the chief axis of the embryo. 

When we consider that in the majority of embryos there are no signs 
of ingression, and that in the cases where it does occur the immigrating 
cells in some instances degenerate early, and in others persist undivided 
throughout the process of gastrulation, and that they at no time show evi- 
dences of even sharing in the formation of an entoderm, — and when we 
further reflect that all the conditions shown in Goette’s Figures 6-9 can 
easily be reproduced from sections of invaginating gastrule of a single 
stage of development, — it seems improbable that the entoderm of Au- 
relia develops even occasionally by ingression, At present, therefore, 
there seems to me to be no evidence that in this genus gastrulation 
occurs by both methods, invagination and ingression. 

The Scyphomedus® present several interesting variations in gastru- 
lation. The anomalous development occurring in Lucernaria is as far 
removed from the usual process as that group itself is from the other 
Seyphomedusw. According to MeMurrich (’91, p. 314), the solid plan- 
ula in Cyanea arctica is formed by the immigration of certain of the 
blastula cells. This planula is subsequently hollowed out, and gives 
rise to a structure like an invaginate gastrula, but it is formed without 
any invagination. In Cyanea capillata (Hamann, ’90, pp. 16, 17) there 
seems to be a solid ingrowth of cells from one pole of the embryo, and a 
simultaneous development of the cwlenteron. The entoderm of Chry- 
saora (Claus, ’83, p. 5, Taf. I. Fig. 21 4) is developed in away which is 
somewhat similar to that described by Hamann for Cyanea capillata, 
According to Claus (’83, p. 2, and ’90, p. 4), the gastrulation of Aurelia 
aurita approximates the method by invagination a little more closely 
than that of Chrysaora, since its cells are arranged in a single layer 
about the fissure-like coolenteron. Aurelia flavidula exhibits a still more 
nearly typical invagination, since the coolenteron is from the beginning 
an open sac-like cavity. Cotylorhiza tuberculata (Cassiopea Borbonica) 
has an invaginate gastrula which closely resembles that of Aurelia 
flavidula (Glaus, ’90, Taf. I. Figs, 2 and 3; Kowalevsky, ’73, Taf. II. 
Fig. 1). Finally, in Pelagia noctiluca and Nansithoé marginata, as 


124 BULLETIN OF THE 


shown by Metschnikoff (86, pp. 66-68, Taf. X.), there is a typical in- 
vagination. 

If the observations of McMurrich (’91, p. 314) on Cyanea arctica are 
substantiated, we have among the Scyphomeduse one example of the 
formation of a sterrula by ingression, with the subsequent formation of a 
gastrula-like structure, without an invagination. From the preceding 
summary it is to be seen that there are in Scyphomedus® two cases in 
which the mode of gastrulation appears to be intermediate between 
ingression and invagination, and at least four cases of unquestionable 
invagination. If, in the light of so much variation in the mode of 
gastrulation in this group as is shown by the few forms studied, it is 
safe to conclude that any one mode is typical, that mode would cer- 
tainly appear to be invagination, and not, as Hamann and McMurrich 
have recently maintained, ingression. 


CAMBRIDGE, June 20, 1891. 


MUSEUM OF COMPARATIVE ZOOLOGY. 125 


BIBLIOGRAPHY. 


Claus, C. 

’83. Untersuchungen über die Organization und Entwieklung der Medusen. 
Prag u. Leipzig, 96 pp. 

’90. Ueber die Entwicklung des Seyphostoma von Cotylorhiza, Aurelia und 
Chrysaora, sowie über die systematische Stellung der Seyphomedusen. 1. 
Arbeiten a. d. zool. Inst. Wien, Tom. IX. p. 85. 

Goette, A. 

’87. Abhandlungen zur Entwicklungsgeschichte der Tiere. Viertes Heft. 
Entwicklungsgeschichte der Aurelia aurita und Cotylorhiza tuberculata. 
Hamburg u. Leipzig, 79 pp. 

Hamann, O. 

’90. Ueber die Entstehung der Keimblätter. Ein Erklärungsversuch. In- 

ternat. Monatsschr. f. Anat. u. Physiol., Bd. VII. pp. 1-28. 
Kowalevsky, A. 

"73. Untersuchungen über die Entwicklung der Coelenteraten. 
Gesellsch. Freunde Naturerkennt., Anthropol. u. Ethnog. Moskau, 1873. 
(Russian.) 

See also Hoffmann u. Schwalbe, Jahresbericht, Bd. IL p 279. 


Nachrichten 


McMurrich, J. P. 
’91. Contributions on the Morphology of the Actinozoa. II. On the Devel- 


opment of the Hexactiniw. Jour. Morphol., Vol. IV. p. 303. 
°91", The Gastrea Theory and its Successors. Biological Lectures delivered 
at the Marine Biol. Laboratory, Wood’s Holl. Boston, p. 79. 


Metschnikoff, E. 
’86. Embryologischestudien an Medusen. Wien, 159 pp. 


EXPLANATION OF FIGURES. 


All the figures were drawn from sections with the aid of an Abbé camera. The 


sections from which the figures were made were 5 u in thickness. 


Fig. 


‘ 


Samira. — Gastrulation in Aurelia, 


1. 


pw» 


PLATE I. 

ABBREVIATIONS. 
bl’ po. Blastopore. 
cav. sq. Segmentation cavity. 
Cl Immigrated cell. 
ceelent. Coelenteron. 
cog. Coagulum. 
ec’drm. Ectoderm. 
en’drm. Entoderm. 
nl, Chromatic portion of degenerated nucleus. 


nl. ec’drm. Nuclei of deeper portion of ectoderm. 


Figures 1-4. Sections to illustrate the nature of the invagination. 


An early stage of invagination. 460. 

A slightly later stage than that of Figure 1. x 540. 

A stage in which the invagination is well advanced. 885, 

A gastrula with invagination completed. 410. 

Section of a gastrula cut in a plane (equator) perpendicular to the axis of 
the blastoporic canal. X 385. 

Section of an older individual through the equator, showing increase in 

thickness of the entoderm. x 885. 


"RULATION IN AURELIA. 


Jae Sh 


SMITH.- Gas 
bl po. Be" 


bi po. 
eoelent 


endrm., 


endrm,,, 


PLI 


Smira, — Gastrulation in Aurelia, 


ia 


8. 


9. 


10. 


11. 


12. 


PLATE IL 

ABBREVIATIONS. 
bl’ po. Blastopore. 
cav. 8g. Segmentation cavity. 
cl. Immigrated cell. 
coclent. Coelenteron, 
cog. Coagulum. 
ec’drm. Ectoderm. 
en’drm. Entoderm. 
nl. Chromatic portion of degenerated nucleus. 


nl. ece’drm. Nuclei of deeper portion of ectoderm. 


Figures 7, 9, and 10 are from different sections of the same individual. 
Fig. 


Section through the blastoporic canal and nearly perpendicular to it. 
x 410. 

Section at a stage preceding invagination. It shows an immigrated cell 
in which the nucleus has degenerated. X 385. 

Section before the close of gastrulation, showing an immigrated cell in 
the segmentation cavity. X 410. 

Section from the same individual as Figure 9. It contains an immigrated 
cell in the colenteric cavity. X 410. 

Section of a gastrula with two immigrated cells contained in the colen- 
teric cavity. X 385. 

Section from the same individual as Figure 3, to show the appearance 
when the gastrula is cut parallel to, but at one side of, the blastoporic 


canal. X 885. 


geen OP Om, 


g” RER. o ® 


<a 


pA Ak 
fe Pe eto 


Piy ETA j 


No. 3.— Amitosis in the Embryonal Envelopes of the Scorpion. 
By H. P. Jonson." 


In the fall of 1889, at the suggestion of my instructor, Prof. E. L. 
Mark, I decided to work upon the problem of the so-called “ direct ” or 
amitotic division of nuclei. While in search of suitable material, my 
attention was called to a brief article by Blochmann (785), describing a 
very well marked amitotie division for the large nuclei of the embry- 
onal membrane of the scorpion. A number of Centrurus embryos 
were kindly given to me by my friend, Dr. G. H. Parker. These em- 
bryos had lain in 90% alcohol since the summer of 1886. The mode 
of fixation (for the purpose of studying the development of the eyes) 
was somewhat unusual; for, immediately after their removal from the 
mother, they were immersed in 35% alcohol, and thence carried up 
quite rapidly, through 50 and 70%, to 90%. Notwithstanding this 
rather crude method, the membranes were in excellent histological 
condition, in no way inferior to material afterwards prepared by the 
most approved methods of fixation. 

In addition to the material above mentioned, I received from Mr. 
Richard Goeth, of Burnet County, Texas, during the following winter 
and spring, about three dozen live specimens of Centrwrus (sp. incog.).? 
A lot that arrived in the latter part of May contained several pregnant 
females, with embryos in different stages. The scorpions were chloro- 
formed, and the ovarian tubes with the embryos enclosed were dissected 
out as quickly as possible. A number of killing agents were used, 
including Flemming’s weaker chrom-aceto-osmic, Rabl’s chrom-formic, 
Perenyi’s fluid, Kleinenberg’s picro-sulphuric, and Merkel’s fluid. 

For staining, I have used chiefly Ehrlich’s hematoxylin. Grena- 
cher’s alcoholic borax-carmine and Czokor’s alum-cochineal have given 
fair results. Safranin, employed according to Flemming’s method, I 


1 Contributions from the Zoölogical Laboratory of the Museum of Comparative 
Zoölogy, under the direction of E. L. Mark, No. XXX. 

2 This is the species used by G. H. Parker in his study on the development of 
the eyes (see Bull. Mus. Comp. Zoöl., Vol. XII. No. 6, p. 173, 1887), and was then 
undescribed. I am not aware that it has since received a name. 

VOL. XXII. — NO. 3. 


128 BULLETIN OF THE 


have found less serviceable than the stains above mentioned. After 
staining, the preparations were dehydrated, cleared with oil of cloves, 
and mounted in benzole-balsam. 

The embryo is enveloped by three epithelial membranes, the ovarian 
capsule, the membrana serosa, and the amnion, —named in order from 
without inward. 

The serosa and amnion are strictly embryonic structures, analogous 
to the footal membranes of the higher Vertebrates. There are two 
contradictory accounts as to the manner of their formation. Possibly 
they do not arise in the same way in all genera of scorpions. In a 
brief communication by Kowalevsky und Schulgin (’86, p. 526) upon 
the development of Androctonus ornatus, it is stated that they originate 
as a fold from the edge of the blastoderm, the outer layer of the fold 
forming the serosa, the inner the amnion. The fold grows up over the 
blastoderm, the edges coalesce, and the membranes finally separate from 
the ovum. ‘The more recent account by Laurie (’90, p. 114) states that 
in Huscorpius the serosa arises by a proliferation of the peripheral cells 
of the blastoderm, extends as a delicate membrane forward and back- 
ward over the egg, which it finally covers completely, and then becomes 
entirely separate from the blastoderm. The formation of the amnion 
begins when the serosa has covered about two thirds of the embryo, and, 
like the serosa, its origin is ectodermic. The amnion, however, “ never 
loses its connection with the epiblast as the serous membrane has now 
done, but remains attached to its edges and only extends round the 
egg as the epiblast extends” (p. 116). Unfortunately, I have not 
obtained sufficiently early stages of Centrurus to ascertain how its mem- 
branes arise, but, in removing the latter from the embryo, I have never 
found the amnion attached to the ectoderm. The membrane which I 


” Tat first wrongly took to be the 


have called the “ovarian capsule 
follicular epithelium, and under this supposition it was indicated as e’th. 
fol. in Figure 2. Like the follicular epithelium, it arises from the 
ovarian tube; but the follicle is formed as a diverticulum of the tube, 
previous to the maturation of the ovum, and serves as a nutritive cap- 
sule for the latter during its growth. The ovarian capsule, on the 
contrary, is that part of the ovarian tube which receives the ovum after 
fertilization, and enlarges to accommodate the growth of the embryo. 
The feetal membranes fit so loosely over the embryo that they can be 
easily removed in a single piece. In late stages, the ovarian capsule is 
readily separable from the membranes ; in earlier stages, it adheres closely 
to them. It is rarely possible to separate the serosa from the amnion, 


MUSEUM OF COMPARATIVE ZOÖLOGY. 129 


and a transverse section (see Fig. 2) shows only a trace of a dividing 
wall between them, although in surface view the cell walls of both mem- 
branes are clearly seen (Fig. 1). Metschnikoff (71, p. 219) describes 
the membranes of Scorpio (Euscorpius) italicus as connected with each 
other by delicate fibres, which terminate just over the amniotic nuclei, 
I have found such fibres in the earlier stages of my material, but not in 
the older ones, nor are they everywhere present in the younger mem- 
branes. The membranes of the Brazilian scorpion examined by Bloch- 
mann (’85, p. 481) were found closely applied to each other. 


I. The Serosa. 
Plate I.; Plate II. Figs. 14, 15; Plate III. 


The cells of the serosa have great superficial extent, measuring 
half a millimeter or more in diameter; but proportionally they are 
very thin. Their size is exceedingly variable, as may be seen by com- 
paring Figure 3 with Figures 11 and 13 of the same magnification, 
although the last two represent cells of only average size. Both 
small and large cells are,apt to .be aggregated in certain parts of the 
serosa, yet very small cells often occur sporadically in the midst of 
large ones. The cell walls are extremely distinct in late stages of the 
embryo, but in earlier stages are often difficult to trace in an ordinary 
stained preparation. As remarked by Blochmann, they have a distinct 
fibrous structure. The cells are irregularly polygonal in shape, usually 
elongated, sometimes nearly square or triangular, Not infrequently 
they are bounded by curved outlines (Fig. 13). 

The nuclei of the serosa measure from 25u to 60u or more in 
diameter, but as a rule are small in proportion to the cells (Figs. 1-3 
and 11-15). In the membranes of young embryos the nuclei are larger 
absolutely and in proportion to the cells than in old membranes. In 
face view the resting nucleus is nearly circular ; in section, it is seen 
to be considerably flattened, in accordance with the thinness of the cell 
(Fig. 2, nl. sr.). It occupies the full thickness of the serosa, and some- 
times causes a bulging of the cell at the point where it lies, as is shown 
in Figure 2. Blochmann states (’85, p. 480) that the nuclei of the se- 
rosa always cause that membrane to encroach inward upon the amnion ; 
but a dividing line between amnion and serosa is so seldom visible in 
Oentrurus, that T am unable to say whether such is the case. 

The nuclear membrane is thin, but clearly visible, except in nuclei that 
have undergone degeneration. The chromatic substance, or nuclein, is 


VOL. XXIIT, — NO. 8. 9 


130 BULLETIN OF THE 


for the most part in the form of granules distributed evenly throughout 
the nucleus. Indications of a reticular or filamentous structure are, 
however, frequently present. I believe there is a chromatic network 
throughout the nucleus, but the abundance of granular chromatin pre- 
vents one from tracing it. Several nucleoli are always present. They 
are extremely variable in size and shape, and in many cases appear to 
be only aggregations of granular chromatin. They take a stain with 
hematoxylin and carmine in no way different from the rest of the 
chromatin, except that it is more intense. 

A very large proportion (about four to one) of the cells of the serosa 
contain two nuclei. These pairs of nuclei have all arisen from single 
nuclei by amitotic division. It is obvious that division of the cell is not 
contemporaneous with, aud does not immediately follow, the division of 
the nucleus. In many cases, especially when the embryo is far ad- 
vanced, cell division probably does not occur at all. Very few cells out 
of the thousands I have examined have had more than two nuclei; but 
I have found several with three nuclei, and two cells with four. ‘This 
seems to be the maximum number. These cells of the serosa, therefore, 
are not to be classed with multinucleate cells in which the nucleus 
divides into a great number of irregular and unequal fragments. Here 
the division takes place in an orderly fashion, and division of the cell 
follows nuclear division in regular sequence, though not immediately. 

In every serosa examined, nuclei were found in process of division. 
Some preparations furnish many more examples of-division than others ; 
and occasionally three or four adjacent cells will contain dividing nuclei 
(Fig. 15). Very frequently, however, only one or two dividing nuclei 
will be found in the whole serosa. It cannot therefore be supposed that 
nuclear division is frequent; and I have found that there are more cells 
with dividing nuclei in the membranes of late stages of the embryo 
than in the earlier ones. 

The first sign of approaching division is an elongation of the nucleus 
(Fig. 4), almost always parallel to the long axis of the cell. Naturally, 
the clongation progresses by insensible gradations from the nearly circu- 
lar form of the resting nucleus, so that one cannot say positively that 
the nucleus is going to divide until the elongation has become marked. 
The absolute amount of elongation varies greatly, and is less in the 
membranes of young embryos than in those of older ones. The example 
represented in Figure 4 is from an old membrane, and shows almost the 
extreme of elongation. This stage, while giving not the slightest evi- 
dence of ordinary mitosis, is characterized by a longitudinal arrange- 


MUSEUM OF COMPARATIVE ZOÖLOGY. 131 


ment of the chromatic substance, as indicated in Figure 4. The effect 
is most marked upon the nucleoli. Blochmann (85, p. 482) found only 
two nucleoli at this stage, and these were usually situated one at each 
end of the elliptical nucleus. Where there are several nucleoli, as is 
usually the case with the nuclei I have studied, there is an approxi- 
mately equal distribution of them to the daughter nuclei. The nucleoli 
vary so much in size and shape, that it is impossible to say how precise 
is the apportionment of chromatin by this method. 

Most nuclei in the elongated condition already show a slight constric- 
tion, generally more marked on one edge than on the other (Fig. 4). 
If no further elongation takes place, the constrietion becomes deep and 
narrow, as represented in Figures 5 and 12. This style of division is 
characteristic of young membranes, and gives rise to daughter nuclei 
which lie close together, or even in contact (Fig. 13). It is doubtless a 
more vigorous and rapid type of division than that found in the older 
membranes, to be described directly. If the nucleus continues to elon- 
gate while constricting, it assumes the dumb-bell form represented in 
Figures 6 and 7. The daughter nuclei, at first ovate or pyriform, be- 
come rounder as the connecting thread becomes thinner. Division of 
this type is almost confined to old membranes; I have rarely found it 
in those from young embryos. 

The nuclei represented by Figures 6 and 7 show more clearly than 
usual a peculiar arrangement of the chromatic threads. The filaments 
have the appearance of a fascicle of slender rods, which lie very close 
together in the connecting bridge, and thence radiate into both daughter 
nuclei. They are stainable both with carmine and hematoxylin. Some- 
times these threads can be resolved into rows of granules (Fig. 7, right- 
hand daughter nucleus). The later stages also show traces of these 
longitudinal threads (Figs. 8, 9, 10). In the example represented by 
Figure 6, the nucleoli partook of the general longitudinal disposition of 
the chromatic substance, but were probably arranged in this manner at 
an earlier stage of division, as explained for Figure 4. In the later stages 
of division, this arrangement of the nucleoli is gradually lost. 

The final stages, represented in Figures 8, 9, 10, may be briefly de- 
scribed. These stages are far commoner than the early ones; hence, it 
must be supposed that they require more time. The constricted por- 
tion is drawn out into a thin, deeply staining thread. This thread 
undoubtedly contains chromatin, and in a peculiarly condensed form. 
In this respect these nuclei differ from the nuclei of the Malpighian 
vessels of Aphrophora spumaria, as described and figured by Carnoy 


132 BULLETIN OF THE 


(85, Plate I. Fig. 7); for the connecting thread in the dividing nucleus 
of Aphrophora remains unstained, and therefore contains no chromatin. 

The dividing nucleus represented by Figure 8 is peculiar in several 
respects. In the first place, the daughter nuclei are very unlike in 
form, though this is by no means unusual with dividing nuclei from old 
membranes. All the stainable nucleoli are in one daughter nucleus, 
while the other still shows a faint longitudinal arrangement of its 
chromatic threads. The sharply stained connecting thread is notched 
at a point midway between the daughter nuclei, probably indicating 
the place where, at a later stage, rapture would have occurred. The 
daughter nucleus on the left is nearly destitute of chromatin in the 
crescent-shaped space lying next the connecting thread, and an inner 
contour line is visible (x), from the central point of which a stainable 
cord extends to the proximal end of the connecting thread. I have seen 
a similar appearance in the late stages of other dividing nuclei, and it 
undoubtedly indicates the manner in which the daughter nuclei some- 
times attain a rounded form. Occasionally, however, daughter nuclei 
entirely separate from each other have a conical or tapered form. 

In the last stages of division, the connecting thread is drawn out to 
extreme tenuity (Figs. 9 and 10). So exceedingly fine does this thread 
become, that, with the highest power accessible to me (Zeiss’s homoge- 
neous immersion objective yg), I could barely trace its course through 
the cytoplasm, though in most cases I made out that it was continuous 
from nucleus to nucleus. It is finally broken at or near the centre, and 
the proximal tips, as Blochmann suggests, are probably absorbed by the 
daughter nuclei. In even so late a stage as that shown by Figure 10, 
the longitudinal chromatic filaments are still perceptible. The right- 
hand daughter nucleus contains four loop-shaped bodies that strongly 
resemble chromosomes. They are, however, almost unstained by hama- 
toxylin. 

Blochmann states (85, p. 482) that ın no case did he find a division 
of the cell following the division of the nucleus. As already said, the 
great proportion of binucleate cells renders it certain that cell division 
is not an immediate consequence of nuclear division. Although I have 
carefully examined great numbers of binucleate cells, I have only 
once seen a cell wall in process of formation (Fig. 27). Yet one finds 
plenty of evidence that cell division does take place. Pairs of cells like 
those in Figure 11 are of frequent occurrence. It is safe to infer, I 
think, from the arrangement of the binucleate cells which surround 
these, as well as from the correspondence in size and shape of this pair, 


MUSEUM OF COMPARATIVE ZOÖLOGY. 133 


that they have arisen from an elongated binucleate cell by the forma- 
tion of a divisional cell wall. In one instance, I have found a cell wall 
fully formed before division of the nucleus was completed (Wig. 27). It 
cuts across the fine connecting thread at about the middle point of the 
latter. This must be considered as in some degree abnormal, especially 
since it was found in a serosa the nuclei of which had evidently degen- 
erated. 

Although division of the cell is almost always accomplished by the 
formation of a cell wall, I have found several constricted cells, showing 
that division may be partly, or even wholly, effected in this manner. 
Sometimes the constriction is so deep that the opposite walls meet 
(Fig. 28) ; but it is more usual to find that, after the cell has become 
considerably constricted, a cell wall is formed joining the inward curves 
of the constriction, and completing the division, At first, I thought it 
possible that the constriction was mechanically produced by the pres- 
sure of growing cells on either side. But this would not explain the 
invariable occurrence of the constriction at precisely the point where it 
would take place in a free cell, — equidistant from the daughter nuclei. 
Furthermore, the curvature of cell walls (see Fig. 13), which is almost 
certainly caused by the growth of cells and consequent tension, has no 
reference to the position of the nuclei. 

As far as can be judged, the daughter nuclei are, as a rule, of equal 
size, and alike in shape. I have found many instances of beautifully 
symmetrical division (Figs. 9 and 1 0); but the nuclei of the serosa are 
not altogether exempt from the irregularities that seem to be inseparable 
from amitotic division wherever it occurs. Sometimes the resulting 
nuclei are obviously unequal (Fig. 13), even in young membranes; and 
in old membranes, where the nuclei have undergone degeneration, not 
only are the daughter nuclei extremely irregular in shape, but often 
very dissimilar in size. 

