VOLS9 


[PES 


rg Z 4 | 
Pibrary of the THuseum 


OF 


COMPARATIVE ZOOLOGY, | 
AT HARVARD COLLEGE, CAMBRIDGE, MASS. 


Founded by private sudscription, tn 1861. 


Deposited by ALEX. AGASSIZ. 


No. 


FF? 
fin 


QUARTERLY JOURNAL 


OF 


MICROSCOPICAL SULENCE. 


EDITED BY 


E. RAY LANKESTER, M.A., LL.D., F.R.S., 


Linacre Professor of Comparative Anatomy, Fellow of Merton College, and Honorary 
Fellow of Exeter College, Oxford ; Corresponding Member of the Imperial 
Academy of Sciences of St. Petersburg, and of the Academy of 
Sciences of Philadelphia; Foreign Member of the 
Royal Bohemian Society of Sciences. 


WITH THE CO-OPERATION OF 


ADAM SEDGWICK, M.A., F.R.S., 
Fellow and Lecturer of Trinity College, Cambridge ; 


AND 


W. F. R. WELDON, M.A., F.B.S., 
Jodrell Professor of Zoology and Comparative Anatomy in University College, London; 
late Fellow of St. John’s College, Cambridge. 


VOLUME 39.—NeEw Series. 
With Xithographic Plates and Engrabings on Wood. 


LONDON: 
J. & A. CHURCHILL, 7, GREAT MARLBOROUGH STREET. 
~ 1897, 


CONTENTS. 


CONTENTS OF No. 153, N.S., MAY, 1896. 


MEMOIRS : PAGE 
The Blood of Magelona. By W. Buaxtanp Benuaw, D.Se.Lond., 
Hon.M.A.Oxon., Aldrichian Demonstrator in Comparative Ana- 

tomy in the University of Oxford. (With Plate 1) : : i 


Fission in Nemertines. By W. Buaxtanp Brnuam, D.Sc.Lond., 
Hon.M.A.Oxon., Aldrichian Demonstrator in Comparative Ana- 
tomy in the University of Oxford. (With Plates 2 and 3) eae) 


Studies on the Nervous System of Crustacea. By Epear J. ALLEN, 
B.Se., Director of the Plymouth Laboratory of the Marine Biolo- 
gical Association. (With Plate 4) : : : cae 


Notes on Oligochetes, with the Description of a New Species. 
By Evwin 8. Goopricu, B.A., Assistant to the Linacre Pro- 
fessor of Comparative Anatomy, Oxford. (With Plates5and6) 51 
On the Development of Lichenopora verrucaria, Fabr. By 
Sipney F, Harmer, M.A., B.S8c., Fellow of King’s College, 
Cambridge; Superintendent of the University Museum of 
Zoology. (With Plates 7—10, and two figures in text) nee 


CONTENTS OF No. 154, N.S., AUGUST, 1896. 


MEMOIRS: 
Letters from New Guinea on Nautilus and some other Organisms. 
By Artuur Witty, D.Sc. : : : . 145 


The Brain of a Foetal Ornithorhynchus. Part I.—The Fore-brain. 
By G, Exxior Smitu, M.D., Ch.M., Demonstrator of Anatomy, 
University of Sydney, N.S.W. (With Plate 11) 3 5 eH 


On Arhynchus hemignathi, a New Genus of Acanthocephala. 
By Artuur HE. Surerezy, Fellow and Tutor of Christ’s College, 
Cambridge, and University Lecturer in the Advanced Morpho- 
logy of the Invertebrata. (With Plate 12) 5 : . 207 


iv CONTENTS. 


PAGE 

Zoological Observations in the South Pacific. By ARTHUR WILLEY, 
D.Se.Lond., Balfour Student of the ome of Cambridge. 
(With Plate 13) ‘ - é : . 219 


Chlamydomyxa montana, n. sp., one of the Protozoa Gymno- 
myxa. By E. Ray Lanxester, M.A., LL.D., F.R.S., Linacre 
Professor in the University of Oxford. (With Plates 14 and 15) 233 


CONTENTS OF No. 155, N.S., NOVEMBER, 1896. 


MEMOIRS : 

The Constitution and Development of the Society of Termites: Ob- 
servations on their Habits; with Appendices on the Parasitic 
Protozoa of Termitide, and on the Embiide. By Professor B. 
GrassI in collaboration with Dr. A. Sanpias. oe Plates 16 
—20) . : ; : : . 245 


On Ctenoplana. By ArtHuR Witter, D.Sc. (With Plate 21) . 323 


An Attempt to deduce the Vertebrate Eyes from the Skin. By 
H. M. Bernarp, M.A.Cantab., F.L.S. (With Plate 22) . 343 


The Reproduction and Metamorphosis of the Common Eel (An- 
guilla vulgaris). By Professor Grassi ‘ : s ovk 


CONTENTS OF No. 156, N.S., JANUARY, 1897. 


MEMOIR: 
Changes in the Cell-organs of Drosera rotundifolia, pro- 
duced by Feeding with Egg-albumen. By Lity Huts, Physio- 
logical Laboratory, Oxford. (With Plates 23 and 24) . . 387 


Observations upon the Development and Succession of the Teeth in 
Perameles ; together with a Contribution to the Discussion of 
the Homologies of the Teeth in Marsupial Animals. By J. T. 
Witson, M.B., Professor of Anatomy, and J. P. Hirt, F.LS., 
Demonstrator of Biology in the University of ces New 
South Wales. With Plates 25—32) : F . 427 


TITLE anpD INDEX. 


JUN Le 1896 


The Blood of Magelona. 


By 


W. Blaxland Benham, D.Sc.Lond., Hon. M.A.Oxon., 
Aldrichian Demonstrator in Comparative Anatomy in the University of Oxford. 


With Plate 1. 


THe small Annelid first discovered and briefly described by 
Fr. Miller! in 1858 has more recently received more elaborate 
treatment at the hands of Professor W. M‘Intosh.? It presents 
several peculiarities deserving of still more detailed research ; 
and, at Professor Lankester’s request, Dr. M‘Intosh has kindly 
sent to me from time to time living specimens, obtained at St. 
Andrews, which have served as the starting-point of my own 
observations. 

Amongst these peculiarities, the most striking and astonishing 
is afforded by the contents of the blood-vessels ; and I propose 
at present to limit my remarks to an enumeration and descrip- 
tion of the observations and experiments which I have made 
on this fluid. 

In order to obtain a greater quantity of quite fresh material 
I spent a week at the Marine Laboratory at St. Andrews 
during the summer of 1894. Professor M‘Intosh was good 
enough to send me a list of low tides—at which time only can 
Magelona be obtained in any considerable number. My 
very hearty thanks are due to Dr. M‘Intosh for his permission 
to make use of the laboratory and appliances for the purpose 

1 Fr. Miller, “ Hiniges iiber d. Annelidenfauna d. Insel 8. Catharina,” &c., 
‘ Arch. f. Naturgesch.,’ 1858, p. 211. 

2 W. C. M‘Intosh, “ Beitrage zur Anat. von Magelona,” ‘Zeit f. wiss. 
Zool.,’ 31, 1878. 


VoL. 39, PART 1.—NEW SER. A 


2 W. BLAXLAND BENHAM. 


of my study, and for his advice and kinduess in aid of my work 
there. 

Magelona lives in the sand at and below ordinary low-water 
mark, and the lower the tide the more abundant are the worms. 
My sojourn at St. Andrews was timed so as to cover par- 
ticularly low tides occurring between August 30th and Septem- 
ber 4th. I found the worms limited to a comparatively small 
area on the sands near the harbour, close to the Laboratory, 
though they occur more sparingly over a larger area. The 
work involved in their capture is rather heavy, as they occur 
at a depth from eight to sixteen inches below the surface of the 
sand. I was fortunately able to obtain the help of the laboratory 
attendant, A. W. Brown, who accompanied me in my search. 
The process was as follows:—Having selected a spot which he 
believed from experience to be likely, Brown dug as deeply into 
the sand as possible. As he raised the spade-load the sand na- 
turally broke across, and if Magelona was present the worms 
would be seen stretching across the gaps thus formed in the 
spadeful of sand. The worms are rather brittle, and consider- 
able care had to be exercised in raising the mass of sand, and 
further separating it, so as to liberate the worms. Sometimes 
such a spadeful would produce only one or two worms, at other 
times considerable numbers might be obtained; but the work 
of sorting through such masses of wet sand was no light task. 
We usually spent a couple of hours in this manner till the tide 
drove us back from the Magelona area. 

The body of Magelona papillicornis presents two 
regions,—a short “ thorax,” with thick, muscular wall; and an 
“abdomen,” through whose thin walls the gut and its sandy 
contents can beseen. The general appearance then of the worm 
is sand-colour, for there is no pigment in the skin or body 
wall, but the very short thorax is of a faint madder- pink tint, 
or in some cases dull white, according, as later observations of 
captive specimens showed, to the condition of oxidation of the 
blood. When the worms, separated more or less from the 
sand, are placed in clean aérated sea water the colour of the 
thorax soon becomes a deeper madder-rose colour; as the water 


THE BLOOD OF MAGELONA. 3 


becomes foul the tint of the worm becomes fainter and fainter, 
till the thorax is white. I first observed this change while 
working upon worms sent from St. Andrews to me at Oxford. 
The worms had been put into clean sea water on their arrival, 
and had been put aside overnight. Next morning they were 
no longer madder-rose, but white. On pointing this out to 
Professor Lankester, he at once suggested aérating the water 
by squirting air into it; this I did, and soon the madder-rose 
tint reappeared. It was this small but extremely interesting 
change in colour which called our attention to the blood, for it 
was evident that we had to deal with a respiratory pigment of 
a kind unusual in the Chetopoda. 

The fact that the tint of the worm is due to contained blood, 
and not to any pigment in the skin, is readily recognised by 
the unaided eye when the alternating processes of eversion 
and retraction of the ‘ proboscis” or “introvert” is watched. 
So long as the introvert is at rest within the body the thorax 
is coloured; when eversion takes place the tint becomes quite 
faint (Pl. 1, fig. 2),—in fact, frequently the thorax becomes 
white. The blood in the thorax is contained in greatly dilated 
vessels, which block up and obliterate nearly the whole of the 
ceelom (fig. 3); the introvert is a hollow sac traversed by thin 
bundles of retractor muscles, the cavity of the sac being con- 
tinuous with the dilated vessels, so that on eversion nearly the 
whole of the blood in the thorax is driven into the introvert— 
this flow of blood is, of course, the cause of the eversion,—and 
the thorax is more or less completely deprived of its colour. 
The fact that the abdomen is not tinted by the blood to any great 
degree is due to the small size of the blood-vessels in this region. 

The arrangement of these blood-vessels has been minutely 
described by M‘Intosh, and need not detain us at present. 
But the remarks of this author, so far as they relate to the 
blood itself, may here be quoted :—‘‘ The blood is a coagulable 
pale rose-red fluid, containing numerous corpuscles. On 
being shed these group themselves in different clumps. The 
size of the spherules is nearly constant, although variations 
may occur. They exhibit molecular movement, and their 


4, W. BLAXLAND BENHAM. 


contour is changed by pressure. Many are ovoid, others 
circular or irregular. If studied within the blood-vessels the 
corpuscles exhibit a refringent body or nucleus-like central 
structure. In addition to these spherules various other elements 
exist, which may perhaps be explained as a development of 
spherules inside cells. Here and there a coagulum may be 
seen between the corpuscles. Treated with strong acetic acid, 
very noticeable changes occur. The whole field is now covered 
with a granular débris (Pl. 38, fig. 9), and in place of the 
corpuscles we see masses of granular cells of larger size. The 
acid probably dissolves the envelope of the corpuscles or 
alters their walls, so that the (probably fatty) contents run 
together to form larger masses.” 

It will be seen later that this description, though true to a 
certain extent, does not exhaust the subject. I may mention 
that nearly the whole of my own observations had been made 
before I had acquainted myself with the above account of the 
blood given by M‘Intosh,—that is, in reading his paper some 
time back I had not paid special attention to his statements 
as to the blood itself. 

1. Spectroscopic Examination.—A fter observing the re- 
markable change of colour on oxidation I wished to examine 
the blood spectroscopically, and to this end I sought for some 
reagent which would extract the pigment; but though I was 
unsuccessful in my attempts, yet these experiments are, I 
believe, worthy of record, as the action upon the blood of 
some of the reagents is sufficiently peculiar. These will be 
described below. 

Having been unsuccessful in extracting the colouring matter, 
I examined the blood itself by means of Zeiss’s spectroscope. 
For this purpose I compressed a worm with its introvert everted, 
so as to give a thin film of pink fluid, and examined it through 
the micro-spectroscope; but no absorption bands were visible. 
I then had short pieces of glass tubing cut and fixed to glass 
slips, so that I could fill them to a greater or less extent, and 
get a column of fluid of different depths. The process of 
obtaining sufficient blood to fill even a tube one eighth of an 


THE BLOOD OF MAGELONA. 5 


inch deep was tedious. ‘The method I adopted was as follows :— 
A number of living worms were taken, one by one, laid on 
blotting-paper, so as to dry them as much as possible; then, 
while holding them just above the stout piece of tube, I 
snipped the everted introvert or the thorax itself; the worms 
were then left head downwards over the tubing, and the blood 
gradually flowed out of the cut end down the side of the tube 
and collected at the bottom. A considerable number of worms 
had to be thus treated, and as the blood tends to coagulate the 
process occupied considerable time ; nevertheless I was able to 
collect what appeared to be a sufficient quantity of blood for 
spectroscopic examination. The result, however, was nil. 
I could obtain no absorption bands, though I used both bright 
sunlight and gaslight, and took all precautions to keep the 
light out of the apparatus except that which passed through 
the blood. I also employed solutions of blood in salt solu- 
tions and in water. These solutions were colourless, but I 
thought it possible that even then the colouring matter might 
give some result; but I was equally unsuccessful. 

I believe, then, that we must conclude that the colouring 
matter of the blood of Magelona causes no absorption bands 
when a beam of light passes through it, and is spectroscopically 
analysed. 

2. The Histology of the Blood.—The fact that all the 
reagents which I employed had greater or less action on the 
blood rendered difficult the ascertainment of the real structural 
condition of the fluid. The only way in which the true con- 
ditions could be studied was by compressing the worm and 
examining the fluid as it flows along the vessels ; for even the 
mere shedding of the blood, without the addition of reagents, 
might be supposed to produce certain alterations in its con- 
dition, though from what follows I do not thiuk this to be the 
case. The blood thus examined in situ is seen to consist of 
very small madder-rose coloured globules, varying in size within 
certain small limits. These globules choke the vessels, and 
there appears to be very little plasma. The appearance pre- 
sented in sections of worms fixed in Hermann’s solution is 


6 W. RLAXLAND BENHAM. 


very similar to that exhibited by the living subject. Further, 
these corpuscles or globules do not flow freely in the vessels as 
the walls contract, but adhere together in clumps, and some-- 
times a mass or clump will be separated by some considerable 
space from neighbouring masses, appearing to indicate a colour- 
less plasma. This massing of the globules also occurs in the 
freshly shed blood, as will be seen hereafter, but its occurrence in 
the blood-vessels is extremely peculiar, and M‘Intosh expresses 
the phenomenon by speaking of the blood as “ coagulable ”— 
a term which perhaps is scarcely true in its original sense, — 
for there is no evidence of the separation of “fibrin” or 
anything of that nature; the globules or corpuscles rather 
adhere together. 

The general shape of the corpuscles is spherical ; in size they 
average 0°002 mm. in diameter in the living condition; they 
are not amceboid; they are homogeneous, and rather oily in 
appearance. They are not nucleated, as the addition of re- 
agents will show; and I did not observe the refringent body 
within them mentioned by M‘Intosh. Nuclei, however, do exist 
in the blood, though usually in an isolated condition, and are not 
recognisable with certainty in the living worm ; but in stained 
sections, as well as in preparations treated with osmic acid, fol- 
lowed by picro-carmine, such nuclei are more or less abundant 
(fig. 4), and though usually quite deprived of any protoplasmic 
envelope, yet in some cases I noted a granular mass surrounding 
the nucleus, without any boundary or recognisable limit, and I 
believe this to be the fluid plasma of the blood, coagulated by the 
reagents. In a preparation of the blood mounted in glycerine, 
in which the globules have undergone a considerable amount of 
fusion, so that the globules are of very varied sizes, I note several 
instances of what appear to be nucleated corpuscles. In these 
cases the nucleus, of the usual size, is deeply stained, and lies at 
one side of an unstained “ globule” (figs. 8, 9); sometimes, in 
fact, the outline of the nucleus seems to project slightly beyond 
the general outline of the globule, but this may be due to the 
greater refractive index of the former. These globules, of which 
I give some figures (fig. 6), vary in size, but are much larger 


THE BLOOD OF MAGELONA. a 


than the unaltered corpuscle as seen in absolutely fresh blood, 
and I suggest that two or more normal globules have fused 
together and involved the nucleus. The isolated nuclei may 
occur singly or in groups of three or four (fig. 7). 

The size of the nucleus, as seen in sections, is rather greater 
than that of the globules themselves (Pl. 1, fig. 4). The chro- 
matin granules vary considerably in number and arrangement ; 
‘sometimes there is but one such granule, at other times several. 
In the glycerine preparation the former condition is more 
frequent, and the chromosome is then rod-like (fig. 7, a, 0, c). 

A. The readiest way to obtain the blood as free as possible 
from other matters is to dry a worm by placing it on blotting- 
paper, place it ona dry slide, and pierce with a sharp needle the 
proboscis, which is pretty sure to be everted. The blood now 
oozes out slowly, but does not flow freely, the globules adhere 
together in masses, and the masses are connected by narrower 
bridges of corpuscles, giving rise to a coarse network. This 
mass changes its shape and arrangement, as neighbouring 
nodes or clumps run together by the shortening of the bridges ; 
further, very delicate threads (really due to strings of globules) 
are seen (¢.) traversing the meshes, and owing to the continued 
stretching of the narrowing bridges isolated clumps of cor- 
puscles may come to occupy the meshes as these threads 
snap asunder. This change of arrangement may be due to 
evaporation of the plasm and to surface tension. The appear- 
ance presented by such a drop of blood shed in this way, 
and examined without a cover-slip, is shown in fig. 10. 
The whole mass is pink in colour, highly refringent and oily- 
looking; no outlines of the separate corpuscles are visible 
in the larger masses, though in the smaller groups they may 
be recognised. 

When such a drop of blood is covered (fig. 11) the outlines of 
the corpuscles become evident; and though the majority are 
similar in size, here and there larger ones are met with, and it 
is quite possible that the latter are nuclei. The change in 
arrangement of the bridges and meshes is seen to continue. The 
appearance presented by this mass of corpuscles is strikingly 


8 W. BLAXLAND BENHAM. 


similar to the “soapy foams” manufactured by Butschli’s 
method; the “alveolar layer” is formed by regular arrange- 
ment of the corpuscles along the margin of the mass, and the 
slow movement recalls that described by him. 

B. The blood when freshly drawn by rupture of the “ intro- 
vert ” in sea water (after removal of as much of the sea water 
as possible from the slide), and at once covered, presented the 
following appearance. Instead of a mass of similar small glo- 
bules, these were seen to be of various sizes,—most of them 
much larger than the corpuscles seen within the vessels, being 
as much as 0:04 mm. in diameter (fig. 15). 

From later experiments, carried out with greater care so 
that no water should be present, I am led to believe that the 
result here obtained was due to the presence of a certain 
amount of sea water on the slide; this causes the globules to 
fuse with one another, hence the varied sizes of the globules. 


3. Appearances presented by the Blood on the 
Addition of Various Reagents. 


1. Normal salt solution, passed under the cover-slip of 
such a preparation as described above (A.), causes the globules 
to separate from the masses if there has been no great amount 
of pressure. If the latter has occurred, only the corpuscles 
lying at the edge of the mass separate. At first the colour is 
unchanged, but soon it disappears. 

This disappearance of the colour was much more evident 
when the tubes used for spectroscopic analysis containing a 
little salt solution, the worms cut as above described, and 
the blood allowed to flow into the salt solution. I hoped to 
get a pink solution, but the colour very soon changed and 
became colourless. This may be due to deoxidation rather 
than to any destructive power of the salt; but whatever the 
cause, the result was a disappearance of colour. 

2. Distilled Water.—When water is added to blood 
already treated with salt solution, the following extraordinary 
series of phenomena ensues:—As the water displaces the 
salt solution the outlines of the globules disappear: this 


THE BLOOD OF MAGELONA. 9 


takes place naturally at the edge of the masses first, and then 
the mass presents the appearance of a colourless, oily-looking 
drop, in which more or less numerous rounded granules are 
embedded (fig. 12). These granules are the globules, and the 
longer the water acts the fewer distinct globules remain, till 
finally the mass of globules is replaced by a homogeneous, oily- 
looking drop (fig. 13). Meanwhile the movement already de- 
scribed in the fresh blood is going on, and is aided by the current 
of water as it is drawn under the cover-slip ; the narrow bridges 
uniting the masses become drawn out into longer and thinner 
threads—the globules of which they were composed have fused, 
and the threads, like the masses at the nodes of the network, 
become homogeneous. These threads may thin out till they 
break, the two ends are drawn into the masses at the nodes, 
and the latter become rounded off. In such a way the network 
may be resolved into a number of larger and smaller droplets, 
clear, colourless, and without any indication of the globules of 
which they were originally composed. 

If a group of globules or corpuscles not forming part of a 
network be examined, it can be readily seen that the neighbour- 
ing globules do fuse with one another, forming larger 
globules: these larger ones fuse with others, and thus, as a 
greater and greater number of globules become involved, a 
great oily-looking droplet is formed, in which all trace of the 
original globules has disappeared. In this reaction there is no 
swelling up of the original globule—neighbouring globules 
fuse with one another (fig. 14). 

The addition of normal salt solution to such a mass, so as 
to replace the water, did not lead to the reconstitution of 
separate globules, so that there can be no mistake about the 
actual fusion. 

Such a droplet is shown in fig. 16; after salt solution has 
been added, it lost its circular shape, and became irregular 
there being a marked movement—almost amoeboid in general 
appearance; and small irregular outgrowths comparable to 
pseudopodia became formed, and some of these separated from 
the mass as pear-shaped, and later circular droplets (z.). It 


10 W. BLAXLAND BENHAM. 


may be suggested that these are the original globules, but such 
is not the case ; they are very much smaller, and vary more in 
size. Further, only a few such separations occurred, even 
after prolonged action of the salt solution. 

When water is added to freshly drawn blood the same 
series of reactions occur (fig. 14). The colour of the corpuscles 
is quickly changed by the water, the masses becoming colour- 
less. I also “ bled” worms into a few drops of water in a 
tube, in the hope of obtaining a solution of colouring matter. 

The reaction of the corpuscles or globules to water indicates 
that these coloured globules are not provided with any mem- 
brane or envelope. They appear to be droplets of coloured 
material, and their general appearance suggested some oily 
substance. 