Relations of the Nuclei to the Cell. — A very brief examination of a 
preparation of the serosa convinces one that the nuclei are symmetri- 
sally arranged in the cells. When there is but one nucleus, it occupies 
the centre of the cell; when there are two or three nuclei, each presides 
over a half or a third of the cytoplasm. ‘This arrangement is so con- 
stant, that any marked deviation from it catches the eye at once. In- 
stances of decidedly unsymmetrical arrangement of nuclei, one of which 
Figure 13 represents, are very unusual, As regards elongated cells, the 
daughter nuclei lie in the long axis of the cell, and at approximately 
equal distances from its ends. Occasionally, however, the nuclei lie in 


134 BULLETIN OF THE 


the short axis (Fig. 12), and much more frequently are placed obliquely, 
as in cell a, Figure 14. We would suppose that, in the event of division 
of an elongated cell with nuclei lying transversely, the cell wall would 
pass longitudinally between the nuclei; but I have not been able to find 
evidence of longitudinal divisions. From the large number of cells with 
nuclei lying obliquely, one would infer that oblique division of the eell 
often took place. I am unable to discover, however, that such is the 
case ; and it seems extremely probable that the divisional plane of the 
cell does not always coincide with that of the nucleus, 

I have found about 25 cells of the serosa with three nuclei. This 
seems to be a matter of individual variation in the make-up of the 
membrane, for all but three of the trinucleate cells were in membranes 
from the brood of a single scorpion, and membranes from some broods 
appear to have none, I have in one instance found a group of tri- 
nucleate cells (Fig. 14, 4, 2, 3, 4). At this spot nuclear multiplica- 
tion has outstripped cell multiplication. It is nearly always easy to see 
which of the two original nuclei has divided, for we find two of the 
nuclei smaller than the third, and nearer to each other than to the 
latter. In cell 2, for instance, the pair of nuclei on the left have arisen 
from a nucleus occupying a position about midway between them. The 
same statement would doubtless hold true for the two nuclei on the right 
in cell 3, and here the odd nucleus is elongated. When the cell is long 
and the nuclei all lie in the longitudinal axis, as is the case in cell if. 
it is usually impossible to determine which of the two original nuclei 
has divided ; for the nuclei are equidistant, and nearly alike in size. 
Another type of equidistant nuclei is shown in cell 4, — a distribution 
quite as characteristic of very large, broad cells as the linear arrange- 
ment is of elongated cells. I have spoken of the division of one of the 
two original nuclei as though it always took place after the nuclei were 
completely separate, and had taken their positions in the cell. This 
seems to be the usual method, for I have several times found one of 
the original nuclei in the act of dividing. But it is possible, of course, 
for thern to arise by a tripartite division, in which the three nuclei 
would be formed simultaneously. I have found only one instance of 
a true triple division, represented in Figures 29 and 30, and as this 
occurred in a serosa which had plainly undergone degencration, I do 
not consider it as altogether normal. It will be noticed that the origi- 
nal nuclens became trilobed, and that the lobes became daughter nuclei 
of approximately equal size by the formation of three divisional planes, 


mecting at the centre of the original nucleus, The daughter nuclei on 


MUSEUM OF COMPARATIVE ZOÖLOGY. 135 


the right are still united to each other by strands at the corners. Very 
similar tripartite divisions were found by Overlach (85, Plate XI. Figs. 
35 and 41) in the epithelium of the cervix uteri. In two other cases, 
I have found one of the daughter nuclei in a late stage of division 
(Figs. 31, 32) itself elongating and undergoing constriction, It will 
be noticed that the constricted daughter nucleus is considerably larger 
than its mate. 

I have found but two cells with more than three nuclei, and these both 
contained four. This condition is brought about by the division of both 
nuclei of a binucleate cell. On a priori grounds, one would reason that 
quadrinucleate cells would be nearly as abundant as those with three 
nuclei, for, apparently, it must often happen that a pair of daughter 
nuclei, arising as they do by a symmetrical and accurate constriction, 
are ready to divide at almost the same moment. Yet there are doubt- 
less influences which operate to prevent the division of one of the 
nuclei. Although it is of course impossible to generalize on the char- 
acteristics of quadrinucleate cells, it may be of interest to mention the 
peculiarities of the two found. They are both large cells, of nearly 
equal width at the ends, and the breadth of both exceeds half the 
length. In one, both pairs of nuclei lie transversely, showing that the 
second divisional plane was at right angles to the first. In the other, 
represented in Figure 33, the lower pair of nuclei lie in the longitudinal 
axis, the upper pair almost transversely. One of the quadrinucleate 
cells is considerably larger than any cell near it, while the other (Fig. 
33) though by no means small, is of much less dimensions than the im- 
menso bi- and uninucleate cells around it. I am unable to assign any 
reason for the multinuclear condition of this cell. One fact, however, 
is worthy of note. The united volume of its four nuclei does not 
exceed the bulk of the single nucleus of a neighboring cell. One can- 
not, of course, ascertain what the size of the primitive nucleus of the 
multinucleate cell was, but it is very improbable that it exceeded in 
volume the nucleus of the uninucleate cell in question, for the latter 
cell is considerably the larger of the two, and throughout this serosa the 
sizo of the nuclei bears a direct ratio to the size of the cells. 

As regards the influence or influences impelling nuclei to divide 
independently of the division of the cell, nothing very definite can be 
stated. It is certain that the absolute or relative size of the cell has little 
or no influence upon the division of the nucleus. There are cells of all 
sizes, from the largest to the very smallest. (Fig. 3), which are binucle- 
ate; and it is usual to find, side by side with bi- or multinucleate cells, 


136 BULLETIN OF THE 


others with a single nucleus that are actually larger than the former 
(compare the cells in Figure 14). In such cases, the single nucleus is 
always larger than the daughter nucleus of the other cells. I am unable 
to see that multiplication of nuclei in the cell leads to any immediate 
increase of nuclear material. The more they divide, the smaller they 
become. Probably the most important office of division is a more 
extensive distribution of nuclei throughout the cytoplasm, with correspond- 
ing increase of nuclear surface; and this, considering the great superfi- 
cial extent of the cells, and the comparatively small size of the nuclei 
(at least in the older membranes) must be a matter of some importance 
for the activities of the cell. It is especially so in the case of elongated 
cells. If such cells have but a single nucleus, a large part of the 
cytoplasm must be remote from it; and if the nucleus is at the centre 
of the cell, the cytoplasm at the ends of the cell will be most remote, 
So, to restore the equilibrium between cytoplasm and nuclei, the nucleus 
must elongate in the longitudinal axis of the cell, and the daughter 
nuclei move toward the ends of the cell. 

As a matter of fact, nearly all elongated cells have two nuclei, and 
these lie in the long axis of the cell, usually rather nearer its ends than 
toeach other. It cannot be denied that many short or squarish cells 
also contain two nuclei; and, conversely, a few much elongated cells 
can be found that have but one. In the latter case, it is interesting 
to observe that almost invariably the nucleus has begun to elongate 
in the longitudinal axis of the cell, and is often far advanced towards 
division. We can say almost with certainty, then, that such cells are 
of recent formation, and that the equilibrium between cytoplasm and 
nucleus is promptly restored by division of the latter. Itis true that 
cases like that represented in Figure 12, where nuclear division takes 
place in the short axis of an elongated cell, cannot be explained in this 
manner. Such instances are so rare that they might almost be con- 
sidered as abnormal; but the difficulty of the matter lies in the fact 
that we get all gradations between nuclei ranged in the true longitudi- 
nal axis, and those placed in the transverse axis. It is common to find 
them lying more or less obliquely in the cell, though the obliquity is 
seldom so great as to prevent them from practically fulfilling the con- 
ditions of the hypothesis, 

It is not supposable that all the agencies impelling nuclei to divide, 
and controlling the direction in which division shall take place, reside 
in the cytoplasm ; possibly the most potent of them exist in the nucleus 
itself. That axial differentiation, with definite pole and antipole, is as 


MUSEUM OF COMPARATIVE ZOÖLOGY. 137 


characteristic of the resting nucleus as of the mitotic nucleus, was 
postulated by Rabl (’85, p. 323) from a careful study of the chromatic 
network in the “skein stage ” of mitosis. In a recent paper (’89, pp. 23, 
24), the same writer states that the “polar depression,” usually visible 
in young daughter nuclei, persists much longer than usual in the epi- 
thelial nuclei of the Triton ; so that for these mitotically dividing nu- 
clei it is highly probable that polar differentiation is always present in 
the resting state. Carnoy (’85) has shown that, in the resting nuclei 
of the testicular cells of certain Arachnids, the chromatic filaments are 
distinctly arranged with reference to a definite axis (Planche V. Figs. 
165-169), and Van Gehuchten (’89) has found the same in glandular 
cells of a Dipterous insect, Piycoptera contaminata. 

It is obvious that the discovery of an “organic axis,” as Van Gehuch- 
ten calls it, in amitotically dividing nuclei is more difficult, for here there 
is no polar depression or longitudinal arrangement of chromatic fila- 
ments to indicate its direction in the resting nucleus. It is usual for 
each division of the nuclei of the serosa to take place at right angles, or 
nearly so, to the plane of the previous division. This is well seen in 
many multinuclear cells,,where one or both pairs of nuclei lie trans- 
versely in the cell, and therefore at right ‘angles, or nearly so, to the 
direction of the first division (see cells 2 and 3, Fig. 14). In other 
cases, however, two consecutive divisions take place in the same direc- 
tion (Fig. 14, cell 2). It occurred to me that possibly there was an 
organic axis in the nuclei of the serosa which in some cases exerted a 
controlling influence upon the direction in which division took place, 
but which in most instances was counteracted by influences resident in 
the cytoplasm. Transverse divisions of the nucleus (Fig. 12) could then 
be accounted for by assuming that the influence of the organic axis is 
dominant in these cases, while oblique divisions would be explainable on 
the ground that neither influence was predominant, but that both acted 
with about equal force in directions at right angles to each other. A 
question of interest in this connection is, whether, when the cytoplasmic 
influence is dominant, and tends to make the nucleus divide in a plane 
parallel to its organic axis, division actually does take place in that 
direction. If such were the case, an organic axis would be a fact of 
slight morphological importance, and the longitudinal arrangement. of 
chromatin, which takes place in the earlier stages of constriction (Figs. 
4, 6, 7), might occur in any direction, without reference to an organic 
axis. If, on the contrary, it were necessary that the longitudinal fila- 
ments should be arranged parallel to the organic axis, in order that 


138 BULLETIN OF THE 


division might take place transversely to the axis, this result could still 
be attained by a rotation of the nucleus, even when the tendency was for 
the nucleus to divide at right angles to the previous division. It is 
obvious that rotation would occasionally be apparent, provided it took 
place soon after division, and previous to the absorption of the proximal 
end of the connecting filament. I examined a large number of prepara- 
tions to find evidence of rotation, but I must admit that the evidence 
was slight, and hardly sufficient to establish the hypothesis which I had 
formulated. It is therefore put forth provisionally, in the hope that it 
may lead to further investigations in this line. 

The most striking instance of rotation was found in one of the quadri- 
nucleate cells (Fig. 33, nuclei a and b). It is evident that three nuclear 
divisions have taken place without any division of the cell, producing 
two, three, and four nuclei. The arrangement of nuclei makes it rea- 
sonably certain that the dower pair arose by division of one, and the 
upper pair by division of the other nucleus of the binuclear stage. 
Only under this supposition could the daughter nuclei of that stage 
have had the normal arrangement, to which all the neighboring cells 
rigidly conform. We further find, that, while the upper pair of nuclei 
has arisen by a division in the long axis of the cell, the lower pair has 
been produced by division in the transverse axis, and therefore in con- 
formity with the law previously stated (p. 136). One nucleus of each 
pair (a and b) retains a remnant of the connecting filament, which is 
directed, not toward the sister nucleus, but to a point 90° distant from 
it. This condition could have been brought about only by rotation of 
the nuclei, which in both cases has been through an are of 90°. 

In the serosæ from older embryos, the daughter nuclei almost inva- 
riably recede from each other in the course of division, The amount 
of recession is governed by the length of the cell (Fig. 15). In the 
younger membranes, as already stated, the constriction is deep and 
narrow, so that the nuclei not infrequently lie very near together 
(Fig. 13). In these young membranes, however, the nuclei are larger, 
and the cells are usually smaller, than in the old membranes. Since, 
moreover, the large binucleate cells of young membranes almost always 
have their nuclei symmetrically placed at the ends, it is probable that 
the nuclei gradually move apart after division, as the cell increases in 
size. 

It will be seen that my interpretation of the primary cause of the 
division of these nuclei agrees in part with the hypothesis advanced by 
Chun (’90) for the explanation of amitotic division in general, This is, 


MUSEUM OF COMPARATIVE ZOÖLOGY. 139 


in brief, that the object of amitotie division is the distribution of nuclear 
material throughout the cytoplasm, with corresponding increase of nu- 
clear surface. He considers it the final phase of a series of conditions 
which begins with a simple lobed nucleus, and includes branched nuclei 
of various degrees of complication. In support of this interpretation, 
Chun lays stress on the statement that cell division, after an amitotic 
division of the nucleus, has seldom or never been observed with cer- 
tainty, thereby implying that amitosis cannot have in view the multi- 
plication of cells, I do not consider this as essential to the hypothesis, 
nor, in fact, do I believe him correct on this point. The evidence of 
cell division after amitosis seems to me abundant and conclusive. It 
was observed by F. E. Schulze (75) in Amoeba polypodia; by Ranvier 
(75), Bütschli (76), Flemming (82), Arnold (787), and others, in leu- 
cocytes; by Kükenthal (85), in the lymphoid cells of Annelids ; and by 
Carnoy (’85), in various cells of Arthropods. As the foregoing shows, 
there is abundant evidence that, in the serosa of the scorpion, division 
of the cell sometimes, at least, follows amitotie division of the nucleus. 
Furthermore, the extremely regular and well ordered manner in which 
the nuclei divide, and the similarity as to size and shape of the daughter 
nuclei, seem to me decidedly against the notion that the sole object of 
the division is to disseminate nuclear substance in the cytoplasm ; for 
in those cases where amitosis is not followed by division of the cell, and 
assumably takes place simply for the purpose of dissemination, the 
nuclear products are very variable as to number, size, and shape. 


II. The Amnion. 
Plate I. Figs. 1 and 2; Plate II. Figs. 16-20. 


The amnion is much thinner than the serosa, and like it is composed 
of a single layer of flat, polygonal cells (Fig. 1, am.). But, while both 
the cells and nuclei of the serosa have become enormously larger than 
the blastodermic cells from which they originated, those of the amnion 
have changed little as regards size. The boundaries of the amniotic 
cells are not always visible, and I find that preparations, even when 
hardened and stained in the same manner, show the greatest variation 
in this respect. As a rule, the cell walls in the amnion are sharply and 
clearly defined only in preparations of membranes from advanced em- 
bryos. The same is true of the cell walls of the serosa. 

In general, the amniotic cell has but one nucleus, which usually occu- 
pies the centre of the cell. Blochmann makes the same statement as to 


140 BULLETIN OF THE 


the number of nuclei in each cell, and he found no evidence of division 
among them. The outline of the nuclei, which measure about 15 p in 
diameter, is frequently somewhat irregular or lobed. Like the nuclei of 
the serosa, they are flattened tangentially (Fig. 2, nd. am.) ; but not- 
withstanding this, they cause an outward bulging of the cell upon the 
serosa, as shown in Figure 2. They contain always one or more highly 
refractive, deeply staining nucleoli. The rest of the scanty chromatic 
substance is in the form of minute granules, occasionally arranged partly 
in a very faint network (Fig. 18, 6 and c). As in the nuclei of the 
serosa, chromatic threads frequently unite the nucleoli. 

Division of the amniotic nuclei is of rare occurrence, Im only one of 
my preparations are dividing nuclei at all abundant. The division 
takes place without mitosis, but is of a different type from that of the 
nuclei of the serosa. The only alteration of the chromatin is possibly 
a change in the position of the nucleoli; I have not been able to detect 
any modification of the reticulum, The first sign of approaching divis- 
ion is elongation of the nucleus (Figs. 16 and 18, a). A deep narrow 
constriction appears at the equator of the nucleus (Fig. 17). This is 
followed by the formation of an equatorial septum, at once partition- 
ing off the nucleus into two daughter nuclei (Fig. 18, b). If there are 
but two nucleoli, it is the rule to find one in each daughter nucleus ; but 
where there are several, they are often unequally apportioned. After 
the formation of the septum, the daughter nuclei still adhere to each 
other, and division seems always to be attained by deepening of the 
equatorial constriction in the plane of the septum (Migs. 18, b, c, and 19). 
I have not found any evidence of a recession of the nuclei before 
division of the cell. Furthermore, the rarity of binucleate cells makes 
it very probable that cell division follows nuclear division promptly. 
As in the serosa, division of the cell takes place by the formation of a 
cell wall without marked constriction (Fig 20). The position of the 
nuclei in this figure, and the frequency with which nuclei are found 
near the boundaries of the cells (Fig. 1, am.) is evidence of the prompt- 
ness of cell division after the division of the nucleus. 

It is clear that Chun’s hypothesis will not hold in this case, for there 
is even less tendency than in the serosa to accumulate nuclei in the 
cell. This may be owing in part to the shape of the cell, for it is sel- 
dom elongated. It would seem that, in case the cell becomes elongated, 
nuclear division takes place and the cell divides immediately after the 
nucleus. The orientation of the nuclei with reference to the cytoplasm 
of their respective cells would then be accomplished by their migration 
to the centre of the cells. 


MUSEUM OF COMPARATIVE ZOOLOGY. 141 


III The Ovarian Capsule. 
Plate II. Figs. 21-26. 


The epithelium of the ovarian capsule is not often easily made out in 
ordinary stained preparations, for the nuclei of muscle fibres and con- 
nective-tissue cells lie not only just external to the epithelial nuclei, 
but frequently in the same plane with them. In most of my prepara- 
tions the boundaries of the epithelial cells cannot be seen at all, and I 
have therefore confined my attention mainly to those which show them 
distinctly. In shape, the cells are more or less irregular, oblong hexa- 
gons (Figures 24 and 25 represent typical shapes). The cell walls are 
broad and fibrillated, like those of the serosa, though the cells them- 
selves are smaller even than those of the amnion. The nuclei are not 
only larger in proportion to the cells, but often larger absolutely, than 
the amniotic nuclei. The amount and arrangement of the chromatin 
in the capsular nuclei (except in a certain phase) is almost precisely 
like that already described for the nuclei of the amnion, but there is usu- 
ally only one conspicuous nucleolus. The small amount of chromatic 
substance, aside from the nucleolus, has a granular appearance, but 
sometimes shows indications of a filamentous or reticular arrangement 
(see Figs. 21, 23, 24). Seen in face view, the nuclei are circular, and 
have a distinct nuclear membrane. The section (Fig. 2, nl. fol.) shows 
that they are less flattened than the amniotic nuclei. 

Here, again, we have amitotic division, and of precisely the same 
type as prevails in the amnion, Apparently, division is not of common 
occurrence, for I have been able to find only a few instances, and have, 
Figures 21, 22, and 23 show 


unfortunately, not seen its earliest stages. 
As each daughter nucleus 


the simple manner in which it is effected. 
contains a nucleolus, and the ordinary resting nucleus has but one, 
division of the nucleolus must precede division of the nucleus. In one 
important respect the division of these nuclei differs from that of the 
amniotic nuclei, The cell does not divide immediately after the nucleus, 
and consequently a great number of cells are binucleate. Some even 


contain three nuclei. I have obtained no evidence whatever of cell 


division. 


BULLETIN OF THE 


IV. Degenerative Changes. 
Plate II. Figs. 14, 24-26 ; Plate III. Figs. 28, 34. 


The striking difference in the appearance of cells and nuclei, and 
the different manner of division of the nuclei, exhibited by serosze of 
different ages, have frequently been referred to. Such changes, in 
part at least, I believe to be due to degeneration of the membranes, 
which, with the exception of the ovarian capsule, are temporary struc- 
tures, soon to be cast off by the embryo. Hence it is not surprising to 
find them undergoing degeneration in toto, The degenerative changes 
are about equally well marked in all three membranes; but on account 
of the great size of cells and nuclei, the changes are most conspicuous in 
the serosa. If the membrane comes from a young embryo, the walls 
of the cells are unstainable, and therefore often difficult to make out. 
The nuclei have a vesicular appearance, with smooth, rounded contour, 
abundant karyoplasm, and scanty chromatic substance. For this reason 
the nuclei seldom stand out clearly from the cytoplasm in a stained 
preparation, often being no darker than the rest of the cell. 

Serose from somewhat older embryos, while giving no sure signs of 
degeneration, have nuclei slightly different from those of the youngest 
membranes. The amount of chromatic substance appears to be larger. 
It is gathered into denser and more deeply staining masses, and the 
nucleoli become larger and more stainable (compare Figures 4 and 5, the 
former from an older membrane than the latter). Many nuclei at this 
stage become irregular in outline, and are more or less shrunken in 
appearance, changes which prepare the way for complete degeneration, 
found in membranes from the oldest embryos. The nucleus here 
becomes shrunken into a formless mass, which stains deeply and uni- 
formly. This condition seems to be due almost wholly to loss of the 
karyoplasm, for the nuclear membrane is seen to be drawn closely 
over the much condensed chromatic substance. The uniformly staining 
effect, however, is generally believed to be produced by the solution of 
a part of the chromatin in the karyoplasm ; this is best seen in nuclei 
that have not completely degenerated, where the deeply stainable solid 
chromatin is immersed in the less stainable matrix. Not all the nuclei 
in a membrane are affected to the same degree by the degenerative 
change. This is shown in Figure 14, where the nuclei of cell a, and that 
of the cell farthest to the left, are more affected than any others. But 
in the oldest membranes almost every nucleus has undergone extreme 


degeneration. 


MUSEUM OF COMPARATIVE ZOOLOGY. 143 


It is an interesting fact, tnat even the most thoroughly degenerated 
membranes have numerous nuclei in all stages of division, The divid- 
ing nuclei have undergone the same degenerative alteration as the rest. 
It is impossible to state whether these nuclei had begun to divide after 
the regressive change, or had been overtaken by these changes while 
undergoing division; and it is equally impossible to say whether 
degeneration would have prevented the nuclei from completing their 
division, The division is essentially like that of younger nuclei, but 
often unsymmetrical. 

Not all the degenerative changes are confined to the nuclei. The 
cells also give evidence of modification, Their walls become more 
distinct, not only because they are denser and thicker, but on account 
of their stainability with hematoxylin. The cytoplasm frequently has 
a reticulated structure, which. is densest about the nucleus. In the 
oldest membranes, certain large groups of cells have nuclei surrounded 
by a narrow bright ring, and outside this a much broader halo of a 
radiating structure, which takes a deeper stain than the rest of the 
cytoplasm (see Fig. 34). The appearance of the whole is strikingly 
like that of the “attraction spheres” of ovarian and other cells, but in 
this case has certainly nothing to do with mitosis. If the cell contains 
two nuclei, or a dividing nucleus, each daughter nucleus is surrounded 
by a halo. In early stages of division, however, the elongated nucleus 
has a single halo. I am unable to account for these appearances ; I do 
not regard them as attraction spheres, but rather as a result of degener- 
ation. The attraction sphere should radiate from a centrosome ; here 
it radiates from the nucleus as a centre. I may state, in passing, that 
my search for centrosomes in the serosa has been wholly unsuccessful. 
The pale ring is very generally present around nuclei that have under- 
gone degeneration. It seems to have no intimate connection with the 
radiating zone, being frequently found where the latter is absent. 

The life history of the serosa cells corresponds closely with that of 
certain cells in the Malpighian vessels of Aphrophora spumaria de- 
scribed by Carnoy (’85, p. 219). The cells at the two extremities of 
the tubes contain nuclei not greatly different from those of young 
serosæ, but the nuclei of the middle portion are irregular, jagged, 
and filled with amorphous chromatin. They therefore bear a strong 
resemblance to the degenerated nuclei of the serosa. Furthermore, the 
origin of the peculiar nuclei of the middle portion of the Malpighian 
vessel agrees closely with that of. the degenerated nuclei of an old 
serosa. It is thus described by Carnoy- (p. 220): “Sur les petites 


144 BULLETIN OF THE 


larves on rencontre tous les intermédiaires entre les noyaux des extrémi- 
tes et ceux du milieu. Peu à peu le boyau s'efface, le noyau lui-même 
se rétrécit et perd la régularité de ses contours à cause du plissement 
de sa membrane; à la fin la nucleine ne forme plus à Vintérieur qu’une 
masse compacte et homogene, & peu pres comme cela se présente dans 
la tête des spermatozoïdes.” In both cases the degenerated nuclei are 
found in stages of division ; in both, the cytoplasmic reticulum is distinct 
only in old cells, and where these cells are binucleate it is dicentric, 
with filaments radiating from the nuclei. The dicentricity of the binu- 
cleate cells is a point to which Carnoy calls special attention (p. 229). 
He considers that here the radiating filaments of the cytoplasmic retic- 
ulum answer to the polar asters of karyokinesis, and that the nucleus 
has the function of a centrosome. The same reasoning would apply 
to the degenerated cells of the scorpion’s serosa. 

The regressive metamorphosis undergone by the epithelial cells of 
the ovarian capsule (Figs. 24-26) is very peculiar. Here, again, the 
cell walls are affected in the same way as in the serosa and amnion, for 
they are not distinctly seen until after the nuclei have degenerated. 
Nearly all of the epithelial cells of an old capsule have two nuclei, which 
are dissimilar in size and appearance (Figs. 24 and 25). The smaller 
takes a rather deep, uniform stain, almost as dark as that of the chro- 
matin of the other. A nucleolus is always present, and frequently 
minute granules of chromatic substance. The uniformly staining char- 
acter of the nucleus is doubtless produced by chromatic substance held 
in solution by the karyoplasm, a condition of common occurrence with 
degenerating nuclei. The larger nucleus (Figs. 24 and 25) takes only a 
slight stain, owing to the scantiness of its chromatic substance, which is 
present in the usual form of isolated granules and an imperfect network. 
By examination of a large number of cells, I found nuclear differentia- 
tion of every degree, beginning with nuclei almost alike in size and 
stainability (Fig. 24), then passing to examples of marked dissimilarity 
(Fig. 25), where the pale nucleus has become almost invisible, and the 
smaller deeply staining one has attained a very sharp, definite outline. 
As the pale nucleus becomes more and more shadowy, its shape becomes 
irregular. Near cells of this sort others can be found which contain only 
a single deeply staining nucleus (Fig. 26), the other having disappeared 
altogether. In case of trinucleate cells, I have invariably found two of 
them to be of the pale sort. 