Chloroform.—Employing freshly shed blood on a dry 
slide, and running in chloroform as soon as possible, I observed 
the gradual disappearance of the corpuscles. Each corpuscle 
as the chloroform reached it first lost its colour, then rapidly 
became smaller and smaller, till nothing was left but a few 
very minute granules. Amongst these a number of small 
ovoid, highly refringent bodies, being a faint olive-green in 
tint, make their appearance. 

Here, too, I have no doubt as to the fact that the corpuscles 
are dissolved : there is no mere “ diminution in size,’ such as 
is observable in the case of mammalian corpuscles treated 
with chloroform ; there is an absolute disappearance of the 
globules of Magelona, the only trace of them that remains 
being a few granules. 

Ether produced effects similar to those produced by chloro- 
form. Both these reactions seem to point to fat or oil of some 
kind. 

Absolute alcohol gave the same reaction. I added nitric 
acid (10 per cent.) to the small olive-green refringent bodies, 
and found that they are insoluble in the acid. But both these 
and the smaller granules were dissolved in strong nitric acid. 

Alcohol (70 per cent.) had similar action. 

In all the above cases the colour of the corpuscle disappeared. 


THE BLOOD OF MAGELONA. 11 


Osmic acid coagulates the substance of the corpuscles, but 
does not otherwise change them. They only become distinctly 
brown after some minutes, and do not give the dark brown or 
black colour characteristic of fat. 

Nitric acid (25 per cent.) and Hydrochloric acid 
(20 per cent.) cause the corpuscles to become granular, 
though they do not actually disappear. The colour is changed. 

Potash (80 per cent.) has a similar action. 

Glacial acetic acid dissolves the globules, leaving minute 
granules, though the action is very slow. 

Further, picro-carmine stains the globules, as also does 
eosin: they retain their homogeneous appearance, however. 

These tests are unfortunately inadequate for the identifi- 
cation of the substance of which these globules are composed ; 
but from the ether, chloroform, and alcohol reactions, I think 
we may conclude that it is one of the “ fats ” which are soluble 
in these reagents, and moreover, give rise to crystalline bodies 
on evaporation. And amongst these fatty bodies, lecithin 
appears to give a similar reaction with water, in that it swells 
up in a peculiar way, forming droplets and threads, as I have 
described above. 


4. General Remarks. 


From these observations it will be seen that the blood of 
Magelona is totally different in structure from that of any 
other Chetopod, in that it consists mainly of very small 
madder-rose coloured, non-nucleated globules, embedded (rather 
than floating) in a very small amount of colourless plasma: 
amongst the corpuscles occur isolated nuclei. It was originally 
demonstrated by Professor Lankester that nuclei occur in the 
red fluid of the common earthworm, and this observation has 
been extended to sundry other Annelids by various observers. 
In these cases, as in Magelona, the nucleus is surrounded by 
very little, if any, protoplasm, and floats freely in the perfectly 
liquid plasma, which is coloured red by hemoglobin, or ina few 
cases green by chlorocruorin or chlorochromin ; while in some 
Oligochzetes the plasma is colourless. No other structural 


12 W. BLAXLAND BENHAM. 


element has been described in the fluid contained in the blood- 
vessels of any Annelid. But a few worms are known in which 
the blood-vessels are absent, whilst the fluid in the celom is 
coloured ; and in these cases (Glycera, Capitellide, Poly- 
cirrus) the hemoglobin exists within the corpuscles, the plasma 
of the coelomic fluid being colourless. The coloured corpuscles 
of this fluid in Glycera and Capitellide are, however, 
entirely unlike the coloured globules in the blood of Magelona. 

In the case of Capitella I took the opportunity of examin- 
ing the corpuscles during my stay at St. Andrews, as I was 
able to get abundant material. My observations are not new; 
Claparéde and Eisig have already sufficiently demonstrated 
the character and constitution of these corpuscles. Lankester! 
was the first to show by spectroscopic examination that they 
contain hemoglobin. The corpuscles when drawn fresh from 
the worm are circular, somewhat flattened discs, yellowish red in 
colour. A distinct membrane is visible, and granules of various 
small sizes are present. Stained with picro-carmine, a small 
round nucleus is seen in each corpuscle. When fresh cor- 
puscles are treated with salt solution no effect is noticeable. 
On the addition of distilled water the hemoglobin is dissolved 
out and the nucleus is rendered distinct, and the corpuscle 
itself becomes slightly smaller. This is no doubt due to os- 
mosis, the water passing into the corpuscle and causing it to 
become spherical: it thus loses in diameter in one direction as 
it gains in the other,—a process exactly like that occurring 
when Vertebrate corpuscles are similarly treated. 

Prolonged action of water causes the outlime of the cor- 
puscles to become less and less distinct, till it is almost 
impossible to recognise it, the nucleus and refringent bodies 
alone remaining. This reaction is quite different from what 
we have seen to occur in the caseof Mageiona, but resembles 
the corresponding reaction with human blood. 

Chloroform dissolves the hemoglobin, the corpuscles become 
granular, and the outline becomes gradually less distinct, but 
the membrane remains for a long time. But there is no 


1 * Proc. Roy, Soc.,’ xxi, 1873, p. 70. 


THE BLOOD OF MAGELONA. 13 


marked change in size; there is nothing approaching the 
gradual though rapid disappearance of the globules seen in 
the case of Magelona. 

Ultimately, though after a considerable time, the corpuscles 
appear to become dissolved, breaking up into granules. 

In Glycera, too, the corpuscles, as Claparéde, Lankester, 
and others have shown, are nucleated—the nucleus being oval 
and the corpuscle larger than that of Capitella. I was unable 
to obtain Glycera in any quantity, —in fact, I only came across 
two small specimens while collecting Magelona, and was unable 
to make many observations on them. But previous authors 
have compared these corpuscles with those of Vertebrates. 

The so-called “corpuscles” or coloured globules of Ma- 
gelona thus differ from the coloured corpuscles observed in 
other Annelids, not only in position, viz. within blood-vessels 
instead of in the ccelom, but also in structure and in their 
behaviour to chemicals, and I believe that they are different 
in constitution. 

These coloured globules in Magelona are to be compared 
with the coloured plasma of the ordinary Cheetopod blood, 
rather than with the coloured corpuscles of Capitella,Glycera, 
and Polycirrus. These globules, in fact, though recognisable 
as separate elements, do adhere together in the blood-vessels to 
form a fluid of the consistency of thick oil. If we suppose 
that a little more water were present in their composition we 
may imagine the globules to fuse with one another, and so form 
a more freely flowing plasma—such as is obtainable by ex- 
periment. This plasma would then be quite comparable with 
that of the blood of ordinary Annelids. But if we imagine the 
individual globules to have a firmer envelope—more resistent 
to reagents than it is in fact—we should have more perfectly 
defined corpuscles, comparable to the hematids in the blood of 
mammals. These adhere together, just as those of Magelona 
do, on being shed from the vessels. 

The globules of Magelona stand, as it were, midway between 
the coloured liquid plasma of Annelids generally and the 
coloured corpuscles of mammalian blood. The latter are 


14, W. BLAXLAND BENHAM. 


known not to be cells, but to be parts of cells, produced in 
many cases within cells by the modification of the protoplasm, 
its coloration and subsequent separation into globules, which 
are then set free by the breaking up of the parent cell. 

With regard to the development of the globulesin Magelona 
I have not yet come to any definite conclusion, though I 
expect to find that a similar origin of the corpuscles obtains.! 
The formation of the plasma in ordinary Annelids is practically 
on the same lines, but the fluid, in place of separating as 
globules or corpuscles, is discharged en masse (see Lankester’s 
observations “On the Connective and Vasifactive Tissues of 
the Medicinal Leech,” ‘Q. J. Micr. Sci.,’? vol. xx, 1880, 
p- 307). 

5. The Colouring Matter of the Blood.—We have 
already seen that no absorption bands were obtainable by means 
of the spectroscope, nor have I been able to separate any pig- 
ment by the use of various reagents. But from mere tint alone, 
and its change of colour on deoxidation, it is evident that the 
pigment is quite different from any hitherto recorded amongst 
the Chetopods; it is most evidently not hemoglobin. The 
other colouring matters known in the Chetopods are the greens 
of the Sabellids and Chlorhzmids, known as chlorocruorin. 

In no other Chetopod do we find the madder-rose colour 
characteristic of Magelona; but in Sipunculids a similar blood- 
pigment exists, presenting a similar loss of colour on deoxida- 
tion. This colouring matter, as seen in Sipunculus nudus, 
closely resembles that of Magelona. Its peculiarities were 
first described by Lankester,” and subsequently by Krukenberg,’ 
who gave to it the name ‘‘ Hemerythrin.” 

The blood-pigment in the Sipunculids occurs in the nucleated 
circular disc-shaped corpuscles of the fluid contained in the 

1 Miss Buchanan describes some of the early phenomena of the formation 
of the blood of Magelona (‘ Rept. Brit. Assoc.,’ 1895, p. 469), in which she 
suggests that the “corpuscles” arise from subdivision of multinucleate cor- 
puscles derived, apparently, from a special blood-forming organ in the dorsal 


vessel. 
2 Lankester, ‘ Proc. Roy. Soc.,’ xxi, p. 80. 
3 Krukenberg, ‘ Vergl. Physiol. Stud.,’ lst reihe, 3te Abth, p. 82. 


THE BLOOD OF MAGELONA. te 


body-cavity. In Phoronis the pigment is hemoglobin, as 
Lankester showed by its spectrum, and is not hemerythrin, 
as Krukenberg assumes. In Phascolosoma elongatum 
Schwalbe describes the tint as light rose or faint greyish red 
(matt graurothlich), which on oxidation grows darker, and finally 
becomes “ burgunderrothe.” This description of the colour 
scarcely agrees with that of Magelona, nor does that given 
by Ehlers and Keferstein for other members of the group— 
Sipunculus and Priapulus,—for they term the colour of 
the body-fluid ‘ wine-red;” but in Sipunculus nudus at 
any rate, as I can testify from my own observations, the tint 
is very close to that seen in Magelona. 

6. Relation of Vessels to Celom.—The fact that in 
the thorax the ventral vessel dilates to such an enormous extent 
as to nearly obliterate the ccelom is a very interesting and im- 
portant piece of evidence in favour of Professor Lankester’s 
suggestions (‘ Quart. Journ. Mic. Sci., xxxiv, 1893) as to the 
manner in which the heart of Arthropods, with its series of 
ostia for the entrance of blood, has been derived from a Cheeto- 
pod vascular system, owing to the enlargement of the afferent 
vessels or veins, till the latter fuse with one another to form a 
great “ pericardial blood sinus,’’ to the exclusion of the ceelom. 
Here, in Magelona, we have a great dilatation of the ventral 
vessel and lower part of the “ afferent” vessel leading upwards 
to the dorsal vessel. This great lateral dilatation (J. ext.) com- 
municates with the ventral vessel just behind the septum, 
between consecutive segments, by a comparatively narrow 
opening, as shown in the figure. If this dilatation extended 
further upwards to the dorsal vessel we should have a condition 
closely approaching that of the Arthropod vascular system. 
This ventral vessel of Annelids no doubt corresponds to the 
great ventral blood sinus of Astacus, &c., whence the blood 
makes its way up to the heart, either directly or by way of 
the gills. 

I hope, in a later contribution, to give a further account of 
the vascular system of this interesting worm, Magelona. 


16 W. BLAXLAND BENHAM. 


EXPLANATION OF PLATE 1, 


Illustrating Dr. W. Blaxland Benham’s Paper on ‘‘ The 
Blood of Magelona.” 


EXPLANATION OF FIGURES. 


Fie. 1.—Side view of the anterior region of Magelona, with the “‘pro- 
boscis” withdrawn. The thorax is coloured by the blood contained in its 
vessels. 


Fic. 2.—Side view of the anterior region of Magelona, with “ proboscis”’ 
(a) everted. The thorax is now almost colourless, the blood being driven 
into the proboscis. 


Fie. 3.—A transverse section through the thorax of Magelona, showing 
the great dilatation of the ventral blood-trunk and its extension into the 
“lateral chamber” of the body. On the right side one of the dorso-ventral 
muscles is represented separating the lateral extension of the vessel from the 
median portion. This dilated vessel comes to occupy almost the entire ccelom. 
The epidermis (ep.) is somewhat diagrammatically represented; the remainder 
is drawn with the camera. D.v. Dorsal blood-vessel, with thick muscular 
coat. V.v. The dilated ventral vessel, with thin nucleated wall. Jat. ext. 
and Z. ext. Lateral extension of the ventral vessel. /.v. Vessel from ventral 
to dorsal vessel. Ca. Celom. wn. Nerve-cord. G. The gut. cére. Circular 
coat of muscles of the body-wall. ¢.m. Longitudinal muscles. dor.v.m. 
Dorso-ventral muscle. 0d/. Oblique transverse muscle. (Camera Zeiss B 2,) 

Fic. 4.—A group of blood-globules. (x 550.) Camera drawing of 
contents of a blood-vessel from a section stained in borax carmine. The size 
of the corpuscles is practically unaltered by the preservatives. . Nucleus. 

Fic. 5.—Two globules from same section, greatly magnified. The con- 
tents have been coagulated by reagents. 

Fic. 6.—A group of globules (x 550) from a preparation of blood killed 
with Fleming’s solution, stained in picro-carmine, and mounted in glycerine. 
The globules have ‘‘fused” with one another, so that the field is filled by 
larger globules of various sizes. A coagulated plasma (p/.) is represented. 
A nucleus (z.) shows the relative size of the globules. (Camera drawing.) 

. Fic. 7.—Some isolated nuclei, showing varying numbers of chromatic 
masses. From same preparation. 

Fie, 8.—An example of nuclei apparently contained within large globules 
(from picro-carmine, glycerine preparation). #. Nucleus. pl. Plasma. 
Camera. xX 550. 


THE BLOOD OF MAGELONA. ily? 


Fie. 9.—Other examples of the same more highly magnified. In a the 
nucleus, overlying the globule, projects beyond its margin. 

Fic. 10.—Freshly drawn blood, without reagent, uncovered (Zeiss, F 2), 
The globules do not freely separate from one another, but adhere together, 
forming an irregular mass or network. c. A few isolated globules. ¢. De- 
licate thread-like bridges of corpuscles. #’. The same broken across by the 
movement of the mass, resulting in the isolation of some globules. 

Fic. 11.—A portion of the same preparation covered; the same magnificas 
tion (Zeiss, F. 2). The irregular mass is now seen to consist of masses of 
corpuscles or globules, forming a network with irregular meshes. 


Fig. 12.—A portion of a similar drop of blood in salt solution, to which 
water has been added. The threads of the network have become greatly 
elongated and more delicate. The margins (a) of the masses have become 
homogeneous, owing to the fusion of neighbouring globules; these still exist 
unaltered in the centre of the nodes (4). 

Fic. 13.—Further action of water (lower magnification). 


Fic. 14.—A small drop of blood after addition of water, showing the 
fusion of a large number of globules to form larger and larger clear masses. 


Fie. 15.—Another drop of blood, with a trace of sea water. (Camera 
drawing. xX 550.) Various sized drops have been formed by fusion of 
globules. 


Fic. 16.—Action of salt solution after water. (a) Mass in water, action 
commencing. (4) A mass of globules in water have fused to form a round 
mass. (c) Salt solution added; change of shape and movement of the mass; 
protrusion of processes. (d) The changes are more marked, and small drop- 
lets (w) are separating from the ends of processes. (This figure is drawn 
on a slightly larger scale than the other two.) In the mass there is a hole (7) 
which in @ has made its way to margin. 


VOL. 39, PART 1,—NEW SER. B 


ged. : Ns (ie Sa AP Ain Mb 
7h. >» SA Fe oy =k eae; coy SX r 


any so A 


- 


abate. Ole WP Sate Sine at 
: ‘ ewes iv glk ale plat tira 
o£ § ug e ie} b a est 5 veo VS Bre 


a Day 


aa y ; le 
At hb 0 7] 708 Cee ae 


; 
= alt Yaer ee Ire a ee Ais: . 
= ii yra ¥4 = FS" ia Wale ‘ : i BAT ae al Sie i 
CR orf aaah eee 
oie ss130 Sry ge os. + a » EAS tcp ie oe © +i 
f at t tv = way a om ae utp? A) Sol ; wile - 
x ton © ts, x = oe “ 
fee yf aa SVK a) cit =i fra) +aig@yrr an | 
rarze (© gd tp ee wR » UDO aan 
: : ert & 19 hana 
res 7 6 > 1a i. ; ‘ ! 
4 De, 
= e rs 
7) 
. way hi 
Fd a 
% rt hes i" ‘aa¢ + % = 
a — ’ * 
= : y i ' 
2 ee 
“| f 4 ¥ fia, 
22° i + te Ae The r Spy ser 6S ‘ 4 ' 
7 ~ { 
45751 Scat oT i’ f 
tH SIRR eis Swth sh ihe wien’ aie 
oo FMW UE CLT SEER aie 
s 
é 
_ 4 » et 
F ee a AY Ear Fi 3 
> ‘ —_ 
G \ .. 
¥ fe 


FISSION IN NEMERTINES. 19 


Fission in Nemertines. 
By 


W. Blaxland Benham, D.S¢.Lond., Hon. M.A.Oxon., 
Aldrichian Demonstrator in Comparative Anatomy in the University 
of Oxford. 


With Plates 2 and 3. 


Ir is a well-known fact that many Nemertines break up into 
pieces when irritated; but although this statement is current 
in text-books as well as in special memoirs the matter has 
received very little attention, and I find no account of any 
internal changes which may take place in the tissues of the 
body previous to the process of fragmentation. Indeed, it 
appears that the statement is made rather vaguely, and its 
application is scarcely so general as might be inferred. There 
is no doubt but that Carinella, Lineus, and other elongated 
and comparatively slender forms do fragment, and that these 
fragments can remain alive for a considerable time, as Dalyell, 
and later M‘Intosh have recorded ; and further, it appears that 
the anterior end of such a fragmented Lineus may produce a 
new posterior end, and M‘Intosh figures various stages in the 
formation of a head to amore posterior piece. As the specimen 
which he kept under observation lived in very unfavorable 
conditions of food, &c., he did not observe the completion of 
the head even after some months, but remarks that in a state 
of nature such a regeneration most probably takes place fre- 
quently.1. Further, we are left somewhat in doubt as to 
whether a Nemertine does of its own accord, and independently 
of irritation, normally break up into pieces. 

1 M‘Intosh, ‘Mar. Brit, Annelids: Nemertea,’ p. 125, 


20 W. BLAXLAND BENHAM. 


My attention was called to this matter in 1894 on examining 
a small Carinella from St. Andrews that happened to be 
amongst a number of specimens of Magelona, which Pro- 
fessor M‘Intosh had most kindly sent to me for the purpose 
of certain work on the latter worm on which I was engaged. 

This Carinella presented, in the hinder region of its body, 
two conspicuous constrictions at nearly equal distances from 
one another, A B, and each piece thus marked out was further 
constricted, though much less deeply, C D, into two nearly 
equal portions (Pl. 2, fig. 1).° So striking was this apparent 
“segmentation ”’ of the body, that I preserved the worm for 
sectionising. It was killed in corrosive sublimate to which 
1 per cent. acetic acid was added. 

Later on in the year, while working in the Marine Labora- 
tory at St. Andrews, with a view to collect and observe living 
Magelona, I came across two more specimens of this same 
Nemertine, which I found living in the sand with the Poly- 
cheete below the ordinary low-water mark. Both these worms 
were about the same size as the one I had obtained previously, 
and both presented similar phenomena of ‘ segmentation.” 
One of these was unfortunately mislaid; the other I killed, 
stained, and mounted entire. 

The Carinella to which these observations refer is a small 
worm about two inches (55 mm.) in length when in moderate 
extension. It is pure white, without eye-spots, and without 
pigment except for a girdle of a faint yellowish-brown colour, 
about a quarter of the worm’s length from the anterior end 
(fig. 1, 6). This is very faint even in life, and is due to pig- 
ment in the epidermis, Further, the proboscis is coloured 
faint blood red for a short region just in front of this pigmented 
band (p). 

The body is nearly cylindrical in section; the head is pointed, 
and slightly marked off from the body by a lateral vertical 
furrow on each side at the level of the mouth; the head is not 
noticeably wider than the body (fig. 2). 

The posterior end of the worm, as longitudinal sections 
showed, is not complete—a portion has already been separated, 


FISSION IN NEMERTINES. 21 


so that I cannot state the true length; but at first sight this 
end does appear entire and rounded off (PI. 2, fig. 1). 

In transverse sections (fig. 10) probably the most noticeable 
feature is the thinness of the circular muscles; so thin is this 
layer that it cannot be distinguished in transverse section 
through the middle of the body until longitudinal sections 
are called in to help:! the basement membrane is also thin. 

The two nerve-cords lie between the basement membrane 
and the circular muscles, and this position at once shows that 
the worm belongs to the genus Carinella; but as to its 
specific identity I cannot be so sure. I hesitate to confer a 
new name on the worm, as it bears a certain likeness to 
C. linearis, Montagu, which is described by M‘Intosh; and 
one of his two woodcuts illustrating this species (p. 206) also 
shows constrictions of the hinder end, recalling at once the 
phenomenon presented by the present worm, although the 
“ segments ” are not so nearly equal in size. But this point 
would not be of any moment, for in the stained entire worm 
(Specimen I1) there is less equality in the size of these “ seg- 
ments” than in the one I first noted and drew when alive. 
There is, however, a statement in M‘Intosh’s description 
which renders an identification doubtful, for he states that 
C. linearis is “ flattened,” and this is certainly not the case 
with my worm at any rate—transverse sections do not exhibit 
any flattening. M‘Intosh states that “the inner muscle of the 
body-wall shows a marked tendency to separate in the dorsum.” 
If this means that the longitudinal coat is traversed by a 
band of tissue in middle line my specimen agrees with his. 
Nevertheless no great weight can be placed on this fact, for 
Birger has shown that this occurs in probably all species of 
Carinella. But from the general description, and from 
quotations from Montagu’s MS., it appears at any rate 
probable that the present worm is C. linearis, for it agrees 
in habitat, and in the fact that it ‘‘ secretes a tenacious slime 
from its body, which, collecting sand, readily forms a covering 


1 The thickness of the circular muscle-coat has been intentionally exag- 
gerated in the figure. 


22 W. BLAXLAND BENHAM. 


like that of Sabella.” It agrees more closely with this species 
than with any more recently described, such as C. albida, 
Biirger, so that I must leave the matter of species open. 

The hardened specimen was sectionised as follows :—(1) The 
anterior end was’cut sagittally (vertical longitudinal); (2) the 
hinder end was cut horizontally, though rather obliquely 
to the horizontal plane; (3) the middle region of this body 
was transversely sectionised. The stain used was carmalum, 
though some sections were stained with picro-carmine. 

With regard to the posterior “segmented” region of the 
body, one is struck at once by the fact that at the level of each 
of the four constrictions there is a transversely arranged 
row of small nuclei traversing the longitudinal muscles. They 
are readily seen even with a low power (fig. 3), and are more 
evident at the deeper constrictions a, B, than at the shallower 
ones ¢, D. 