I am unable to offer any other explanation of these changes than that 
they are the result of degeneration or of decreased activity of the tissue, 


MUSEUM OF COMPARATIVE ZOOLOGY. 145 


But why one nucleus should become altered in one way, and the other 
in an entirely different manner, is difficult to say. A very similar dif- 
ferentiation of nuclei has been observed by Chun (90) in the egg germs 
of a Siphonophore (Stephanophys). He found only one nucleus in the 
youngest germs, while the middle-sized and larger egg cells contained 
two of different size, the larger being pale, and the smaller staining in- 
tensely. The smaller nucleus moves to the periphery of the egg and is 
no longer visible when the latter is ripe. The larger nucleus persists 
as the germinative vesicle. In only one instance did he see a stage 
that showed that the smaller nucleus budded out of the larger, Chun 
compares the small, deeply staining nucleus to the “ Stoffwechselkern” 
(macronucleus), and the pale one to the “ Fortpflanzungskern ” (micro- 


nucleus) of the ciliate Infusoria. 


Summary. 


1. The embryo of the scorpion is enveloped by three membranes, the 
ovarian capsule, the serosa, and the amnion. 

2. The ovarian capsule is-an enlargement of the ovarian tube ; the 
serosa and amnion arise from the blastoderm of the egg. 

3. Serosa and amnion are at first distinct, and joined to each other 
by minute fibres. These afterwards disappear, and the membranes 


coalesce. 


4, Tho serosa is composed of immense flat cells, very variable in size 
and-shape. The cell walls are fibrillated. 

5. The majority of the serosa cells have two large nuclei of equal 
size. There are rarely more than two. 

6. The nuclei are disk-shaped, have a distinct nuclear membrane, and 
chromatin in the form of granules and filaments, the latter forming an 
indistinct reticulum. There are usually several nucleoli. 

7. The cytoplasm of the serosa has a distinct reticular structure. 

8. Nuclear division in the serosa is amitotic, and takes place by con- 
striction, preceded by elongation of the nucleus. It is followed or ac- 
companied by recession of the daughter nuclei, which remain for some 
time connected by a fine strand, 

9, Constriction of the nucleus is usually accompanied by a longitudi- 
nal arrangement of some of the chromatic threads, radiating from the 
constricted part. The nucleoli are distributed about equally to the 
daughter nuclei. 

VOL. XXII, — NO. 3. 10 


146 BULLETIN OF THE 


10. Nuclear division may be followed by division of the cell, but not 
often immediately. The cell divides by the formation of a cell wall, 
either with or without constriction. 

11. The binucleate condition of cells is independent of their size ; 
but, in general, the size of the nucleus, or nuclei, is proportional to the 
size of the cell. 

12. Elongated cells of the serosa are generally binucleate. The 
nuclei almost invariably lie in the long axis of the cell, near the ends. 

13. A binucleate cell becomes trinucleate by division of one of its 
nuclei, and quadrinucleate by the division of both. Very rarely the 
division is tripartite, and the three nuclei are produced simultaneously 
from a single one. 


14. Division of the amniotic nuclei is also amitotic, but the constric- 
tion is supplemented by a septum at the equator of the elongated 
nucleus. 

15. There is apparently no rearrangement of the chromatic substance. 
Nucleoli are apportioned equally to the daughter nuclei. 

16. Division of the nucleus is quickly followed by division of the cell, 
so that binucleate cells are not common. 


17. The epithelium of the ovarian capsule is composed of small 
hexagonal or rectangular cells, which frequently contain two or more 
nuclei. 

18. The nuclei are very similar to those of the amnion, but usually 
contain only one nucleolus. 

19. Nuclear division is amitotic, and precisely like that of the amni- 
otic nuclei. Each daughter nucleus contains one nucleolus. 

20. No instance of cell division was observed. 


21. All three membranes undergo degeneration as the embryos ap- 
proach maturity. 

22. In the serosa the cytoplasmic reticulum becomes more distinct, 
and is seen to radiate from the nuclei. The cell-walls become stainable. 

23. The chromatic substance of the nuclei becomes grouped into 
dense masses ; the reticulum and nucleoli become more distinct. The 
outlines of the nuclei become irregular. 

24, As degeneration proceeds, the cytoplasm frequently forms a halo 
of radial structure around the nucleus. 


MUSEUM OF COMPARATIVE ZOÖLOGY. 147 


25. The nuclei finally become reduced to uniformly staining, irregu- 
lar masses of chromatin, which has partly entered into solution. Such 
nuclei are found in all stages of division. 

%6. In binucleate cells of the ovarian epithelium the nuclei become 
dimorphic. 

27. The chromatic substance of one of the nuclei enters into solution 
in the karyoplasm, and the nucleus becomes reduced in size. 

28. The other nucleus loses its stainability, and increases in size. It 
finally disappears. 


V. Discussion of Amitosis. 


As long as karyokinesis was supposed to be a uniform process, all the 
complicated details of which were carried out with the greatest exact- 
ness and in the same sequence, wherever it occurred, no one sought to 
homologize it with the little known and far simpler “direct” division. 
The latter had, apparently, so restricted a range, and had received so 
little attention, that its very existence was denied ; and it was generally 
anticipated that, in the few kinds of cells in which it was stated to oc- 
cur, a better technique and more careful study would reveal mitotic 
phenomena. This opinion seemed to receive confirmation by the dis- 
covery of mitotic division in leucocytes and the Protozoa, thus carrying 
mitosis back to the simplest types of cells and to the lowest forms 
of life. The ascertainment of two facts has brought about a radical 
change in our views regarding amitosis : (1) the variability of karyo- 
kinesis, including, in some cases, the omission of apparently essential 
steps ; and (2) the wide occurrence of amitosis, new instances of which 
are constantly coming to light in various parts of the Animal Kingdom. 
Inasmuch as it became necessary to recognize the existence of direct 
division, efforts were naturally made to find links connecting it with 
mitosis; the variability of both mitosis and amitosis seemed to lend 
strength to the theory which refers them to a single fundamental plan 
of division. In this scheme, amitosis is considered either as a primitive 
method from which mitosis was evolved, or else is looked upon as a 
degenerate form of mitosis, occurring in nuclei which, from their patho- 
logic or exhausted condition, have lost the power of dividing by the more 
complicated process. By fixing epithelium of the salamander larva with 
osmic acid, then treating it with Miiller’s fluid, and finally staining with 
hematoxylin, Pfitzner (’86") has shown conclusively that, even in cases of 
very perfect mitosis, the karyoplasm maintains its integrity, and divides 


148 BULLETIN OF THE 


by a simple constriction, as in direct nuclear division. This fact has led 
Waldeyer (’88) to the conclusion that karyokinesis in based upon the 
simple scheme of division conceived by Remak. He says: “I would 
interpret the facts in such a way that we have to regard as the funda- 
mental form the simple amitotic division, which is now proved for many 
cases ; it always takes place where the nucleus either is poor in chroma- 
tin, or when it does not matter about strict bipartition of the chromatic 
material. Should the latter be required, then we shall find mitosis, 
since it ig the most direct, most certain, and most simple manner in 
which an exact bipartition of chromatic substance is brought about.” 

It seems to me, however, that there are differences of so fundamental 
a character between mitosis and amitosis, as at present understood, 
that it is impossible to refer them to a single plan of division. Both, 
indeed, achieve the same result, — division of the nucleus, including its 
two constituents, chromatin and karyoplasm. In both cases, the karyo- 
plasm divides by constriction. In amitosis, the chromatin undergoes 
little if any change in preparation for division ; in mitosis it becomes con- 
solidated into a limited number of thickened rods or loops (chromosomes), 
which arrange themselves in the plane of division (“mother star,” 
“ couronne équatoriale”) and segment either longitudinally or trans- 
versely, the halves moving to opposite poles (“ diaster ”), and undergo- 
ing a reversed metamorphosis to form two daughter nuclei. If this 
were all there is to karyokinesis, — and in some cases the process is 
much simpler, — we might hope to find transitions between it and ami- 
tosis ; for there are examples of amitosis in which the chromatic net- 
work undergoes changes during division, and it would be conceivable 
that the highly organized changes of the chromatic substance during 
mitosis were either evolved from them, or that they were a simplifica- 
tion of the more detailed changes. In mitosis, however, other struc- 
tures besides chromosomes make their appearance, — the centrosomes, 
attraction spheres, and spindle. These structures are not known to take 
any part whatever in amitosis, and in this respect at least the two kinds 
of division are fundamentally different. The most recent workers upon 
karyokinesis agree in ussigning to the spindle rays the function of 
separating or dividing the chromosomes, and drawing (or pushing) the seg- 
ments towards the poles. The centrosomes are focal points towards which 
the spindle rays converge, and lie entirely outside the nucleus. The for- 
mation of the spindle has been carefully studied by many investigators 
of karyokinesis, and, while there are very divergent views as to its ori- 
gin and mode of action, the most recent workers in this field (of whom 


MUSEUM OF COMPARATIVE ZOOLOGY. 149 


E. van Beneden, Boveri, and Watase may be mentioned) are agreed that 
the spindle arises from the cytoplasm. The same view with regard to 
the spindle in the mitosis of vegetable cells was expressed by Stras- 
burger, Guignard, and other botanists. 

The centrosome, as a converging point for the spindle fibres and 
polar rays, plays a most important part in karyokinesis, and, so far as 
known, none at all in amitosis. The centrosome has indeed been found 
by Flemming (91) in leucocytes, which certainly divide amitotically ; 
but there it is a single structure, and as Flemming’s figures show, 
takes no part in the amitotie division of the nucleus, Whether it also 
remains passive during the mitotic division of leucocytes and in amitosis 
followed by division of the cell, is not known, It has been supposed by 
Carnoy (’85) that spindle rays were present in certain nuclei which 
divide amitotically, but this seems extremely doubtful, especially since 
they have no perceptible action on the chromatic substance. I believe 
it can be shown in every case of amitosis known, that the division of 
the chromatin is accomplished independently of chromosomes, spindle rays, 
or any other visible influence outside of the nucleus. 

The persistence of the nuclear membrane in amitosis, and its dis- 
appearance in mitosis, were formerly considered points of distinction 
between the two kinds of division; but, as is well known, more recent 
studies have shown that the membrane persists in many cases of un- 
doubted karyokinesis, especially among the Arthropods (Carnoy, ’85) and 
Protozoa (Gruber, ’83, R. Hertwig, ’84, Pfitzner, ’86°, and Schewiakoff, 
’88). Its presence seems to offer no obstacle to the karyokinetic 
changes, and Watase (’91) has pointed out that it need not prevent 
the formation of an extra-nuclear spindle, the rays of which may pene- 
trate the membrane. In the nuclei of Opalina ranarum, and in the 
micronuclei of Infusoria generally, where, according to all observers, the 
nuclear membrane persists, the mitotic division is accompanied by con- 
striction ; but the fact that constriction is here visible may be considered 
as in some measure a result of the persistence of the membrane, thereby 
making evident the outline of the karyoplasm. Yet constriction does 
not always take place when the membrane persists, for in the spermatic 
cells of Pagurus striatus, figured by Carnoy (’85, Plate VII. Fig. 244), 
the nuclear membrane is visible at all stages, and gives no evidence of 
constriction. 

The modification of the chromatic substance into chromosomes is 
usually the most conspicuous feature. of karyokinesis, and in most cases 
serves to distinguish mitotic nuclei from any of the amitotic ones. The 


150 BULLETIN OF THE 


chromosomes invariably include al the stainable substance of the nu- 
cleus, so that the presence of nucleoli in a nucleus undergoing constric- 
tion may be taken as perhaps the strongest evidence of direct division. 
The behavior of nucleoli in amitosis is of peculiar interest. Where 
there is a single nucleolus, it constricts previous to the constriction of 
the nucleus, according thus with the Remakian scheme. The division 
of the nucleolus, however, has rarely been observed. It was first de- 
scribed, I believe, by F. E. Schulze (’75), in the division of Ameba poly- 
podia; has since been figured by Carnoy (’85, Plate I. Figs. 10, 12, 13) 
for various amitotically dividing Arthropod cells, and by Hoyer (’90) 
for the intestinal epithelium of Zrhabdonema nigrovenosum. A peculiar 
modification of the nucleolus, and its division into four segments pre- 
vious to the constriction of the nucleus, was observed by Platner (’89, 
pp. 145-149) in the Malpighian vessels of Dytiscus marginalis. It is 
extremely probable that, whenever the nucleolus is a single and defi- 
nitely organized structure, it always divides previously to or during con- 
striction of the nucleus. Where there are several small nucleoli, they 
may indeed arrange themselves so as to be equally apportioned to the 
daughter nuclei ; but they are not known to divide, as the chromosomes 
in mitosis do. 

Amitotic division, even more than karyokinesis, is variable in its 
phenomena. It takes place by constriction, by formation of division 
planes, by gemmation, and by enlargement of one or more perforations 
(Arnold, ’88, Flemming, ’89). It is either simple or multiple, and it 
may or may not be accompanied by division of the cell. The resulting 
nuclei may be equal or unequal. Amitosis occurs throughout both the 
Animal and Vegetable Kingdoms ; but as far as animals are concerned, 
it is far the most frequent among unicellular organisms, amaboid cells 
(leucocytes), and epithelial tissues. There seem to be no authentic instances 
of it in connective tissues (except possibly the fat-cells of Arthropods, 
described by Carnoy), none in nervous tissue, and but one or two in 
muscle fibres (Carnoy, ’85, p,221). Not only the nuclei of fixed tissues 
divide by the direct method, but also those of nascent tissues, at least 
among the Arthropods. Direct division is, however, of rare occurrence 
in the embryo. I believe there are only two authentic instances of 
it, — that discovered by Carnoy in the ventral plate of an embryo of 
Hydrophilus piceus (85, p. 224, Plate I. Fig. 11), and that found by 
Wheeler (89, p. 313) in the formation of the blastoderm of Blatta 
germanica, where no instance of mitosis was detected. The embryonal 
membranes of the scorpion I do not include under this head, because 
they are temporary structures forming no vital part of the embryo. 


MUSEUM OF COMPARATIVE ZOOLOGY. 151 


Among the Metazoa, epithelial tissues offer by far the greatest num- 
ber and the most interesting cases of amitosis. Furthermore, as Ziegler 
C91) has very recently shown, epithelial cells of unusual size, with some 
peculiar functional activity (generally secretion) are most apt to exhibit 
this method of division. Cell division has seldom been observed to fol- 
low amitosis in such large cells, which therefore become multinucleate. 
Other epithelial cells which frequently furnish instances of amitosis are 
those which are near the end of their Functional activity. Cells of the outer 
layer of a stratified epithelium sometimes divide amitotically, while 
those of the deeper (and therefore younger) layers of the same epithelium 
divide by mitosis. A good instance of this was recently described by 
Dogiel (’90) in the epithelium of the bladder of Mammals. The nuclei 
of the large epithelial cells lining the intestine of Arthropods very com- 
monly divide by amitosis, as was found by Frenzel (’85) in the midgut of 
Astacus and Maja ; by Carnoy (’85) in the intestinal epithelium of Iso- 
pods; and by Faussek (’87) in the digestive tract of a Cricket (Hremobia 
muricata) and in the larva of Kschna. The intestinal epithelium in all 
Arthropods has an important secretory function. Cells whose function 
is excretory likewise exhibit amitotic division of the nucleus, as in the 
Malpighian vessels of Insects. ‘The occurrence of amitosis in glandular 
and excretory epithelium is readily explainable on Chun’s hypothesis, 
for the functional activities of such cells are peculiarly intense, and it is 
easy to see that a distribution of nuclear material in the cytoplasm is 
of advantage to the cell. The occurrence of nuclei of unusual size (as 
compared with the nuclei of other cells of the same animal) seems to 
me likewise referable to the peculiar needs of the cytoplasm in these 
cells. 

Cases of amitosis peculiarly difficult of explanation are those pre- 
sented by the germinal epithelium of the testis. So many observers 
have reported direct division in sperm mother-cells, that there seems 
no reasonable doubt of its occurrence. It has been suggested that the 
cells which divide amitotically never produce spermatozoa, but merely 
serve to secrete a fluid. This explanation, however, will not serve in 
the case of certain Isopods (Oniscus asellus and /dotea sp.) in the testes 
of which Carnoy (’85, p. 222) found amitosis the prevailing type of di- 
vision, and mitosis of very rare occurrence. Direct division is found 
more or less frequently in the testicular cells of many other Crustacea, 
as the extensive work of Gilson (’84-87), and the investigations of 
Sabatier (’85) show, and occasionally in the other groups of the Arthro- 
pods. Among Vermes, it was found by Lee (’87) in Nemertians, and 


152 BULLETIN OF THE 


by Lowenthal (’89) in a Nematode (Oxyuris ambigua). It need hardly 
be said that amitosis in sexual cells is unexplained by any hypothesis 
yet offered regarding the biological significance of this type of division, 
and further investigations on this point are absolutely necessary before 
we can form any general opinion in regard to it. 

In the maturation and segmentation of the ovum no instance of direct 
division is known, and it is here that karyokinesis is exhibited in its 
most complete form. The well known observations of Boveri (’87) on 
the segmentation of the egg of Ascaris megalocephala are of special in- 
terest on this point. He found a modification of the chromatic threads 
as early as the two-blastomere stage, one of them (cell A) retaining the 
four chromosomes characteristic of the nucleus after fertilization, the 
other (cell B) undergoing a reduction of its chromosomes into the form 
of granules, The two blastomeres arising by division of cell A undergo 
the same differentiation, the nucleus of one (cell A?) retaining the 
chromatic loops, the other (cell A*) undergoing reduction, so that in 
the four-cell stage only one nucleus has retained its chromatic loops. 
The systematic reduction of chromosomes was observed up to the 64- 
cell stage. The important deduction Boveri makes from these facts 
is, that the cells retaining their ancestral nuclear characters are the 
Anlage of the sexual cells of the developing animal, and that the cells 
whose nuclei undergo a modification of the chromosomes are all somatic 
cells. In accordance with this hypothesis, the division of both male 
and female sexual cells ought always to be karyokinetic, and of a 
somewhat different type from the karyokinesis of the somatic cells of 
the same animal, The latter statement, indeed, holds true for the 
testicular cells of the salamander, as was discovered by Flemming (’87). 
It also appears from the work of Carnoy, that in the post-embryonic 
life of Arthropods mitotic division is of rare occurrence in the tissue 
cells, but is of constant occurrence in the reproductive cells of the same 
forms. 

As has already been stated (p. 147), attempts have been made to 
find a morphological connection between karyokinesis and direct divis- 
ion, and thus to solve the puzzling question of the relations they bear 
to each other. Carnoy (’85, p. 398) believes he has found transitions 
between them in the division of the numerous nuclei of Opalina rana- 
rum. Some of these show a distinct spindle, others none ; in both cases 
the nuclear membrane persists, and division is accomplished by constric- 
tion. Pfitzner (’86”), however, found only mitosis in O. ranarum. Car- 
noy has also seen transitional forms of division in the spermatic cells of 


MUSEUM OF COMPARATIVE ZOÖLOGY. 195 


Pagurus striatus, and P. callidus (Planche VII. Figs. 244, 245). A nu- 
clear plate is here formed, both in perfect mitosis and in degenerated 
mitosis; but in the former instance & spindle is formed, and the chromo- 
somes segment individually, while in the latter the plate divides in éto 
by constriction, without the help of a spindle, This modified type of 
mitosis, if we may so regard it, Carnoy considered as the result of 
degradation (pp. 316, 317), inasmuch as it appeared only in old sperm 
mother-cells after spermatozoa had become numerous in the testis. 
This accords with the earlier view that direct division is concomitant 
with senescence of the nuclei, based especially upon nuclear division in 
plants (Schmitz, ’79, Johow, ’81). I have regarded this as a possible 
explanation of the occurrence of amitotic division in the embryonal 
envelopes of the scorpion, for these tissues are temporary structures 
which obviously are near the end of their functional activity. This 
explanation, however, will not fit all cases ; for instance, the occurrence 
of amitosis in embryonic cells, and its prevalence in the testicular cells 
of some Isopods, already mentioned. 

The hypothesis advanced by Chun seems to throw light upon many 
of the cases of amitotic division which are referable to a sort of bud- 
ding or branching of the nucleus, carried to such a point that the 
buds or branches become constricted off as separate nuclear elements. 
These cases are, of course, not to be confounded with a disintegration 
of the nucleus, such as takes place in the macronucleus of Infusoria 
after conjugation, and sometimes in the degeneration of tissues. The 
distribution or extension of nuclear substance in the cytoplasm, whereby 
the surface of the nucleus is increased, is an event of frequent occur- 
rence. It is seen in the many forms of lobed nuclei, such as those of 
the ovarian capsules of Amphibia (see Flemming, ’82), and in those of 
leucocytes ; in hollow or perforated nuclei (giant cells); in branched 
nuclei (spinning glands and Malpighian vessels of Lepidoptera) ; and in 
the band-shaped and moniliform nuclei of many Infusoria. These pecu- 
liar shapes are evidently produced by the activity of the nucleus itself, 
probably correlated with a special function of the cytoplasm. From 
the deeply incised lobation or band-shape of such nuclei it is an easy 
stop to the formation of separate smaller nuclei by the deepening of a 
constriction already formed. Such daughter nuclei will as a rule be 
irregular in shape and unequal in size ; but if their production subserves 
a definite and important function, we should expect that in some cases 
their formation would become a regular process, governed by definite 
laws. It is possible that the more symmetrical kinds of direct division 


154 BULLETIN OF THE 


are to be explained in this way, and such an explanation seems to apply 
well, as suggested on a preceding page, in the case of the scorpion’s 
serosa. Division of the cell does not follow as a rule, and upon this fact 
Chun lays stress. But, so far as we know, there is nothing to exclude 
the subsequent occurrence of cell division, and it is even probable that 
cell division is induced by the presence of more than one nucleus. This 
I take to be the case in the scorpion’s serosa, where I believe the division 
of the cell is due in part to the dicentricity set up in the cytoplasm by 
the division of the nucleus. 

The study of nuclear division among the Protozoa seems likely to 
throw much light upon the relations of amitosis to mitosis, for there can 
be little doubt bat that this group presents the most primitive types of 
nuclear division. So far as known, the very lowest forms of animal cells 
(Amebe®) always divide by the direct method, as the study of Amaba 
polypodia by F. E. Schulze (75), and of Pelomywxa villosa, Amoeba secunda, 
and A. proteus by Gruber (’83 and ’85), has shown. The division of the 
nucleus of Amaba proteus takes place by a sharp equatorial cleft, passing 
through the large, centrally placed nucleolus, and dividing that and the 
peripheral zone of chromatin into two exactly equal halves, which after- 
wards move apart. This is regarded by Gruber (’83, p. 385) as a simple 
type of karyokinesis, because an exact division of the chromatin is accom- 
plished. No kinetic change of the chromatic substance is necessary to 
bring this about, hence none occurs. It seems to me that the absence 
of centrosomes and a spindle effectually separates this type of division 
from true karyokinesis, and until these are discovered, the nuclear di- 
vision of Ameba proteus must be relegated to amitosis. The presence of 
so perfect a type of karyokinesis as that found in Luglypha alveolata, 
worked out so completely by Schewiakoff (’88), is strong evidence against 
the hypothesis that karyokinesis was gradually evolved from direct di- 
vision. For here, among the lowest forms of animal life, we have nuclei 
dividing both by a simple constriction, and by the most highly developed 
kinetic changes. 

Nuclear division among the Infusoria is of special interest, for we 
regularly find in the same individual nuclei very different in structure 
and function, — macro- and micronuclei. The former divide directly, 
the latter by karyokinesis. Apparent exceptions are seen in Spirochona 
gemmipara, where, according to R. Hertwig (’77) the macronucleus 
divides by karyokinesis ; and in Opalina ranarum, studied most carefully 
by Pfitzner (86>). As only one kind of nucleus is found in Opalina, 
it is probable, as Biitschli suggests (’88, p. 1500), that these are of 


MUSEUM OF COMPARATIVE ZOÖLOGY. 155 


the micronuclear type, inasmuch as the division is in all essential re- 
spects like that of micronuclei, and in the resting state the nuclei bear 
no resemblance to macronuclei. The direct division of macronuclei is 
often accompanied by a longitudinal arrangement of the chromatic fila- 
ments, resembling that found in the scorpion’s serosa (see Figs. 6, 7, 8). 
It seems to me that Carnoy is wrong in speaking of these longitudinal 
filaments as a “spindle,” for it has never been shown that they converge 
to the poles of the nucleus, and frequently they can be resolved into 
granules, which is never the case with spindle fibres. Their resemblance 
to the spindle of karyokinesis is deceptive. From their behavior with 
stains, I regard them as consisting of chromatin, and Bütschli (88, 
p. 1526) speaks of this stage of the macronucleus as the “ Knäuelsta- 
dium,” implying that the parallel filaments are chromatic threads. 

Among the Vertebrates, amitosis is unusual, and where it exists kary- 
okinesis is generally found to occur in cells of the same kind. It is 
almost confined to cells which do not form fixed tissues, as leucocytes of 
all kinds, and “ giant cells,” especially those of the red marrow. It also 
occurs in testicular cells of Vertebrates. In leucocytes, according to 
all observers, the nuclear division takes place by constriction, and is 
frequently accompanied by division of the cytoplasm (Ranvier, ’75; 
Flemming, ’82, p. 344; Arnold, ’87). But, as the recent work of 
Flemming (91) and others shows beyond a doubt, leucocytes also di- 
vide by karyokinesis. It is difficult to say whether there is more than 
a single kind of leucocyte, one dividing directly, the other indirectly, 
or whether cells of the same kind divide in two different ways. In 
case of giant cells, it has been shown by Arnold (84), Denys (’86), 
Demarbaix (’89), and others, that division occurs both directly and by 
multiple karyokinesis. Both kinds of division are followed by division 
of the cytoplasm, leading to the formation of a brood of daughter cells 
within the mother cell. 

After going over the literature of amitosis, taking especial note of 
the manner of its occurrence and distribution in the Animal Kingdom, I 
have become convinced that it is not derived from mitosis, and, on the 
other hand, is not the forerunner of the more complicated process. I con- 
sider it another type of division altogether, which, along with karyoki- 
nesis, has been transmitted from the simplest forms of life to the most 
highly organized. While apparently every kind of nucleus may, at 
some stage of its existence, divide by karyokinesis, many afterwards 
exchange this type of division for the simpler process. The special 
conditions which evoke the exchange are very imperfectly understood, 


156 BULLETIN OF THE 


and no hypothesis has yet been offered that will explain all the known 
instances. Some of the hypotheses that have been suggested I have 
already dwelt upon at length; others, as scantiness of chromatin, and 
even its entire absence in the nucleus (Löwit, ’90), seem to me still more 
inadequate. 

One fact in favor of the independence of the two types of division is 
the sudden change from mitosis to amitosis, without any visible interme- 
diate stages. Phylogenetically, this is seen in the abrupt transition from 
the amitotic division of Ameba to the very perfect karyokinesis of the 
nearly related Huglypha. Ontogenetically, of course, the exchange is far 
more abrupt. In the conjugation of Infusoria, all divisions of the micro- 
nucleus are undoubtedly mitotic, while the first (after conjugation) and 
all subsequent divisions of the macronucleus, itself formed from modified 
micronuclei, are by direct division. Again, the amitosis of the blasto- 
dermic nuclei of Blatta (Wheeler, ’89) is an abrupt change from the 
perfect mitosis of segmentation. Other instances are the sudden change 
from mitosis to amitosis in the layers of stratified epithelium, and in 
the generations of spermatic cells. 

Another fact in favor of my view is the almost universal distribution of 
amitosis, and its occurrence in many kinds of cells with widely different 
functions. It seeihs more reasonable to suppose that a process so widely 
extended is inherited, and exists potentially in all cells, rather than to 
look upon it as independently assumed in a multitude of special cases. 
The latter supposition is opposed to all we know of the transmission of 
fundamental characters. 

While it is evident that both mitosis and amitosis appeared at a very 
early period of organic life, it is impossible to say which appeared first. 
But, on a priori grounds, we may conclude that the simpler type pre- 
ceded the more complex. 


CAMBRIDGE, September 28, 1891. 


It was not until this paper had gone to press that I had access 
to the recent communications on amitosis by Flemming (’91*), Löwit 
(91), Verson (’91), Frenzel (’91), and O. vom Rath (91). In his review 
of recent work on cell division, Flemming says (p. 139): “Es ist also 
nicht nur als feststehend anzusehen, dass Amitose vorkommt, sondern 
auch, dass sie in normal lebenden Geweben vorkommt, und dass sie zur 


MUSEUM OF COMPARATIVE ZOÖLOGY. 157 
Zellenvermehrung führen kann.” When, however, both mitosis and 
amitosis occur in the same tissue, he considers it probable that only the 
former is the normal method of regeneration and of growth. 

The brief papers by Löwit, Verson, and Frenzel are replies to Ziegler’s 
C91) recent article on amitosis, and contain little that is new. Verson 
describes briefly the early stages in the spermatogenesis of the silkworm 
(Bombyx mori). He states that the spermatocytes originate from a 
single large nucleus (“ Riesenkern”), which divides repeatedly and 
unequally by amitosis. The small daughter nuclei thus produced divide 
by mitosis, and at length form the spermatocytes. Frenzel adduces 
instances of amitosis in the intestinal epithelium of Crustacea and 
Insects which do not fall within Ziegler’s generalizations. 

Vom Rath’s paper is a valuable contribution to our scanty knowledge 
of the occurrence of amitosis in spermatogenesis. He shows very con- 
clusively that, in the testis of the crayfish, amitosis does not oceur in 
the generations of sperm-forming cells, but only in abortive nuclei 
(“ Randkerne ”), which soon degenerate into an amorphous mass. If 
such a fate could be established for all amitotically dividing nuclei in 
the testes of animals, it would be much easier to form a logical estimate 


of amitosis. 


BULLETIN OF THE 


BIBLIOGRAPHY. 


Arnold, J. 
84. Ueber Kerntheilung und vielkernige Zellen. Virchow’s Arch. f. pathol. 


Anat., Bd. XCVIII. p. 501. 

’87. Ueber Theilungsvorgäuge an den Wanderzellen, ihre progressiven und 
regressiven Metamorphosen. Arch. f. mikr. Anat., Bd. XXX. p. 205. 

88. Weitere Mittheilungen über Kern- und Zelltheilungen in der Milz., 
ete. Arch. f. mikr. Anat., Bd. XXXI. p. 541. 

Blochmann, F. 

'85. Ueber directe Kerntheilung in der Embryonal-hülle der Scorpione. 

Morph. Jahrb., Bd. X. p. 480. 
Boveri, T. 

'87. Ueber Differenzirung der Zellkerne während der Furchung des Eies von 

Ascaris megalocephala. Anat. Anz., Jahrg. II. p. 688. 
Bütschli, O. 

76. Studien über die ersten Entwickelungs-vorgänge der Eizelle, die Zell- 
theilung, und die Conjugation der Infusoric&. Abhandl, Senckenberg. 
Naturf, Gesellsch., Bd. X. p- 218. 

‚88. Protozoa. Abth. III. Infusoria und System der Radiolaria. Bronn’s 
Classen u. Ord. des Thier-Reichs, Bd. I. Abth. 3, p. 1098. 


Carnoy, J. B. 
‚85. La Cytodiérése chez les Arthropodes. La Cellule, Tom. I. p. 191. 


Chun, C. 
90. Ueber die Bedeutung der direkten Kernteilung. Sitzungsber. Physik- 


ökonom. Gesellsch. Königsberg i. Pr., Jahrg. 31, 6 pp. 


Demarbaix, H. 
‚89. Division et dégénérescence des cellules géantes de la moélle des os. 


La Cellule, Tom. V. p. 27. 


Denys, J. 
'86. La Cytodiérése des cellules géantes et des petites cellules incolores de 


la moélle des os. La Cellule, Tom. II. p. 245. 


Dogiel, A. S. 
‚90. Zur Frage über das Epithel der Harnblase. Arch. f. mikr. Anat., 


Bd. XXXV. p. 389. 


Faussek, V. 
‚87. Beiträge zur Histologie des Darmkanals der Insekten. Zeitschr. f. 


wiss. Zool, Bd. XLV. p. 694. 


MUSEUM OF COMPARATIVE ZOOLOGY. 159 


Flemming, W. 
82. Zellsubstanz, Kern- und Zelltheilung. Leipzig, F. C. W. Vogel. 
'87. Neue Beiträge zur Kenntniss der Zelle. Arch. f. mikr. Anat., Ba. 
XXIX. p. 389. 
'89. Amitotische Kerntheilung im Blasenepithel des Salamanders. Arch. f. 
mikr. Anat., Bd. XXXIV. p. 437. 
'91. Ueber Theilung und Kernformen bei Leucocyten, und über deren At- 
tractionssphären. Arch f. mikr. Anat., Bd. XXXVI. p. 249. 
‘914, Usher Zellteilung. Verbandl. Anat. Gesellsch., 5. Versamm., p. 125. 
Frenzel, J. 
'85. Ueber den Darmkanal der Crustaceen, nebst Bemerkungen zur Epithel- 
regeneration. Arch. f. mikr. Anat., Bd. XXV. p. 137. 
’91. Zur Bedeutung der amitotischen (direkten) Kernteilung. Biol. Cen- 
tralb , Bd. XI. p. 558. 
Gehuchten, A. van. 
’89. L’Axe organique du noyau. La Cellule, Tom. V. p. 177. 
Gilson, G. 
'84-'87. Étude comparée de la spermatogenese chez les Arthropodes. La 
Cellule, Tom. I. p 11; Tom. 11. p. 83; Tom IV. p. 1. 
Göppert, E. 
’91. Kerntheilung durch indirekte Fragmentirung in der lymphatischen 
Randschicht der Salamandrinenleber. Arch. f. mikr. Anat., Bd. XXXVll. 
p 375. 
Gruber, A. 
'83. Ueber Kerntheilungsvorgänge bei einigen Protogoen. Zeitsehr. f. wiss. 
Zool., Bd. XXXVIII. p. 372. 
’85. Studien über Amöben. Zeitschr. f. wiss. Zool., Bd. XLI. p. 186. 
Hamann, O 
’90. Monographie der Acanthocephalen. Teil I. Jena. Zeitschr., Bd. XXV. 
p 113. 
Hertwig, R. 
’77. Ueber den Bau und die Entwickelung der Spirochona gemmipara. 
Jena. Zeitschr., Bd. XI. p. 149 
'84. Ueber die Kerntheilung bei Actinosphaerium eichhorni. Jena. Zeitschr., 
Bd XVII. p. 490. 
Hoyer, H. 
'90. Ueber ein für das Studium der direkten Kemtheilung vorzüglich 
geeignetes Objekt. Anat. Anz, Jahrg. V. p. 26. 
Johow, F 
'81. Die Zellkerne von Chara foetida. Bot. Zeit., Jahre. XX XIX. Nos. 
45, 46. 
Kowalevsky, A., und M. Schulgin. 
'86. Zur Entwicklungsgeschichte des Skorpions (Androctonus omatus). 
Biol. Centralb , Bd. VI. p. 525. 


Kükenthal, W. 
'85. Ueber die lymphoid Zellen der Anneliden. Jena. Zeitschr., Bd. XVIII. 
p- 319. 
Laurie, M. 
’90. Embryology of a Scorpion (Euscorpius Italicus). Quart. Jour. Micr. 
Sci, Vol. XXXI. p. 105. 
Lee, A. B. 
'87. La Spermatogénèse chez les Némertiens. Recueil zool. Suisse, Tom. 
TV. p. 409. 
Löwenthal, N. 
’89. Ueber die Spermatogénèse von Oxyuris ambigua. Internat. Monatsschr. 
f. Anat. u. Physiol., Bd. VI. p. 364. 
Löwit, M. 
’90. Ueber Amitose (directe Theilung). Centralb. f. allgem. Pathol. u. 
pathol. Anat., Bd. I, p. 281. 
’91. Ueber amitotische Kernteilung. Biol. Centralb., Bd. XI. p. 513. 
Metschnikoff, E. 
’71. Imbryologie des Skorpions. Zeitschr. f. wiss. Zool., Bd. XXI. p. 204. 
Overlach, M. 
’85. Die pseudomenstruierende Mucosa uteri nach acuter Phosphorvergift- 
ung. Arch. f. mikr Anat., Bd. XXV; p. 191. 
Pfitzner, W. 
86°. Zur morphologischen Bedeutung des Zellkerns. Morph. Jahrb., 
Bd. XI. p. 54. 
’86°. Zur Kenntniss des Kerntheilung bei den Protozoen. Morph. Jahrb., 
Bd. XI. p. 454. 
Platner, G. 
’89. Beiträge zur Kenntniss der Zelle und ihre Theilungserscheinungen, 
T-III. Arch. f. mikr. Anat., Bd. XXXIII. p. 125. 
Rabl, C. 
’85. Ueber Zelltheilung. Morph. Jahrb., Bd. X. p. 214. 
’89. Ueber Zelltheilung. Anat. Anz., Jahrg. IV. p. 21. 


| 
160 BULLETIN OF THE 
f 


Ranvier, L. 
755 Recherches sur les éléments du sang. Travaux Lab. d’ Histol., p. 1. 


Rath, O. vom 
’91. Ueber die Bedeutung der amitotischen Kerntheilung im Hoden. Zool, 
Anz., XIV. Jahrg. Nos. 373, 374, 375. 
Sabatier, A. 
’85. Sur la spermatogénése des Crustacés décapodes. Compt. Rend., Tom. C. 
p. 391. 
Schewiakoff, W. 
'88. Ueber die karyokinetische Kerntheilung der Euglypha alveolata. 
Morph Jahrb., Bd. XIII p. 193. 


| 
| 
| 
| 


MUSEUM OF COMPARATIVE ZOOLOGY. 161 


Schmitz, F. 
'79. Beobachtungen über die vielkernigen Zellen der Siphonocladiaceen. 
Halle. 
Schulze, F. E. 
75. Rhlizopoden-studien, V. Arch. f. mikr. Anat., Bd. XI. p. 583. 
Verson, E. 
’91. Zur Beurteilung der amitotischen Kernteilung. Biol. Centralb., Bd. 
XI. p. 556. 
Waldeyer, W. 
’88. Ueber Karyokinese und ihre Beziehungen zu den Befruchtungsvor- 
gängen. Arch. f. mikr. Anat., Bd. XXXII. p. 1. [Trans], Quart. Jour. 
Mier. Sci, Vol. XXX. p. 159 ] 
Watase, S. 
'91. Studies on Cephalopods. I. Cleavage of the Ovum. Jour. Morph., 
Vol. IV. p. 247. 
Wheeler, W. M. 
'89. The Embryology of Blatta germanica and Doryphora decemlineata. 
Jour. Morph., Vol. III. p. 291. 
Woodworth, W. M. 
91. On the Structure of Phagocata gracilis, Leidy. Bull. Mus. Comp. 
Zoöl., Vol. XXI. p. 1. 
Ziegler, H. E. 
’87. Die Entstehung des Blutes bei Knochenfischembryonen. Arch. f. mikr. 
Anat, Bd. XXX. p. 596. 
’91. Die biologische Bedeutung der amitotischen (direkten) Kernteilung im 
Tierreich. Biol. Centralb., Bd. XI. p. 372. 


EXPLANATION OF FIGURES. 


All figures are from drawings made with the aid of an Abbé camera. 


Jounson, — Nuclear Division. 


PLATE I. 


Vig. 1. Five cells of the serosa, two of them covered by the amnion, which is 
omitted from the rest of the figure for the sake of clearness. am., 
amnion; sr., serosa. X 150. 

Fig. 2. Section through the embryonal membranes and ovarian capsule. ‘The 
fibrous appearance of the ovarian capsule is due to the presence of 
muscle fibres and connective tissue. The boundary line between 
amnion and serosa is visible only in the vicinity of the amniotic 
nuclei. eth. fol., epithelium of ovarian capsule (when the plates were 
engraved I still took this to be the follicular epithelium, hence the 
error in the abbreviation); nl. fol., nucleus of capsular epithelium ; 
nl. sr., nucleus of serosa; nl. am., nucleus of amnion. X 630. 

Figs. 3-15 are all from the serosa. 

Fig. 3. Very small, binucleate cell. X 180. 

Figs. 4-10. Nuclei at different stages of division. vac., vacuole; x, new nuclear 
wall within the old one. X 550. 

Fig. 11. Two cells produced by division of a binucleate cell. X 130. 

Fig. 12. Cell from the serosa of a young embryo, with dividing nucleus; the axis 

‚ of elongation corresponds with the short axis of the cell. X 130. 

Fig. 13. Cell from serosa of a young embryo, with nucleus unequally divided and 

daughter nuclei eccentric in position. X 130. 


‘LEAR DIVISION 


nil fol. 


eth. fol 


nlam 


‚JOHNSON. — Nuclear Division. 


PLATE I. 


Fig. 14. Piece of the serosa from an advanced embryo, with four adjacent tri- 
nucleate cells (1, 2, 3,4); nuclei of cell a and the large cell farthest 
to left have undergone degeneration. X 90. 

Fig. 15. Three cells of the serosa from an old embryo to show recession of daugh- 
ter nuclei towards the ends of the cells. X 90. 

Figs. 16-20 are from the amnion. 

Figs. 16-19. Stages in the division of amniotic nuclei. In Figure 18 three stages 
are shown, a, b,c. X 800. 

Fig. 20. Two amniotic cells, apparently formed by recent division. X 875. 

Figs. 21-26 are from the capsular epithelium. 

Figs. 21-23, Cells showing successive stages of nuclear division. 800. 

Figs. 24-26, Cells to show the degeneration of nuclei. In Figure 24 the nuclei 

a.e but slightly differentiated; in Figure 25 the pale nucleus has 

become much larger and very faint; in Figure 26 it has disappeared 

altogether. X 800. 


16. 


w 7 f 
A3. 
& 
è &> 
® i > 
i 
r è. ET 
& e Q 
® 
PR) 
\ 
© 
5 
© 
¢€ 
J. 
| 
\ 
20. 
19. 
| 
05 
22 
` 
a } 
y i r | 
\ \ e. 
I 
25 
j 
i 26 
2 u. 
J = 
d F. 


JOHNSON. — Nuclear Division 


PLATE I. 


Figs. 27-34 are all from the serosa. 

Fig. 27. A cell undergoing division by formation of a cell plate. The daughter 
nuclei are still united by a connecting thread. The dotted line on the 
left indicates the edge of the fragment of membrane in which this 
cell occurs. From the serosa of an advanced embryo. X 304. 

Fig. 28. A cell divided by constriction, without the formation of a cell plate. The 
nuclei have undergone degeneration. From the serosa of an advanced 
embryo. X 150. 

Fig. 29. A cell, the nucleus of which has undergone tripartite division. From an 
old serosa. X 150. 

Fig. 30. Nucleus of the same, more highly magnified. The chromatin is grouped 
in granular masses. Two of the daughter nuclei are still united by 
strands of the nuclear membrane. X 680. 

Figs. 81-32. Constricted nuclei from a young serosa. One of the daughter nuclei 
of each is larger than its mate, and has itself become elongated and 
constricted. X 804. 

Fig. 83. Quadrinucleate cell. The upper of the two original nuclei has divided 
in a longitudinal, the lower in a transverse plane. Nucleus a still 
shows a remnant of the connecting thread, and nucleus b retains the 
conical form it had in division. Both nuclei have rotated 90° from 
the plane of elongation. X 804. 

Fig. 34. Cell from the serosa of a far advanced embryo. The nuclei have under- 

gone extreme degeneration. Each nucleus is surrounded by a bright 

ring, outside of which is a broad zone of a radiate structure, more 

stainable than the rest of the cytoplasm. X 150. 


A 


u 


No. 4. — A Fourth Supplement to the Fifth Volume of the Terres- 
trial Air-Breathing Mollusks of the United States and Adjacent 
Territories. By W. Q. Binney.l 


Tun following pages are believed to contain all that has been added 
to our knowledge of the subject prior to date. 

Students are requested to note that in the Third Supplement, p. 214, 
the figures of Arionta Diabloensis and Bridgesi ave reversed, On p. 225, 
Explanation of Plate VII., the references E and F are reversed: on 
p. 226, Explanation of Plate XI., Figures D and G are reversed. 


BURLINGTON, New JERSEY, July 1, 1891. 


Glandina decussata, Desu. 
Plate I. Fig 4. 


Under the name of decussata, specimens are found in most collections which 
I have figured one of them, and its 
| figured in my Third Supplement. 
Should it prove 


can hardly be referred to that species. 
dentition has already been described anc 
The shell is readily recognized by its more cylindrical form. 
distinct from decussata, I would suggest for it the specific name of Singleyana. 
I received it from Bexar County, Texas, collected by Mr. Wetherby. 


Selenites Vancouverensis, Lea, var Keepi, HEMPHILL. 
Plate II. Fig. 5. 


Shell umbilicated, greatly depressed, thin, smooth, shining, transparent, scarcely 
marked by the delicate wrinkles; very light horn-color; whorls over four, some- 
what flattened above and beneath, and scarcely descending at the aperture ; spire 


1 The Terrestrial Air-Breathing Mollusks of the United States and the Adjacent 
Territories of North America, described and illustrated by Amos Binney. Edited 
by A, A. Gould. Boston, Little and Brown, Vols. I., II., 1851; Vol. III., 1857. 
Vol. IV., by W. G. Binney, New York, B. Westermann, 1859 (from Boston Journ. 
Nat. Hist.). Vol. V., forming Bull. Mus. Comp. Zoöl., Vol. IV., 1878. Supplement 
to same, in same, Vol. IX. No. 8, 1888. Second Supplement, in same, Vol. XIII. 
No. 2, 1886. Third Supplement, in same, Vol. XIX. No. 4, May, 1890. 


VOL. XXIL — NO. 4. 


164 BULLETIN OF THE 


flat, not rising above the body whorl; suture well impressed; umbilicus moderately 
large, exhibiting most of the volutions; aperture transversely subcircular, wider 
than high; lip simple, thickened, sinuous above, very slightly reflected at the base, 
ends scarcely approached Width % inch, height 5 inch. 

Hills near Oakland, California. One specimen only. 

This rare and interesting little shell I collected some years ago. It is a perfect 
miniature form, in every respect, of S. Vancouwverensis, I regard it as an extremely 
small variety of that so called species. It is about the size of the variety of 
S. Durant, lately described as S. cwlatus, Mazyck, but differs very materially in form, 
sculptnre, and the general texture of the shell. It differs from var. Catalinensis in 
being more robust, larger, and has a smaller umbilicus. I dedicate this pretty little 
shell to Prof. Josiah Keep, of Mills College, California, who has done so much 
through his interesting little book to stimulate the study of West Coast shells. 


The above is Mr. Hemphill’s description, from “The Nautilus,” Vol. IV. 
p. 42, 1890. My figure is drawn from an authentic specimen, 


Selenites Vancouverensis, var. hybridus, Humeutte. 


Shell broadly umbilicated, depressed, slightly convex above, surface shining, 
polished, of a dark yellowish green color, lines of growth coarse, rib-like and regu- 
lar on the spire, finer and more irregular on the body whorl, crossed by fine revolv- 
ing lines that become fainter on the last whorl, suture well impressed; aperture 
rounded, broader than high, greatly indented above; lip simple, very little reflected 
below at its junction with the columella, very sinuous above, its terminations joined 
by a very thin callus. Height § inch, breadth 1 inch. 

Astoria, Oregon. 

In the strong rib-like sculpturing of the spire, depressed form, and sinuous lip, it 
resembles sportellus. In its greater diameter, dark greenish color, and the absence 
of the decussating sculpture on the last whorl, it approaches Vancouverensis. 

All our American Selenites commence life with a finely granulated shell. When 
they have attained about two whorls, the stria begin to appear, and increase in 
strength as the shell increases in size. 

It is well known that all shell-bearing mollusks construct their shells in obedi- 
ence to the laws of their constitutional characteristics and the environment, among 
which I include affinity of matter and mechanical skill, the latter a faculty pos- 
sessed to a greater or less degree by all animals. Some individuals in a colony of 
shells display greater mechanical skill than others, or possess stronger imitative 
powers, and closely follow the lines and styles of their forefathers, strictly attend- 
ing to the details of sculpturing, not omitting a rib or line, Other individuals of 
the same colony, not having this imitative faculty so strongly developed, may 
change or vary the form of the shell by constructing it with more convex whorls, 
generally resulting in a narrower or more elevated shell; or they may flatten the 
whorls, resulting in a broader and depressed form. Some modification of the um- 
bilieus generally follows the change in the form of the shell. In both cases the 
sculpturing may be what we call characteristic of the species, or may be more or 
less modified by the omission of one, two, or more ribs, or the ribs may be more 


MUSEUM OF COMPARATIVE ZOOLOGY. 165 


irregular in shape. A few lines may also be dropped, perhaps some added, or the 
entire surface may be modified in obedience to the laws of the mechanical skill 
possessed by the individual, and the affinity of matter secreted by the animal, for 
the purpose of constructing the shell. An examination of a large number of 
Selenites concanus, and of our West Coast forms, convinces one that the entire group 
of American Selenites is the offspring of a single common type. 