What appears to be the commencement of the phenomenon 
is illustrated by the sections across the constriction c (fig. 4). 
At this point the tissues are as yet perfect—there is no 
rupture. The section figured passes through the lateral line 
of the body, on one side cutting along the nerve (NV); at the 
opposite side the nerve is notinvolved. The nuclei constituting 
the transverse row are not as yet as definitely arranged as they 
will be; nevertheless they are more numerous along the level 
of the constrictions than elsewhere, and take a deeper tint in 
the stain. It may here be mentioned that the accompanying 
figures are drawn with the camera, and so far as was possible 
every nucleus is put in its true position, except in the case of the 
epidermis and intestinal epithelium where these are involved. 
The nuclei which are thus conspicuous appear to belong to 
the connective tissue in which the longitudinal muscles are 
wrapped ; they are smaller and more deeply stained than those 
of the muscles. Although the impression given by the section 
fig. 4, through the region c, appears to indicate an irregular 
transverse line, yet at a further stage this line becomes very 
much more distinctly transverse, and, moreover, it is in reality 
double; the nuclei are arranged into two rows close to one 


FISSION IN NEMERTINES. 23 


another, as is shown in fig. 5, through the region p. Here, 
however, we have a further phenomenon: the double row of 
connective-tissue nuclei having been established, the longi- 
tudinal muscles break across between the two rows of nuclei. 
This rupture appears to start just below the epidermis, and 
then travels inwards. But already the epidermis exhibits a 
furrow even at an earlier stage, as at c (fig. 4), so that it is not 
merely the result of a contraction of the muscles. 

The muscles, once ruptured, naturally contract, and leave a 
space which appears in sections to be partially occupied by a 
coagulum (2), though it may bea pre-formed material—part of 
the connective tissue. This appears to be the more likely, as is 
seen at a later stage in the process, as in fig. 6, which passes 
through the middle of the region a. Here the rupture of the 
muscle has gone a step further, and is extending towards the 
opposite side; this figure also illustrates another point—viz. 
that the rupture may commence at one point of a plane and 
extend in all directions in a radial direction, so that while it 
is complete at one side of the body it may not have yet com- 
menced on the opposite side. 

In this figure, for instance, the muscles have shrunk con- 
siderably at the left side; yet, though the rupture is extending 
towards the right, the muscles are still entire below the epi- 
dermis of the right side. By the rupture of the muscles the 
linear arrangement of the connective-tissue nuclei is to some 
extent destroyed. 

Finally, the epidermis, which during this process has become 
thinner, as well as furrowed at the plane of constriction, gives 
way; but the basement membrane still appears to resist the 
rupture, for in fig. 7 there is a distinct membrane (dm.) left 
after shrinkage of the epidermis. We must consider this 
basement membrane, as Hubrecht and others have suggested, 
as a firm skeletal tissue, and its resistance to the process here 
described is only what would be expected. 

When the rupture is complete, or even before it has travellod 
all the way round the worm, the circular muscles come into play 
(right side of fig. 8), and drawing together the margins of the 


24 W. BLAXLAND BENHAM. 


wound formed by the rupture, form a boundary to the portion 
separated as well as to the anterior region. The gonads are thus 
held in place, and do not project through the wound, and hence 
the rounded end of the body looks like the true anal end. During 
the above process the intestine has, of course, become nipped. 

After the epidermis has been ruptured one notices the 
surfaces of the wound to be covered with a flat epithelium 
(fig. 8, ep’). Whence comes this? Is it derived from the 
connective tissue? It looks as if the row of nuclei before 
mentioned had flattened out, and so given rise to the flat nuclei 
of this membrane. 

This appears to be the general history of the process of 
fission. But we are no nearer to the answer—how do the 
muscles become ruptured? why do they all give way at this 
particular plane? Further, is it merely a rupture of the 
fibrille, or is there a degeneration of these fibrillz over a cer- 
tain small area, viz. between the two rows of nuclei? Have 
these nuclei, or rather the cells belonging to these nuclei, any 
part to play in this rupture or degeneration, or are they merely 
a preparation for the new membrane which forms the ends of 
the fragments of the body ? 

At present I am unable to answer these questions with 
certainty. It appears to me, after a careful study of my pre- 
parations, that the muscle-fibrils do rupture; at any rate, if 
there is a degeneration of tissue, it occurs over only a very 
minute distance: the ends of the ruptured fibrils appear clean- 
cut (fig. 6), and I see no sign of any modification in their 
substance which would point to a degeneration. 

With regard to the second part of the query, viz. the action 
of the row of nuclei, I would suggest that they do play an 
active part in the process. The muscle-fibrils are wrapped 
together by connective tissue, which has the form of a network, 
through the meshes of which the fibrillz pass (fig. 9). At the 
point of future rupture this network is denser and the meshes 
smaller, and it may be that the cell-substance becomes actively 
contractile, and really nips the muscle-fibrils in two. On 
the other hand, we must not overlook the possibility of the 


FISSION IN NEMBERTINES. 25 


occurrence of some kind of solvent action—the fibrils may be 
“eaten through,” as it were; but I know of no means of 
deciding the matter. 

The process, whatever its details, seems to be different from 
that which occurs in Planarians, such as Microstoma, which 
reproduce asexually by fission. 

In his account of the process, van Wagner (‘ Zool. Jahr- 
buch,’ iv, 1891) expressly states that there are no nuclei in 
the “ septa,’ which make their appearance along the lines of 
the future separation of the individual into zooids. Each septum 
is represented in v. Wagner’s illustrations as being of some 
homogeneous tissue, passing from the subepidermic muscles 
to the muscularis of the intestine. He gives no details as to 
how the “ parenchyma” separates, which bear any resemblance 
to those detailed in the process of rupture of the muscles seen 
inthis Nemertine. Further, there is a great deal of ‘‘ regene- 
ration ”’—i.e. formation of new brain, new pharynx, &c.—in 
Microstoma before separation takes place. Of this there is 
no trace in Carinella. I need not enter here upon the vexed 
question as to the distinction between “ division” and “ gem- 
mation.” In Carinella there seems no doubt that the process 
is “ division.” 

A second part of the problem relates to the meaning of the 
whole process in regard to the life of the animal. It is 
generally assumed, and apparently rightly assumed, that the 
process of rupture will be of advantage to the Nemertine, in 
that each of the pieces thus formed will be able to give rise to 
gonads, as M‘Intosh observed; in most cases, however, there 
does not seem to be any direct relation between rupture and 
genital maturity. But in the present instance there is such a 
relation, for in both the specimens the genital elements are 
only present in the hinder part of the worm, which 
is being separated off from the rest. 

In the specimen I (fig. 1), which is a female, the gonads are 
present in each of the four “segments,” and no trace of ova 
exists in front of the line B, while behind the line A the ova are 
larger and more numerous than in front of it. 


26 W. BLAXLAND BENHAM. 


The specimen II, which was mounted entire, is a male, and 
the spermatozoa exist only in the hindermost part of the body. 

It would appear, therefore, that the gonads make their appear- 
ance simultaneously over a certain stretch of the body, com- 
mencing at the hindermost region; that as they ripen this 
region begins to be constricted off from the rest of the body. 
A second tract now commences to ripen, and in its turn becomes 
marked off from the non-sexual anterior region of the body ; 
and so the process goes on up to the point of origin of the 
proboscis. 

Further, each such region may become restricted into two 
(or more?) pieces ; and it will be very interesting to ascertain 
whether this segmentation goes any further than in the speci- 
mens to hand. It is possible that the segmentation of this sexual 
region of the body, which is already beginning in our specimens, 
may ultimately become so far repeated that each “‘ segment ” 
will contain only a single pair of ovaries. In that case we 
should have a strong case in support of Hubrecht’s theory of 
segmentation, as elaborated in his ‘ Report on the Nemertines 
of the Challenger Expedition.’ 

However that may be, we have in this Carinella what 
appears to be an undoubted preparation for the spontaneous 
subdivision of the sexual region of the body into a number of 
isolated portions, and this subdivision is due to the develop- 
ment of the gonads, and is thus directly related to the pro- 
pagation of the species. 

We have yet to learn whether each of the pieces so isolated 
will produce a new head; but it appears to me a needless 
assumption that this would happen. As each “segment” 
drops off, the tissues after decomposition might set free the 
generative products; but since genital ducts (fig. 3, g.d.) are 
already forming, and are in the hindmost segments nearly 
completed, it appears that the segments live an independent 
life for some time after separation. This fact, therefore, mili- 
tates against the idea of any greatly extended “‘ segmentation.” 

Anatomical.—The Nemertine which forms the subject of 
the present contribution exhibits one or two points of interest 


FISSION IN NEMERTINES. 27 


in its histology, which differs from that described by Birger! 
for Carinella and its immediate allies. 

The lateral nerve-stems do not exhibit the characters 
described and figured by that author, who finds an inner and an 
outer neurilemma separated by a considerable space. In this 
lie numerous ganglion-cells, chiefly collected above and below 
the “punktsubstanz,’ which is surrounded by the inner 
neurilemma, and to which the processes of the ganglion-cells 
pass. In the present worm the ganglion-cells are very few, at 
any rate towards the middle of the worm, though they are 
more numerous anteriorly. But these cells lie not only above 
and below, but at the side of the nerve (fig. 11). Further, I 
do not find distinct inner and outer neurilemma. The nerve- 
cord is surrounded by a distinct but delicate membrane, with 
flattened nuclei (m’ m’”); this appears to correspond to Biirger’s 
outer neurilemma. The ganglion-cells are situated in a net- 
work of tissue, resembling, in general, the substance of an 
ordinary Invertebrate nerve-cord, and the meshes of this net- 
work are occupied by fine dots—the “ punktsubstanz ”’ (p.). 

Lying near the centre of the nerve-cord is a distinct tube, 
with well-marked wall, against which lie here and there nuclei 
resembling those of the ganglion-cells. This tube (ch.) usually 
appears empty, but in some sections a faint coagulum with 
very minute scarcely visible granules init. There is no net- 
work, such as occurs in a nerve-cord, and I take this tube to 
be a “ giant fibre” or “‘ neurochord,” such as Biirger finds in 
Nemertine nerve-trunks generally. 

In Carinella, according to that author, there are none, 
whereas in the Schizonemerti he finds several of them. It 
might be suggested that what I regard as the wall of a tube or 
neurochord is, in reality, Biirger’s “ inner neurilemma;” but 
the contents of this tube are certainly not like the structure 
seen in an ordinary Invertebrate nerve-cord. My sections 
were stained in picro-carmine, and others in alum carmine. 
I did not employ hematoxylin, and it is possible that a 

1¢ Unters. ub. d. Anat. u. Histol. d. Nemertinen,” ‘Zeit. fiir wiss. 
Zool.,’ 50. 


28 W. BLAXLAND BENHAM. 


renewed investigation with other preservatives and stains may 
lead to a different interpretation of this “tube ;” but, so far 
as my preparations go, | cannot think that this is an “inner 
neurilemma,” or that the contents form the “ punktsubstanz.” 

The muscles of the body-wall are chiefly interesting for the 
great thickness of the longitudinal coat, and the very feeble 
development of the circular coat in the greater part of the 
body, though anteriorly this coat becomes rather thicker. 

The ‘inner coat of circular muscles,” which has so often 
been described for the Paleonemertini, has been identified 
by Birger as the muscle of the rhynchocele (proboscis sac) 
and gut, which, owing to the slight development of the paren- 
chyma, comes to lie immediately within the Jongitudinal coat. 
This explanation brings the muscular system of the Palzone- 
mertini into agreement with the general scheme of Nemertine 
musculature. 

This inner coat (r/.c.) is in the present worm very thin indeed, 
but can easily be recognised in longitudinal sections. I have 
exaggerated its thickness in the figure of a transverse section, 
This inner coat can be traced as a sheath round the rhynchoceele 
and intestine (éné. c.), and a second coat surrounding the rhyn- 
choceele alone. 

Birger describes a decussation of this inner coat in the 
median dorsal line. I am unable to satisfy myself whether 
this exists in the present worm; but traversing the longi- 
tudinal muscles is a vertical strand of tissue (fig. 10, a.) stain- 
ing like the basement tissue and continuous with it dorsally, 
and with a similar tissue surrounding the rhynchoceele ; 
whether the circular muscles accompany this strand—as I 
believe they do—I was unable to determine. The intestine, as 
I have just stated, is partially—i.e. on the ventral side— 
surrounded by a layer of circular muscles; in addition, its 
dorsal wall is formed by a fairly thick layer of longitudinal 
fibres (int. lg.), lying between the intestinal epithelium and 
the wall of the rhynchoceele. 

The proboscis presents throughout the greater part of its 
extent a glandular epithelium, surrounded by circular muscles, 


FISSION IN NEMERTINES. 29 


and then by a thick coat of longitudinal muscles. But the 
epithelium changes its character for a short space near its 
anterior end; the gland-cells, so conspicuous elsewhere, no 
longer catch the eye in section, if indeed they are present ; 
but the cells are long, narrow, with flat nuclei at the sides, and 
close to the free end there is a series of round nuclei arranged 
in a well-defined row. 

In longitudinal section this region occurs between 2 and 
3mm. from the anterior end; whether it corresponds to the 
blood-red tract on the proboscis, noticeable in the living worm, 
I am not certain. 

In all the Palzonemertini the parenchyma is but very 
feebly developed, but in the present worm, at any rate behind 
the middle of the body, it is practically absent. I can detect 
no space between the wall of the gut-muscle, and body-wall, 
i.e. between the inner circular muscles and the longitudinal 
muscles, except the longitudinal blood-vessel on each side and 
a nearly homogeneous tissue (y.) passing up from it to the inner 
wall of the intestine; this tissue resembles the ordinary con- 
nective tissue, which occurs in some sections in sheets quite 
like this little piece. This is all that represents the “ paren- 
chyma”’ posteriorly. 


There is one point in which this Carinella does not agree 
with Biirger’s statement as to the arrangement of the gonads. 
He states that in the Palzeonemertini there is usually more 
than a single pair of gonads in a transverse plane, so that a 
transverse section through this region would exhibit more than 
two gonads and ducts. ‘This is not the case with the present 
Carinella. The gonads are strictly paired; there is one, 
and only one, on each side,—each with its duct formed in the 
usual way. ach ovary when nearly ripe consists of three or 
four large ova, surrounded by a common membrane of flat 
cells, which is continuous with a heap of indifferent rounded 
cells on the outer side of the ripe cells (fig. 12). Amongst this 
mass of rounded cells one can distinguish one or more cells, 
still small, but distinctly larger than the rest, with a distinct 
germinal vesicle (ov.); these are evidently young ova. 


30 W. BLAXLAND BENHAM. 


EXPLANATION OF PLATES 2 and 3, 


Illustrating Dr. W. Blaxland Benham’s Paper on “ Fission in 
Nemertines.”’ 


Fie, 1.—View of Carinella (No. 1) immediately after death, showing the 
“segmentation ”’ of the hinder part of the body. (x 6.) A, B, GC, D, indi- 
cate the constrictions in order of depth, 4 being deepest. The drawing also 
shows the general appearance of the worm. pp is the red region of the 
proboscis. The body is coloured orange in the region just in front of 4. 


Fic. 2.—Ventral view of the anterior end, to show the shape of the head. 
M. Mouth. &. Aperture of the rhynchodeum. JZ. Ciliated furrows. 


Fie. 3.—Camera drawing of the hinder end of the worm. A, B, C, D. 
The constricted regions, each marked out internally by a double row of 
nuclei. The section is nearly horizontal. ov. Ovaries, fully developed in the 
hinder “‘segments.” ov’. Less developed ovarian sacs in anterior regions. 
g.d. Genital duct. zt. Intestine. ep. Epidermis. /y. Longitudinal muscle. 
Mag. Zeiss AA 2. 


Fic. 4.—The region C enlarged. The section is at a different level from 
that of the preceding figure, as it cuts the nerve-cord on the left, and does 
not involve the intestine. ep. Epidermis. 4.m. Basement tissue. circ. 
Circular muscles. J/g. Longitudinal muscles. 2. The transverse row of 
nuclei. WV. Lateral nerve-stem. Mag. Zeiss B 4. 


Fic. 5.—Horizontal section through the region D, to show the next stage in 
the process of fission. The longitudinal muscles have ruptured, contracted, 
and left a cavity just below the nerve-cord. 2. A coagulum. 


Fic. 6.—The next stage in the process, illustrated by section through the 
region 4. The ruptured muscles are more numerous, and the cavity (sp.) 
therefore more extensive, and is traversed by connective tissue (c¢.¢.). Mag. 
Zeiss B 4. 


Fic. 7.—Another portion of the same region, to show the coagulum (z.), 
the broken ends of the longitudinal muscle-fibres, and the unruptured base- 
ment membrane (4.m.). 4v. is the lateral blood-vessel. 


Fie. 8.—Portion of A (fig. 3) at a different level, enlarged. The rupture 
has extended nearly right across, and the ends of the two pieces are bending 
inwards owing to the contraction of the circular coat of muscles. This figure 
also shows the flattened epithelium (ep’.) which covers the broken ends of the 
muscles. 


FISSION IN NEMERTINES. 31 


Fic. 9.—A somewhat diagrammatised view of the muscle-fibres and con- 
nective tissue in the region of one of the “septa.” m. Muscle-fibres. 
c.f, Connective-tissue network, through the meshes of which the fibrils pass: 
the network is denser in the region of the “septum” than elsewhere. 7! 
Nucleus of ‘‘ septum.” 


Fic. 10.—Transverse section of the worm (mag. B 2,cam.). ep. Epi- 
dermis. 4.¢. Basement tissue. 4v. Blood-vessel. circ. Circular muscles 
of body-wall. 7g. Longitudinal muscles. zz¢. Intestine. izé.c. Circular muscle 
of intestine. zz¢./y. Longitudinal muscle of intestine. Pr. Proboscis. P.ep. 
Outer epithelium of proboscis. RA. Rhynchocel. R.ep. Epithelium of 
rhynchocel. rh.c. Its coat of circular muscles. WV. Lateral nerve-stem. 
x. Streak of tissue resembling connective tissue interrupting the longitudinal 
muscles dorsal (? decussation of r/.c.). y. ‘Connective tissue” near blood- 
vessel. 


Fic. 11.—View of section across lateral nerve-stem. 4.¢. Subepidermic 
basement tissue. circ. Circular coat of muscles. ch. Neurochord, with 
distinct membrane and (?) coagulum within. m!. Outer membrane (? neuri- 
lemma). #', Inner membrane, separated at one point by a space (sp.), probably 
due to shrinkage. gy. Nuclei of nerve-cells. .s. Nuclei of sheath. p. 
Punktsubstanz in the meshes of nerve-cord. D.V. indicate the dorsal and 
ventral surfaces of the nerve-cord. 


Fie. 12.—A portion of a section across the region A to show the ripe 
ovary. ov. Ovary. O. A ripe ovum. w. Wall of ovary. s. “Septum.” 
b.m. Basement tissue exposed after the rupture of the epidermis. 


a Roti Phir smear ey 


tea mins Emr eSiiabieand sty eel bea enbtar sat aul 
an een nlp ind te Oa Sre mitt 7h Age Ty ning we 
¢ - omuiiy SAT Sst @ De petals Wank AY oad) Vbe—é a iad th 
‘ae, cro wollte aut) “ abhagas HD Wipe, gee 1: ow" 

nae ern 


7 neck uy . ead a ; Cu), ure ny ny ity " vebvrewtcheme 
- Z “of ee re ca 2) 2 oe a} th io% af inaif {ov eped® a) 
‘| i viru? vases Pyllinl Jo a t iia oo ~ 
7 Viv ( atratyl 1 aa. ch Link yeaa an cont 
sl i a - . i so ios wil awh 
é ’ > ¢ = 
: Vip 4 F ; A ri A dat OOS ia 
2 i lint iy nde os f rr 
€ iY 7 . : 
tw Se cl" ae ; 
= al he sles} See le. al Tea 
; ~-"e 
? yo» urea} 
> . e} * ‘wit 
, : - ae. a 
t j if 0 ts, Py “nts 
, \ ree ifs xe 
; ey olf ho OO ae 
tt SPL Ripetecdl 
7 nae we aes Pe 
aS 
i? ; 
A 
4 . 
, . 
f ’ 
. Qghmaeny 
-— AY Oe 
a 


STUDIES ON THE NERVOUS SYSTEM OF CRUSTACEA. 33 


Studies on the Nervous System of Crustacea. 
By 


Edgar J. Allen, B.Sc., 
Director of the Plymouth Laboratory of the Marine Biological Association. 


With Plate 4. 


IV. Further Observations on the Nerve Elements of the 
Embryonic Lobster. 


In the first part of these studies! an account was given of 
certain nerve elements which had been demonstrated by the 
action of dilute solutions of methylene blue upon the ganglia 
of the embryonic lobster (Homarus vulgaris). The elements 
then described arose, in most cases, from cells which lay in the 
anterior ganglia of the thorax. In the present paper a number 
of additional elements occurring in these ganglia will be 
noticed, together with those found in the posterior thoracic and 
in the abdominal ganglia. 

The method of investigation has been the same as that 
described in my former communication, the results in this 
case, however, having been obtained principally from embryos 
which were near the point of hatching. Several of the new 
elements differ in essential points from any of those previously 
described, and are of interest in throwing additional light upon 
the manner in which different portions of the central nervous 
system, or different movements of the body, are co-ordinated 
one with another. 

1 This Journal, vol. 36, pt. 4, 1894; a preliminary notice occurs in ‘ Pro- 
ceed. Roy. Soc.,’ vol. lv. 

VOL. 39, PART 1.—NEW SER. C 


34 EDGAR J. ALLEN. 


The observations will be dealt with in the following order: 
I. Elements arising from cells situated in the THoracic 
GANGLIA. 


(2) Elements of types the same as or similar to those 
previously described. 
(6) Elements belonging to types not previously described. 
I]. Elements arising from cells situated in the ABDOMINAL 
GANGLIA. 


III. Elements arising from cells outside the central nervous 
system. 


I. Tuoracic GANGLIA. 
(a) ELEMENTS SIMILAR TO THOSE PREVIOUSLY DESCRIBED. 


B, Tuorax xt! (fig. 1).—The cell lies in the anterior 
portion of the lateral mass of ganglion-cells of Thorax x1, 
and gives off a fine fibre, which passes into the neuropile 
( punkt-substance”’), and there breaks up into three main 
branches. Of these branches one passes forwards into the 
next ganglion in front, another backwards to the ganglion 
behind, whilst the third curves round the central ganglion- 
cells, and, keeping close to these, runs forward towards the 
brain on the opposite side of the cord to that on which its 
cell lies. 