The above is Mr. Hemphill’s description, from “ The Nautilus,” Vol. IV. 
p- 42, 1890. 


Selenites Duranti, var. Catalinensis, Hrmrmtr. 
Plate II. Fig. 3. 


I figure an authentic specimen. See Third Suppl., p. 221. 


Selenites Vancouverensis, var. transfuga, HEMPHILL. 


Shell very much depressed, planulate, broadly umbilicated, of a dirty white 
color; whorls 34 or 4, flattened above, more rounded beneath, with regular strong 
rib-like striæ; suture well impressed, becoming deeper and channel-like as it ap- 
proaches the aperture; aperture hardly oblique, slightly flattened above, with a 
tendency to a corresponding depression below ; lip simple, roundly thickened inter- 
nally, its terminations approaching, forming in some specimens a short columellar 
lip, joined by a heavy raised callus in very adult specimens. Height 7; inch, 
greatest diameter 7%, lesser 77; inch. 

San Diego, California, to Todos Santos Bay, Lower California. 

This is the small flat shell that has been distributed as a variety of sportella, and 
also as a variety of Voyanus. I find, however, on comparing it with the typical 
Voyanus collected by me last fall, that it is quite a different shell. The ribs are 
closer and finer than either sportellus or Voyanus, the umbilicus is much larger, 
and it is a very much more depressed shell. I consider it, however, a deserter from 
the Northern forms, and name it accordingly. It is a much Irrger and a more 
globose form than simplilabris of Ansey. 


The above is Mr. Hemphill’s description. 


Selenites Vancouverensis, Lea. 


The only differences that I can detect between this shell and Selenites concava, 
Say, are these. The umbilicus in the California shells is a little more contracted, the 
color is a shade darker, the striw are a little closer, stronger, and more regular, and 
the body whorl is a little more flattened at the aperture. Height § inch, breadth 
4 inch. 

Sonoma Co. to Santa Cruz Co., California. 


The above is Mr. Hemphill’s description of what he calls S. concavus, var. 


occidentalis. 


BULLETIN 


OF THE 


Selenites Vancouverenis, var. tenuis, HEMPHILL. 


Shell broadly umbilicated, depressed, nearly planulate ; of a dirty greenish brown 
color; whorls 5, flattened above, more rounded beneath, the last expanding later- 
ally as it approaches the aperture, and crowded with fine oblique stri® ; suture well 
impressed ; aperture rounded, slightly flattened above; peristome simple, hardly 
reflected below. Height 4 inch, breadth „7 inch. 

Napa Co., California. 

The small size, nearly planulate form, and thin, lean body whorl as it emerges 
from the aperture, will serve to distinguish this shell from the other forms of 
concavus found on the West Coast. 


The above is Mr. Hemphill’s description. He refers ali these varieties to 
concavus, but I use the specific name Vancouverensis for all Pacific Region forms, 


Limax Hemphilli. 
Plate III. Fig. 1. 


Length (contracted) 19 mm. Mantle long, 9 mm. End of mantle to end of 
body 9 mm. Foot wide 2 mm. Median tract of foot gray, lateral tracts brown. 
Median area of foot rather wider than either lateral area. Mantle free an- 
teriorly as far as respiratory orifice. Body tapering posteriorly, not carinate. 
Mantle somewhat granulose, not concentrically striate. Color dark brown, 
obscurely marbled with gray ; sides anteriorly grayish and paler. 


Limax Hemphilli, W. G Bınnev, 3d Suppl. T. M. V., p. 205, Plate VIII. Fig. E; 
Plate I. Fig. 13; Plate II. Fig. 3 (1890). 


A species of the Pacific Province, having been found from British Columbia 
to San Tomas River, Lower California, by Mr. Henry Hemphill, in whose 
honor it is named. 

The general outward appearance of this species resembles that of campestris, 
but every specimen examined by me from numerous localities had a peculiarity 
in its lingual dentition which seems to me of specific value, — the presence of 
an inner cutting point to the lateral teeth, very much the same as is found in 
agrestis. The anatomy of this species is specifically distinct from agrestis in 
wanting the trifurcate penis sac of the latter, even did its distribution not 
preclude its being a form of agrestis. Ihave ventured therefore on giving it a 
specific name. 

The penis sac is large, long, gradually tapering to the apex; the genital 
bladder ıs globular, on a short, stout duct. 

I figure on the plate a variety from San Tomas River, Lower California, 
called pictus by Mr. Cockerell. Its body is pale, reticulated with gray spots; 
mantle with black or gray spots. Resembling L. Berendti, Strebel, from 
Guatemala. 

For lingual dentition, ete., see Third Supplement. 


MUSEUM OF COMPARATIVE ZOOLOGY. 167 


Zonites Shepardi, Hrmraıte. 


Shell umbilicated, very small, depressed ; whorls 3 or 34, shining, transparent, 
smooth, somewhat flattened; spire scarcely elevated above the body whorl; aper- 
ture oblique, oval; peristome simple, acute, its ends hardly approaching ; suture 
well impressed; umbilicus pervious, and moderately large for so small a shell. 
Great diameter, 2 mm. Height, 1 mm. 

Santa Catalina Island, California. 

This little shell belongs to the planulate forms, and somewhat resembles a minute 
Z. Whitneyi. 

I dedicate it to Miss Ida Shepard in recognition of her active services among the 
mollusks of Long Beach, Cal., where she resides. 


The above is Mr. Hemphill’s description. 


Zonites Lawæ. 


Shell small, umbilicated, globose, flatter below, shining, light horn-colored, 
marked with coarse wrinkles of growth; spire rounded; whorls 8, gradually 
increasing, slightly convex, the last excavated below around the umbilicus ; 
aperture oblique, rounded; peristome simple, acute, thickened with callus 
within. Greater diameter 9,mm., lesser 7 mm.; height 4 mm. 


Zonites placentula, part, W. G. Binney, formerly, Terr. Moll. U.S. V., p. 124, Fig. 
44; Plate III. Fig. L (dentition). 

Zonites Lawi, W. G. Binney, Suppl. to Vol. V. p. 142; Plate II. Fig. E (also, 
Ann. N. Y. Ac. Sci., Vol. I., Plate XV. Fig. E, as undetermined). 


Mountains of Tennessee (Miss Law); a species of the Cumberland Subregion, 

Readily distinguished from placentula by its larger size, higher rounded spire, 
greater number of whorls, and more widely excavated umbilical region. 

Jaw as nsual in the genus. 

Lingual membrane (Vol. V. Plate III. Fig. L, as placentula) with 25-1-25 
teeth; three laterals and one transition tooth. 


Zonites Caroliniensis, COCKERELL. 
Plate III. Fig. 7%. 


Among the specimens of Zonites sculptilis collected in the mountains of 
North Carolina are many which differ from the type widely enough to be 
considered a distinct species. Mr. Cockerell suggests for it the name Caro- 
linensis, thus describing it : — 


This species differs from sculptilis in its fewer whorls, straighter columellar 
margin, less iunate aperture, fewer radiating strie, and other points. It is figured 
as sculptilis in Manual of American Land Shells, Fig. 231. 


BULLETIN OF THE 


Zonites sculptilis. 
Plate III. Fig. 9. 


For the sake of comparison with the preceeding species, I have given other 
figures here of the true Z. sculptilis. 


Zonites Simpsoni, PILSBRY. 
Plate I. Fig. 8, 


I give an enlarged figure of an authentic individual of this species. For 
the description see Third Suppl., p- 218. 


Zonites Diegoensis, HEMPHILL. 


Plate III. Fig. 2. 


Shell minute, umbilicated, thin, light horn-colored, with delicate incremental 
striæ, globose; whorls 34, convex; base swollen; suture deep; umbilicus broad ; 
aperture narrow, rounded ; peristome thin, acute, its ends approximated, the inner 
one slightly reflected. Greater diameter 34 mm., lesser 14; height 14 mm. 

Near Julian City, San Diego Co., California. On Cuyamaca Mountain, 4,500 
feet elevation. 


The above is Hemphill’s description. My figure is drawn from an authentic 
specimen. 


Zonites cuspidatus, Lewis. 
Vol. V., Fig. in text; Suppl., Plate II. Fig. C. 


Shell imperforate, small, slightly convex above, flattened below ; light horn 
color, shining; whorls 6, gradually increasing in size, with wrinkles of growth, 
the last not descending at the aperture; peristome thin, acute; aperture 
rounded, bearing within behind the peristome a white callus, on which is 
one subcentral and a second basal, erect, recurved tooth-like process, sepa- 
rated by a rounded sinus ; base often blackish, showing ‘the white callus 
prominently. Greater diameter 8 mm., lesser 6; height 4 mm. 

Zonites cerinoideus, var. cuspidatus, Lewis, Proc. Phila. Ac. Nat, Sci., 1875, p. 334. 


Zonites cuspidatus, W. G. Binney, Ann. N. Y. Ac. Nat. Sci., Vol. I. p. 359, Plate 
XV. Fig. C; Suppl. to Terr. Moll. V., Plate II. Fig. C. 


Mountains of Tennessee and North Carolina: a species of the Cumberland 


Subregion. 
The tooth-like processes within the aperture, strongly curved towards each 
other, form an arched space. 


MUSEUM OF COMPARATIVE ZOÖLOGY. 169 


Miss Law thus wrote from Philadelphia, Tenn., of this species: “ Unlike 
gularis, it seems to be a rare shell, and I find it only by scraping off the sur- 
face of the ground in the vieinity of damp mossy rocks. Its habits are more 
like placentula than gularis. I never mistake one for a gularis, even before 
picking it up; the thickened yellow splotch near the lip, and the thinner spot 
behind, showing the dark animal through it, as well as its more globular form, 
particularly on the base, make it look very different when alive.” 


Zonites macilentus, Snurtı. 
Plate III. Fig. 3. 


The individuals of this group are very often difficult to identify, on account 
of the blending of their specific characters. The typical macilentus is distin- 
guished by a very wide umbilicus and a single revolving lamina starting from 
near the basal termination of the peristome. The figure of macilentus in Vol- 
ume V. shows a second revolving lamina and a much smaller umbilicus. I 
give here another figure of what appears to me to be the shell described as 
macilentus. How constant are the characters of the species can be shown only 
by a large suite of individuals. 


Tebennophorus Hemphilli. 


Plate III. Fig. 4. 


I give a figure of the jaw already described by me. 


Patula strigosa, Gourp, var. jugalis, Hemrmite. 


Shell umbilicated, depressed with numerous prominent oblique striae; spire very 
moderately elevated or depressed ; whorls 53, somewhat flattened above, but more 
convex beneath, the last falling in front, with two dark revolving bands, one at the 
periphery and the other above; the body whorl subcarinated at its beginning, but 
more rounded as it approaches the aperture; suture well impressed; color ashy 
white, with occasional horn-colored stains; umbilicus large, pervious, showing the 
volutions; aperture oblique, ovate, but in very depressed specimens the aperture 
is at right angles with the axis of the shell; peristome simple, thickened, its ter- 
minations approaching and joined by a thick heavy callus, making the peristome 
in very adult specimens continuous. Height of the largest specimens + inch, 
breadth 1 inch. Height of the smallest specimens ;% inch, breadth 44 inch. 


Patula strigosa, var. jugalis, HemPHILL, The Nautilus, 1890, p. 134, in Binney’s 
3d Suppl, p. 215, figure. 


Banks of Salmon River, Idaho. 

This is another interesting form of the very variable strigosa. It inhabits stone 
piles, and other places where it can find shelter and protection against the fatal 
rays of the summer’s sun, close along the banks of the river. It is interesting on 


170 BULLETIN OF THE 


account of its very depressed form and the ovate form of the aperture, the heavy 
callus joining or “ yoking ” together the extremities of the peristome. 


The above is Hemphill’s description. 
The figure in the Third Supplement is drawn from an authentic specimen. 


Patula strigosa, Gourp, var. intersum, HEMPHILL. 


Shell umbilicated, sublenticular, depressed, thin, dark horn-color, more or less 
stained with darker chestnut. Whorls 54 or 6, somewhat flattened above, more 
convex beneath, obtusely carinated at the periphery, and bearing numerous coarse 
oblique rib-like strie, and two dark revolving bands; suture well impressed; um- 
bilicus large, pervious; aperture oblique, subangulated ; peristome simple, thick- 
ened, its terminations joined by a thick callus. Height of the largest specimen 4 
inch, breadth ¢ inch. Height of the smallest specimen $ inch, breadth 77 inch. 


Patula strigosa, var. intersum, HEMPHILL, The Nautilus, 1890, p. 135. 


Bluffs along the banks of Little Salmon River, Idaho. 

This shell inhabits stone piles at the foot of a steep bluff back some distance 
from the river. It seems to be quite rare, as I found but few specimens during 
the two or three days of my stay in its vicinity, and many of them were dead. I 
regard it as one of the most interesting shells found by me during the season, for 
it combines the depressed angulated or keeled forms of the Haydeni side of the 
series with the sculpturing of /dahoensis, two shells representing opposii 2 charac- 
ters in every respect. It thus becomes the companion of Wahsatchensis, a beautiful 
shell, combining the same characters, but much more developed, and connected 
with the large elevated forms. Var. intersum fills the opposite office, by uniting 
these characters with the small depressed forms. Taken as a whole, this series of 
shells, as now completed, seems to me to offer the best guide or key to the study 
of species that the student can have. Every known external character belonging 
to the genus Helix is so gradually modified and blended with opposite characters, 
that, if one had the moulding or making of the many and various intermediate 
forms, he could scarcely make the series more complete than Nature has done 
herself. 


The above is Hemphill’s description. 


Patula strigosa, Gourp, var. globulosa, CockrreEtt. 


Small, globose, dark above periphery, with two bands, transverse grooved striæ 
rather well marked. Diameter 114, alt. 84 mm. Black Lake Creek, Summit Co. 
The specimen seems immature, but is remarkable as being the only form I have 
seen in Colorado that is nearer to strigosa than Cooperi. It is doubtless allied to 
var. Gouldi, Hemphill. (Cockerell.) 

Patula strigosa, var. globulosa, CockERELL, The Nautilus, 1890, p. 102. 


The above is Cockerell’s description. 


The above varieties of Patula strigosa are transversely ribbed. The following 
are smooth or striate 


MUSEUM OF COMPARATIVE ZOÖLOGY. 171 


Patula strigosa, Gour, var. Buttoni, Hemruitı. 
Plate I. Figs. 2 and 10. 


I figure the typical and the toothed forms. See 3d Suppl., p. 220. 


Patula strigosa, Gourp, var. albofasciata, HemPHILL. 
Plate IV. Fig. 9. 


Shell globose, elevated or depressed; whorls six, convex, with a broad white 
band at the periphery, which shows just above the suture on two or three whorls 
of the spire as it passes towards the summit or apex, separating two variable 
chestnut-colored zones; the upper one in some specimens is often very dark, in 
others very light passing into horn-color, and broken into blotches, stains, or 
irregular lines, which pass up a few whorls of the spire and blend with the 
horn-colored summit; the lower zone spreads towards the umbilicus in irregular 
stains, often beautifully clouding the base of the shell, or is often broken into 
irregular revolving lines, and other varied patterns of coloring; striæ rib-like, 
quite coarse in some specimens, in others finer and closely set together; aperture 
circular, ovate, and occasionally pupæform; peristome simple, thickened, sub- 
refiected at its junction with the columella, and partially covering the umbilicus, 
the ends approached and often joined by a callus, the peristome sometimes bearing 
a tooth-like process ;'umbilicus deep, moderately large, narrower in elevated and 
broader in depressed specimens; suture well defined. Greater diameter of the 
largest specimen 17 mm., height, 12 mm.; greater diameter of the smallest 12 
mm., height 7 mm.; with all the intermediate sizes. 

Box Elder Co., Utah. 

Among leaves, brush, and grass, on limestone rock. Altitude, about 4,600 feet 
above the sea. 

This variety of strigosa is so very variable in all its characters I find it quite 
difficult to draw a description that will cover all the individuals which I include in 
it. Ihave given the measurements of the largest and smallest specimens, but there 
are all the intermediates between those figures. 


The above is Mr. Hemphill’s description. An authentic individual is figured 
on the plate. 


Patula strigosa, Gout, var. subcarinata, HEMPHILL. 


Among the shells recently collected by Mr. Hemphill at Old Mission, Cœur 
d’Alene, Idaho, was a marked variety of this species, for which Mr, Hemphill 
suggests the name subcarinata. The specimens vary greatly in elevation of 
the spire, and in the number and disposition of the revolving bands, often 


quite wanting, as in the specimen figured in the Third Supplement. All have 
a very heavy shell, the body whorl of which has an obsolete carina which 
is well marked at the aperture, modifying the peristome very decidedly. See 


the figure. 


172 BULLETIN OF THE 


In examining the genitalia I find the base of the duct of the genital bladder 
greatly swollen along a fifth of the total length of the duct, 
Mr. Hemphill (The Nautilus, 1890, p. 133) thus describes it :— 


The shell in general form resembles a large, coarse elevated or depressed Cooperi’ 
It has six whorls, well rounded above and beneath, and subcarinated at the periph- 
ery. The body whorl has two revolving dark bands, one above and the other 
below the periphery; sometimes the upper band spreads over the shell to the su- 
ture, forming a dark chestnut zone that fades out as it passes toward the apex. 
The peristome is simple, thickened, its terminations joined by a callus; aperture 
obliquely subangulate; the suture is well impressed. Height of the largest speci- 
men 1 inch, breadth 14 inches; height of the smallest specimen 4 inch, breadth 
1 inch. 

Rathdrum, Idaho. 


An authentic specimen is figured in the Third Supplement. 


Patula strigosa, Gour», var. bicolor, HEMPHILL. 
Plate IV. Fig. 7. 


This shell is a colored variety of the last. It may be characterized as being of 
a general dark horn-color mingled with dirty white; there are occasional zones of 
dark horn-color above and fine dark lines beneath, but no defined bands. In some 
of the specimens the light color prevails, in others the horn-color spreads over 
the shell in irregular patches. Height } inch, breadth 1} inches. 

Rathdrum, Idaho. (Hemphill.) 


Patula strigosa, var. bicolor, Hrmruıtr, The Nautilus, 1890, p. 183. 


An authentic specimen is figured. 


Patula strigosa, Goutp, var. lactea, HEMPHILL. 
Plate IV. Fig. 8. 


This is a beautiful clear milk-white shell, with 53 whorls, subcarinated at the 
periphery. In the elevated forms the aperture is nearly circular, as broad as high ; 
but in the depressed forms the aperture is broader than high, obliquely suban- 
gulate. The lip is simple, thickened, its terminations joined by a heavy callus, — 
the thickening of the lip and callus is a shade darker than the body of the shell. 
Height of the largest specimen 1 inch, breadth 14 inches. 

Rathdrum, Idaho. 

The above varieties represent a colony of the largest specimens of the strigosa 
group that I have collected. They are an important and very interesting addition 
to the series, and serve to confirm my previous views on the relationship of what I 
call the strigosa group. This colony inhabits open places in the dense pine forests 
of the mountains, overgrown with deciduous bushes. They hibernate among 


MUSEUM OF COMPARATIVE ZOÖLOGY. 178 


leaves, brush, and roots of trees, and in protected and secure places, generally 
on the north slopes of the mountains. (Hemphill.) 


Patula strigosa, var. lactea, Humeni1, The Nautilus, 1890, p. 134. 


An authentic specimen is figured. 


Patula strigosa, var. Utahensis, Hrmruınn. 


For locality, see 2d Supplment, p. 30. This is a rough, coarse, carinated variety, 
figured in Terr, Moll. V., p. 158, Fig. 66. The peristome is sometimes continuous 
by a heavy raised callus connecting its terminations. It is sometimes smaller 
and more elevated. (2d Suppl., p. 88.) 


Patula strigosa, Goutp, var. depressa, CocKERELL. 


Shell flattish, maximum diameter 214, altitude 124 mm. Specimens of this 
variety were sent to me by Miss A. Eastwood, who found them in a cañon near 
Durango, Colorado. The same variety is figured by Binney, Man. Amer. Land 
Shells (1885), p. 166, Fig. 153. . (Cockerell.) 


Patula strigosa, var. depressa, Cockurnii, The Nautilus, 1890, p. 102. 


Patula strigosa, var. albida, Hemrimur. 


Shell broadly umbilicated, greatly depressed, white, tinged with horn-color; sur- 
face covered with fine oblique strie and fine microscopic revolving lines; whorls 
6, convex, the last falling in front; spire very little elevated, apex obtuse, aperture 
oblique, nearly round; peristome simple, thickened, subreflected at the columella, 
its terminations approaching, joined by a thin callus. Height 4 inch, greatest di- 
ameter 1 inch, lesser $ inch. 

Near Logan, Utah. 

Patula strigosa, var. albida, Hnmenrir, The Nautilus, IV. p. 17, June, 1890. 


The above is Hemphill’s description. 


Patula strigosa, var. parma, HEMPHILL. 


Shell broadly umbilicated, greatly depressed, of a dark dirty horn-color, surface 
somewhat rough, covered with coarse irregular striæ, and microscopic revolving 
lines ; whorls 54 or 6, subcarinated throughout, somewhat flattened above, rounded 
beneath, and striped with two chestnut-colored bands, one above and the other 
just at the periphery; spire very little elevated, umbilicus moderately large and 
deep; aperture ovately round, oblique; peristome simple, subrefiected, its termi- 
nations approaching and joined by a thin callus. Height $ inch, breadth 1 inch. 

Near Spokane Falls, Washington. 


Patula strigosa, var. parma, Hsmeniue, The Nautilus, IV. p. 17, June, 1890. 


The above is Hemphill’s description, 


BULLETIN OF THE 


Patula strigosa, var. rugosa, HEMPHILL. 


Shell umbilicated, elevated or globosely depressed, of a dull brown ash-color; 
surface rough, covered with coarse irregular oblique striæ, and microscopic re- 
volving lines; whorls 5, convex, with or without one or two narrow faint revolv- 
ing bands. In most of the specimens the bands are obsolete; spire elevated, 
obtusely conical; suture well impressed; umbilicus large, deep; aperture nearly 
round; peristome simple, thickened, its terminations approaching and joined by 
a thin callus. Height of the largest specimen $ inch, greatest diameter 1 inch. 
Height of the smallest specimen 4 inch, greatest diameter ¢ inch. 

New Brigham City, Utah. 

A large rough robust form, with very convex whorls. Some of the specimens 
so closely resemble solitaria, Say, that one not well acquainted with both forms 
would be easily deceived, and refer it to that species. In its adolescent state the 
lip is very thin or easily broken, and on the surface of the adult shells these frac- 
tures give it a rough and uneven appearance. 


Patula strigosa, var. rugosa, HEMPHILL, The Nautilus, 1890, Vol. IV. p. 16. 


The above is Hemphill’s description. 


Patula strigosa, var. carnea, HEMPHILL. 


Shell umbilicated, greatly depressed, dark horn-color, rather solid, shining, sur- 
face somewhat uneven and covered with irregular oblique striæ; whorls 54, con- 
vex, the last faintly subcarinated in the depressed specimens, falling in front, 
sometimes faintly banded, but most of the specimens are plain and without bands ; 
spire subconical, apex obtuse; suture well impressed, umbilicus large; aperture 
circular; peristome simple, thickened, its terminations well approached and joined 
by a callus. Height % inch, greater diameter #, lesser $ inch. 

Near Salt Lake, Utah. 


Patula strigosa, var. carnea, HemPHILL, The Nautilus, Vol. IV. p. 15, June, 1890. 


The above is Hemphill’s description. 


Patula strigosa, var. fragilis, HemPnHILL. 


Shell umbilicated, elevated or globosely depressed, translucent, thin, fragile, 
somewhat shining, of a dark horn-color, surface covered by fine oblique striæ; 
whorls 5, convex, the last descending in front and striped by two dark chestnut 
bands, one above and the other below the periphery ; suture well impressed ; aper- 
ture oblique; peristome simple, thickened; umbilicus moderate, deep, partially 
covered by the reflected peristome at the columella. Height of the largest speci- 
men % inch, greatest diameter 4 inch, lesser Ẹ inch. 

Near Franklin, Idaho, among red sandstone. 