By actual observation this fibre has been traced as far for- 
wards as Th. viii, but there can be no doubt that it behaves 
in the same way as B, Th. v, and B, Th. vii (Pt. I, pl. 35, 
fig. 3),? and runs forward to the brain. A comparison of the 
three elements makes it evident that they belong to one and 
the same system, B, Th. vitt, and B, Th. x1, especially resem- 


1 The denomination of the elements to be adopted here will be a continua- 
tion of the system used in the first paper. On comparing fig. 1 of this paper 
with fig. 1 on Pl. 35 of Part I (this Journal, vol. 36), it will be seen that in 
the present case the lateral mass of ganglion-cells in each ganglion (shaded 
blue) is more completely divided into an anterior and a posterior portion by 
the neuropile (left white in the figure) and the fibres of the anterior nerve-root. 
The present diagram represents more nearly the appearance presented by very 


late embryos, whilst the former figure represents that found in earlier ones. 
2 This Journal, vol. 36. 


STUDIES ON THE NERVOUS SYSTEM OF CRUSTACEA. 35 


bling one another in details of arrangement. Also, in a pre- 
paration of an earlier embryo, in which the fibres B, Th. v, 
and B, Th. vir1, were completely stained, a third fibre was 
observed ending with them in the brain, running close to them 
along the cord, but proceeding backwards beyond Th. vir. 
Its destination at that time remained undetermined, but there 
can be little doubt that it was the fibre of this element, B, Th. xt. 

It will be observed, too, that Th. xr is exactly the ganglion 
in which we should expect to find another element of this class. 
Such elements have already been described (see Pt. I, pl. 36, 
fig. 3) in Th. 11 (with branches to Th. 1 and Th. 111), in Th. v 
(with branches to Th. 1v and Th. vi), in Th. vitr (with 
branches to Th. vir and Th. 1x), and now we find a similar 
one in Th. xi (with branches to Th. x and Abd. 1). In this 
way each ganglion of the thorax must be influenced by these 
elements, which end in a particular region of the brain. 

The series of elements of which the element B, Th. x1, is a 
type resembles in many respects the series of elements de- 
scribed by Retzius,' in Amphioxus lanceolatus, the cells of 
which give rise to the giant fibres in the nerve-cord of that 
animal. These fibres, after leaving the cell and crossing to 
the opposite side of the cord, are described as running forwards 
for some distance, but their ultimate fate was not determined. 
A knowledge of the nature and position of their endings would 
be of great interest. 

PropaBLe Motor Exements or Anterior Roots.—The 
elements E, G, and H (fig. 1. Compare also Pt. I, pl. 35, 
fig. 1), the fibres of which leave the ganglia by the anterior 
nerve-roots, and which have already been described for the 
anterior ganglia, occur also in Th. 1x, x, and xt. 

The elements E (4) and H (e) (fig. 1) resemble in general 
features the element E, but differ from it in the number and 
mode of branching of the fibres which they give off to the 
neuropile. Their characteristic appearance and the situation of 
the cells may be seen from the figure. They occur upon both 
sides in the ganglia from Th. vii1 to Th. x1. 


1 Retzius, ‘ Biol. Untersuch. Neue Folge,’ ii, 1891. 


36 EDGAR J. ALLEN. 


The element F (0) (fig. 1), which occurs in Th. x and x1, 
resembles the element F (Pt. I, pl. 35, fig. 1), and may be the 
representative of that element in these two ganglia. 

The element P, Th. vi (fig. 1), presents a striking appearance 
when stained, and differs in important respects from the 
elements previously described. ‘The cell lies in the anterior 
portion of the lateral mass of ganglion-cells. The fibre curves 
backwards to the neuropile, and after giving off what, from its 
relatively small diameter, may be regarded as a subsidiary 
branch, divides into two main branches, one of which passes 
immediately out of the ganglion through the anterior nerve- 
root, whilst the other runs across as a stout transverse fibre to 
the opposite side, where it turns back again forming a loop 
with itself, and was traced as far as the centre of the ganglion. 
When the element has stained upon both sides the two trans- 
verse branches lie close together and appear as one stout fibre. 
The subsidiary branch, which leaves the fibre before it bifur- 
cates, takes the somewhat complicated curved course repre- 
sented in the figure, and was traced to the neuropile of the 
opposite side. As, however, neither this nor the main branch 
was observed to break up into finer branches, it seems pro- 
bable that the staining was in all cases somewhat incomplete. 
Similar elements have not been found in any other ganglion. 

ProspaBLeE Motor ELemeEnts or Posterior Roots.—The 
cells of the elements just described all send out fibres through 
the anterior nerve-root of the ganglion in which the cell is 
situated. In the two following cases, however, the fibre leaves 
by the posterior root. 

In Th. vist, an element (fig. 1, K, Th. vir) similar to K, 
Th. 111 (Pt. I, pl. 35, fig. 1), has stained. The cell is situated 
in the central mass of ganglion cells, the fibre decussates with 
its fellow of the opposite side and passes out at the posterior 
root of the ganglion. This element has been found only in 
Then and Th. vars. 

In Th. x1 is an element which most nearly resembles the 
element J of the anterior ganglia. This is denominated J (3) 
(fig. 1). The cell is situated at the posterior end of the central 


STUDIES ON THE NERVOUS SYSTEM OF CRUSTACKA. 37 


mass of ganglion cells of the ganglion, and is relatively large. 
From the cell the fibre passes forwards and slightly outwards 
to the centre of the neuropile, where after turning downwards 
it bifurcates, sending one branch out of the ganglion through 
the posterior root, whilst the other runs as a transverse fibre 
to the neuropile of the opposite side. The element has not 
stained in any other ganglion of the thorax, but as will be 
seen later, what is probably the same element occurs in the 
abdominal ganglia. 

The element O in Th. vit (fig. 1) also sends its fibre through 
the posterior nerve-root. The cell lies in the anterior portion 
of the central mass of ganglion cells and pursues the course 
indicated in the figure, giving off branches to the neuropile 
on the side only on which the cell lies. The element occurs 
on both sides of the ganglion Th. vir, but corresponding 
elements have never stained in any other ganglion. 


ELEMENTS BELONGING TO NEW TYPES.—The motor elements 
referred to in Part I, and those already noticed in the present 
paper, are all characterised by the fact that the fibre leaves 
the central nervous system through one of the roots of that 
ganglion in which the cell attached to it is situated. The 
portion of the element which lies within the central nervous 
system is therefore entirely confined to one ganglion. In the 
cases now to be described, whilst the cell lies in one ganglion 
the fibre passes out of the cord by the nerve-root of some 
other ganglion. 

The element Q, Th. vir, is an instance of this class. The 
cell lies in the anterior portion of the central mass of ganglion 
cells of Th. vir, gives off a fibre which runs outwards and 
then upwards to Th. v1, where it passes out by the posterior 
root of the ganglion. The fibre gives off a stout arborescent 
branch in Th. vir, and a straight transverse branch in Th. v1, 
which passes across to the opposite side of that ganglion. 
This pair of elements has only stained in Th. vi. 

Three pairs of elements having many of the characteristics 
of the above, but differing in detail, are found in Th. tv, v, 


38 EDGAR J. ALLEN. 


and vi. The element R, Th. v1, shows their principal cha- 
racters. The cell in this case lies in the posterior portion of 
the central mass of ganglion cells of Th. v1.1 It is somewhat 
smaller than the cell of the element Q, and its fibre runs 
forwards for some distance before turning outwards. After 
proceeding in the outward direction, the fibre again turns 
forwards and runs into Th. v, leaving the central nervous 
system by the posterior root of that ganglion. Two principal 
branches are given off during its course, one arborescing in 
the neuropile of Th. vi, the other in that of Th. v. 

Another pair of elements, whose cells lie in one ganglion 
whilst the fibres pass out through the posterior roots of the 
ganglion in front, is that lettered S in Th. x1 (fig. 1). In this 
case, however, the elements of the opposite sides decussate, 
giving rise to the characteristic figure shown (fig. 1). The 
cell of this element is small, and lies on the central surface of 
the ganglion. The element occurs also in the abdominal 
ganglia, and will be subsequently described in more detail. It 
has not stained in any other ganglion of the thorax, except 
Th. x1. 


ELEMENTS HAVING TWO OR MORE BRANCHES WHICH PASS OUT 
OF THE CENTRAL NERVOUS SYSTEM BY THE NERVE-ROOTS OF 
DIFFERENT GANGLIA.—T wo pairs of elements of this kind have 
stained in the thorax, the cells of one pair being found in 
Th. vir, those of the other in Th. vi1rt. In the latter case 
(Fig. 1, T, Th virz) the cell les in the anterior portion of the 
lateral mass of ganglion cells, near the point at which the 
anterior nerve-root leaves the ganglion. The cell gives off a 
moderately fine fibre, which very soon bifurcates, one branch 
passing immediately out of the ganglion through the anterior 
nerve-root, whilst the other runs forwards along the ganglionic 
cord. The forward branch, keeping close to the lateral masses 


1 As was mentioned in Part I (see this Journal, vol. 36, p. 465) the central 
masses of ganglion-cells of the embryo, shaded blue in fig. 1, divide at a later 
stage into an anterior and a posterior portion, the two portions belonging to 
two adjacent ganglia. The line of demarcation is seen in Th, IX (fig. 1). 


STUDIES ON THE NERVOUS SYSTEM OF CRUSTACEA. 39 


of ganglion cells, pursues a perfectly straight course until it 
reaches Th. 111, where it gives off a branch, which passes out 
through the posterior root of that ganglion. After giving off 
this branch the fibre continues to Th. 11, where it turns and 
leaves the ganglion through the posterior root. In one or two 
preparations, another branch appeared to be given off from the 
fibre in Th. tv, and to pass through the posterior root of that 
ganglion, but I was never able to make myself quite sure of 
this point. 

From the preceding description and the figure (fig. 1), it will 
be seen therefore that the element T, Th. viir supplies fibres 
to three (possibly four) nerve-roots of different ganglia, namely, 
the anterior nerve-root of Th. vir1, the posterior root of Th. rv 
(probable), the posterior root of Th. 111, and the posterior root 
of Th. 11, and that all these fibres have their origin in a single 
cell. 

A corresponding element occurs also in Th. vir (fig. 1, T, 
Th. vit). The cell is similarly situated to that of Th. vii1, and 
a fibre passes almost immediately to the anterior root of 
Th. vir. A second fibre pursues a straight course forwards, 
keeping close to the corresponding fibre of Th. vii1, until it 
reaches Th. 111, where it gives off a branch to the posterior 
nerve-root of that ganglion. The fibre then continues to run. 
forwards, but its ultimate destination has not been satisfactorily 
determined. It appears to turn inwards in Th. 11. 

Mention may here be made of a fibre which stains in almost 
every preparation from the earliest stages onwards. It is 
inserted in fig. 1, T (a). The fibre appears to run longitudinally 
through the ganglionic cord from the level of Th. 11 to the 
posterior region of the brain, and to give off branches to two 
nerve-roots. One branch passes out through the posterior 
root of Th. 11, whilst the second leaves the brain by a small 
nerve which arises immediately anterior to the nerve which 
supplies Antenna 11. A few small fibres are given off to the 
neuropile of the brain at the angle which the latter branch 
makes with the longitudinal fibre. 

No cell has ever been seen to stain in connection with the 


40 EDGAR J. ALLEN. 


fibre just described, and whether the originating cell lies inside 
or outside the central nervous system remains unknown. It is 
possible that we are here dealing with an element similar to 
the elements T in Th. vir and vii, the cell of which however 
has never taken up the methylene blue. 

Elements in which one ganglion cell gives rise to two or 
more fibres passing out of the ganglion by different nerve-roots 
have been described by Retzius in Aulastomum gulo.! In 
that case the two fibres traced passed out by the two roots 
of the ganglion in which the cell lay, whilst a third, whose 
ultimate fate was not determined, passed into the general mass 
of longitudinal fibres of the ganglionic chain. 


Misce,tnanzous Exements.—In Th. iv a system of 
elements occurs, which offers certain difficulties in resolving 
it into its component parts. It is found to stain either com- 
pletely or incompletely in embryos at almost all stages of 
development. Its position and relations to the neighbouring 
ganglia are shown in fig. 1 (U, Th. iv), whilst the details will 
be best seen in fig. 4. As may be gathered from the latter 
figure only two cells, belonging to corresponding elements of 
the opposite sides, have stained, but it appears to me to be most 
probable that the system contains several sets of elements, the 
cells of some of which have never taken up the methylene blue. 
The only clue which it has been possible to obtain as to the 
course pursued by the individual elements is the fact that in a 
number of preparations of late embryos (near hatching) the 
portion of the system represented in fig. 5 has been alone 
stained, and may therefore represent a single element. If this 
be the case, the course of the element may be described as 
follows (cf. fig. 1, U., figs. 4 and 5): Starting from a cell in 
the anterior portion of the lateral mass of ganglion cells of Th. 
iv, the fibre runs near the dorsal surface of the ganglion with 
an almost straight course inwards towards its centre, where it 
gives off a pair of branches (figs. 4 and 5, c.) which run down- 


1 Retzius, ‘ Biol. Untersuch. Neue Folge,’ ii, ‘ Zur Kenntniss des centralen 
nervensystems der Wiirmer.’ 


STUDIES ON THE NERVOUS SYSTEM OF CRUSTACEA. Al 


wards and break up into tufts of fibres near the ventral surface. 
The original fibre continues its course to a point a little beyond 
the centre of the dorsal surface of the ganglion, where it 
bifurcates, one branch turning upwards and curving outwards 
to the posterior root of Th. 111 (figs. 4 and 5, a), whilst the 
second branch continues in a transverse direction for some 
distance, and then turns suddenly forwards and runs into the 
neuropile of Th. 111 (figs. 4 and 5,6). Its fate here is a little 
uncertain, but it often presents an appearance which suggests 
that it ends in a tuft of fine branches, as in fig. 4, b. (on the 
right hand side of the figure). It should be mentioned that 
there is one difficulty in regarding the portion represented in 
fig. 5 asa single element, namely, that two central branches 
(c) are stained, whereas there appear to be only two when the 
complete system (fig. 4) is stained. 

Returning to fig. 4, it will be seen that the whole system 
there represented contains, in the first place, two elements 
similar to that in fig. 5, and lying upon opposite sides of the 
ganglion. In addition to this pair of elements, there is a fibre 
upon each side (fig. 4, d.), which appears to start from the 
point where the transverse fibre turns forwards to form the 
branch 6. This fibre (d) runs backwards for a short distance 
and then turns outwards to the posterior root of Th.1v. It is, 
however, probably not simply a branch of the element already 
described, but has a transverse portion of its own running 
parallel and close to the transverse portion of that element. 
In some preparations it is clear that the main transverse fibre 
of the whole system is of a composite nature. This is indicated 
on the right-hand side of fig. 4. 

A third fibre belonging to the system (fig. 4, e) appears to 
spring from near the centre of the transverse fibre, to curve 
forwards, and finally to pass out at the posterior nerve-root of 
Th. 111. It seems to be impossible to determine from embryos 
whether the fibres d and e are independent elements whose 
cells have never stained, or merely branches of the other 
elements of the system. It is to be hoped that the study of 
young adults may throw light upon this point, 


42 EDGAR J. ALLEN. 


Two other fibres in the same region often stain, and are 
shown in fig. 4, fandg. The fibre f appears to enter by the 
posterior nerve-root of Th. 111, to pass transversely across the 
ganglion, and leave by the posterior root of the opposite side, 
the straight course being broken by a slight indentation at the 
middle line. It is not unlikely that there are in reality two 
elements which decussate in the middle line, and are there 
connected with cells which have remained unstained. The 
fibre g has a similar relation to the posterior roots of Th. rv. 

Traces of a system of elements similar to that just described 
occur in Th. 111. 

A pair of elements exists in each ganglion from Th. vi to 
Th. x, which unfortunately seldom stains, but when stained 
presents several points of interest. The appearance is generally 
that shown in fig. 1, Th. rx, W., but in one preparation the 
element stained on one side only of the ganglion, and fortu- 
nately showed the position of the cell. This element, which 
actually occurred in Th. vr is inserted in fig. 1 for the sake of 
clearness in Th. x. (W.). Since in another preparation in 
which the element was stained, a cell just commenced to take 
up a similar position, I feel little doubt that this figure (fig. 1, 
Th. x., W.) represents the true condition of the element. Its 
course may then he described as follows :—Starting from a 
cell, which lies in the anterior portion of the central mass of 
ganglion cells, the fibre passes first backwards, decussates with 
its fellow of the opposite side, and then turns outwards to the 
neuropile, upon entering which it gives off a small tuft of fine 
branches. After giving off this tuft, the fibre continues to run 
transversely through the neuropile to about the level of the 
lateral mass of ganglion cells, where it bends suddenly forwards, 
giving off a short richly arborescent branch at the angle, and 
then runs to the next ganglion in front, in the neuropile of 
which it ends in another tuft of fine branches. No branch 
has stained which passes out through either of the nerve-roots. 
If the above description represents in reality the entire course 
of the element, it must serve to put different parts of two 
adjacent ganglia into communication with each other. 


STUDIES ON THE NERVOUS SYSTEM OF CRUSTACEA, 43 


II. ABDOMINAL GANGLIA. 


Mertuop or Preparation.—Staining of elements in the 
abdominal ganglia can be observed in two ways. In the case 
of embryos at an early or medium stage of development which 
have been prepared, as described in Part I, for the staining of 
elements in the thorax and in which the abdomen is allowed 
to remain undisturbed, fibres which have taken up the methylene 
blue in the thorax often continue to absorb the colouring 
matter in the abdomen, and the cells with which they are con- 
nected are thus brought to light. The best results for the 
abdominal ganglia can, however, be obtained by special pre- 
paration of embryos which are very near the hatching point. 
In such embryos the abdominal ganglia may be dissected out 
from the surrounding tissue by careful manipulation with 
needles. Special care must be taken not to injure or stretch 
the ganglia and their continuity with the ganglia of the thorax 
should be maintained. Ifthe embryos, thus prepared, be placed 
with the dorsal surface uppermost in very dilute methylene 
blue (1: 100,000 may be used to commence with, and the 
strength gradually increased), satisfactory staining of many of 
the elements of the abdomen will soon take place. 

A variation of this method, which often gives good results, 
consists in removing, or even simply tearing, with needles the 
cuticle on the ventral surface of the abdomen (after having 
first turned the abdomen backwards, and caused it to lie in a 
line with the thorax), and allowing the embryo to lie in the 
methylene blue with the ventral surface uppermost. In this 
case, as soon as the staining is thought to be satisfactory, the 
ganglia must be dissected out before they are examined. A 
cover-glass may be placed upon them, and the elements rapidly 
drawn, or the preparation may be fixed with ammonium picrate, 
and mounted in glycerine diluted with an equal volume of 
saturated solution of the picrate. Such preparations, however, 
only retain their full colour for a comparatively short time. 
For practical purposes a large number of fresh preparations, 
examined with a cover-glass, have been found more instructive 


44, EDGAR J. ALLEN. 


and satisfactory. The same elements are seen again and again 
with such frequency, that little doubt can remain as to the 
accuracy of an observation. 

The elements of the abdomen are similar in kind to those 
described for the thorax, and will be considered under the 
same general headings. 


ELEMENTS OF WHICH BOTH CELL AND FIBRE LIE ENTIRELY IN 
THE Gane ionic Cua1n.—In each of the abdominal ganglia, 
from Abd. 11 to Abd. v1, a pair of elements similar to that 
represented in fig. 6, B (2) has been found to exist. Each 
element has its origin in a large cell situated in the anterior 
lateral portion of the ganglion near its ventral surface. From 
this cell a fibre arises which runs inwards and backwards, de- 
cussating with its fellow of the opposite side at the middle line 
and subsequently giving rise to two fibres, one of which runs 
backwards, whilst the other runs forwards towards the brain. 
The backward branch breaks up into fine fibres in the neuro- 
pile of the ganglion in which the cell lies. The forward branch 
runs as a longitudinal fibre along the ganglionic cord and in 
all probability enters the brain. Unfortunately, owing to 
difficulties of technique, it has never been possible to trace a 
single longitudinal fibre of one of these elements through its 
entire course, but a consideration of the whole evidence leaves 
little doubt as to their entrance into the brain. One of the 
most satisfactory direct observations was made upon an embryo 
near the point of hatching, in which the abdominal ganglia 
had been dissected out, as described above. In such prepara- 
tions the pair of elements under consideration most frequently 
stains in Ab. 11, and the fibres can generally be traced forwards 
through Th. x and 1x. In one preparation, however, where 
there was a wound at the level of Th. 111, the fibres of the 
elements upon both sides were distinctly visible as far forwards 
as this point. 

But it is in embryos at an earlier stage, in which the 
abdomen has been left uninjured, that these elements most 
frequently appear. In order to make the evidence clear it is 


STUDIES ON THE NERVOUS SYSTEM OF ORUSTACEA. 495 


necessary to again draw attention to the method of preparation 
adopted for such embryos, when it is desired to obtain staining 
of the longitudinal fibres. The yolk is removed with needles, 
and the embryo placed with the exposed thoracic ganglia upper- 
most in the dilute methylene blue, the abdomen being allowed 
to remain turned in underneath the thorax. The ganglionic 
cord is then cut across, generally at tlie level of the cesophagus, 
and the methylene blue enters the fibres at the wound. In 
preparations so made the elements under consideration (B (4), 
figs. 6 and 7) have frequently taken up the blue, and can be 
well seen on turning the embryo over and lifting back the 
abdomen. The longitudinal fibres of the posterior ganglia 
can be traced through the anterior ones and show no sign of 
terminating. They are clearly continued forward into the 
thorax, although they cannot be individually followed on 
account of the flexure of the abdomen. On turniug the 
embryo back again so that the thoracic ganglia can be ex- 
amined, all the longitudinal fibres coming from the abdomen 
are seen to commence at the wound which has been made, 
and when this wound lies at the level of the cesophagus it is 
obvious that they must be fibres which enter the brain. No 
endings of longitudinal fibres coming from the abdomen have 
ever been observed in the thorax, and preparations of the 
kind just described have been so frequently made that there 
can be practically no doubt that the fibres in question enter 
the brain. 

These elements (B (0), figs. 6 and 7) are evidently the same 
as those seen by Retzius in the abdominal ganglia of the adult 
Astacus. (See ‘ Biol. Untersuch.’? Neue Folge I, pl. xi, fig. 1; 
pl. ix, fig. 4.) 

A pair of elements of this kind, as has been already stated, 
occurs in Abd. vi, and is represented in fig. 7. In this ganglion 
a second pair of elements, whose cells are situated in the 
posterior portion of the ganglion, also frequently stains. The 
fibres from these cells (fig. 7, B (c)) have been traced as far 
forwards as the abdominal flexure, in preparations in which 
the only wound in the thorax has been at the level of the 


46 EDGAR J. ALLEN. 


cesophagus, and they probably therefore enter the brain. The 
fact that two pairs of elements in Abd. vi send fibres to the 
brain, whilst only one pair has been found in the other abdominal 
ganglia is in accordance with the known composite nature of 
this ganglion. 


PropasLe Moror Evements.—Of elements consisting of a 
cell in a ganglion and a fibre passing out of the cord by one 
of the nerve-roots, two principal kinds have been found in 
the abdomen, as in the thorax, namely (1) those in which the 
fibre passes out through one of the roots of the ganglion in 
which the cell lies, the whole of the element within the central 
nervous system being confined to one ganglion; and (2) those 
in which the fibre passes out through a nerve-root of some 
ganglion other than that in which the cell lies. 

These elements are represented in figs. 2and 3; fig. 2 giving 
the appearance presented when the elements upon both sides 
of a ganglion are stained ; fig. 3, that when they are stained 
upon one side only, and indicating therefore the course of the 
individual elements. Each of the elements to be described 
has been found in all the ganglia from Abd. 1 to Abd. v. 


ELEMENTS CONFINED TO ONE GaNnGLION.—The element a, 
shown in figs. 2 and 3, takes its origin in a cell, which lies 
near the centre of the ganglion and at its ventral surface. 
The fibre passes first upwards and outwards, turns inwards, 
and after a short course divides into two branches, one of which 
passes out of the ganglion by the posterior nerve-roots, whilst 
the other runs across to the opposite side, keeping close to the 
corresponding branch from the element of that side (fig. 2) 
and subsequently breaks up on the opposite side of the gang- 
lion. The fibre gives off numerous fine branches during its 
course through the ganglion, which have not been represented 
in the somewhat diagrammatic figs. 2 and 3, but which may 
be seen in fig. 8, drawn from a preparation fixed with am- 
monium picrate (fig. 8, a; Abd. 1 and 111). 

Element 6 (figs. 2, 3 and 8) is similar in most respects to 


STUDIES ON THE NERVOUS SYSTEM OF CRUSTACEA. 47 


the element just described. It differs chiefly in the position 
of the cell, which lies in the posterior lateral portion of the 
ganglion. The fibre curves first upwards and forwards, and 
then downwards, when it divides into two branches similar to 
those of element a. 

The element J (4), which has already been described as 
occurring in the last ganglion of the thorax (fig. 1, Th. x1) is 
found also in the ganglion of the abdomen (Abd. 1-v), and is 
shown in figs. 2 and 3. The cell lies near the middle of the 
posterior end of the ganglion. The fibre runs at first forwards 
and outwards, and then turns inwards and backwards, giving 
off a small branch to the neuropile at the angle. Subsequently 
it divides into two branches, one of which passes out at the 
posterior root, whilst the other runs over to the opposite side 
of the ganglion, where it breaks up in the neuropile. 

The element D (figs. 2 and 3) arises from a comparatively 
small cell, which lies in the anterior half of the ganglion near 
the middle line. The fibre runs backwards and slightly out- 
wards, decussates with its fellow of the opposite side forming 
the characteristic figure represented in fig. 2, gives off a small 
arborescent branch to the neuropile, and then runs backwards 
to the posterior root of the ganglion through which it passes. 


ELEMENTS NOT CONFINED TO ONE GaNnGLion.—Two pairs of 
elements have taken up the stain in each of the abdominal 
ganglia, whose fibres pass out through one of the roots of the 
ganglion immediately anterior to that in which the cell lies. 

The element e (figs. 2 and 3) has the cell situated near the 
middle line at the anterior end of the ganglion. From the 
cell, which is small, the fibre runs outwards and backwards, 
and then, turning sharply inwards, runs as a transverse fibre 
close to its fellow of the opposite side. On reaching the other 
side it turns forwards, giving off an arborescent branch at the 
angle, runs into the next ganglion in front, and there passes 
out at the posterior root. When the elements upon opposite 
sides are stained it is generally impossible to distinguish the 
two transverse fibres, which lie so close together in this and 


48 EDGAR J. ALLEN. 


in the other similar cases in the abdominal ganglia, that they 
appear as one. 

The element S (figs. 2 and 3) is similar to the element S 
already described in the thorax (Th. x1, fig. 1). The cell, 
which is very small, lies near the centre of the ganglion and 
at the ventral surface. The fibre runs first upwards, back- 
wards, and outwards, and then curves forwards and inwards. 
After running for some distance in this direction it decussates 
with its fellow of the opposite side, the two fibres lying for a 
short distance close together. The fibre then turns forwards 
and outwards, gives off an arborescent branch to the neuropile, 
and then proceeds forwards in a straight line to the ganglion 
in front, where it passes out by the middle root of the three 
which spring from that ganglion. 

From the foregoing description of the motor elements found 
in the abdomen, it will be noticed that they, in nearly every 
case, supply fibres to the posterior nerve-roots of the ganglia, 
whilst the greater number of those described for the thorax 
in this and the previous paper send their fibres to the anterior 
nerve-roots, a few only supplying the posterior roots. The 
probable reason for this difference is not difficult to find. 
The anterior roots chiefly supply the appendages, which are 
well developed in the thorax of the embryo, whilst in the 
abdomen they are wanting. The posterior roots, on the other 
hand, supply the muscles of the body itself, which are well 
developed in the abdomen, but less so in the thorax owing to 
the considerable space occupied by the still unabsorbed yolk. 


III. ELements Arisinc FRoM CELLS OUTSIDE THE 
CENTRAL Nervous SystEmM. 


In Part I a number of elements were described, arising from 
cells which lay in the ectoderm of the dorsal surface of the 
abdomen. These cells give off fibres which enter the abdo- 
minal ganglia and there bifurcate in the Y-shaped figure, 
which is characteristic of sensory nerve-fibres in all divisions 
of the animal kingdom, sending one branch forwards and the 


STUDIES ON THE NERVOUS SYSTEM OF ORUSTACEA. 49 


other backwards along the ganglionic cord. Numerous ele- 
ments of this kind have stained in embryos near the point of 
hatching, and it has been possible to trace the fibres inside 
the central nervous system for considerably greater distances 
than was done before. 

In the thorax, elements having the characteristic Y-shaped 
bifurcation enter the various ganglia by the posterior roots. 
One such fibre is represented in fig. 1 at M, Th. x. From this 
ganglion (Th. x) it has been traced forwards by direct ob- 
servations as far as Th. 1, but there was no indication of a 
definite ending there, the blue colour becoming gradually less 
distinct and suggesting that the true termination had not been 
reached.! From Th. x1 the element has also been traced as 
far as Th. 1. 

In the abdomen, elements of this kind enter by the middle 
root of the three which belong to each ganglion. From Abd. 
1 the forward branch has been actually traced as far as Th. 11, 
but gave no indication of a definite ending there. 

From the three ganglia already mentioned the forward fibre 
has been directly observed to pass through at least nine or ten 
ganglia. Cases in which the fibre could be traced from these 
and neighbouring ganglia through five, six, and seven ganglia 
occurred in a large proportion of the numerous preparations of 
late embryos that were made. The fibres from spindle-shaped 
cells lying in the telson can, in embryos of moderate age, be 
seen to enter the last abdominal ganglion and then to run 
forward through the anterior ganglia. Such fibres have been 
traced as far as Abd. 1, but could not be followed further on 
account of the flexure of the abdomen. 

The fact mentioned in Part I, that elements of the kind 


1 If an element stains at all frequently it is generally possible to form an 
opinion as to whether the actual termination has been reached, or whether 
the staining is incomplete. In the former case the terminal portion is, gene- 
rally, at least as deeply stained as any other part of the element, and it may 
continue to take on a stronger colour after there has been a considerable 
fading of the other parts. In the case of incomplete staining, on the other 
hand, the colour becomes more and more faint towards the end of the fibre 
until it finally disappears. 


VoL. 39, PART 1.—NEW SER. D 


50 EDGAR J. ALLEN. 


under consideration frequently stain in the abdomen, when 
the longitudinal fibres commence to take up the methylene 
blue at a wound made at the level of the esophagus, from 
considerations of a similar nature to those already adduced in 
the case of the abdominal elements B (4), leaves little doubt 
that the fibres actually go to the brain. The observations 
now recorded, although they do not directly demonstrate the 
point, render it still more probable. 

With regard to the second branch given off by these ele- 
ments after entering a ganglion, which is directed backwards 
along the ganglionic cord, no definite termination has been 
found. It has never been seen to pass through more than 
three ganglia, and can generally only be followed through two. 
This, however, is probably due to incomplete staining, and the 
entire course of this branch remains yet to be determined. 


EXPLANATION OF PLATE 4. 


Illustrating Mr. Edgar J. Allen’s “ Studies on the Nervous 
System of Crustacea.” 


Fic. 1.—Brain and thoracic ganglia of Homarus embryo. es, (so- 
phagus. ¢r.dr. Transverse bridge behind cesophagus. com. Csophageal 
commissure. azt.11. Ganglion of Antennaul. Th.1—1x. Thoracic gan- 
glia. For individual elements see text. Somewhat diagrammatic. 

Fic. 2.—Three abdominal ganglia of Homarus embryo. Motor elements 
inserted upon both sides. For individual elements see text. Somewhat 
diagrammatic. 

Fic. 3.—Ditto. Motor elements inserted upon one side only. 

Fics. 4 and 5.—System of nerve elements in Th. m1 and Th. Iv of Ho- 
marus embryo. 

Fic. 6.—Abdominal ganglion of Homarus embryo. B (4). Element 
sending fibre to brain. 

Fie. 7.—Sixth abdominal ganglion of Homarus embryo. B (4) and 
B (c). Two elements sending fibres to brain. 

Fic. 8.—Second and third abdominal ganglia of Homarus embryo. 
Camera drawing from preparation preserved in ammonium picrate. a. 0. 
motor elements. 


NOTES ON OLIGOCHATES. 51 


Notes on Oligochetes, with the Description of 
a New Species. 


By 


Edwin 8. Goodrich, B.A., 
Assistant to the Linacre Professor of Comparative Anatomy, Oxfoad. 


With Plates 5 and 6. 


Tue following observations were chiefly made on some 
worms sent to me by Mr. Damon from a garden at Weymouth. 
I have to thank Professor Ray Lankester for much help during 
my researches. 


On the Structure of Enchytreus hortensis, n. sp. 


At first sight this little Enchytreid presents no very dis- 
tinctive characters, It is, when full grown, about 15 mm. in 
length, and milky white in colour, the anterior end being 
sometimes yellowish. The chetz (fig. 18) have a straight 
shaft and a hooked inner end, but this hook is not always so 
pronounced or so sharp as in the cheta figured. The bundles 
contain from three to four chet, generally three in the 
dorsal and four in the ventral bundle. There are no dorsal 
pores. A small dorsal head-pore is situated between the pro- 
stomium and first segment. The brain (fig. 1, 67.) is larger 
behind than in front ; its posterior margin is slightly indented 
or rounded. The dorsal blood-vessel arises from the wall of 
the intestine about the 17th segment. Three pairs of septal 
glands (fig. 1, sept. g/.) are present; two sets of bulky glands 
in front, and a third pair of smaller glands behind, each of 
which is subdivided into two lobes. The well-developed sali- 


52 EDWIN S. GOODRICH. 


vary glands (fig. 1, sa/. gl.) are unbranched and of considerable 
length. The nephridium (fig. 2) is compact, with a portion of 
the canal slightly coiled in the pre-septal region. The funnel 
of the sperm-duct is of moderate length and very thick-walled. 
The spermatheca has a distinct duct beset with glandular cells 
near its external aperture, and possesses a posterior blind 
sac (fig. 23, post. sac). Three chief varieties of corpuscles are 
found in the celom ; one of these bears a peculiar product in 
the shape of a coiled thread, to be described in detail further 
on. In a favourable light the cuticle can be seen in most 
specimens to be covered throughout with fine hair-like pro 
cesses (fig. 8,i.p.), similar to those described in Vermiculus(8). 
This species is probably distributed all over England, as I 
have found it near Oxford, London, and Weymouth. 


On the Nephridium of Enchytreus hortensis. 


I have thought it worth while to give in detail the results 
of observations carried on for some years on the nephridia, 
since not only does a great deal remain to be described in 
these interesting organs, notwithstanding the large number of 
authors who have written on the subject, but also on account 
of the recent publication by M. Bolsius of descriptions totally 
at variance with the observations of other naturalists and my 
own (4). 

As M. Bolsius gives an account of the literature, I need 
only say that former writers (Claparéde, Eisen, Vejdovsky, 
Michaelsen, Ude, &c.) all agree in describing the nephridium 
of Enchytreeids as consisting of a more or less compact mass 
of cells pierced by a continuous ciliated convoluted canal 
leading from the nephrostome to the exterior. Bolsius, on 
the contrary, declares that the lumen of the canal forms a 
complex network “ des canalicules anastomosés dans toutes les 
directions ;” and, moreover, that it is totally devoid of cilia: 
‘Les canalicules et le canal collecteur ont été dessinés par 
nous sans cils vibratiles : c’est que nous n’en avons pas trouvés, 
ou plutdt c’est que nous en avons constaté l’absence compléte 
dans nos nombreuses préparations.” Since the first of these 


NOTES ON OLIGOCHATES. ba 


statements has been included in the excellent monograph Mr. 
Beddard has just published (2), it seems all the more important 
that it should be criticised before it becomes generally 
accepted. 

Let us now examine the structure of the nephridium in En- 
chytrzus hortensis as seen in the living animal. The organ 
(fig. 2) consists of an oval main body, flattened from side to 
side, and narrowing in front to a neck, which piercing the 
septum ends in an open funnel (fig. 2, neph. st.). Behind the 
nephridium passes into a lobe leading to the body-wall. The 
protoplasm of the nephridial cells, whose boundaries I have 
been unable to distinguish, is very granular, especially in the 
anterior region, which is frequently of a brownish tinge. The 
funnel (figs. 8—5) has a protruding upper lip (up. J.) and 
a truncated lower lip (/./.). At the extremity and along the 
edges of the upper lip are set numerous very long extremely 
fine waving cilia (ezt. cil., figs. 4 and 5), while from the inner 
surface springs a large bunch of powerful undulating cilia (fl.). 
As far as I have been able to determine, cilia are not present 
on the lower lip; but rising from its middle region is a peculiar 
sharp protoplasmic process (m. p., figs. 4 and 5), sometimes 
branched and ragged. 

Already in Nereis (7) and Vermiculus (8) I have drawn 
attention to the two kinds of cilia attached to the lip of the 
nephrostome, which indeed have not escaped the notice of 
previous authors, but the exact disposition and differences 
of which have not heen sufficiently insisted upon. Whilst the 
inner beat rapidly, simultaneously, and rhythmically, forming 
an undulating bunch! or “ flame,” which, no doubt, propels 
liquid down the lumen of the canal; the long external cilia, 
on the other hand, do not beat in unison: each cilium seems 
to move independently with trembling motion, sweeping down 
towards the protruding process of the lower lip. These cilia 
seem rather to guard the mouth of the funnel and select 


1 The seat of the motion seems to be at the base of the cilia, the wave 
passing along them as an undulation passes along a cord shaken at one ex- 
tremity. 


54 EDWIN S. GOODRICH. 


particles from the celomic fluid than to propel it down the 
canal, 

Passing backwards from the nephrostome is the nephridial 
canal (neph. c., figs. 2, 3, and 5). I may say at once that, 
although I have not been able to follow it throughout its whole 
extent with certainty, owing to the opaque and granular nature 
of the post-septal region and the small dimensions of the 
tortuous lumen, I believe the main duct to be continuous and 
single (i.e. without anastomosing branches), except in that 
region which leads directly to the exterior, where the lumen 
occasionally undergoes secondary subdivision (figs. 2 and 7). 
Although, therefore, I cannot claim to be able to prove the 
unbranched character of the canal, yet amongst the number of 
Enchytreid nephridia I have had under observation during 
the last three or four years I have never come across a single 
undoubted case of such anastomosing. The course of the canal 
drawn in fig. 2 does not profess to be accurately represented in 
all its details, nevertheless I have been able to observe that at 
least the main features of the canal’s course, if not the less 
important of its convolutions, are constant, not only in the 
different nephridia of the same worm, but also in the nephridia 
of different individuals of the same species. For the first part 
of its course it has a very narrow lumen, and is not ciliated. 
It then widens slightly, and undergoes several complicated 
twists in the more posterior and dorsal region of the main body 
of the nephridium. It is along this second region of the canal 
that here and there we find a slight oval enlargement of the 
lumen, in which is situated a bunch or “ flame” of cilia (ce. a., 
fig. 2), similar to the “flame” in the nephrostome, and con- 
stantly driving the liquid in one direction, namely, to the 
exterior. As far as I can make out, there are no more than 


1 Anyone examining the living worm could hardly miss seeing these cilia. 
It is, of course, just possible that the worms observed by M. Bolsius were of 
exceptional structure, but I am more inclined to think that he was misled, 
both with regard to the absence of cilia and the anastomosing of the canal by 
employing the method of sections alone. 


NOTES ON OLIGOCHMTES. 55 


five such ciliated regions in each nephridium.' Their essential 
similarity to the ‘ciliated ampullz ” found in the nephridium 
of Vermiculus (8) scarcely needs pointing out. From the 
second region the canal passes to a third, where the lumen is 
often wider still, and is entirely devoid of cilia. The course of 
this, the terminal region of the canal, lies along the postero- 
dorsal edge of the main body, and from thence passes through 
the posterior limb to the external pore. It will be noticed that 
the blind sacculations of the canal (mentioned by Michaelsen, 
9a) are here very large (fig. 2, d/. b7.), and frequently a sub- 
division of the lumen takes place, forming a more or less com- 
plicated network (figs. 2 and 7), which must not, however, be 
confused with the supposed anastomosing branches between 
the different regions of the canal. Finally the lumen 
expands to form a vesicle (e. v.), which opens to the exterior 
by a small transversely elongated pore in front of the ventral 
chete. 

Sections only confirm these observations. No limits are 
distinguishable between the cells. The funnel has two nuclei 
in its upper lip near the point of insertion of the cilia (fig. 3), 
and one at the base. Small deeply staining nuclei are seen 
in the main body (n., fig. 3), and larger nuclei are distributed 
apparently in more intimate connection with the second and 
third region of the nephridial canal (n., fig. 6). 


On the Celomic Corpuscles of Enchytrzus 
hortensis. 


Most characteristic of this worm are the celomic corpuscles ; 
these are of three kinds. Floating in the ccelomic fluid, or 
creeping on the surfaces exposed to it, may be seen a small 
number of ameeboid corpuscles of very variable character. 
Some (fig. 28, a.) are more or less spherical with short- pointed 
processes ; their protoplasm is pretty clear, and contains, as a 
rule, a few highly refringent round granules, apparently of a 

1 M. Bolsius (4) justly criticises previous authors for describing the canal 


as ciliated, and yet not figuring any cilia. M. Roule (10) alone does so, and 
represents them erroneously along the entire length of the canal. 


56 EDWIN 8S. GOODRICH. 


fatty nature. Others are formed of more opaque and finely 
granular protoplasm, prolonged into fine branching processes 
of great length (fig. 28,.): corpuscles intermediate in character 
between these two forms also occur. 

The second variety is very characteristic of the Enchy- 
treide in general, and has been described by almost every 
writer on this group of the Oligocheta. Much larger than 
the amceboid form (nearly twice the size), these corpuscles 
are of regular and constant oval shape, somewhat flattened, 
thicker in the middle than at the periphery, and frequently 
bluntly pointed at one or both ends. The body of the cell is 
made up of large spherical, clear, and refringent granules, 
embedded in a delicate meshwork of protoplasm (fig. 15, a). 
The nucleus is round or slightly oval, in which case its long 
axis is invariably situated at right angles to that of the cor- 
puscle ; staining reveals a dark nucleolus lying in the less 
deeply staining nucleus, n (fig. 12, a). These corpuscles are 
found either freely floating in the ccelom, or attached to the 
body-wall or septa by means of a short stalk, apparently of 
cuticular nature. In the latter case the corpuscle is always 
fixed by its more pointed extremity. The remains of the stalk 
of attachment may generally be detected on those corpuscles 
which float freely (st., fig. 15, a). Although my observations 
as to the origin of these corpuscles are unfortunately neither 
complete nor conclusive, yet I am strongly inclined to believe 
that they arise from the coelomic epithelium lining the body- 
wall and septa. Certain it is that smaller, flatter, and younger- 
looking corpuscles of similar structure may sometimes be seen 
attached to the epithelium. The fact that similar apparently 
young corpuscles may occasionally be seen floating freely in 
the fluid does not seem to me fatal to the view that normally 
they remain attached to their place of origin until they have 
reached their complete development. Never have I found 
intermediate stages between these oval corpuscles and the 
ameeboid variety, or the granular cells of the modified epithe- 
lium covering the gut, from which they differ markedly, as 
already noticed by Vejdovsky (13). In the youngest worms I 


NOTES ON OLIGOCHATES. ay) 


have been able to examine, the oval corpuscles are relatively 
about as abundant as in the adult; but the free ones appear 
to be more rounded and flattened, and with fewer granules, 
thus resembling what I have taken for the younger stages in 
the adult. As no definite signs of extensive disintegration of 
the corpuscles are apparent, and as I have never observed any 
case of their multiplication by division, it would seem that the 
life of each individual corpuscle must be of considerable 
length, and that they are but rarely renewed. 

The third and most interesting kind of ceelomic corpuscle 
should probably be reckoned as a variety of that just described. 
The body of the cell, though smaller, is formed of a similar 
meshwork, enclosing granules of identical appearance and pro- 
perties ; the nucleus also resembles that of the ordinary oval 
corpuscle (n., fig. 12), but lying on one side is a colourless 
refringent body of peculiar structure. This body, in the 
fully developed form, is of the shape of a thick disc or of a 
truncated cone, thicker at the edge than in the centre (thr., 
figs. 12—15), the flat surface being next to the nucleus. 
When the celomic corpuscles are squeezed out of the worm 
under a cover-glass, and they come into contact with some 
strange fluid, such as distilled water, or even salt solution, it 
becomes apparent that the refringent body is formed of a 
long thread of transparent homogeneous substance, 
closely coiled like a rope (fig. 13). Surrounded by a thin layer 
of protoplasm the thread-coil is obviously an endoplastic pro- 
duct of the ccelomic corpuscle. At first I took these strange 
bodies for parasites, but subsequent observations have proved 
this view to be false. 

As already mentioned, immediately the corpuscle comes 
into contact with some foreign liquid, such as distilled water, 
weak acids, or alkalies, the body of the cell begins to dis- 
integrate, the coil of thread swells and gradually unwinds, 
forming a tangle of loops, amongst which I have never de- 
tected free ends (fig. 13). The thread appears to be thicker 
in some places than in others, or perhaps it would be more 
correct to say that the coils stick together in certain regions, 


58 EDWIN S. GOODRICH. 


Like the ordinary oval cells, the thread-corpuscles are found 
both floating and attached by a stalk to the body-wall or septa. 
In the latter case the thread-coil is invariably situated on 
that side of the nucleus which is far from the point of attach- 
ment. Looking in the adult worm for possible stages in the 
development of thread-corpuscles, smaller cells can be found 
in which the coil is smaller relatively to the granular cell 
body (figs. 10, 11,d, e, f, and 12,6). Such small and appa- 
rently young cells are scarce, and generally attached. Turning 
now to very young worms, we find in them the thread- 
corpuscles scarcely less abundant than in the adult relatively 
to the whole number of ccelomic corpuscles. They are, how- 
ever, as a rule, flatter and more oval in shape (fig. 9), with 
more loosely packed granules and smaller thread-coils. In- 
deed, these refringent bodies differ from those of the adult, 
not so much in the circumference of the disc as in its thickness 
(cp. figs. 9, d, and 138, a, both side views). Similar shallow 
discs occasionally occur in the adult (fig. 10, a). It will be 
seen, moreover, that in what I have taken to be the younger 
stage of development the disc presents its edge to the nucleus 
(figs. 9 and 10) ; whilst in the later stages the nucleus is lodged 
on its flat surface (figs. 12, c, and 13). Whether the disc has 
during growth turned round or merely changed shape is a 
doubtful point. These facts lead to the conclusion that the 
thread-corpuscles originate from the coelomic epithelium lining 
the body-wall or septa, passing through stages essentially 
similar to those of the young corpuscles of the ordinary oval 
type, and are not formed by the development of a thread 
within full-grown oval corpuscles. The thread itself would 
appear to be formed at the expense of the granules, cp. (figs. 
9,10, and 13); but this appearance may be deceptive, and 
merely due to the increase in size of the thread alone. The 
chemical nature of the substance composing the thread is 
discussed in detail below; it differs widely in its chemical 
properties from that of which the granules are formed, and 
from chitin. 