A very thin and almost transparent variety of the very variable strigosa. By its 


MUSEUM OF COMPARATIVE ZOOLOGY. 175 


peculiar shade, it is very evident that the animal has drawn largely from the red 
sandstone for the material to build its shell. 


Patula strigosa, var. fragilis, HsmrHıur, The Nautilus, Vol. IV. p. 17, June, 1890. 


The above is Hemphill’s description. 


Patula strigosa, var. picta, HEMPHILL. 


Shell umbilicated, elevated or globosely depressed, of a dirty white color, stained 
more or less with chestnut; surface somewhat rough and uneven, covered with 
moderately coarse oblique striæ, and fine revolving lines ; whorls 6, convex, sub- 
carinated, with a broad white band at the periphery, and a dark zone of chestnut 
on the upper side, extending from the peripheral band to the suture, fading out as 
it traverses the whorls of the spire; beneath, on the base of the shell, it is striped 
with numerous bands that sometimes extend into the umbilicus, and also into the 
aperture; spire elevated; apex obtuse; suture well impressed; umbilicus moder- 
ately large and deep, broader in the depressed than in the elevated forms; aper- 
ture nearly circular; lip simple, subreflected, its terminations approaching and 
joined by a thin callus. Height § inch, greatest diameter 14 inches, lesser 1 inch. 

Rathdrum, Idaho. 


Patula strigosa, var. picta, Humpuiii, The Nautilus, Vol. IV. p. 16, June, 1890. 


The above is Hemphill’s description. 


| Patula strigosa, var. hybrida, HEMPHILL. 


| Shell umbilicated, depressed, white, spire horn-color, surface of the shell cov- 
ered with fine oblique striæ, and widely separated revolving raised lines ; whorls 5, 
flattened above, rounded beneath, the last falling in front, and striped with two 
faint chestnut bands; suture well impressed ; umbilicus large, showing nearly all 
the volutions; aperture nearly circular; peristome simple, thickened, its termina- 
tions approaching and joined by a thin callus. Height § inch, diameter $ inch, 
lesser § inch. 

Near Logan, Utah. 

This is an interesting shell, as it is the beginning of the forms of strigosa that 
finally develop the revolving lines into prominent ribs, as seen on the surface of 
var. Haydeni, Gabb. 


| Patula strigosa, var. hybrida, HempuirL, The Nautilus, Vol. IV. p. 17, June, 1890. 


The above is Hemphill’s description. 


Mr. Cockerell (The Nautilus, 1890, p. 102) mentions by name only the fol- 
lowing Colorado forms: — 

P. strigosa Cooperi, form trifasciata, Ckll. Mesa Co. 

P. strigosa Cooperi, form confluens, Ckll. West Mountain Valley, Custer Co.; 
Garfield Co.; Mesa Co. 


176 BULLETIN OF THE 


P. strigosa Cooperi, torm elevata, Ckll. Delta Co. 

P. strigosa Cooperi, form major, nov. Shell with diam. 25 mm. Near head 
of North Mam Creek, Mesa Co., Sept. 14, 1887. 

P. strigosa Cooperi, var. minor, Ckll. Near Egeria, Routt Co., abundant. It 
is quite a distinct local race. 


Pristiloma, Ancer. 

Animal as in Patula. 

Shell small, imperforate, horn-color, shining, many whorled ; spire de- 
pressed conic; aperture sometimes armed with radiating, rather crowded, 
palatal lamelle. 

Northern and Arctic North America, 

Types: Zonites Stearnsi and Lansingi, BLAND. 

Formerly Pristina, Ancey, and Anceyia, PILSBRY, preoc. 

Jaw low, wide, slightly arcuate, ends little attenuated, blunt, with numer- 
ous crowded broad ribs, denticulating either margin. 

Lingual membrane with tricuspid centrals, bicuspid laterals, aculeate mar- 
ginals, as in Zonvtes. 

Separated from Microphysa by the ribbed jaw combined with the lingual 
membrane of Zonites: a very unusual occurrence. 


Pristina Lansingi, BLAND. 


Plate III. Fig. 6. 


I give a better figure of this species. 


Pristiloma Stearnsi, Brann. 
Vol. V., figures in text. Suppl., Plate I. Figs. N (dentition) and O (jaw). 


Shell minute, imperforate, globose conic, striate, shining, horn-colored ; 
suture impressed ; whorls 7, regularly increasing, the last not descending, very 
globose, swollen below, excavated closely around the imperforate umbilical 
region; aperture rounded; peristome simple, acute. Greater diameter 4 mm., 

y that 1 
lesser 34 ; height 2} mm. 

Zonites Stearnsi, Brand, Ann. N. Y. Lyc., XI. 74, Figs. 1, 2 (1875). 

Microphysa Stearnsi, W. G. Binney, Terr. Moll. V., figs. in text; Suppl., Plate II. 

Figs. N (dentition) and O (jaw). 

Astoria, Portland, Oregon ; Olympia, Washington; Alaska. A species of 
the Oregonian region. 

It is larger, more elevated, and more distinctly striated than Lansingi, with 
wider, more rounded, unarmed aperture. 


MUSEUM OF COMPARATIVE ZOÖLOGY. 107 


The jaw is of the same type as described under P. Lansing?, with over 19 
ribs. (Suppl., Plate II. Fig. O.) 

The peculiar lingual membrane also is the same as in that species, with four 
laterals on each side of the central tooth. (Suppl., Plate I. Fig. N.) 


Punctum, Morse. 


Animal as in Patula. 

Shell minute, umbilicated, thin, horn-colored, depressed globose; whorls 4, 
the last not descending ; spire slightly elevated ; aperture rounded ; peristome 
thin, acute. 

Europe and North America. 

Jaw slightly arcuate, ends blunt, not acuminated, composed of numerous , 
subequal, overlapping distinct plates. 

Lingual membrane as usual in the Helicide ; bases of attachment sub- 
quadrate, reflection small, tricuspid in the centrals, bicuspid in the laterals, 
marginals irregularly denticulated. 

Distinguished by the peculiar free plates of the jaw. 

There are two species of Punctum, conspectum and pygmaum. 


Helicodiscus fimbriatus, Wrernersy, var. salmonaceus, HEMPHILL. 
Plate III. Fig. 8. 


I give a figure of this variety from an authentic specimen. See 3d Suppl., 
p. 189. 


Anadenus, HEYNEMANN. 


Animal limaciform, subeylindrical, tapering behind ; tentacles simple; man- 
tle anterior, concealing an internal shell-plate ; no longitudinal furrows above 
the margin of the foot, and no caudal mucus pore ; a distinct locomotive disk ; 
external respiratory and anal orifices on the right posterior margin of the 
mantle; orifice of combined genital system behind and below the light eye- 
peduncle. (See Plate I. Fig. 1.) 

Internal shell-plate small, oval, flat, with posterior nucleus and concentric 
striæ. (See Plate.) 

Jaw with numerous ribs. See Plate TII. Fig. 5. 

Lingual membrane with tricuspid centrals, bicuspid laterals, and quadrated 


marginals. (See same.) 

Differs from Prophysaon by its posterior respiratory orifice, by the position 
of the genital orifice, and by its locomotive disk. 

Himalaya Mountains ; recently found in San Diego County, California, by 
Mr. Hemphill. 


VOL XXI. — NO. 4. 12 


178 BULLETIN OF THE 


It will be remembered that Fischer considers Prophysaon a subgenus of 
Anadenus. 

The geographical distribution of Anadenus would seem to preclude its being 
found in California, but to that genus only can I refer the species whose de- 
scription here follows. 


Anadenus Cockerelli, Hemruıtr. 
Plate I. Fig. 1; Plate III. Fig. 5. 


Length (contracted) 134 mm. ; mantle, length 43, breadth 2 mm. End of 
mantle to end of body, 8 mm. Foot, breadth 2 mm. Foot with the locomotive 
disk, being distinctly differentiated into median and lateral tracts. Respira- 
tory orifice slightly posterior on right side of mantle. Genital orifice below 
right tentacle. No caudal mucus pore. Locomotive disk about half as wide 
as either lateral area. Sides of foot wrinkled, but not differentiated from 
lateral areas, nor specially marked, the wrinkles being a continuation of the 
transverse grooves of the lateral areas. Mantle tuberculate-rugose, oval in 
outline, bluntly rounded at either end ; not grooved as in Amalia. Mantle 
free in front as far as respiratory orifice, Back rather bluntly keeled its 
whole length; rugæ rather flattened and obscure, consisting of grooves en- 
closing mostly hexagonal lozenge-shaped spaces, which are themselves rugose. 
Solor uniform brown-black, without markings, except some dark marbling on 
the lighter sides. The portion beneath and in front of the mantle is pale, and 
the head and neck have a gray tinge. Foot brown. Shell internal, thinnish, 
white, oval in outline. Stomach large, swollen, broad. Liver pale ochrey. 


Anadenus Cockerelli, Hemen, The Nautilus, Vol. IV. No. 1, May, 1890, p. 2. 
Anadenulus, CocKERELL, Ann. Mag. Nat. Hist., Oct., 1890, p. 279. 


Cuyamaca Mountains of San Diego Co., California. Mr. Henry Hemphill. 

Jaw low, wide, slightly arcuate, ends blunt, anterior surface with about 
twenty wide, flat ribs, squarely denticulating either margin. (Plate III. 
Fig. 5.) : 

Lingual membrane short and narrow. Teeth 20-1-20, of which eight only 
on either side are laterals. Centrals tricuspid, laterals bicuspid, marginals 
quadrate, bluntly bicuspid. (Same Plate.) 


Prophysaon Hemphilli. 


From Portland, Oregon, Mr. Hemphill brought seventy-seven individuals of 
a slug which may prove a variety of P. Hemphilli. They have the tawny color 
of flavum. The internal shell is so delicate, it is impossible to remove it 
without breaking it. The penis sac is as in P. Hemphilli. The mantle is 
sometimes smooth, sometimes tuberculate; its fuscous lateral bands are some- 
times united by a transverse posterior band. Some of the individuals had the 
tail constricted preparatory to excision. (See below, under Phenacarion.) 


MUSEUM OF COMPARATIVE ZOÖLOGY. 179 


Prophysaon Andersoni, J. Q. Coormr. 


8d Suppl., Plate III. Fig. 1? Plate VII. Fig. C; Plate I. Fig. 8 (dentition) ; 
Plate IX. Figs. I, J (enlarged surface). 


Shield strongly granular-rugose, the respiratory orifice nearly median on its 
right margin ; tail acute, with small gland; reddish gray, the body somewhat 
clouded with black, the shield paler, clouded, or more usually with a dark 
band on each side above the respiratory orifice, converging in an elliptic form ; 
a pale dorsal streak ; head uniform pale brown, tentacles darker ; foot and 
often the mantle tinged with olive. Length 2.5 inches (Cooper). 


Arion Andersoni, J. Q. Coorer, Proc. Phila. Ac. Nat. Sci., Plate III. Fig. F. 

Prophysaon Andersoni, J. G. COOPER, Pr. Amer. Phil. Soc., 1879, p. 288. 

Prophysaon Andersoni, W. G. Binney, Terr. Moll. V., 8d Suppl., Plate III. Fig. 
1? Pl. VII. Fig. C; Plate I. Fig. 3 (dentition); Plate IX. Figs. I, J 
(surface). 


A species of the Pacific Province, Straits of De Fuca to Oakland, California. 

The characteristic of this species is the light dorsal band, which is not 
present in P. Hemphilli. It has the broad vagina, stout, short, cylindri- 
cal penis sac, and genital bladder of P. Hemphilli, as well as the foliated 
reticulations. 

In the many living and alcoholic specimens which I have examined, I have 
failed to detect any appearance of a caudal mucus pore, which Dr. Cooper is 
confident of having observed, excepting in eight individuals out of thirty col- 
lected by Mr. Hemphill on San Juan Island. 

Many individuals examined by me are excided as described under Phena- 
carton foliolatus. 

Figure 1 of Plate IIT. of 3d Suppl. was drawn from a specimen received from 
Dr. Cooper. It represents the true Andersom, distinguished by a light dorsal 
band, and by genitalia such as I have described for P. Hemphilli. The same 
form, also received from Dr. Cooper, is drawn by Mr. Cockerell on Plate VII. 
Fig. ©. Mr. Cockerell has shown me that I have confounded with it another 
species, which he proposes to call P. fasciatum. See next species. 

Specimens collected by Mr. Hemphill at Old Mission, Coeur d’Alene, Idaho, 
appear to agree with specimens of this species received from Dr. Cooper. 
The jaw is low, wide, slightly arcuate, with over 12 broad, stout ribs, denticu- 
lating either margin. The lingual membrane is given in Plate IT. Fig. 2, of 
3d Suppl. The central and lateral teeth are slender and graceful. The latter 
have, apparently, a second inner cutting point, as is found in Limax agrestis. 
I have so figured it, hoping to draw attention to it, and thus settle the question 
of its being there. On Plate IX. I have given enlarged views of the surface, 
drawn by Mr. Arthur F. Gray. (See Explanation of Plate IX. Figs. I and J 


of 3d Suppl.) 


BULLETIN OF THE 


Prophysaon fasciatum, COCKERELL. 


Length (in alcohol) 19 mm. Mantle black, with indistinct pale subdorsal 
bands, — an effect due to the excessive development of the three dark bands of 
the mantle. Body with a blackish dorsal band, commencing broadly behind 
mantle and tapering to tail, and blackish subdorsal bands. No pale dorsal 
line. Reticulations on body squarer, smaller, more regular, and more sub- 
divided than in P. Andersoni, Cooper. Penis sac tapering, slender. Testicle 
large. Jaw ribbed. (Cockerell.) 

Prophysaon fasciatum, CocgereLL, The Nautilus, 1890. 

Prophysaon fasciatum, W. G. Binney, 3d Suppl. to Terr. Moll. V., p. 209, Plate 

VIL Fig. A. 

Cœur d’Alene Mountains, Idaho; a species of the Central Region. 

This species is described by Mr. Cockerell as distinct from Andersoni, with 
which I have formerly confounded it. (2d Suppl. to Vol. V., p. 42.) It hasa 
dark band on each side of the body, running from the mouth to the foot, and a 
central dorsal dark band. To this must be referred the descriptions of animal, 
dentition, jaw, and genitalia formerly published by me as of Andersoni. 

I am indebted to Mr. Theo. D- A. Cockerell for a figure and description of 
this species. The former is given on Plate VII. Fig. A, while the latter is 
given here in the words of Mr. Cockerell, whose name must consequently be 
associated with it as authority. 

The animal extends itself into a long, cylindrical worm-like body with ob- 
tuse ends; the mantle is covered with minute tubercles. 

Jaw low, arcuate, ends blunt; with numerous (over 15) irregularly devel- 
oped broad, stout ribs, denticulating either margin. 

The lingual membrane has 30-1-30 teeth, with about 12 perfect laterals, 
Centrals tricuspid ; laterals bicuspid; marginals with one long, stout, oblique 
inner cutting point, and one outer short, blunt, sometimes bifid cutting point. 
Resembling that of P. Hemphilli. Another membrane has 50-1-50 teeth. 

Mr. Cockerell describes the penis sac as tapering; in specimens examined by 
me it is cylindrical, as in Hemphilli. 

The internal shell is thick, easily extracted without breaking. 


Phenacarion, CockerkuL.! 


Animal limaciform, cylindrical, blunt before, tapering behind; tentacles 
simple; mantle large, anterior, pointed behind, concealing a delicate, thin, 
subrudimentary calcareous shell-plate, easily fractured; no longitudinal fur- 
rows along the margin of the foot; a caudal mucus pore; no distinct locomo- 
tive disk; external respiratory and anal orifices on the right anterior margin 

1 Phenax = an impostor, and Arion. Cockerell, The Nautilus, Vol. III. p. 126, 
March, 1890. 


MUSEUM OF COMPARATIVE ZOOLOGY. 181 


of the mantle; orifices of the combined generative organs behind and below 
the right eye-peduncle. (See 3d Suppl., Plate VIII. Fig. A.) 

Jaw arcuate, with numerous ribs. (Plate IX. Fig. B of same.) 

Lingual membrane with tricuspid centrals, bicuspid laterals, and quadrate 
denticulated marginals. (Plate IX. Fig. C of same.) 

Northwestern parts of North America, in the Oregon Region, 

Allied to Prophysaon, but distinguished by its more anterior respiratory 
orifice, its rudimentary shell-plate, and decided caudal pore. 


Phenacarion foliolatus, Govt. 


Color a reddish fawn, coarsely and obliquely reticulated with slate-colored 
lines, forming areolw, which are indented at the sides, when viewed by a mag- 
nifier, so as to resemble leaflets; the mantle is concentrically mottled with 
slate-color, and the projecting border of the foot is also obliquely lineated. 
The body is rather depressed, nearly uniform throughout, and somewhat trun- 
cated at the tip, exhibiting a conspicuous pit, which was probably occupied by 
a mucus gland. The mantle is very long, smooth, and has the respiratory ori- 
fice very small, situated a little in front of the middle. The eye-peduncles are 
small and short. Length 85 mm. 

Arion foliolatus, Goutp, Moll. U. S. Exp., page 2, Fig. 2, a, b (1852); BINNEY, 

Terr. Moll., II. 30, Plate LX VI. Fig. 2 (1851); W. G. Binney, Terr. 
Moll., IV. 6; copied also by Tryon and W. G. Binney, L. & Fr. W. 
Sh, 0.877, 

Phenacarion foliolatus, Coceres, The Nautilus, 1890, III. 126; W. G. BINNEY, 
8d Suppl. to Terr. Moll. V., p. 206, Plate VIII. Fig. A; Fig. B (shell- 
plate); Plate IX. Fig. B (jaw); Fig. C (dentition); Fig. D (genitalia). 

Discovery Harbor, Puget Sound (Pickering) ; Olympia and Seattle, Wash- 
ington (Hemphill). 

Dr. Gould adds to the above description these words (Vol. IT. p. 31): “ That 
this animal belongs to the genus Arion there can be little doubt, from the 
peculiar structure of the tail, as represented in Mr. Drayton’s figure, and from 
the anterior position of the respiratory orifice. It is a well marked species, 
characterized especially by the leaf-like areole by which the surface is 
marked.” 

It is. with the greatest pleasure that I announce the rediscovery by Mr. 
Henry Hemphill of this species, which has hitherto escaped all search by 
recent collectors. p It has till now been known to us only by the description 
and figure of the specimen collected by the Wilkes Exploring Expedition, 
almost fifty years ago, and given in Vols. II. and III. of Terrestrial Mollusks. 
A. single individual was found in December, 1889, at Olympia, Washington, 
and sent to me living by Mr. Hemphill. It can thus be described. (See 
Fig. A of Plate VIII. of 3d Suppl.) 

Animal in motion fully extended over 100 millimeters. Color a reddish 


182 BULLETIN OF THE 


fawn, darkest on the upper surface of the body, mantle, top of head, and eye- 
peduncles, gradually shaded off to a dirty white on the edge of the animal, 
side of foot, back of neck, and lower edge of mantle, and with a similar light 
line down the centre of back; foot dirty white, without any distinct locomo- 
tive disk ; edge of foot with numerous perpendicular fuscous lines, alternating 
broad and narrow; mantle minutely tuberculated, showing the form of the 
internal aggregated particles of lime, the substitute of a shell-plate, reddish 
fawn-color, with a central longitudinal interrupted darker band and a circular 
marginal similar band, broken in front, where it is replaced by small, irregu- 
larly disposed dots of same color; these dots occur also in the submarginal 
band of light color. Body reticulated with darker colored lines, running 
almost longitudinally, scarcely obliquely, toward the end of the tail, and con- 
nected by obliquely transverse lines of similar color, the areas included in 
the meshes of this network covered with crowded tubercles, as in Prophysaon 
Andersoni, shown in Plate IX. Figs I, J. Tail cut off by the animal. (See 
below.) Excepting its being of a deeper red, it agrees perfectly with Dr. 
Gould’s description. 

Mr. Hemphill writes of it: “I have to record a peculiar habit that is quite 
remarkable for this class of animals. When I found the specimen, I noticed 
a constriction about one third of the distance between the end of the tail and 
the mantle. I placed the specimen in a box with wet moss and leaves, where 
it remained for twenty-four hours. When I opened the box to examine the 
specimen, I found I had two specimens instead of one. Upon examination of 
both, I found my large slug had cut off his own tail at the place where I no- 
ticed the constriction, and I was further surprised to find the severed tail piece 
possessed as much vitality as the other part of the animal. The ends of both 
parts at the point of separation were drawn in as if they were undergoing a 
healing process. On account of the vitality of the tail piece, I felt greatly 
interested to know if a head would be produced from it, and that thus it would 
become a separate and distinct individual.” The animal on reaching me still 
plainly showed the point of separation from its tail (see Fig. A). The tail 
piece was in an advanced stage of decomposition, I have noticed the con- 
striction towards the tail in many individuals. The edges of the cut were 
drawn in like the fingers of a glove, after the excision. 

The tail of the foliolatus having been cut off, I was unable to verify the 
presence of a caudal pore from this individual. It was plainly visible in an- 
other specimen from Seattle. 

In the large Olympia individual, the irregularly disposed «particles of lime 
in the mantle, of unequal size, seemed attached to a transparent membranous 
plate. With care I removed this entire, and figure it. It is suboctagonal in 
shape (Plate VIII. Fig. B). Under the microscope it appears that the par- 
ticles of lime do not cover the whole plate; at many points they are widely 
separated. This aggregation of separate particles is the distinctive character of 
the subgenus Prolepis, to which foliolatus would belong if retained in Arion, 


MUSEUM OF COMPARATIVE ZOOLOGY. 183 


The genitalia of the large individual from Olympia is figured on Plate IX. 
Fig. D. The ovary is tongue-shaped, white, very long and narrow ; the ovi- 
duct is greatly convoluted ; the testicle is black in several groups of cœca ; 
the vagina is very broad, square at the top with the terminus of the oviduct, 
and the duct of the genital bladder entering it side by side; the genital blad- 
der is small, oval, on a short narrow duct; the penis sac is of a shining white 
color, apparently without retractor muscle; it is short, very stout, blunt at the 
upper end where the extremely long vas deferens enters, and gradually narrow- 
ing to the lower end. There are no accessory organs. The external orifice of 
the generative organs is behind the right tentacle. (See 3d Suppl., Plate IX. 
Fig. D.) 

The jaw is very low, wide, slightly arcuate, with ends attenuated and both 
surfaces closely covered with stout, broad separated ribs, whose ends squarely 
denticulate either margin. There are about 20 of these ribs. (See Plate IX. 
Fig. B.) 

The lingual membrane is long and narrow, composed of numerous longitu- 
dinal rows of about 50-1-50 teeth, of which about 16 on each side (Plate IX. 
Fig. C) may be called laterals. Centrals tricuspid, laterals bicuspid, marginals 
with one long inner stout cutting point, and one outer short side cutting point. 
The figure shows a central tooth with its adjacent first lateral, and four extreme 
marginals. 


Phenacarion Hemphilli. 


This form is figured on Plate VIII. Fig. © of 3d Suppl. When extended 
fully, it is 70 mm. long. It is more slender and more pointed at the tail than 
foliolatus. The body is a bright yellow, with bluish black reticulations. The 
edge of the foot and the foot itself are almost black ; shield irregularly 
mottled with fuscous ; the body also is irregularly mottled with fuscous, and 
has one broad fuscous band down the centre of the back, spreading as it joins 
the mantle, with a narrower band on each side of the body. The other charac- 
ters, external and internal, are given below. It loses its color on being placed 
in spirits, becoming a uniform dull slate-color. Mantle lengthened oval. 
Shell-plate represented by a group of calcareous grains concealed in the mantle; 
it is impossible to remove it as one shell-plate. A decided caudal pore. 


Phenacarion foliolatus, var. Hemphilli, W. G. Binney, 8d Suppl. to Terr. Moll. V., 
p. 208; Plate VIII. Fig. C; Plate X. Fig. H (genitalia). 


Gray’s Harbor and Chehalis, Washington, and Portland, Oregon (Hemphill); 
a species of the Oregon Region. 

On the only living one of the lot from Gray’s Harbor, the pore was dis- 
tinctly visible, and is figured on Plate VIII. Fig. ©. Usually it seemed more 
“a conspicuous pit” than a longitudinal slit, as in Zonites. At one time I 
distinctly saw a bubble of mucus exuding from it. It opened and shut, and is 


BULLETIN OF THE 


184 


still plainly visible on the same individual, which I have preserved in alcohol 
and added to the Binney Collection of American Land Shells in the National 
Museum at Washington. 

Jaw low, wide, arcuate, ends attenuated, anterior surface with 16 ribs, den- 
ticulating either margin. 

Lingual membrane as in foliolatus ; teeth 50-1-50, with 19 laterals on each 
side. 