On the function or fate of the thread in the ccelomic cor- 


NOTES ON OLIGOCHATES. 59 


puscle of this Enchytreid I am sorry to say that my observa- 
tions throw no light. It occurs in immature and mature 
worms at all seasons of the year. It has been searched for in 
vain, either discharged into the coelomic fluid or amongst the 
tissues of the body. Nor have empty cells been found which 
could have discharged the thread. It might be suggested that 
this thread is an excretory product ; however, as far as I have 
been able to make out, it is never got rid of,—indeed, it would 
seem to be most ill adapted for passing through the compli- 
cated coils of the nephridial canal. Possibly it may be a waste 
product of harmless nature, which is allowed to remain and 
accumulate throughout life, but here also facts are wanting to 
enable us to come to any definite conclusion. 

Amongst the other Enchytreeids or the Oligochetes in gene- 
ral no products in the celomic corpuscles which could be 
compared to these threads seem to have been hitherto observed. 
Amongst the Polychetes, on the other hand, there have been 
described in Ophelia radiata ameeboid celomic corpuscles 
containing a chitinous rod! of unknown function, first noticed 
by Claparéde (5), and recently studied in great detail by Dr. 
Shaepi (11). The latter adopts the view urged by Dr. Hisig (6), 
that these rods of chitin are of the nature of an excretory 
product.? 

In the epidermis of many fish and amphibia cells have been 
described which secrete a thread-like substance. 

Perhaps the thread-cells of Myxine, described by Johannes 
Miller, Blomfield (8), and others, resemble more closely the 
thread-corpuscles of Enchytrzus hortensis than any other 
cell hitherto known, especially in the earlier stages of their 
development. But here, again, the product appears to be of 
different chemical nature, and the cells in their origin and 
situation are so different that they scarcely help us to solve 
the problem. 


1 T have found similar but longer chitinous rods in Ophelia limacina. 

2 As the thread in Enchytreus hortensis is certainly not of chitin, a 
comparison with Ophelia (even supposing the excretory nature of the rod had 
been proved) would not warrant our coming to a similar conclusion, 


60 EDWIN S. GOODRICH. 


On the Spermatheca of Enchytreids. 


In 1885 (9) Dr. Michaelsen first made known the remarkable 
discovery that the spermatheca in the Enchytreidz, which lies 
in the 5th segment, communicates directly with the alimen- 
tary canal. More recently Mr. Beddard (1) has shown that 
the same thing occurs in Sutroa, and he discusses the subject 
in the monograph already mentioned (2, pp. 127-8). As such 
a case of the direct communication of the cavity of the ceso- 
phagus with the exterior by means of a sac invaginated from 
or derived from the epidermis‘is unknown in any other group 
of Invertebrates, and is of great morphological interest, I give 
a series of longitudinal and transverse sections through these 
organs. Fig. 27 represents a longitudinal horizontal section 
through the 5th segment of Marionia enchytreoides 
St. Loup.! 

As the section is slightly oblique it shows on one side the 
opening of the spermatheca to the exterior (sp. op.), and on the 
other the opening into the alimentary canal (éné. op.). In 
Pachydrilus, and forms closely allied to it, such as Marionia, 
the spermatheca is simple, and opens into the cesophagus at 
its hinder extremity. The spermatheca figured contains no 
spermatozoa, showing that the internal opening is formed 
before the spermatozoa are received, and the point at which 
the fusion has taken place between the two organs can be 
detected, since the lining of the csophagus is ciliated (cs.), 
while that of the spermatheca is not (sp.). 

In Enchytreus hortensis the spermatheca fuses with 
and opens into the esophagus at a point more nearly opposite 
its external opening, sending backwards a large blind sac in 
which are lodged the bulk of the spermatozoa. 

It will be seen that the communication between the sperma- 
theca and the alimentary canal is very wide (figs. 21 and 25, 
and the diagram, fig. 23); the external spermathecal pore 
remains of normal size (sp. op.). The spermatozoa (spiz., figs. 

1 Professor Marion kindly sent me a number of these worms from Mar- 
seilles. 


NOTES ON OLIGOCHATES. 61 


20—25) arrange themselves round the walls of the sperma- 
theca, with their long heads deeply sunk in or between the 
lining cells, and their tails hanging into the cavity. Many 
bundles of spermatozoa also lie loose in the cavity, and may 
get carried into the alimentary canal (Michaelsen, 9). 


On the Celomic Corpuscles of Pachydrilus, sp. (?). 


It is not without hesitation that I put on record the following 
very incomplete observations on a small species of Pachydrilus, 
of which I unfortunately only found two specimens amongst 
the Enchytreids sent to me by Mr. Damon from Weymouth. 

On examining the worm under the microscope, the coelom 
was seen to contain a number of remarkably long and slender 
cells, attached to the body wall by one extremity (fig. 16). The 
longest of these cells reached nearly across a segment, being 
abont ten times as long as an ordinary celomic corpuscle 
(cp., a., and c., fig. 17). Near the middle was an oval clear 
region, indicating the position of the nucleus (n., fig. 17), whilst 
the body of the cell is formed of granular protoplasm, in which 
the granules could be observed circulating. No thick or 
definite wall was apparent covering the somewhat irregular 
surface, and the whole cell appeared to be very flexible, being 
twisted about backwards and forwards with the motion of the 
coelomic fluid. An expanded sucker-like foot or base ( ft.) 
served to fix it to the body-wall. In stained sections the 
nucleus appears small and rounded as in the ordinary oval 
corpuscles. 

Although it may seem extremely probable that these strangely 
shaped cells were parasitic Protozoa, yet the fact that there 
were present cells representing in appearance every stage 
between the elongated form and the ordinary ceelomic corpuscle 
(fig. 16) suggests that the former are derived from the latter. 


62 EDWIN S8. GOODRICH. 


Notes on the Action of Chemical Reagents on the 
Celomic Corpuscles of Enchytreus hortensis 
and Vermiculus pilosus.} 


In a former paper I described the characteristic coelomic 
corpuscles of Vermiculus pilosus (8, fig. 28, pl. 28); they 
are spherical cells, filled with large oval refringent granules. 

If these be watched under the high power of the microscope 
in distilled water, or even in normal salt solution, the large 
white granules will be seen to dissolve, whilst minute and 
yellower-looking granules come into view.? Under the in- 
fluence of weak alcohol, a solution of iodine, very weak acids, 
or alkalies the same thing occurs. With osmic acid (1 per 
cent.) the dissolution of the large granules takes place with 
almost explosive rapidity, the small granules in this, as in the 
other cases, remaining sharply defined. Whether the smaller 
granules are hidden by the larger, and only become visible 
when these have been dissolved, or whether they remain as an 
actual residue from the large granules, is a point difficult to 
settle. I feel convinced from appearances observed that the 
large granules really include the smaller. 

It is difficult, if not impossible, to permanently fix the large 
granules. Sometimes they could be fixed for a time by a 
sudden drenching with absolute alcohol, and they would then 
be unaffected by ether. 

The small highly refringent granules, which remain when 
the large have disappeared, are insoluble in water, iodine solu- 
tion, alcohol, and ether. They are soluble in a solution of 
caustic potash, the mineral acids, in oxalic acid, and with 
difficulty in acetic acid. 


1 T am indebted to Professor Gotch for many useful hints while conducting 
these experiments. 

2 The difficulty of conducting chemical experiments on small bodies which 
have to be watched under the high power is very great, and often renders the 
result uncertain. I have, therefore, always repeated the experiments many 
times over, and used strong solutions of the reagents (unless otherwise 
stated). 


NOTES ON OLIGOCHATES. 63 


Turning now to the oval corpuscles of Enchytreus hor- 
tensis, we find here also white granules soluble in distilled 
water, in potash, lime water, acetic and tannic acid, with diffi- 
culty (perhaps not entirely) in oxalic acid, and also with diffi- 
culty in weak mineral acids, but easily in strong. When 
treated with osmic acid (2 per cent.) these granules are gene- 
rally reduced to smaller and more refringent bodies. This 
also takes place under the influence of alcohol, strongly recall- 
ing the much more pronounced action in the large granules in 
Vermiculus. The white granules are insoluble in ether, and 
weak mineral acids are unaffected by silver nitrate and by 
iodine. They do not answer to the xanthoproteic test, and 
are not doubly refractile. 

It would seem possible that we have in these granules, 
more firmly combined, the same two substances which are so 
easily separable in the corpuscles of Vermiculus. 

The Thread.—The coiled thread in the celomic corpuscles 
of Enchytreus hortensis is insoluble in distilled water (hot 
or cold) and salt solution. Iodine solution shrivels it slightly, 
and stains it yellowish brown. In strong alcohol the thread 
contracts to an irregular mass, which is not farther acted upon 
by ether. In strong potash (30 per cent. sol. of KOH) the 
thread is not dissolved in the cold; but as soon as the cell 
has been destroyed it uncoils, forming very characteristic 
U-shaped loops, the two limbs of the U being closely applied 
(fig. 14). When heated in strong potash portions of the 
thread are dissolved or broken up into minute globules, a 
highly refringent thick U-shaped loop remaining. On boiling 
this portion appears to be dissolved.! A solution of silver 
nitrate, oxalic and picric acids produces no effect. Glacial 
acetic acid, on the contrary, has a most marked influence. 
Under its action the thread swells up and becomes transparent, 
then melts down to an irregular coagulated mass. On further 
action this refringent mass swells up again, and begins vio- 

1 It is almost impossible to make quite certain as to whether the thread is 


really dissolved in boiling (in this and other cases), as the operation cannot 
be conducted under the microscope. 


64 EDWIN S. GOODRICH. 


lently to “boil up,” and hecomes dissolved, leaving only one 
(or two) minute yellowish and very refringent coagulated 
granules, which undergo no further change. Strong HCl and 
H,SO, have at first somewhat the same action as acetic acid, 
swelling up the thread to a transparent viscid mass, which, 
however, is not dissolved in the cold. Strong HNO, acts in 
the same way, the swollen thread being further reduced to a 
small granular mass, which, on heating, is broken up into 
globules, and finally dissolved on boiling (?). Osmic acid 
shrivels the thread a little, and stains it a pale greyish tint. 
With tannic acid it becomes reduced to an irregular shrivelled 
mass. The thread is insoluble in lime water,—in fact, the 
addition of lime water to the celomic fluid is one of the 
best methods .of showing up the threads in a mass of cells! 
(figs. 15, 16). 

The xanthoproteic test gives no result as with the white 
granules. The thread does not stain readily with eosin; but, 
on the other hand, it stains readily with methyl blue, and 
especially with methyl green. In an acidified solution of 
Victoria blue it stains dark purple, and again dark blue in an 
alcoholic solution of cyanin. 


The chief results are shown below in tabular form. The 
sign + means that the substance is soluble in the reagent, 
the sign — that it is not obviously soluble without heating. 


3 : | | | 
ducts in the celomic] :| S |x $4 sHESM Whe Re I Lol ie 5 
me meaertclee $3 3 a a3 z8 ac ea sig = = zd 3 
Ael| 4 ROB Re asel/Os|<4s/m m = Ssis 
Of Vermiculus— | | 
1. Large white | | 
granules . J + |+—/—| + |... | oe | + | + [+] + 4! + [+ 
2. Small residual | | 
granules . .| — | — |—| + | «| «. | + | + J+) + J+] = J= 
Of Enchytreus— | | | 
3. White granules) + | — /—| + | + | +/+] + /+! + [4/4—J— 
par- | 
tially, 
4. Thread. . J — | — |J—| —| - | —/] —] + ‘| a i a 


1 The thread does not give rise to double refraction. 


NOTES ON OLIGOCHATES. 65 


From this table it will be seen that we have at least four 
distinct endoplastic products and perhaps three distinct 
chemical substances in the ccelomic corpuscles of Vermiculus 
and Enchytrzeus, which are probably of an albuminoid nature.! 
That none of the products are mucin is proved by their 
solubility in acetic acid ; that none of them are chitin is proved 
by the first three being soluble in caustic potash and the 
fourth in acetic acid. That they are neither of a fatty nor of 
an amyloid nature is shown by the action of ether, osmic acid, 
and iodine (which does not stain them dark). 


Notes on the Cuticle and Chete of Oligochetes. 


It has long been known that the cuticle of the Oligocheta is 
not made of chitin (Timm, 12; LHisig, 6), since it is readily 
soluble in caustic potash; but few, if any, observations seem 
to have been made with a view to acquiring a more definite 
knowledge of the character of the substance of which it is 
formed. Below are given the results of some experiments 
performed on the cuticle of the common earthworm, Lum- 
bricus herculeus. The worms were killed in 30 per cent. 
alcohol, and placed for a time in water until the cuticle could 
be easily peeled off. The cuticle, having been thoroughly 
washed in salt solution and distilled water, was then experi- 
mented upon in test-tubes. It is insoluble in alcohol and 
ether; soluble in hot distilled water (before boiling-point is 
reached), in solutions of ammonia or weak potash, in lime 
water, in weak mineral acids, in oxalic acid, and with some 
difficulty in acetic acid.” Cuticle dissolved in distilled water, 
boiled and filtered, gives a slightly viscid solution of whitish 
opalescent colour, which does not thicken on cooling. On 


1 In a recent paper on “The Process of Secretion in the Skin of the 
Common Eel,” Professor E. Waymouth Reid comes to the conclusion that 
the thread-like secrections from the epidermis are formed of an albuminoid 
possibly allied to keratin (‘ Phil. Trans.,’ v., 185, B. 1894). The substance 
presents no great resemblance to that found in Enchytreus. 

2 The same results were obtained with the cuticle of Vermiculus and 
Enchytreeus. 


VOL. 39, PART 1,—NEW SER. E 


66 EDWIN S. GOODRICH. 


evaporation the solution leaves a film of transparent substance 
without crystallisation. Boiled with caustic potash and cupric 
sulphate we get no reduction. ‘The solution answers to a 
certain exteut to the proteid tests, turning straw-coloured 
with the xanthoproteic test, and pale lilac with the biuret 
test. Boiling with Millon’s reagent produces no more than a 
pinkish hue. Tannic acid in the solution acidified with acetic 
acid gives a white precipitate.’ Sodium sulphate and po- 
tassium ferrocyanide give no precipitate. If we treat the 
solution with absolute alcohol we get a white precipitate, and 
the filtered liquid answers no more to the above tests. 

The Chete.—It is generally assumed that the cheetz are 
chitinous, but this seems to be by no means always the case. The 
cheetee of Vermiculus and Enchytreeus are insoluble in distilled 
water (hot or cold), alcohol, ether, caustic potash (even boiling), 
acetic and oxalic acids, hydrochloric acid (even after prolonged 
boiling for some minutes). In strong sulphuric acid they 
become swollen and soft, finally dissolving on heating. 

The cheetz of the earthworm are insoluble in distilled water, 
alcohol, ether, caustic potash, lime water, acetic and oxalic 
acids; soluble in strong hot sulphuric acid. Boiled in strong 
caustic potash the chete become somewhat swollen and broken 
up, but do not dissolve. If they be now washed and placed in 
concentrated hydrochloric acid the proximal region (that part 
which lies in the body-wall) is rapidly dissolved, partially in 
the largest cheetz, and almost entirely in the smaller. If the 
acid be now boiled the resisting distal region is soon dissolved, 
leaving only a thin colourless layer which formed a sort of 
cap over the cheta. This shell appears to be insoluble in 
hydrochloric acid like the entire cheta in Vermiculus. The 
chetz of Vermiculus, Enchytreeus, and Lumbricus become 
orange with xanthoproteic test, and crimson with Millon’s 
test. 

Below are set forth in tabular form the chief of these 
results : 


1 It behaves, in fact, somewhat like a peptone. 


NOTES ON OLIGOCHMTES. 67 


wa ee ee eee 
3S mld 2 : ae < 
Ssh1[4/ 8 |/eele8/35/2] = | £8. | 2. 
Sas | 8 | 2 1es|8e (se) 5 don | Se 
AFS | 4)/ RA /OS/AeEl/das\|H) Aas As 
1. Cuticle of Lum- 
IDM WI Q All Ge —;/—}] +] +] + [+l + Pale Pale 
2. Cheete of Vermi-; yellow. | pink. 
culus and En-' 
chytreus). =. .| = —-!—-/]-]-—-]-— —-| — | Orange.| Red. 
3. Chete of Lum- 
WINGER > o « of = ~!|—|—|{—]-— j\-| +. | Orange. | Red. 
with 
| insol. 
resi- 
due. 


From this table it would again appear that we have three 
albuminoid substances distinct from each other, and from those 
found in the coelomic corpuscles. The chete of the small 
worms (Vermiculus and Enchytreus) are formed of a substance 
which is not chitin, since it is insolublein HCl. The chetz 
of Lumbricus, on the contrary, are probably chiefly composed 
of chitin, or some substance closely allied to it, since they are 
insoluble in caustic potash, and partially soluble in HCl. On 
the other hand, so far as the solubilities show, the cuticle 
appears to be formed of a substance closely allied neither to 
chitin nor to mucin. 


List oF WORKS REFERRED TO. 


1. Bepparp, F. E.—‘‘ A Contribution to the Anatomy of Sutroa,” ‘Trans. 
Roy. Soc. Edinb.,’ vol. xxxvii, 1892. 

2. Bepparp, F. E.—‘ Monograph of the Order Oligocheta,’ Oxford, 1895. 

3. Biomrretp, J. E.—‘The Thread-cells and Epidermis of Myxine,” 
‘Quart. Journ. Mier. Sci.,’ vol. xxii, 1882. 

4. Bousius, H.—* L’Organe segmentaire d’un Enchytreide,” ‘Mem. d. 
Acad. Pont. d. Nuovei Lincei,’ vol. ix, 1893. 

5. CuarparkpE, E.—‘ Annélides, Chétopodes du Golfe de Naples,’ Geneva, 
1868. 

6. E1stc, H.—‘ Die Capitelliden des Golfes von Neapel,’ Monog., Berlin, 
1887, 


68 EDWIN 8S. GOODRICH. 


7. Goopricu, E. §8.—‘On a New Organ in the Lycoridea, &c.,” ‘ Quart. 
Journ. Micros. Sci.,’ vol. xxxiv, 1893. 
8. Goopricn, EH. 8.—“On the Structure of Vermiculus pilosus,” 
‘Quart. Journ. Micros. Sci.,’ vol. xxxvii, 1895. 
9. Micuartsen, W.— Untersuchung iiber Enchytreus MObii,” ‘In. 
Diss.,’ Kiel, 1886; and ‘ Vorlaufige Mittheilung,” ‘ Zool. Anz.,’ 1885. 
9a. MicuaEtsen, W.—“ Enchytreeiden-Studien,” ‘Arch. f. Mikros. Anat., 
Bd. xxx, 1888. 
10. Route, L.—“ Etudes sur le développement des Annélides,” ‘Ann. Sci. 
Nat.’ (7), vol. vii, 1889. 
11. SHarprri, Th.—‘‘ Das Chloragogen von Ophelia radiata,” ‘Jen. Zeit. 
f. Naturwiss.,’ Bd. xxviii. 
12. Timm, R.—‘* Beobachtungen an Phreoryctes, &.,” ‘Arb. Zool.-Zoot.,’ 
Wurzburg, vol. vi. 
13. Vespovsky, F.—‘“‘Beitrige z. vergl. Morph. d. Anneliden,” ‘Monog. 
d. Enchytreiden,’ Prag, 1879. 


EXPLANATION OF PLATES 5 and 6, 


Illustrating Mr. Edwin S. Goodrich’s “ Notes on Oligochetes, 
with the Description of a New Species.” 


List oF REFERENCE LETTERS. 


él. bv. Blind branch of the nephridial canal. 7. Brain. 3.0. Body-wall. 
e.a@. Ciliated enlargement of nephridial canal. circ. m. Circular muscles. 
cel. Celom. cel. corp. Celomic corpuscle. cal. epith. Colomic epithelium. 
d. bl. v. Dorsal blood-vessel. epid. Epidermis. e.v. End vesicle. ect. cil. 
External cilia. ~ Point at which the spermatheca has fused with the ceso- 
phagus. jl. Flame-like bunch of cilia. /. Expanded foot whereby the cell is 
fixed to the body-wall. g/. Spermathecal gland. 4. p. Hair-like processes. 
7. e. Internal hooked end. iz¢. op. Internal opening of the spermatheca into 
the cesophagus. 7,7. Lower lip. Jong. m. Longitudinal muscles. m. p. 
Median process. 2. Nucleus. z.c. Ventral nerve-cord. xeph.c. Nephridial 
canal. xeph.p. Nephridiopore. zeph.st. Nephridiostome. @s. (sophagus. 
op. sp.d. Opening of spermathecal duct into the main chamber of the sperma- 
theca. post. sac. Posterior sac of the spermatheca. s. Septum. — sal. gl. 
Salivary gland. sept. g/. Septal gland. sp. Spermatheca. sp. d. Sperma- 
thecal duct. sp.op. Spermathecal opening. spéz. Spermatozoa. st. Stalk of 
attachment. hr. Thread in the celomic corpuscle. w.p/. Upper lip. v. d2.v. 
Ventral blood-vessel. 

All the figures, except figs. 16, 17, and 28, refer to Enchytreus hor- 
tensis, n. sp. 


NOTES ON OLIGOCHATES. 69 


Fic. 1.—Enlarged view of the anterior region, showing the shape of the 
brain, salivary glands, and septal glands. 
Fic. 2.—Enlarged view of the nephridium, drawn from the living. 


Fie. 3.—Section through part of the nephridium and nephrostome. Z. 
apoch. 4 mm., oc. 4 c. cam. 


Fic. 4.—Enlarged ventral view of the nephrostome, from the living. 
Fic. 5.—Hularged side view of the same. 


Fic. 6.—Transverse section through the posterior region of the nephridium. 
Z., D, oc. 4, cam. 


Fic. 7.—Enlarged view of a portion of the last region of the nephridial 
canal, from the living. 


Fic. 8.—Optical section through the body-wall. Z., D, oc. 3. 


Fic. 9.—Thread-corpuscles from a young worm (fresh). a, 0,¢. Front 
views. d. Side view. Z., D, oc. 4. 