Genitalia (3d Suppl., Plate X. Fig. H) ; the form from Gray’s Harbor has 
its generative system very much the same as described for foliolatus above. The 
ovary is much shorter and tipped with brown, and is less tongue-shaped. The 
penis sac tapers to its upper end. The vagina is not squarely truncated above, 
The system much more nearly resembles that of Prophysaon Andersoni (see 
Terr, Moll., V.) than that of the Olympia foliolatus. 


Binneya notabilis, J. G. COOPER. 
Plate I. Fig. 9. 


A new figure is here given, drawn by Mr. Cockerell. 


Triodopsis Mullani, Brayn, var. Blandi, Hemenite. 
Plate II. Fig. 6. 


Shell with the umbilicus partially closed, orbicularly depressed ; dark horn-color, 
obliquely striated; spire short, very slightly elevated, nearly planiform ; aperture 
semilunar, at a right angle with axis of the shell, with a very short nipple-like pari- 
etal tooth; peristome thickened, white, plain, without teeth and roundly reflected. 
Height } inch, breadth 3 inch. 

Post Falls, and banks of Salmon River, Idaho. 

Helix Mullani in form and size resembles very much the common tridentata of 
the Eastern States. Among the various forms it assumes, none are more marked 
than the little depressed shell before me. It can be very readily separated from 
the typical Helix Mullani, or its other varieties, by its very depressed form, small 
size, and the absence of the teeth-like processes on the inner margin of the 
peristome. 

I cannot detect any microscopical revolving lines, or tubercles bearing hairs, 
mentioned by Bland in his description of H. Mullani. 


The above desciption is by Mr. Hemphill, who furnished me with the 
specimen figured. 


Polygyra septemvolva, var. Floridana, Hemruitt. 


Shell deeply umbilicated, elevated, globose conic, light horn-color, with numerous 
fine ribs above, but smooth beneath; whorls 53 or 6, the last subangular at the 
periphery ; suture well impressed; spire greatly elevated with an obtuse apex ; 


MUSEUM OF COMPARATIVE ZOOLOGY. 185 


aperture lunate, well rounded, and nearly circular; peristome reflected, rounded ir 
front, the margins joined by a triangular tooth on the parietal wall. Greater diam- 
eter 6 mm., altitude 5 mm. 

Oyster Bay, Florida. 

This is a small, very elevated form of the P. cereolus group. 


The above is Mr. Hemphill’s description. 


Mesodon ptychophorus, A. D. Brown, var. castaneus, Hermrnitr. 


Shell umbilicated, globosely depressed, of a dark chestnut color; surface covered 
with coarse, irregular, widely separated lines of growth, and crowded, microscopical 
revolving lines; whorls 54, convex, the last slightly descending in front, spire ele- 
vated; suture well impressed, aperture subcircular, lip white, reflected and par- 
tially covering the umbilicus, its terminations approaching ; umbilicus small and 
deep. Height § inch, diameter 1 inch. 

Old Mission and Rathdrum, Idaho. 

I regard H. ptychophorus as the progenitor of what I call the T’ownsendiana group 
of West Coast land shells, and this colored variety seems to still further indicate 
its relationship to Townsendiana, for the spire whorls of nearly all the specimens 
of Townsendiana that I have collected are chestnut-colored. Zownsendiana does not 
begin to put on its wrinkles until it has made about four revolutions of the shell. 
The wrinkles are probably due to its environment. 


The above is Hemphill’s description, from The Nautilus, Vol. IV. p. 41, 
1890. 


Aglaja fidelis, var. flavus, Hempnir4. 


Shell umbilicated, elevated, very faintly subcarinated, of a uniform light yellow 
color throughout, without bands or other stains of coloring; whorls 64, convex, with 
coarse oblique strie, and microscopic irregular revolving lines; peristome reflected 
below, simple above; aperture roundly ovate; umbilicus moderate, and partially 
covered by the reflected peristome; suture distinct. Greater diameter 34 mm., alti- 
tude 23 mm. 

Chehalis and San Juan Islands, Washington ; Port Orford, Oregon. 

This isa rare and beautiful variety of this well known West Coast land snail. 


The above is Mr. Hemphill’s description. 


Aglaja fidelis, var. subcarinata, HEMPHILL. 


Shell orbicularly depressed; umbilicated; of a deep dark chestnut-color without 
bands; whorls 64, convex or somewhat flattened, the last subcarinated at the 
periphery ; strie coarse, oblique, crossed by numerous well defined wavy revolving 
lines; peristome simple, thickened above, reflected below, and nearly covering the 
umbilicus ; umbilicus moderate; aperture roundly ovate; suture well impressed. 
Greater diameter 37 mm., altitude 20 mm. 

Humboldt Co., California. 


186 BULLETIN OF THE 


This is a very dark, intermediate form of jidelis, which in its southern march 
under changed conditions assumes a more carinated form, and is known to con- 
chologists as infumata, Gould. 


The above is Mr. Hemphill’s description. 


Arionta Coloradoensis, STEARNS. 


Shell orbicular, moderately depressed, whorls slightly elevated, apex obtuse, 
number of whorls four to four and a half, rounded. Umbilicus narrow, showing the 
penultimate whorl, though partially covered by the reflection of the lip at the point 
of junction with the base of the shell. Aperture obliquely ovate, nearly circular, 
and almost as broad as high. Lip slightly thickened and reflected, or simple, vary- 
ing in this respect; more reflected and aperture more effuse at the columella. 


Parietal wall in the heavier examples calloused, the callus connecting with the 
inner edges of the outer lip above and below. Shell rather fragile, thin, translu- 
cent; surface smooth and shiny, and sculptured with fine incremental lines. Color 
pale horn to white, and otherwise marked by a single narrow revolving reddish 
brown band just above the periphery, which in some specimens is obscure or 
absent. In some individuals certain faint scars upon the upper whorls imply an 
occasionally hirsute character. 


mm. 
Maximum diameter of largest . . . » 2 2 . . + 15.25 
Minimum diameter of largest . . . . 222... 18.25 
Altitude orlargest «e s oora w o a a a o 210.20 
Maximum diameter of smallest adult . . . . . . 18.75 
Minimum diameter of smallest adult . . . . . . 12.00 
Altitude of smallest adult >a 5 sv. par r y Sid 


Grand Cañon of the Colorado, opposite the Kaibab plateau, at an elevation of 
8,500 feet. (Mus. No. 104,100.) 

The above, while exhibiting a facies or aspect of its own, its nevertheless sug- 
gestive of H. Remondi, Gabb, Mazatlan, in the Mexican State of Sinaloa, and also 
from the high'mesas or table lands in the neighborhood of Mulege, Lower Cali- 
fornia. H. Carpenteri, Newcomb, which is a synonym of H. Remondi, is credited 
by the author to “ Tulare Valley,” and has been found in other localities in Cali- 


MUSEUM OF COMPARATIVE ZOOLOGY. 187 


fornia. A glance at the map will show how widely separated geographically 
H. Coloradoensis is from its nearest allies, and this discovery of Dr. Merriam’s 
extends the distribution of the West Coast type of Helices farther to the eastward 
than heretofore, and adds an area of great extent to that previously known. 


The above description and figure were published by Stearns in Proc. U. S. 
Nat. Mus., Vol. XIII. p. 206, Plate XV. Fig. 6, 7, 8, 1890, all copied above. 

I have examined the jaw and lingual dentition to find them similar to those 
of the other species of Arionta. 


Arionta Traski, var. proles, HEMPHILL. 


Shell umbilicated, very much depressed, thin, shining, of a dark horn-color ; 
whorls 54, somewhat flattened above, convex beneath, the last slightly falling in 
front, with a dark band above the periphery, and crowded with strong oblique 
striæ; suture well impressed; umbilicus moderately large and deep; aperture 
hardly oblique ; peristome simple, thin, subreflected, its terminations approaching. 
Height § inch, breadth $ inch. 

Tulare Co., California, near Fraser’s Mill. 

A much flatter and more depressed form than any of the varieties of Traski that 
I have seen. There are no revolving microscopical lines, as in Traski. 


The above is Mr. Hemphill’s description. 


Arionta tudiculata, var. Tularensis, Hrmrmitr. 


Shell umbilicated, very thin and frail, shining, of a light greenish horn-color, 
globosely depressed; whorls 5}, convex, the surface minutely granulated, and 
crowded with fine oblique striæ, with a single chestnut revolving band; suture 
well impressed; umbilicus very small; aperture oblique, subcircular; peristome 
simple, hardly thickened, its columellar portion expanding and nearly covering the 
small umbilicus. Height § inch, breadth 4 inch. 

Tulare Co., California. 

This is one of those puzzling intermediate forms uniting two species that can be 
with equal propriety placed in one or the other. It has the exact form of the 
typical Traski found at Los Angeles, and along the coast, though much smaller 
and thinner, and it has the sculpturing of tudiculata much modified. It seems to 
fill the gap quite completely between those two species. 


The above is Mr. Hemphill’s description. 


Arionta tudiculata, BINNEY. 
Plate II. Fig. 7, 8. 
New figures are here given of the form cypreophila. 


In The Nautilus, Vol. IV. p. 41, 1890, Mr. Hemphill also describes a 
var, subdolus thus: — 


188 BULLETIN OF THE 


Shell narrowly umbilicated ; globosely depressed, of a dark yellowish color, sur- 
face somewhat shining, covered with oblique striæ, interrupted by numerous wavy 
lines and oblong blister-like wrinkles, hardly perceptible to the naked eye; whorls 
54, convex, striped by a single chestnut band, double margined by lighter ones; 
spire very little elevated, suture well impressed; lip simple, reflected, and nearly 
covering the umbilicus, its terminations approaching and joined by a thin callus ; 
umbilicus narrow and small. Height $ inch, greatest diameter 1 inch, lesser 4 inch. 

San Jacinto Valley, San Diego Co., California. 

A very depressed form, quite variable in size, some of the specimens not being 
more than half the size of the measurements given. It is lighter colored than any 
of the southern varieties of tudiculata except var. Binneyi. 


Arionta Ayresiana, Newcoms. 
Plate I. Fig. 7. 


I give a new figure of this species, 


Arionta intercisa, W. G. Binney. 


In “Zoe,” Vol. I. No. 11, January, 1891, p. 330, Mr. Hemphill describes 
these varieties of A. intercisa : — 

Var. minor. Smallest specimen, greatest diameter 18 mm., altitude 11 mm. 
Uniform light yellowish chestnut-color, with and without a band, and varies 
very much in form and elevation or depression of spire. 

Var. elegans. Uniform ashy buff-color, faintly banded, and variable in form. 

Var. nepos. Uniform ashy white; spire horn-color, variable in form and 
sculpturing. 

Var. albida. Uniform. milk-white, sometimes with a faint band at the 
periphery; sculpture nearly obsolete. 

In the same journal (p. 434) Mr. Hemphill thus describes several varieties 
of redimita, which species he refers, however, to Kelletti: — 

Var. castaneus. Uniform, polished, chestnut-color, darker band at the periph- 
ery, spire sprinkled with fine ashen specks. 

Var. hybrida. Uniform ash-white color, and a dark band at the periphery, 
flecked with transverse markings and specks of dark brown and light chestnut. 


Arionta ruficincta, Gass. 
Plate I. Fig. 3. 


A new figure is given of this species, 


Arionta Kelletti, Fornns. 


Mr. Hemphill, in Terr. Moll. V., 3d Suppl., has thus described several 
varieties. I figure authentic specimens of each. 
Var. albida (Plate IV. Fig. 3). This is a beautiful clear white translucent 


MUSEUM OF COMPARATIVE ZOÖLOGY. 189 


väriety, with no markings or stains of any kind. It is quite thin and frail, 
and a trifle smaller than the average size of Kelletti. 

Santa Catalina Island, California. Two specimens only found by me. 

Var. castanea (Plate IV. Fig. 4). Among the numerous patterns of coloring 
assumed by M. Kelletti, none are more conspicuous than this well marked va- 
riety. The body whorl is of a deep shiny chestnut-color above the periphery, 
and becomes lighter as it follows the whorls of the spire to the apex. The 
band at the periphery is quite variable in the different specimens; it is gener- 
ally light and well defined above, but below it is irregular, and spreads over 
the base of the shell more or less. 

Santa Catalina Island, California. This variety is not rare. 


In “Zoe,” Vol. I. No. 11, pp- 333, 334, Mr. Hemphill has also thus described 
several other forms. 

Var. nitida (Plate IV. Fig. 2). Uniform, translucent, shining, dark horn- 
color, with a poorly defined dark band, coalescing with a poorly defined whit- 
ish band below it, at the periphery; spire faintly flecked with ashen gray. 

Catalina Island. 

Var. multilineata (Plate IV. Fig. 1). Shell marked by alternate shades of 
ashen white, chestnut, or brown, arranged in an irregular series of revolving 
and sometimes wavy lines, with a broader and poorly defined band at the 
periphery; markings finer beneath than above. 

Var. frater. Shell of a beautiful, uniform, horn-buff color, sometimes fad- 
ing into lighter horn-color, with a darker band at the periphery, and numerous 
faint, alternate revolving lines of ashen or dark horn-color above and below ; 
generally, not always, lighter colored beneath, and sometimes with a whitish 
zone beneath the band at the periphery. 

Var. Californica, The shell is colored with a darker shade of uniform buff 
than the above, dark band at the periphery, generally uniform in color above 
and below; sometimes flecked with squarish dots. 

Var. Forbesi. Ground coloring whitish buff, with a revolving series of poorly 
defined and coalescing lines, bands, and blotches. 

Var, bicolor. Color very dark horn or brownish, flecked with numerous re- 
volving very fine dots or irregular lines, with or without a very faint band at 
the periphery. 

Var. tricolor, Trregularly painted with numerous revolving whitish, brown- 
ish, and chestnut flecks, blotches, and stains, with or without a band at the 
periphery. 

Var. albida. (See below.) 

Var. albida, a. Milk white ground, very faintly stained with light horn, 
and with poorly defined and fading lines. 

Mr. Hemphill considers redimita as a form of Kelletti. (See that species.) 


190 BULLETIN OF THE 


Euparypha Tryoni, Newco. 


Mr. Hemphill has thus described several varieties. (See Zoe, Vol. I. pp. 331, 
332.) 

Var. varius. The upper or dark zone is of a lighter shade of bluish brown 
or chestnut than the type, and is flecked and sprinkled with ashen white; 
band at the periphery dirty white beneath. 

Var. nebulosa (Plate IV. Fig. 5). Lighter colored above than var. varius, 
marbled and clouded with various patterns of dark brown and dirty white ; 
dirty white beneath. 

Var. fasciata (Plate IV. Fig. 6). Uniform light chocolate above and be- 
neath, with a dark band at the periphery. 

Var. Californica, Creamy buff-color, darker above than below the periph- 
ery, very faintly banded. 

Var. albida. Uniform creamy, and sometimes milk-white above and be- 
neath, and without band. 

Var. subcarinata. Among the subfossils that occur on Santa Barbara Island 
we find a form of H. Tryoni which adds an interesting link to its history and 
to its present form. It may be characterized as follows. Shell depressed glo- 
bose, consisting of about 53 whorls, the last subcarinated at the periphery; in 
other respects closely resembling the recent form. Greater diameter 23.15 and 
20.11 mm., largest and smallest specimens. 


Pomatia Humboldtiana, Vat. 


Texas, at Altuda, at an elevation of 5,000 feet, where it, a single specimen in fair 
condition, had been thrown out with soil by a prairie dog. (Mus., No. 118,366.) 
William Lloyd. 

This species has not before been reported from any locality within the territory 
of the United States. It was described from Mexico, where it is found in the 
neighborhood of the city of Mexico, and in other localities. The national collec- 
tion contains several examples from the Real del Monte. It has a pretty close 
resemblance to some of the varieties of the European ZZ. (Pomatia) pomatia, and it 
may possibly be an introduced form. H. pomatia has for centuries been esteemed 
as an article of food in various parts of Europe, and was regarded as a dainty by 
the ancient Romans. It was propagated and raised in large quantities for their 
use, and specially fed on certain plants to give the flesh a particular flavor. 

Unmistakable specimens of another favorite edible snail common to Europe, 
H. (Pomatia) aspersa, is found in Mexico, and examples from Puebla, in the prov- 
ince of Puebla, Mexico, were presented to the National Museum by the Mexican 
Geographical Commission a few years ago. The presence of these two forms most 
certainly suggests the question as to whether they were not introduced by the 
Spaniards many years, centuries, ago, either for food purposes or incidentally in 
the routine and accidents of commercial intercourse. 


The above was published by Stearns in Proc. U. S. National Museum, Vol. 
XIV. p. 96, 1891. It will be remembered that Helix Buffoniana was figured 
as aspersa by Dr. Binney in Volume III. 


MUSEUM OF COMPARATIVE ZOÖLOGY. 191 


Bulimulus Ragsdalei, PıLserr. 
Plate II. Fig. 9. 


Tt is about the size and form of B. Mooreanus, but rather more slender and 
elevated. The surface is not smooth, as in the other American Bulimuls, but 
strongly ribbed-striate longitudinally. The apex is blunt; peristome thick- 
ened within ; columella reflexed over the narrow but open umbilicus. The 
aperture is less than half the length of the shell ; color brownish, corneus 
somewhat translucent, the riblets opaque white. Height 22 mm,, diam. 
10 mm. ; height of aperture 10$ mm., diameter 7 mm. 


Bulimulus Ragsdale’, Pinssry, The Nautilus, Vol. III. p. 122, March, 1890. 
Proc. Acad. Nat. Sci. Phila., 1890, p. 296, Plate V. Fig. 3. 


St. Jo, and at Warren’s Bend, twenty-five miles from Gainesville, and in 
Cook and Montague Counties, Texas (Ragsdale). 

A figure of an authentic specimen is given 1} the natural size. The descrip- 
tion is a copy of the original. 


Bulimulis Dormani. 


Plate I. Fig. 6. 
A new figure is given. 


Rhodea Californica. 


This extralimital species has actually been received by Dr. Cooper from. 
Lower California. (Proc, Cal. Acad. Nat. Sci, 1891, p. 102.) It had been 
quoted as an Achatina from Monterey. (See Vol. V.) 


Pupa Californica. 


Dr. Sterki in Nautilus, Vol. IV. page 7, mentions a variety, elongata, from 
San Clemente Island ; on page 18, varieties trinotata, Diegoensis, and cyclops. 


Pupa Coloradensis, COCKERELL. 


Shell brown, shiny, thinnish, striate, especially on penultimate whorl ; out- 
line oblong-oval, barrel-shaped ; apex blunt; whorls 4; aperture pyriform ; 
peristome brown, thick, continuous by a well marked callus on parietal wall ; 
outer lip not constricted. The teeth within the aperture are brown, one long, 
one on parietal wall, one on columella, and two (the lower one largest) on outer 
wall. Long. 1}, lat. 1 mm. Allied to P. corpulenta, but decidedly smaller, 
more striate, and slightly narrower. (Cockerell.) 


192 BULLETIN OF THE, 


Pupa Pilsbryana, Srerkı. 


Shell minute, narrowly perforate, cylindrical-oblong to cylindrical, somewhat 
attenuated towards the rather blunt apex, colorless (when fresh glassy) with a 
very delicate bluish tint, smooth and polished, with few, irregular microscopic 
strie which are more marked near the aperture. Whorls 44-54, moderately 
rounded with a rather deep suture, especially in the upper half, regularly and 
slowly increasing, the embryonal being relatively large, the last somewhat 
ascending toward the aperture; the latter of moderate size, lateral, subovate, 
margins approached, peristome somewhat expanded, without a thickened lip or 
a callus in the palatal wall; outside is a barely perceptible trace of a crest near 
the margin, and behind that a slight impression most marked upon the inferior 
palatal fold. Lamelle 4 or 5; one apertural, rather high, of moderate length, 
simple; one columellar, horizontal, of moderate size, simple; basal very small or 
wanting; palatals the typical, inferior deeper seated, of moderate size, superior 
small or very small. Alt. 1.5-1.7, diam. 0.8-0.9 mm. 


Pupa Pilsbryana, STERKI, The Nautilus, Vol. III. p. 123, March, 1890. 


There is a slight variation ; the example from New Mexico being of lesser diam- 
eter, and having no trace of a basal lamella. 

The soft parts have not been seen so far, but will be of high interest, since, to 
judge from the shell, our species seems to be an intermediate form between the 
hordeacella, ete. group, and P. eurvidens, especially its var. gracilis. 

P. Pilsbryana has much resemblance in shape and size to small albino examples 
of P. hordeacella, Pilsb., but under a glass is at once distinguished by the shorter 
simple apertural lamella not ending at or very near the upper termination of the 
palatal margin, as it does in hordeacella, and by the smooth surface. The fine bluish 
hue may also be a distinguishing character if it prove constant. 


The above is Sterki’s original description. 
g p 


Pupa calamitosa. 
Plate II. Fig. 1, 


See 3d Suppl., p. 219. A reduced copy of one of the original figures is given 
here. 


Pupa Hemphilli, Srerkı. 


In examining a lot of about forty-five specimens of Pupa calamitosa from the 
banks of San Tomas River, Lower California, I found there were two distinct 
forms in them. The author says, in his description of P. calamitosa: “ Several 
specimens have only one lamella on the outer lip, and are rather larger than the 
typical form described,” represented in Plate XII. Fig. 16 (loc. cit., No.7). Probably 
I had a greater number of examples at disposition than Mr. Pilsbry. The two 
forms proved to be distinct by an entirely different formation of the lamella, as 


MUSEUM OF COMPARATIVE ZOOLOGY. 193 


well as of the basal part of the shell. And among the whole number I found not 
one intermediate or doubtful specimen. There is no doubt but that we have to 
consider them as being specifically distinct, the more so since they live together in 
the same locality. For the new species I would propose the name P. Hemphilli, in 
honor of the man to whom we owe so many valuable additions to our malaco- 
logical fauna. 

As in shape and general appearance the two species are almost alike, it may be 
the best way to characterize the one in question by comparing it with P. calamitosa, 
Pilsb. P. Hemphilli averages a trifle larger than its companion, but either is some- 
what variable in size. While calamitosa has a 
minute perforation, Hemphilli is umbilicated in 
quite a peculiar way. ‘There is a nodule-like pro- Ba 
jection on the umbilical part of the last whorl, 
producing a rima beside the umbilicus ; in calam- 
itosa there is nothing of this formation. On the 
other hand, the latter has a small but distinct 
groove-like impression just at the base, near the . 
aperture appearing as a slight projection inside. % 
This feature is wanting in Hemphilli, Lamellæ : 
in the latter species, when looking from front, only one is generally seen in the 
palatal wall, corresponding to the superior one in calamitosa, but longer; i.e. be- 
ginning deeper in the throat, and fairly seen on the outside; also marked there by 
a corresponding impression, ascending in a curve from near the base. A little dis- 
tant from its inner end, just above the projection mentioned, there is another 
lamella beginning, directed toward the base and ending there, also seen on the 
outside. Quite generally there is a very small, thin, but well formed lamella in 
the palatal wall, near the projecting auricle. The columellar fold is quite short 
and small in Hemphilli, yet consisting of a vertical and a horizontal part. The 
(main) apertural lamella is decidedly longer in our species, and the supra- 
apertural higher and entire, while in calamitosa it is evidently composed of two 
parts marked by an indentation in the middle, or even entirely separated, in quite 
mature specimens. 

About twenty examples, collected at San Diego, Cal., by Mr. Hemphill, are all 
P. Hemphilli, no calamitosa among them. They are little different from the San 
Tomas River specimens, except by a somewhat shorter palatal lamella. 


The above is Sterki’s description (The Nautilus, July, 1870, Vol. IV. p. 27). 
My figure was drawn by him from the type. 


Pupa hordeacella, PILSBRY. 
Plate II, Fig. 2. 


The shell is of a long-ovoid shape, smaller and slenderer than P. servilis, Gould, 
translucent, waxen white, finely striate; the aperture is rounded, with a thin, ex- 
panded peristome. Within, there is, on the parietal wall, an entering fold arising 
near the termination of the outer lip, its edge a trifle sinuous or nearly straight; 
the columella has a fold about in the middle. There is a tiny deep-seated fold on 

VOL. XXII. — NO. 4, 13 


194 BULLETIN OF THE 


the base of aperture, near the columella, an entering fold within the outer lip, 
equidistant from the above described parietal and columellar folds, and a tiny 
denticle above it. The columellar fold is not situated so high on the pillar as in 
P. servilis. The latter half of the body whorl is flattened on the outer lower por- 
tion, as the Figure J shows. There is a low wave-like ridge or “ crest” also, but 
scarcely visible in many specimens. Alt. 1.8, diam. 8 mm. 


Pupa hordeacella, Piuspry, Proc. Acad. N. Sci. Phila., 1890, p. 44, Plate I. Figs. 
G, H, I, J, K. 


Arizona to Florida. 