Fic. 10.—Young (?) thread-corpuscles from an adult worm. a. Side view. 
b. Front view. Much enlarged (fresh). 

Fic. 11.—Oval (a, 4, c) and thread-corpuscles (d, e, f/). Flemming and 
carm.-alum. Z., D, oc. 4, cam. 

Fig. 12.—Oval (a) and thread-corpuscles (4, ¢). Corrosive and bor. car- 
mine. (The thread has been acted upon by the reagents leaving the cavity it 
formerly filled.) Z., F, oc. 2, cam. 

Fic. 13.—Series of figures indicating the action of water on the thread- 
corpuscle (a). Much enlarged. 

Fic. 14.—The thread after the action of caustic potash. Enlarged. 

Figs. 15 and 16.—An oval corpuscle before (a) and after (a’), and a thread- 
corpuscle before (4) and after ('), the action of lime water, 

Fic. 17.—Enlarged view of a portion of Pachydrilus, sp. (?), showing the 
long cells in the ceelom; from the living. 

Fie. 17a.—Section through an elongated cell from the ccelom of Pachy- 
drilus, sp. (?). Corrosive, bor. carmine. Z., D, oc. 4, cam. 

Fie. 18.—Cheta. Z., D, oc. 4, cam. 

Figs. 19, 20, 21, and 22.—Series of transverse sections through the sperma- 
theca, taken from before backwards. Z., B, oc. 3, cam. 

Fig. 23.—Diagram showing the communication of the spermathece with 
the exterior and the cesophagus ; in longitudinal section. 

Fries. 24, 25, and 26.—Series of longitudinal horizontal sections taken from 
above downwards, showing the relations of the spermatheca. Z., B, oc. 4 
comp., cam. 

Fig, 27.—Longitudinal horizontal section through the fifth segment of 
Marionia enchytreoides. Z., B, oc. 4, cam. 

Fic. 28.—Ameeboid corpuscles (fresh). Much enlarged. 


ere Te 
ss 

jieth 

i. me ‘ f hee 


A os ek 


ON THE DEVELOPMENT OF LICHENOPORA VERRUCARIA. 71 


On the Development of Lichenopora 
verrucaria, Fabr. 


By 


Sidney F. Harmer, M.A., B.S8c., 
Fellow of King’s College, Cambridge ; Superintendent of the University 
Museum of Zoology. 


With Plates 7—10, and two figures in text. 


I HAVE formerly had occasion to describe certain very re- 
markable phenomena in the development of Crisia (6), and I 
then ventured to suggest that embryonic fission would be found 
to be characteristic of the whole group of Cyclostomatous 
Polyzoa. The development of Crisia being, to the best of my 
belief, without an exact parallel in the entire animal kingdom, 
it was important to test my former results by the study of some 
other Cyclostome. A chance discovery of large numbers of the 
colonies of Lichenopora verrucaria, Fabr., in all stages of 
development, enabled me not only to confirm the main fact of 
the occurrence of embryonic fission, but to discover certain 
previously unsuspected phenomena in the life-history of that 
species. An abstract of my results has been communicated to 
the Royal Society (8). 

It would not be easy to find two genera of Cyclostomata 
which are more unlike than Crisia and Lichenopora; 
and the occurrence of a fundamentally similar mode of develop- 
ment in the two forms may be taken as giving good reason to 
believe that we are really dealing with a normal characteristic 
of the group. Crisia is an erect, branching form, whose 
calcareous branches are interrupted at intervals by chitinous 
joints. Development here takes place in the interior of a 


72 SIDNEY F. HARMER. 


pear-shaped ovicell, which replaces a zocecium at a greater or 
smaller distance from a joint, according to the species. Lichen- 
opora, on the contrary, is a plano-convex disc, usually 
closely attached by its flat surface to some foreign object ; its 
zocecia project from its upper side, and between them is a 
calcareous lamina or “crust,” which forms the roof of a large 
compound ovicell. The development of this ovicell commences 
with the early stages in the growth of the colony. Whilst a 
Crisia-colony may produce a profusion of ovicells at the 
breeding season, or may be entirely devoid of these structures, 
the ovicell of Lichenopora is a single, complicated struc- 
ture, whose growth is intimately connected with the develop- 
ment of the external features of the colony. In dealing with 
this question I am obliged to limit myself to L. verrucaria, 
the only species of which I have obtained an ample supply of 
material. It may, however, be pointed out that, if this species 
does not materially differ from other species of the same genus, 
it is the character of the ovicell which has been taken as the 
distinguishing feature of the genus or even of the family to 
which it belongs. 

According to Hincks (9, p. 471) the zocecia of the Licheno- 
poride are ‘ disposed in more or less distinct series, which 
radiate from a free central area.’’ ‘Thearea here referred to is, 
in L. verrucaria, the roof of the ovicell, and the definition 
applies only to moderately advanced colonies (Pl. 7, fig. 7), 
in which the ovicell possesses a calcareous roof. But when 
development has reached a certain stage, every colony possesses 
an ovicell, and this is in marked contrast to most of the 
Cyclostomata, in which ovicells are not present in a very large 
proportion of the colonies which may be examined. 

A still more noteworthy feature remains to be pointed out. 
The internal processes which precede the external appearance 
of the ovicell commence almost with the beginning of colonial 
life. The individual formed by the metamorphosis of the free 
larva gives rise, almost at the same time, to two new zoccia. 
One of these two commonly becomes fertile, and forms the 
starting-point of the series of stages by which the ovicell 


ON THE DEVELOPMENT OF LICHENOPORA VERRUCARIA. 73 


becomes stocked with numerous larve. The other zocecia of 
the colony remain, for the most part, unfertile, although many 
of them produce spermatozoa. It is clear that they have the 
function, not only ef feeding themselves, but also of providing 
nutritive material at the expense of which the ovicell and its 
brood of larvee may develop. The restriction of the female 
reproductive function to a single individual, or at least to a 
very small number of individuals of the colony, is by no means 
in accordance with what is known of other colonial animals. 
The remarkable nature of the phenomenon will be most clearly 
realised when it is understood that the fertile zocecium is 
usually one of the two “ blastozoites,” which are first formed 
by budding from the “‘ oozoite.”’ } 

My material was collected in Norway at Lervik, in Stordo, 
and at Godosund, off the north coast of Tysnaes6, during the 
end of June and the beginning of July. At that time of the 
year there is no difficulty in collecting an indefinite number of 
colonies from the earliest stages immediately after metamor- 
phosis to the fully developed condition. The entire colonies 
were preserved for the most part in corrosive sublimate, to 
which, in some cases, a few drops of nitric acid and of acetic 
acid were added. ‘The internal details were studied prin- 
cipally from sections, which were prepared for me by my wife, 
but to some extent in entire colonies stained without decal- 
cification. The external features were investigated in dry 
mounts and in Canada-balsam preparations, whether stained 
or unstained. 

Owing to its discoidal or flattened form, the colony will 
obviously tend to rest on one of its flattened surfaces when it 
is mounted. In order to examine certain features which can 
only be made out when the colony is looked at edgeways, it is 
necessary to make some special arrangement. This has been 
successfully accomplished by folding a piece of black paper into 

* These terms, first introduced by Lacaze-Duthiers, have been appro- 
priately employed by Prouho (18) in describing the colonies of Polyzoa. ‘The 
oozoite is the individual developed from an egg, i. e. the metamorphosed 


larva or primary zoccium of the colony. The blastozoite is an individual 
which has been formed as a bud. 


74 SIDNEY F. HARMER. 


a series of ridges whose cross section has the form of MM. The 
iichenopora is then mounted (dry) on the slopes of the 
ridges ; and this allows of the examination of one edge of the 
colony. An improvement on this method, whereby each 
colony could be examined in a series of positions, was effected 
by taking a short length (about 18 mm.) of brass tubing, 
through the axis of which a long needle was passed, the 
interval between the needle and the tubing being then filled 
up with sealing-wax. The outer surface of the brass was 
covered with black paper, on which numerous colonies were 
mounted. The needle, to which a suitable head is attached, 
serves as an axis permitting the revolution of the brass cylinder 
under the microscope. 

The largest colony, among my preparations, has a maximum 
diameter of 5°47 mm. ‘The older colonies are invariably 
altered in appearance by the occurrence of secondary thicken- 
ings of the surface. 


Literature.—A large number of descriptions and figures 
of the ovicells of Cyclostomata are scattered through the litera- 
ture of the Polyzoa, and some of these are alluded to below. 
The only paper which needs special notice here is the second 
of Smitt’s admirable series of memoirs dealing with the Scan- 
dinavian Polyzoa (20). Smitt has thoroughly understood the 
general growth of the colony, and to a large extent that of the 
ovicell of L. verrucaria, the particular species we are con- 
sidering. But as Smitt was not concerned with the embryonic 
development, he paid no attention to the earliest processes in 
the formation of the ovicell. As, moreover, I cannot agree 
with all the statements of this observer, a complete account of 
the development of the entire colony must be given. In further 
justification of this course, I may mention that although it is 
perfectly obvious to any one who has made an independent 
study of L. verrucaria that Smitt had grasped many of the 
important facts, the figures which he has given are insufficient 
to explain his meaning to anyone who is not well acquainted 
with the species. 


ON THE DEVELOPMENT OF LICHENOPORA VERRUCARIA. 795 


Smitt has described correctly the general growth of the 
colony, and he has clearly appreciated the arrangement of the 
alveoli (see below) or interstitial spaces. He has the merit 
of having pointed out that these structures give rise to the 
ovicell, and he calls attention to the important fact that they 
do not reach the basal lamina of the colony. He has further 
stated quite correctly that the central area of the colony, which 
is at the same time an important part of the ovicell, owes its 
character as a space devoid of zoccia to the fact that the area 
originates by the divergence from one another of the most 
centrally placed zocecia. Smitt has also given some account 
of the secondary thickening of the colony. 

The size of the largest colonies (exceeding 8 mm.) recorded 
by Smitt is somewhat greater than that of the largest colonies 
I have found. I have not been able to discover any colony in 
which the ovicell had as many trumpet-shaped openings (one to 
eight) as are described by Smitt. Ido not think that so large a 
number would ever occur in a young colony, in which the 
number seems to be invariably one or two, and in this respect 
I find myself in agreement with Levinsen (12, p. 28) ; but I 
think that it is perfectly possible that a larger number of 
openings are developed in colonies in which a considerable 
number of primary embryos (cf. p. 133) are formed as a second 
generation. I have obtained no actual evidence that this is 
the case, and I have some reason for believing that the larve 
of the second generation make use of the old apertures of the 
ovicells. Ridley has suggested (19, p. 452) that Smitt has 
mistaken zocecia in which incomplete transverse septa have 
been developed for the apertures of ovicells. 

I do not think that Smitt is right when he describes (p. 478) 
the secondary thickening by the formation of the so-called 
“cancelli”’ (cf. my figs. 8 and 9) as giving rise to two or three 
ovicells, arranged in superposed layers. I do not believe that 
there is ever more than a single space which is occupied by 
embryos, although that space can fuse with other spaces by the 
absorption of the calcareous septa which are at first present. 

Smitt describes (p. 476) the first septum formed in the young 


76 SIDNEY F. HARMER. 


colony at the outset of its growth as bending to the left. He 
appears to have missed the curious fact that the growth is 
to the left in some colonies, and to the right in others. He has, 
I think, not quite appreciated the way in which the proximal 
part of the colony becomes covered by the growing edge. His 
fig. 7 (pl. x) omits an essential part of the arrangement,—that 
is to say, the actual growing edge shown in my own figs. 3—5. 

Smitt describes the colour of the ovicell as inclining to blue 
(“ till fargen dragande at blatt’). I suspect that this was due 
to the use of a glass which was not quite achromatic. The 
ovicell. is, according to my observations, perfectly white. 
Finally, I must call attention to the small, scattered tubes 
with an even, round mouth which Smitt describes as some- 
times occurring in the colony. These he compares with the 
small tubes of Diastopora hyalina (= D. obelia, Johnst.), 
and suggests that they may have something to do with the 
production of the male generative organs. ‘This suggestion is 
clearly wrong, since the testes are developed inside ordinary 
zocecia. I have not observed these small tubes as a part of the 
Lichenopora, but I think it possible that Smitt may be 
referring to the Infusorian Folliculina, whose tubes are 
commonly found growing on the Lichenopora. 


External Features of the Colony.—Fig. 7 represents a ma- 
ture specimen, with a diameter of 1:92 mm., in which secondary 
thickenings have hardly commenced. The colony is nearly 
circular, and it is surrounded by the delicate calcareous 
‘basal lamina,” which in this particular case is upturned, 
although more usually it is closely adnate to the seaweed. 
The arrangement of the zocecia may be understood by ima- 
gining a number of the quills of quill pens to be arranged 
in an obliquely vertical position, radiating from a common 
centre. Those near the centre will approach the vertical 
position, while those nearer the margin are more nearly 
horizontal. It must further be supposed that the nib of 
each pen is uppermost, or nearer the centre, so that the orifice 
of the tube is completely concealed when the colony is looked 


ON THE DEVELOPMENT OF LICHENOPORA VERRUCARIA. 77 


at from above, except in those individuals which stand nearly 
vertically. The cylindrical zocecium is, in fact, obliquely cut 
off in the way which has been indicated by the above com- 
parison, and this will be better understood by referring to 
fig. 6. The “nib” of the zocecium is commonly in the form 
of a single spike, but it may be split, so that two, three, or 
even more spikes occur. ‘These are very delicate structures in 
young zocecia, but they are readily broken off, or lost in other 
ways, in old zoecia. In young colonies the zocecia usually 
possess more than one spike each. At the growing margins of 
old specimens the occurrence of a single spike is much more 
common—if, indeed, it is not the rule. 

Between the more centrally placed zocecia stretches a calca- 
reous lamina (fig. 7)—the roof of the ovicell. This lamina is 
not even, but is obviously composed of a number of convex 
portions, separated from one another by shallow grooves. The 
ovicell opens to the exterior by means of a trumpet-shaped 
aperture, within whose expanded mouth is a straight edge, 
turned towards the centre of the colony. The section of the 
trumpet is at this level semicircular. The straight edge cor- 
responds to the level at which the trumpet springs from the 
roof of the ovicell. On the opposite side, however, the wall 
of the trumpet is longer, so that it becomes continuous with 
the roof of the ovicell at a somewhat lower level. Its passage 
into the roof is at the same time more gradual and less angular 
than on the side which faces the centre of the colony. 

This arrangement of the aperture of the ovicell may be 
regarded as characteristic of L. verrucaria. The straight 
edge may correspond with the valve which I have described 
(5) in the ovicell of Crisia, and its function may be to restrain 
the too rapid birth of the larve. 

The early stages in the growth of a Cyclostome colony have 
been well described by Barrois (2 and 8), and my results agree 
with his so far as the general character of the growth is con- 
cerned. It has not, however, been previously noticed that the 
study of the earliest stages of the colony is essential for the 
proper understanding of the ovicell of Lichenopora. 


78 SIDNEY F. HARMER. 


Fig. 1 (Pl. 7) gives two views of an extremely young 
colony of L. verrucaria. The colony is attached by means 
of a circular disc, whose diameter is ‘16 mm. The disc is 
formed, as Barrois has shown, by the calcification of the body- 
wall of the fixed larva. It gives off a calcareous tube which 
lies nearly horizontally, and ends in an open! mouth, whose 
diameter is ‘18 mm. On looking down into this mouth (Fig. 1, 8), 
it is seen that the cavity is divided by a septum which appears 
triradiate in end view. The septum does not in any part 
reach the margin of the aperture of the tube, while proximally 
it passes into the lower wall of the tube in such a way that the 
cavity of the disc is continuous with the part marked 1, and is 
entirely cut off from 2 and 3. 

A curious point may now be noticed. About. half the 
colonies of L. verrucaria may be described as “ right- 
handed,” and about half as “left-handed,” and this will 
become intelligible from a consideration of the woodcut (Fig. 1). 


Fic. 1.—Lichenopora verrucaria; diagram to explain the difference 
between “right-handed” and “left-handed” colonies. The growing 
margin of the colonies is omitted. I, IJ, and ILI are supposed to be 
seen from above, as in PI. 7, fig. 2. (For explanation, see text.) 


1 The mouth appears open (as in later stages) because it is uncalcified. 
The delicate layer of cells which stretches across it, and can readily be 
demonstrated in sections, disappears in the dried colony. 


ON THE DEVELOPMENT OF LICHENOPORA VERRUOARIA. 79 


I shows the end view of a young colony, similar to fig. 1, 
Plate 7. The space or zoccium A is uppermost, while B 
and c are nearer to the seaweed. Assuming it to be possible 
for the whole tube shown in I to roll over, it is obvious that it 
can do so in the direction shown by either of the two arrows. 
Should it roll to the left, we shall have the condition shown in 
II, where a and c have come to the upper surface, while B alone 
is left in contact with the seaweed. III shows the result of 
the rotation of I to the right. 

Returning to Pl. 7, fig. 1, it will be understood that the 
young colony is attached by the flat surface of its plano-convex 
disc, as well as by the flattened wall of the zoccia 2 and 3. 
This precludes any actual rotation of the parts already deve- 
loped, but it does not preclude an alteration in the relative 
position of 1, 2, and 3, as the colony increases in length. This 
it does in such a way as to assume the position which would 
be arrived at if it at first grew quite symmetrically, but subse- 
quently experienced a rotation of the common mouth of the 
tube, while leaving its proximal end unaltered. In other 
words, the zoecium 1 may either grow to the right or to the 
left, as shown in IV and V of the woodcut, the other two 
zocecia experiencing a corresponding change of position. A 
colony in which 1, as seen under the compound. microscope, 
has grown to the right, will be described as “ right-handed ” 
(IV), while a ‘ left-handed” colony has the appearance of V 
in the woodcut. 

It is obvious that a cross-section of either IV or V, imme- 
diately above the constriction which marks the passage of 1 
into the disc, will have the appearance of I; while a section 
taken more distally will show the first zocecium (1) in the 
position of a in IT or III, according as we are dealing with a 
*‘ right-handed ” or a “left-handed” colony. It is clear that 
the zocecium 3 in IV corresponds with B in II; while 3 in V 
corresponds with cin IIT; and the object of giving different 
symbols to these zowcia in 1V and V is to call attention to the 
fact that the individual properties of the zoccia in the older 
colony depend on the direction in which the primary zocecium 


80 SIDNEY F. HARMER. 


grows at the commencement of the formation of the colony. 
The whole colony acquires a certain symmetry with regard to 
3, the axial zocecium, and in studying the characteristics of 2 
and 3 it matters not which of these corresponds with B and 
which with cin I of the diagram. The necessity for a close 
attention to these zocecia will become apparent later. 

The distinction between right- and left-handed colonies per- 
sists throughout life, although it ceases to be obvious in colo- 
nies of more thana certain age, when looked at in the usual 
position, from above. The difference can readily be appreciated 
in young colonies, in which the early stages of the embryonic 
development are taking place. This will be seen by comparing 
fig. 3 (left-handed) with fig. 4 (right-handed). 

The two kinds of colonies seem to occur in about equal 
numbers. Of sixteen chosen at random as a test case, eight 
were right-handed and eight were left-handed ; and this result 
agrees with the other observations I have made on this point. 

The actual position of the three zoccia can be seen from 
fig. 2, which is an end view of a young colony which was 
proved to be left-handed by examining it in another position. 
1 is the primary zoccium, which in future will be designated 
by the symbol z!, the symbols z? and z> being employed for 
the zocecia 2 and 3 respectively in either 1V or V of the wood- 
cut (p. 78). The tubular end of 2? has in fig. 2 reached a 
higher stage of development than that of z). zis in contact 
with the surface of attachment, while z! and z? are on the 
free surface of the colony. The position of these three zoccia 
is also shown in fig. 3, a somewhat later stage seen from above. 
The colony may be described as having the form of a flattened 
funnel, whose narrow end originates in the disc. The funnel 
is attached by one of its flattened sides ; and its lip, which is 
closely attached to the seaweed on this side, here projects 
considerably further than on the opposite side. The wide end 
of the funnel is, in fact, obliquely truncated. 

2° has acquired an axial position, which it retains through- 
out the later development. z! and z’ are, at their upper ends, 
arranged symmetrically with regard to z°, although the wood- 


ON THE DEVELOPMENT OF LICHENOPORA VERRUCARIA. 81 


cut on p. 78 (IV and V) shows that 2? and 2 are at their 
bases symmetrical with regard to z!, and are clearly a pair of 
sister zocecia produced by the primary zocecium of the colony. 

Figs. 2 and 14 illustrate further the manner in which the 
number of zocecia becomes increased. This happens by the 
forking of a pre-existing septum ; and the zoccia are at first 
bounded externally by the common growing rim of the colony. 
The way in which the zocecium is completed externally is shown 
in fig. 14, which represents a part of the growing margin of an 
old Lichenopora. No sooner is the outer wall completed 
than the zocecium commences to grow upwards as a free tube, 
although its inner side has before then commenced its upward 
growth. ‘The obliquely truncated character of the adult orifice 
is thus a marked feature even at the first formation of a 
zocecium. 

Fig. 14 shows that the septa do not reach the actual rim of 
the colony, and that the outer wall of the zocecium always 
grows up from within the edge. Since this happens in the 
entire circuit of the mouth of the funnel-shaped colony, part 
of the rim capable of forming new zoccia is left all round, and 
this is the proximate cause of the assumption of the completely 
discoidal shape which is so characteristic of the genus Lichen- 
opora. The growing edge is well seen in figs. 38 and 4, from 
which it is evident that that part of the rim which is situated 
on the proximal side of the first three zoccia grows at first 
much less energetically than the opposite part. 

We may here notice an important generic difference between 
Lichenopora and such a genus as Tubulipora. In the 
latter, z! and z? are formed in such a way that the proximal 
part of the rim of the young funnel-shaped colony is used up 
in forming their outer walls. There is thus no edge capable 
of giving rise to new zocecia between z! and z? on the one 
hand and the disc of fixation on the other hand. The colony 
thus assumes a fan-like shape instead of the discoidal form 
which is assumed by Lichenopora. This distinction has 
already been pointed out by Smitt (20, p. 476). 

Fig. 4 represents a right-handed colony of Lichenopora 

vou. 39, PART 1.—NEW SER. F 


82 SIDNEY F. HARMER. 


which has a maximum diameter of ‘59 mm. The number of 
zoccia has increased, and it will be noticed that the zocecium 
which we have called z? still maintains its axial position in 
the colony. The zoccium between z? and z! may be termed 
z‘, while that between z® and z? will be called z*. It will be 
unnecessary to give special symbols to the other zoccia. 