The figures were drawn with the aid of the camera lucida. They should be com- 
pared with Gould’s excellent figures of P. servilis in the Boston Journal of Natural 
History, Vol. IV., Plate 16, Fig. 14, and those of P. pellucida, in Strebel’s Beitrag 
zur Kenntniss der Fauna mexikanischer Land- und Süsswasser-Conchylien, Theil 
IV. Plate XV. Fig. 10. The latter are the more valuable in this connection, as 
they are not only faithful drawings on a sufficiently large scale, but are the only 
ones drawn from continental specimens (Vera Cruz, Mexico). The measurements 
given by Strebel and Pfeffer are, alt. 24, diam. of last whorl fully 1 mm., alt. of 
aperture 3 mm. Gould’s P. servilis and Pfeffer’s P. pellucida were both described 
from Cuba. I see no reason for not following W. G. Binney in considering them 
synonymous, pellucidus having precedence. (Pilsbry.) 


The above is Pilsbry’s description. I give also a reduced view of one of his 
figures. 


Pupa Clementina, Srerkt. 


Shell very minute, narrowly perforate, cylindrical, pale horn-colored, transpar- 
ent, with rather obtuse apex ; whorls 54, regularly increasing, moderately rounded, 
with rather deep suture, smooth, with few microscopic strie, somewhat 
shining; last whorl occupying rather more than two fifths of altitude, 
somewhat ascending to the aperture, with a slight, revolving impression 
on the middle of its last one third, ending at the auricle; a very slight, 
flat crest elevation near the margin, only in the lower part; aperture lat- 
eral, scarcely oblique, subovate with the palatal margin slightly flattened, 
upper part of same somewhat sinuous, peristome a little expanded with 

peeved a slightly thickened lip just at the margin; lamellæ 6, white, two on 

the apertural wall, the apertural typical, and a rather long supra-aper- 
tural, ending in a callus at the upper termination of the palatal margin; columel- 
lar one typical, horizontal; basal very small, nodule-like, deep-seated ; palatals 
two, typical, the inferior a little longer. Alt. 1.9, diam. 0.8 mm. ; apert, alt. 6, 
diam. 0.5 mm. 

Three examples of this species were collected by Mr. H. Hemphill on San Cle- 
mente Island, California, among numerous P. Californica, Row. All were exactly 
alike, well formed and fully mature. They cannot be referred to any one of our 
species published, and doubtless represent a form of their own, although so far it 
was not possible to examine the soft parts. 


MUSEUM OF COMPARATIVE ZOOLOGY. 195 


In size, shape, and general appearance it somewhat resembles /sthmia, yet lacks 
the rib-like striation; the lamella would be typical for Vertigo and some of the 
smaller Pupe but for the presence of the well developed supra-apertural which 
P. Clementina has in common with P. calamitosa, Pilsbry, and Hemphill’, Sterki; 
but, on the other hand, there is nothing of the characteristic palatal or gular folds 
of these two species. ‘Thus, in several regards, our form is an intermediate and 
connecting one between different groups, and consequently deserves our special 
interest. 


Pupa Clementina, Srerkı, The Nautilus, Vol. IV. No. 4, Plate I. Fig. 4, August, 
1890. 


The above is a copy of Sterki’s original description and figure. 


Pupa Dalliana, STERKI. 


Shell conic or ovate-conic, of greenish horn-color, transparent, finely irregularly 
striate in the lines of growth, polished ; whorls 44, well rounded, with deep suture 
rather rapidly increasing, the last occupying about } of altitude 
towards the aperture, somewhat ascending on the penultimate. 
Aperture lateral, somewhat oblique, subovate, with just percepti- 
bly flattened palatal margin; margins approximate, the ends pro- 
tracted; peristome shortly but decidedly expanded, with a very fine 
thread-like lip near the margin, the same continuing as a very fine 
callus on the apertural wall inside of the line connecting the ends of 
the margins; palatal wall quite simple; no lamelle. Alt. 1.2, diam. 
1.3 mm. Pupa Dallia na. 

This form has been collected by Mr. Hemphill near Clear Lake, 

Lake Co., Cal., and I propose to name it in honor of Mr. William H. Dall. The 
specimens before me were fifteen, fresh, remarkably uniform in their whole appear- 
ance; all were more or less covered with a dark brown hard crust of slime and 
dirt, generally thickest around the aperture. Doubtless this coating is done 
“ purposely ” by the animals, as in many other species also. When cleaned, it 
shows about the size and shape of a well grown Vertigo ovata, Say; but by a good 
eye, or under a glass, is at once recognized as something else, by the rounded 
aperture and the absence of lamella. (Sterki.) 


Pupa Dalliana, Srerkı, The Nautilus, Vol. IV. No. 2, p. 19, June, 1890. 
Dr. Sterki’s description is copied above, My figure was drawn by him from 
the type. 
Pupa syngenes, PILSBRY. 
Shell subcylindrical but wider above, composed of eight narrow, convex whorls, 


sinistrally convoluted ; texture as in P. muscorum, but color rather lighter brown. 
Last whorl ascending, imperforate, bearing a strong high crest just behind the 


196 BULLETIN OF THE 


outer lip. Aperture shaped as in muscorum, having a single small parietal denticle. 
Altitude 33, diameter 13 mm. 


Pupa syngenes, Prtspry, The Nautilus, 1890, Vol. IIL. p. 296, Plate V. Figs. 1, 2. 


Two specimens of this form are before me, and I am in doubt whether to give 
them a new name, as they may be only sinistral monstrosities of the common 
P. muscorum. The shells are labelled “ Arizona” in the Academy collection, col- 
lector not known. 

(Since the above paragraphs were in type, I have received a communication from 
my friend, Dr. V. Sterki, to whom I sent a specimen of P. syngenes, which I at first 
described as a variety of muscorum. He says: — 

“Tam satisfied that it is a species, and not a var. of muscorum ; the shape of the 
whole shell, the last whorl so considerably flattened, and ascending, the number of 
whorls, seem to me to prove its specifical rank... . After washing out the aper- 
ture of your specimen, I saw a rather strong lamella or tooth on the columella, and 
a barely perceptible trace of an inter-palatal lamella, which, however, is validified 
by the impression on the outside.”) 


The above is Pilsbry’s description. An authentic specimen drawn by Dr. 
Sterki is figured here, 


Vertigo ovata, Say. 


Of V. tridentata Sterki writes (The Nautilus, 1890, p. 135): “It has a 
wide distribution in the northern part of the country ; originally found in 
Illinois, it has been collected in different parts of Ohio and New York, as 
well as in Minnesota and Colorado. In general it is remarkabl y constant in 
its characters ; yet there are slight differences; here I found a few examples 
from low ground, together with V. ovata; they were a trifle larger, with a 
thicker and deeper colored shell than those from upland places,” 


MUSEUM OF COMPARATIVE ZOOLOGY. 197 


Vertigo Oscariana, STERKI. 


This is the most peculiar of our species. It is of the size of milium, but oblong, 
with either end nearly equally pointed, the last whorl being considerably narrowed 
and flattened towards the subtriangular, small aperture; shell thin, delicate, of 
pale horn-color, as is the palatal wall and margin; the latter simple and straight, 
with a very slight, thin callus inside ; lamellæ 3, whitish, rather small; one aper- 
tural, one columellar (longitudinal), and the inferior palatal; some- 
times there is also a very small superior palatal. Length 1.5, diameter 
0.8 mm. 

This remarkable Vertigo has been detected in Eastern Florida, on the 
coast at Mosquito Island, etc., by Mr. Oscar B. Webster and his father, 
Mr. Geo. W. Webster, of Lake Helen, Florida. These gentlemen took 
much pains to ascertain the range of distribution of this form and some V i 
others, and it is consequently only just to name the species in honor of 20 
Mr. Webster. The most striking character of it, besides the narrowed 
last whorl, is the thin and straight palatal wall and margin, so that, indeed, the 
shell appears to be immature. But when seen under a glass of sufficient power, 
the margin is completed, and, as already mentioned, there is a thin callus at a 
little distance from the margin. Moreover, Mr. Webster wrote me that, of more 
than 150 examples he had seen, all were alike. 

A few days ago, in a lot of P. corticaria, Say, from Ithaca, N. Y., sent from 
Texas, there was one example of this species, the shell dead, but in fair condition, 
a little larger and less fragile than the Florida examples, and with a well marked 
callus corresponding to a slight but distinct crest. The specimen may have been 
collected in New York, and from its appearance at least I would ascribe to it an 
origin north of Florida. Since the above was written, I have found a few exam- 
ples in drift from Guadalupe River, Texas, collected by Mr. J. A. Singley, sent by 
Mr. Wm. A. Marsh. 

By the kindess of Mr. Webster I was enabled to see a living example. The foot 
and the lower parts of the head are nearly colorless; head, eye-tentacles, and neck 
light gray. Jaw very tender, thin, pale yellow, consisting of about 14 longitudinal 
plates, shorter and wider in the middle, longer and narrower toward either end; it 
is much like that of V. tridentata, Wolf. Odontophore about 0.36 mm. long, 0.1 mm. 
wide, about 110 square rows in each ?-+ $-++ 7 teeth; central very small; laterals 
gradually passing into marginals; the latter serrate. Different from that of 
V. tridentata. 

In drift with numerous minute shells, from Guadalupe River, Texas, kindly sent 
by Wm. A. Marsh, I found one specimen of this species, which consequently is not 
confined to Eastern Florida, where it was detected by Messrs. Webster, but may be 
widely spread over the southern part of our country. 


Vertigo Oscariana, STERKI, Proc. Ac. Nat. Sci. Phila., 1890, p. 33; The Nautilus, 
1890, p. 136. 


The above is Sterki’s deseription, and the figure is drawn by him from the 
type. 


198 BULLETIN OF THE 


Vertigo Binneyana, STERKI. 


They are of the size and general appearance of V. callosa, very narrowly per- 
forate, cylindrical oblong, light chestnut-colored ; whorls 5, moderately rounded, 
nearly smooth; aperture relatively small, peristome little expanded; outer wall 
with a well formed crest interrupted by a rather long revolving groove ; 
corresponding to the crest there is a callus of lighter color; lamella 6; 
on the apertural wall a small supra-apertural and a well developed 
apertural; columellar appearing rather massive; at the base, one 
rather small but well formed, appearing tooth-like; palatals 2, long, 
especially the inferior. Length 2.0 mm., diameter 1.0 mm. 

Last year, Mr. W. G. Binney kindly presented me with two exam- 
V. Binney- ples of a Vertigo collected at Helena, Montana, by Mr. H. Hemphill, 

ri which seemed to be of a new species; but yet I did not like to publish 
I a description founded upon only these two specimens. Lately among 
a number of small Pupide from different parts of British America sent by Mr. Geo. 
W. Taylor of Ottawa, there were a few examples of this same species, from Win- 
nipeg, Manitoba, dead and weathered, but good enough to be identified. 

Probably there are other examples of this species in collections, and more will 
be found in the Northwest. It is named in honor of Mr. W. G. Binney, to whom I 
owe the two beautiful specimens in my collection. 


Vertigo Binneyana, STERKI, Proc. Ac. Nat. Sci. Phila., 1890, p. 83. 


The above is Sterki’s description. I am also indebted to him for the figure, 


Vertigo callosa, Srerkı. 


There are in collections two different species under the name of V. Gouldii, Binn. 
Their size and coloration is nearly the same, at least in most variations, as are also 
the apertural lamellse as to number and position. Yet they are decidedly and con- 
stantly distinct, especially by the formation of the outer wall at the aperture. 
Judging from the descriptions and more especially from the figures, the true V. 
Gouldii is the one characterized as follows: the last whorl is somewhat predomi- 
nating, thus rendering the whole shell more ovate or conic ovate; the palatal wall 
near the aperture is decidedly flattened, or impressed, the impression comprising 
also the crest and being especially well marked at the “auricle” (as I name the 
more or less projecting part about the middle of the outer margin, to have a con- 
cise expression), forming a roundish groove outside and a decidedly prejecting 
angle inside, thus producing the “two curves meeting in the centre of the peri- 
stome.” A feature not striking, but only seen by careful examination, is the posi- 
tion of the short tooth-like lamella at the base, somewhat nearer the margin than 
the end of the columella, the base perceptibly widened at that place; the said 
lamella is probably an equivalent of the inferior columellar lamella, which in most 
Vertigos stands very low, in many exactly at the base. 

The other species, V. callosa, has the last whorl relatively less wide, so that the 
whole shell is of a more oblong shape. In the palatal wall, only the part behind 


MUSEUM OF COMPARATIVE ZOÖLOGY. 199 


the crest is somewhat flattened, while the latter itself forms one unbroken curve 
from the base up to the suture, and at the moderately projecting auricle there is 
only a slight flattening. The inferior columellar lamella is at the end of the col- 
umella, sometimes wanting or a mere trace. Well worthy of notice is a pecuiiar 
formation of the surface, the epiconch showing microscopic wrinkles or foliations 
in the direction of the lines of growth producing a peculiar silky gloss, especially 
on quite fresh examples, and more in some forms than in others. 

The first two examples of this species I obtained in 1885 from Mr. Henry Moores, 
of Columbus, Ohio, and in 1889 I saw a few more in his collection. In 1887, Mr. 
E. W. Roper sent me some others from Massachusetts. Last year in different 
collections I saw quite a number of specimens from different places in New York 
near the metropolis, under various names: V. Gouldii, milium, ovata, and also mixed 
with Bollesiana. Of the Ohio examples the color is somewhat lighter, the callus and 
the lamella are strong and white, while in the Eastern examples they are somewhat 
thinner and more of the color of the shell. The name callosa was thus mainly 
derived from the Ohio form (which, however, may be regarded as a variety). 

It is with some hesitation, however, that I now bring it under this head ; it is the 
equivalent of the European V. pygmea, Drap., of which I have examples for com- 
parison from different countries of the Old Continent, which I have partly col- 
lected myself there during a number of years. The two may even be identical; 
at least it would be absolutely impossible to distinguish New York examples 
from most Europeans. Both forms agree also in certain variations of the aper- 
tural lamellæ; the inferior columellar lamella may be absent in either, or there 
may be present a small supra-palatal fold, thus rendering the number variable 
from 4 to 6, the typical, however, being 5. An examination of the soft parts will 
probably decide the question; so far I have not had an opportunity to make it. 

On our continent, the range of distribution of the two species —V. Gouldii and 
callosa — seems to be somewhat different, the former having been found in New 
York, Ohio, Illinois, and Colorado, the latter from Massachusetts to Ohio. 


Vertigo callosa, Sterkı, Proc. Ac. Nat. Sci. Phila., 1890, p. 81. 


The above is Sterki’s description. 


Vertigo parvula, STERKI. 


Among several hundred small Pupidæ collected in Northeastern Ohio (Summit 
and Lake Counties) by Mr. A. Pettingell, there were two examples of a doubtless 
new species, which Iin the same way named V. parvula. It is about of the size, 
shape, and appearance of V. (Angustula) milium, Gould; but ranges in quite another 
group, having a quite simple palatal wall and margin, and only three lamellæ. 

In Texas, Vertigos seem to be decidedly rare. In many hundreds of Pupide 
from that State which Mr. J. A. Singley and Mr. Wm. A. Marsh kindly forwarded 
me there were only about half a dozen such; a few milium, one rugosula, one 
Oscariana, as mentioned above, and one specimen of a form which probably will 
prove to be a new species of quite peculiar formation. 


Vertigo parvula, Srerkı, The Nautilus, 1890, p. 186. 


The above is Sterki’s description. 


200 BULLETIN OF THE 


Vertigo approximans, STERKI. 


In 1887, Mr. A. A. Hinkley, of Dubois, Ill., sent me, with other Pupide, one 
specimen of a Vertigo, probably new, and in 1889 another of the same. The said 
gentleman and Mr. William A. Marsh kindly forwarded me all their Pupide for 
examination, but so far I have found no other example, yet I am satisfied such will 
be found. The form is related to Vertigo ovata and Gouldii, but different, and is 
characterized by the two palatal lamella being close together, for which reason I 
gave it the manuscript name V. approximans. 


Vertigo approximans, SrerKt, The Nautilus, 1890, p. 186. 


‘Lhe above is Sterki’s description. 


Vertigo rugosula, STERKI. 


Related to V. ovata and Gouldii; in shape more elongated than the latter, more 
cylindrical, and somewhat larger. Apertural parts and lamelle much like those of 
ovata; but the columella is decidedly longer and straighter, and the inferior colu- 
mellar lamella is distinctly placed on it. Length 1.8-2.0, diameter 1.1 mm. Of a 
peculiar formation is the surface. Of the five well-rounded whorls, about one and 
a half of the upper are nearly smooth; the following, with exception of 
the last, are distinctively and regularly striated; the last is very finely 
but distinctly rugose in the sense of the lines of growth, near the aper- 
ture again striated. Color, dark chestnut. 

This is a beautiful species, of which I saw the first example in the 
collection of Mr. Bryant Walker, who had found it in April last at 
Pass Christian, Mississippi. Last September, Mr. W. G. Mazyck col- 
lected a number of them on Sullivan’s Island, S. C. In either place 
they were in company of Pupa rupicola, Say. Quite lately I have seen 
one example from Lee County, Texas, sent by Mr. J. A. Singley; it was a dead 
shell, and not fully mature, but recognizable. The species consequently seems to 
be widely distributed along the South Atlantic and Gulf coasts. Two specimens 
were sent in by Mr. H. Hemphill, who collected them at Fish Camp, Fresno Co,, Cal. 

In Eastern Florida, Volusia County, etc., a form has been found to be quite com- 
mon which I refer to this species, but as a distinct variety which may be called 
ovulum. It is somewhat smaller, ovate; the striation and rugosity of the surface 
are less marked, and the inferior apertural lamella is wanting. In turn it has in 
most examples a lamella at the base (between inferior columellar and inferior 
palatal), and the callus in the palatal wall is rather strong. The coloration of part 
of them is somewhat lighter. It cannot be confounded with V. ovata, Say, its rela- 
tions to the type of rugosula being evident, and, in addition, ovata has been found 
with it. Nor can it be referred to ventricosa. It is larger and stronger, of much 
darker color, its surface is not so smooth and polished, it has three or even four 
Jamelle more, and the columella is longer. 


Vertigo rugulosa, Sterri, Proc. Acad. N. Sci. Phila., 1890, p. 34. 


V., rugosula. 


The above is Sterki’s description. The figure was drawn by him. 


MUSEUM OF COMPARATIVE ZOOLOGY. 201 


Liguus fasciatus, Motz. 


Plate I. Fig. 5, 


The Vaccas Key variety, noticed in page 435 of the Manual of American 
Land Shells, is figured in the plate. 


Orthalicus undatus, Brue. 
Plate II, Fig, 4, 


I give a new figure of the variety of this species. 


Holospira Arizonensis, STEARNS. 


Shell dextral, elongately cylindrical, pupiform, dingy white to pale horn-color, 
translucent. Number of whorls, twelve to thirteen. Slightly convex, the su- 
tures distinctly defined. The upper six or seven 
whorls rather abruptly tapering towards the obtuse 
apex, which has a slightly twisted and rather a 
papillose aspect. The last whorl'is curved under 
and constricted back of the mouth, forming an 
umbilical notch. The apex and following whorl 
are smooth; the three or four succeeding whorls 
sharply and somewhat obliquely plicated longitu- 
tudinally, the median and following whorls be- 
coming somewhat obscurely sculptured other than 
by distinct growth lines. The basal whorl is 
strongly sculptured below, and back of the mouth, 
and obtusely angulated underneath. Aperture 
ovate, slightly angulated anteriorly, somewhat 
effuse, rimmed and projecting. The dimensions of two examples are as 
follows : — 


mm. 
LOOTO Ste et tines Ve a een 124 
ROU SMUG ie ann ee ae we 
yrei helera are ae ee 
rea a her a aa 


Dos Cabezas, Arizona, where the above two specimens and numerous fragments 
were found in a cave in November, 1889, by V. Bailey, and contributed to the 
United States National Museum (No. 104,392) by Dr. C. Hart Merriam. 

Among the species of this group that are geographically related is 77. Remondi, 
Gabb, described from Arivechi, Province of Sonora, Mexico, a form sharply sculp- 
tured throughout, and in minor ‘features also different; H. Pfeifferi, Menke, col- 
lected by Remond at Hermosillo, in the same province, with the previouslv named 


202 BULLETIN OF THE 


species; and H. (Celocentrum) irregulare of Gabb, from the high table-lands back of 
Mulege, in the peninsula of Lower California. All of these are separable at a 
glance from Arizonensis. 


The above is Stearns’s description and figure from Proc. U. S. National Mus., 
Vol. XIII. p. 208, Plate XV. Figs, 2, 3, 1890. 


Onchidella borealis, Daw. 


Coos Bay, Oregon. 

It is gregarious in its habits. Fifty specimens were taken in a small crevice of 
clay shale, near high tide. Single individuals, or several clustering together, were 
taken afterwards lower down on the tide under loose stones. When in motion, the 
animal moves off quite rapidly for so small a creature, with two short, stout pedun- 
cles protruding in front of the mantle, bearing keen, sharp black eyes. The color is 
dark slate, splashed with blotches and streaks of ashen white. The body when in 
motion is 4 inch long, 33; wide, } high, and oblong-oval in form, a little broader 
behind than before. It is covered with small tubercles, which are larger around 
the edge of the mantle than those higher up on the body, giving the edge of the 
mantle a serrated or tooth-like appearance when the animal is at rest. When 
it is at rest on a smooth surface, the base of the animal is nearly circular, 
or a little longer than wide, the centre of the body is elevated to quite a sharp 
apex, which together with its color resembles some varieties of a very young 
Acmea pelta, and would be very readily taken for such by an inexperienced col- 
lector. The foot is white, and works in rapid undulations when the animal is in 
motion. 


The above remarks are made by Mr. Hemphill in a recent letter. 


MUSEUM OF COMPARATIVE ZOOLOGY. 203 


EXPLANATION OF PLATES. 


PLATE 1. 


Fig. 1. Anadenus Cockerelli. Animal and internal shell. 
Fig. 2. Patula strigosa, var. Buttoni. 
Fig. 3. Arionta ruficincta. 
Fig. 4. Glandina decussata, var. Singleyana. 
Fig. 5. Liguus fasciatus, var. from Key Vaccas. 
Fig. 6. Bulimulus Dormani. 
Fig. 7. Arionta Ayersiana. 
Fig. 8. Zonites Simpsoni, enlarged. 
Fig. 9. Binneya notabilis, enlarged. 
0. 


= 
ag 
En 


Same as Figure 2, toothed variety. 


PLATE II. 


1. Pupa calamitosa, reduced from original figure. 
Fig. 2. Pupa hordeacella, from original figure. 
3. Selenites Duranti, var. Catalinensis, enlarged. 
4. Orthalicus undatus, variety. 
Fig. 5. Selenites Vancouverensis, var. Keepi, enlarged. 
6 
7 


. Triodopsis Mullani, var. Blandi. 
‚8. Arionta tudiculata, var. cypreophila. 
Fig. 9. Bulimylus Ragsdalei, enlarged one half. 


PLATE I. 


Fig. 1. Limax Hemphilli, var. pictus. Animal and internal shell. 
Fig. 2. Zonites Diegoensis, enlarged. 

Fig. 3. Zonites macilentus, enlarged. 

Fig. 4. Tebennophorus Hemphilli, jaw. 

Fig. 5. Anadenus Cockerelli, jaw and tongue. 

Fig. 6. Pristiloma Lansingi, enlarged. 

Fig. 7. Zonites Caroliniensis, enlarged. 

Fig. 8. Helicodiscus fimbriatus, var. salmonaceus, enlarged. 

Fig. 9. Zonites sculptilis, enlarged. 


BULLETIN OF THE MUSEUM OF COMPARATIVE ZOOLOGY. 


PLATE IV. 


Arionta Kelletti, var. multilineata. 
Arionta Kelletti, var, nitida. 
Arionta Kelletti, var. albida. 
Arionta Kelletti, var. castanea. 
Euparypha Tryoni, var. nebulosa. 
Euparypha Tryoni, var. fasciata. 
Patula strigosa, var. bicolor. 
Patula strigosa, var. lactea. 
Patula strigosa, var. albofasciata. 


BINNEY: 4TH SUPPL, TO TERR. MOLL. PLATE b 


BINNEY: 4TH SUPPL. TO TERR. MOLL. PLATE 1I 


BINNEY: 4TH SUPPL. TO TERR. MOLL. PLATE (it. 


| N 


Pi © 


| 
| AN 
SH 
pia 
y wdd 
pi \ BE 
4 ) 70) ri 
= Aa! 
4 ( KAL veel 
A 


RT 


ET 


BINNEY: 4TH SUPPL. TO TERR. MOLL. PLATE IV.