Growth has, so far, proceeded almost exclusively on the 
distal side of the first three zocecia; but the characteristic 
features of the adult Lichenopora are soon acquired by the 
growth of that part of the margin of the colony which inter- 
venes between them and the disc. In fig. 2 the diverging 
zocecia z! and z* are separated by a quadrangular area which 
is the outer wall of a pyramidal space situated between 2}, 
z°, 2°, and the outer wall of the funnel-shaped colony. In 
fig. 4 this space has become divided into two, and at a later 
stage two or three zoccia will be found between the distal 
ends of z! and z?, and directed in such a way that they radiate 
from z® towards the disc of attachment. The position of these 
zocecia is shown in figs. 5 and 12. Their development has 
taken place simultaneously with the extension of the proximal 
part of the rim of the funnel, which grows in such a way 
as to cover that part of the funnel which intervenes between 
the disc and the open mouth. ‘The rim or basal lamina here 
grows horizontally, keeping in close contact with the wall of 
the funnel, and finally covering it and the disc completely. 
Having done this, it comes in contact with the seaweed on 
the proximal side of the disc of attachment ; and the latter is 
thus completely covered. Growth has not, however, been 
confined to this region. The colony has been increasing hori- 
zontally in every direction, z® forming a centre approximately 
equidistant from all parts of the margin. 

The general nature of these processes will be understood 
by comparing figs. 5 and 7. In the fully formed discoidal 
Lichenopora no trace can be seen from the upper side of the 
proximal part of the funnel ; but it is merely necessary to turn 
the colony over to find that that part persists throughout life, 
without any increase of size, A back view of the adult colony, 


ON THE DEVELOPMENT OF LICHENOPORA VERRUCARTA. 83 


showing the disc and the proximal part of the funnel, has been 
well figured by Smitt (20, pl. xi, fig. 6). It is obvious that 
the increase of the diameter of the colony is accompanied by 
an increase in the number of zocecia which are present. 

The divergence of the zocecia from the centre of the discoidal 
colony results in the occurrence of a central space, round which 
the zocecia are arranged; and this space is the commencement 
of the ovicell. Two points must be here expressly noticed : 
firstly, that the ovicell is not restricted to certain colonies, but 
is developed in all colonies which reach the proper stage of 
growth; and secondly, that the description of the young ovicell 
must be given with a word of warning. The account of the 
growth of the colony has been taken from entire specimens, 
and largely from dried specimens. In the latter at least, we 
should not expect to find much trace of any structures which 
are not calcareous; and even in dealing with Canada-balsam 
preparations of entire colonies, it is extremely difficult to see 
much of the delicate, uncalcified membranes which really 
exist. But it must here be noticed that the study of sections 
shows emphatically that the various spaces of the colony are 
roofed in by living membranes. Thus the spaces at the growing 
edge shown in fig. 14 are really closed, and the same is true 
of the interzocecial spaces which give rise to the ovicell. Even 
the orifices of the zocecia are not widely-open tubes, but are closed 
during life by a membranous diaphragm, which can be widened 
to allow of the protrusion of the tentacles. Ofall this, nothing 
appears in the dried colony ; but it must be clearly understood 
that the cavity of a young ovicell is morphologically a body- 
cavity, and is not an external space converted into an ovicell 
by the formation of a calcareous roof. 

The development of the ovicell can be most easily described 
after the structure of the complete ovicell is understood. 
Fig. 6 represents a thick free-hand section showing the whole 
length of a single radius of a colony which has not been decal- 
cified. ‘The zoccia diverge from the centre of the colony. 
Each zoecium reaches the attached basal lamina, a short part 
of which it forms. The older zocecia are very much longer 


84. SIDNEY F. HARMER. 


than the younger ones. The roof of the ovicell connects the 
diverging zocecia, aud the cavity of the ovicell is clearly, in the 
central part of the colony, a space which completely surrounds 
a part of the zocecia. The roof of the ovicell is not perfectly 
even, but is formed of convex portions, shown in surface view 
in fig. 7. 

The ovicell increases in size during the growth of the colony ; 
and the nature of this increase may be understood from fig. 14. 
The marginal spaces, marked out by the vertical septa which 
start from the basal lamina, are all destined to give rise to 
zocecia, which are the only spaces which extend down to the 
base. The zoccium is first clearly indicated by the appearance 
of one or more of the spikes which occur on the upper margin 
of the orifice of all the zocecia. At about this stage (fig. 14, 4), 
one of the septa forming the side walls of the zocecium gives 
off a branch which runs more or less transversely. ‘The cavity 
of the young zocecium is in this way cut off from the cavity on 
the distal side of it. Up to this time the zocecium has been 
a short horizontally-placed tube, whose lower wall is part of 
the basal lamina, and whose free upper wall is parallel to the 
base. No sooner is the cavity of the zocecium completed on its 
distal side than it begins to alter its direction of growth. The 
woodcut Fig. 2 is a diagrammatic representation of the margin 
of acolony. The thick lines represent two young zocecia and 

ji 


8 e Se = : — C - - ‘ 
Fic. 2.—Diagram of the growth of the margin of the colony of Licheno- 
poraverrucaria. (For explanation, see text.) 


part of the basal lamina as they would be seen in a radial 
section, and the dotted lines represent the condition of the 
corresponding parts after a certain amount of growth has taken 


ON THE DEVELOPMENT OF LICHENOPORA VERRUCARIA. 89 


place; acu Bis the basal lamina. The zocecium whose point 
is marked p has its upper end free, even in the earlier con- 
dition, while that marked £ is still incomplete on its distal 
side (cf. fig. 14). It is soon completed on that side by an 
obliquely vertical septum, which starts at Fr; and the pro- 
longation of this septum forms the lower wall of the zocecium. 
By the time that the margin of the colony has extended to BI, 
the two radial septa, which at first formed the lateral parts of 
the zocecium £, have grown in length and in height; and they 
are supposed to have meanwhile given rise to the beginning of 
a new zocecium, G, which has reached the condition of & in 
the earlier phase. 

In examining the woodcut in the condition indicated by the 
dotted lines, it is clear that a space occurs between the two 
older zocecia, and a much younger one between the two younger 
zoecia. Smitt (20) alludes to a space of this nature as an 
“ interstitialrum ;” but it will be convenient to substitute the 
term “ alveolus.” 

The zocecia may be arranged in radial series, as shown in 
the woodcut; or their position may be for the most part 
alternate, as in fig. 14. The upper sides of the zocecia are 
marked with longitudinal ridges, which have been omitted 
from most of the figures for the sake of simplicity, but are 
shown in figs. 8 and 9. The projecting spikes of the zocecia 
are commonly continuous with these ridges, some of which 
become exaggerated at their lower ends to form lateral walls 
for the alveoli. These are thus pyramidal spaces, pointed at 
their lower ends, and at first only closed by membrane above. 
They increase in length by the growth of the zocecia, and by 
the extension of the vertical septa on sides where they are not 
bounded by zocecia. The septa do not, however, grow so high as 
the zocecia; but after they have reached a certain height, each 
alveolus commences to become roofed by the calcification of the 
membrane which is already present in that situation. The 
calcification of the roof of each alveolus begins at the margin 
and gradually advances to the centre, in the same way that an 
opening is gradually closed by the contraction of a sphincter 


86 SIDNEY F. HARMER. 


muscle. An alveolus which is incompletely roofed in is 
shown in fig. 6. 

The original extent of the alveoli is indicated by the grooves, 
which correspond with the original septa, on the surface of the 
ovicell (cf. figs. 7 and 14). It is obvious that the ovicell is 
gradually increased in size by the addition of fresh alveoli all 
round its margin. The vertical septa are absorbed after the 
alveolus has become part of the ovicell, and a continuous space 
thus results, through which the more centrally placed zocecia 
pass as free pillars. Fig. 6 illustrates this point, and the con- 
formation of that part of the roof of the ovicell which lies to 
the left of the right-hand zocecium indicates the loss of the 
vertical septa which were at first present. 

We have still to consider the earliest development of the 
ovicell, which, to the best of my belief, has not been studied in 
detail by any previous observer. Fig. 12, though by no means 
an early stage, will serve to introduce us to a phenomenon 
which is of the highest importance for the understanding of 
the history of the colony. This phenomenon may be spoken 
of as the occlusion of the fertile zoecium. 

In fig. 12, all parts of the ovicell which have received a 
complete roof are shaded. The preparation is a transparent 
one, made in Canada balsam, so that the walls of the zowcia 
can be seen through the roof of the ovicell. Parts which are 
seen through some other part of the colony are indicated in 
dotted lines. The colony is left-handed—z! and z? project in 
the ordinary way beyond the roof of the ovicell, while z?, the 
occluded zocecium, is completely closed by the roof, and would 
probably not have been visible at all in a dried, opaque pre- 
paration. In the immediate neighbourhood of the occluded 
zocecium arises the trumpet-shaped aperture of the ovicell. The 
study of decalcified sections is necessary in order to understand 
the significance of these facts, but it must for the present be 
assumed that we are justified in alluding to z? in this colony as 
the “fertile? zocecium—i.e. as the zocecium in which the 
embryo is first developed. 

The roofing in of the ovicell commences with the 


ON THE DEVELOPMENT OF LICHENOPORA VERRUCARIA. 87 


occlusion of the fertile zowcium. This process is not 
easy to see in a colony which is resting in its position of most 
stable equilibrium on its flat surface; but it can be seen 
without difficulty in a colony of suitable age which has been 
mounted on a rotating cylinder in the manner described on 
p- 74. No general rule can be laid down with regard to the 
size of the entire colony in its relation to the closure of the 
ovicell. The smallest colony in which I have found this 
process actually beginning measured ‘77 mm. in diameter, 
while the largest colony in the same condition as regards its 
ovicell was 1:13 mm. across. In nearly all the cases I have 
actually observed at this stage, 2? was the fertile zocecium, 
although this rule is by no means invariable. A reference to 
the woodcut on p. 78 will show that the orifice of 2? may be 
described as having an axial and an abaxial side. The occlu- 
sion of the zocecium commences with the formation of a calca- 
reous film, starting from the abaxial edge of the zocecium. 
This film grows inwards as a cap to the zocecium (fig. 18) ; 
but while it is closely fitted on to, and indeed arises from the 
abaxial edge of the orifice, it grows over the axial side of the 
edge as an arch which does not fuse with the edge. The film 
is the beginning of the roof of the ovicell; and it is clear 
from fig. 13 that a passage leading from z? into the cavity of 
the ovicell is left between the calcareous film and the axial 
side of the edge of the orifice of z2. Through this passage 
the embryo passes from the fertile zowcium into 
the ovicell. It should perhaps be pointed out that this 
statement is to some extent a matter of inference. I cannot 
claim to have first found this passage in a particular colony, 
and then to have observed the embryo on its way through the 
passage. But I have been able to show that the embryo 
actually passes from the fertile zocecium into the ovicell 
through an aperture close below the roof of the latter; and it 
may fairly be concluded that the passage whose formation we 
have just considered is really formed for this purpose. 

No sooner has the fertile zocecium been occluded than the 
formation of the orifice of the ovicell is commenced. This is 


88 SIDNEY F. HARMER. 


seen in an early stage in fig. 5, where the trumpet-shaped 
mouth is not yet formed. Transparent preparations (fig. 12) 
show that the trumpet is developed in the immediate neigh- 
bourhood of the occluded zocwcium. The appearance of the 
fully formed trumpet is shown in fig. 7. 

The occlusion of a fertile zoecium may or may not result in 
the obliteration of all signs of its presence. In fig. 13 the 
principal spike of z? is still visible externally, in spite of the 
occlusion of that part of the orifice. In other cases, a careful 
examination of a colony may result in the detection of the 
ends of two or three spikes in the immediate neighbourhood 
of the aperture of the ovicell, which clearly belong to a zoecium, 
and from their relation to the trumpet must have belonged to 
the occluded fertile zoccium. What is probably part of the 
orifice of z*? is seen as a zigzag line below the aperture of 
the ovicell in fig. 7. 

The study of sections shows certainly that z? is not neces- 
sarily the fertile zoecium, although the cases in which either 
this or z> is fertile very largely preponderate over all other 
cases put together. The study of entire colonies amply con- 
firms this, if we may for the present assume that the conjunc- 
tion of an occluded zocecium with a trumpet-shaped aperture 
belonging to the ovicell is a strong reason for believing the 
zoecium to have been fertile. In many cases where the open- 
ing of the ovicell is formed in connection with z®, it occupies a 
very characteristic position with regard to that zoecium, being 
placed immediately on the distal side of the original orifice of 
2°, and with its trumpet-like mouth directed towards the disc 
from which the colony started its growth. The opening of 
the ovicell less commonly has a similar relation to 2”. 

It will be remembered that some few of the oldest zoccia of 
each colony form a group which diverge from the centre of 
the discoidal colony. These zocecia are connected with one 
another by vertical septa which do not necessarily reach the 
roof of the ovicell. In other words, a part of the roof which 
starts at one vertical septum does not always become connected 
with the adjacent septum, but may form an arch extending 


ON THE DEVELOPMENT OF LICHENOPORA VERRUCARIA. 89 


completely above that septum to one beyond it. In this way 
is formed what may be called a compound alveolus; i.e. an 
alveolus which at its formation is partially subdivided by one 
or more vertical septa which do not reach the roof of the ovi- 
cell. The closure of the central part of the ovicell commonly 
and probably usually takes place by the formation of one or 
more compound alveoli, whereas the extension of the ovicell 
in a mature colony appears to take place for the most part by 
the formation of simple alveoli. 

The extent of the central area of the colony, devoid of 
zocecia, depends in the first instance on the extent to which 
the older zocecia diverge from one another; but it may be in- 
creased by the occlusion of zoccia. This process, which can 
be best made out in transparent preparations, has nothing to 
do with the occlusion of the fertile zoecium. It is highly 
probable that the occlusion follows the loss of the polypide.1 
The process does not appear to take place to any great extent 
in young and healthy colonies of L. verrucaria, but in old 
colonies a considerable number of the more centrally placed 
zocecia may be occluded. A calcareous lid may in these cases 
be added in the position of the original orifice (Pl. 7, fig. 9), 
and it may or may not be formed in an oblique position corre- 
sponding with the original condition of the orifice. This is 
very different from the state of an occluded fertile zocecium, in 
which the closure takes place at the level of the roof of the 
ovicell. 

All the older colonies of Lichenopora differ from those 
which have so far been described by the occurrence of a very 
striking process of secondary thickening of the surface. Since 
old colonies are further remarkable for producing new broods 
of embryos (cf. p. 132), which are developed in a different way 
from the first brood, the idea suggests itself that the second- 
arily thickened colonies are the result of a second year’s 
growth. I have no actual evidence in favour of this sugges- 
tion, but the fact that secondary broods of larve are produced 


1 Cf. my former paper on Crisia (5, p. 142). 


90 SIDNEY F. HARMER. 


in old colonies at the same time of year as primary broods from 
young colonies is in favour of that view. 

In fig. 14 places will be noticed, as at B, where the neigh- 
bouring alveoli do not meet accurately. These places are the 
** cancelli”” which are so commonly met with in accounts of 
the genus Lichenopera. The cancelli do not at first develop 
to any great extent; but this condition is altered in the old 
colonies (figs. 8, 9). 

Fig. 9 shows three zocecia of an old colony. One of these 
zocecia has been occluded by a calcareous cap at a considerable 
height above the level of the ovicell. This is quite similar to 
the structures which are well known to occur in profusion 
over the more centrally placed zocecia of certain species of 
Diastopora. 

The alveoli which are shown in fig. 9 are separated from 
one another by shallow grooves. At these regions, and at the 
junction of a zocecium with the roof of the ovicell, vertical 
septa are rising up. Fig. 8 shows the appearance of a colony 
after this secondary thickening has been proceeding for a longer 
period. The system of vertical septa, which in the earlier 
stage started only at the intersection of the alveoli with one 
another or with the zoccia, has now extended so as to cover 
the entire surface of the ovicell. The porous roof of the latter 
can be seen at the bottom of the open cancelli, which have 
for the most part no calcareous roof. The base of the right- 
hand zoccium is surrounded by very large cancelli, the walls 
of which (as in fig. 9) are continuous with the longitudinal 
ridges of the outside of the zocecium. These cancelli are 
obviously encroaching on the base of the zocecium, which can 
be seen at the bottom of the spaces. Those marked ‘cancelli”’ 
are beginning to acquire a calcareous roof. The process of 
roofing in is much further advanced at the base of the left-hand 
zocecium, where one of the cancelli is completely closed. In 
this way the bases of the zocecia of old colonies commonly 
become surrounded by a set of blister-like swellings, which 
have been formed above the original roof of the ovicell. These 
are presumably the structures which Smitt regards as forming 


ON THE DEVELOPMENT OF LICHENOPORA VERRUCARIA. 91 


a new ovicell superposed on the old one (cf. p. 75). Judging 
by the analogy of the walls of the alveoli, I think it possible 
that the floor of these blister-like spaces may ultimately be 
absorbed, and the ovicell thereby enlarged. 

It is hardly possible to examine figs. 8 and 9 without feeling 
convinced that there must be a layer of living tissue outside 
the roof of the ovicell. Although I have not obtained complete 
evidence with regard to the morphological character of this 
layer, its presence can be demonstrated in vertical sections of 
stained colonies which have not been decalcified. In pre- 
parations of this nature a film of stained nucleated tissue can 
be seen, passing over the outer side of the calcareous parts of 
the upper surface of the colony, and thus covering the upper 
side of the roof of the ovicell, and extending up the outside of 
the zocecial tubes to the orifices of the latter, where the tissue 
passes round to join the part of the body-wall which is invagi- 
nated into the orifice of each zowcium. The ridges on the 
outer side of the zocecia are as a matter of fact undeveloped 
septa, which can continue their growth under certain circum- 
stances. It is thus natural to find that the outer wall as well 
as the inner wall of the young and active zocecia should be 
covered with living tissue. 

Where the protoplasmic layer passes over the convex surface 
of an alveolus, or up the uniform surface of a zocecium, it is 
usually tightly stretched over the adjacent calcareous matter. 
But in places where it crosses a groove between two alveoli, 
the layer usually appears as a tangent of the two curved surfaces. 
Hence one sees in sections a more or less triangular space, 
roofed only by a delicate membrane, at the junction of two 
alveoli. It is obvious that this is in accord with the arrange- 
ments shown in fig. 9. 

The calcareous part of the upper surface of a Lichenopora 
is thus not the outermost layer; and it becomes a question 
how far it corresponds with the “ectocyst”? of some other 
Polyzoa. In the Entoprocta, the Phylactolemata, and the 
Ctenostomata, in which calcification does not usually occur, 
the outermost layer is a mere cuticle; and it is to this layer 


92 SIDNEY F. HARMER. 


that Allman (1, pp. 8, 13) gave the name of “ ectocyst.” 
There is, however, a good deal of reason for believing that the 
arrangements are not quite so simple in calcareous Polyzoa. 
Ostroumoff (14) announced in 1885 that the use of silver 
nitrate demonstrated the evidence of an epithelial layer external 
to the calcareous skeleton in several genera of calcareous 
Polyzoa; and later (15, p. 12) expressed himself to the effect 
that in calcareous forms, such as Lepralia, the calcareous 
matter separated the ectoderm into two layers, one of which is 
outside the skeleton. Vigelius (22, sep., pp. 3, 4, pl. vi, fig. 24) 
describes and figures an extra-skeletal layer in an encrusting 
Cheilostome, and in a later paper (23) states that the skeleton 
is probably usually formed inside the ectoderm-cells. Pergens 
(17) finds that the skeleton of the calcareous forms examined 
by him is overlaid by a cuticle merely. 

The relation of the calcareous skeleton to the parts of the 
living body-wall appears to need further elucidation; but the 
condition found in Lichenopora makes it probable that, in 
that genus at least, there is more than mere ectoderm outside 
the calcareous layers of the upper surface of the colony. The 
cavities of the cancelli are probably morphologically parts of 
the body-cavity; and as such are no doubt surrounded by more 
than ectoderm. 

It appears to me extremely likely that the relations of the 
calcareous septa in Lichenopora may not be very different 
from those of the septa of a Zoantharian Coral, as shown by 
Koch’s well-known researches. The outer wall of the young 
funnel-shaped colony (fig. 3) is immediately derived from the 
body-wall of the metamorphosed larva; and it is of course a 
young condition of the basal lamina. All the septa which 
divide the zocecia, the alveoli, and the cancelli from one another 
are ultimately derived from the basal lamina; the first forma- 
tion of a young septum at the growing edge always appearing 
as an outgrowth from that layer (fig. 14), and other septa being 
always formed from pre-existing septa. 

I have obtained no evidence of the existence of soft tissues 
on the lower side of the basal lamina, and I do not think it 


ON THE DEVELOPMENT OF LICHENOPORA VERRUOARIA. 93 


likely that any such exist. The view which seems to me to be 
suggested by my observations and by those of previous workers 
is that the basal lamina is the outermost layer of the body- 
wall; that the body-cavity it encloses has been repeatedly 
subdivided by the ingrowth of a series of branching and 
anastomosing ridges or septa from the body-wall. The whole 
upper surface of a Lichenopora colony thus corresponds with 
what would be seen in an end view of a growing point of 
Crisia. But whilst in the latter the growing edge is sub- 
divided by a very simple set of septa, which result in the 
formation of only two series of zocecia, that of Lichenopora 
is divided by an immensely complicated series of septa. The 
parallel is not, however, quite accurate ; since in Crisia, as in 
Tubulipora (cf. p. 81), the septa reach the growing edge of 
the colony, and the mature zoccia are thus completely outside 
the “ growing-point,” because part of the growing edge is used 
up in toto in the formation of each zocecium. The morphology 
of the Lichenopora colony can be most easily understood by 
considering a young colony like fig. 3. It will then be seen 
that while the lateral body-wall is complete from the beginning, 
the mouth of the expanded funnel is closed by a living body- 
wall. While this is constantly encroached on and altered by 
the formation of septa, whether those giving rise to new 
zocecia or those which form alveoli cr cancelli, the body-wall 
is never completely calcified in this region. For this uncalcified 
body-wall, which closes the mouth of the funnel-shaped colony 
of Lichenopora, of the growing points of Crisia, &c., the 
name “terminal membrane” may be suggested. It is this 
layer which gives rise to the polypide-buds, and is invaginated 
into the orifice of every zocecium. 

If the septa are really, as there seems every reason to believe, 
derivatives of the basal lamina, there is no reason for not re- 
garding them as ingrowths of a true calcified ectocyst; and 
they may accordingly be compared, in a general way, with the 
septa of a Coral (as already pointed out), or with the chiti- 
nous ingrowths which form the endoskeleton of a crayfish. 


94, SIDNEY F. HARMER. 


Internal Structure.—The general appearance of a_ section 
through a colony in which the production of the first brood 
of embryos is at its height will be understood from fig. 11, 
representing a section, parallel to the flat surface, of a colony 
in about the same stage as that shown in fig. 12.1 The section 
has a maximum diameter of °80 mm. The ovicell may be ex- 
plained by comparing its contents with an Ameba with a 
series of blunt pseudopodia extending in the intervals be- 
tween the zocecia. The “ pseudopodia”’ contain a considerable 
number of secondary embryos, and the nucleus may be repre- 
sented by the “fertile brown body.” 

This structure, although an inert body which plays no direct 
part in the development, is in fact the centre about which all 
the most important phenomena in the development of the ovi- 
cell take place. The “ pseudopodia”’ form a complex structure 
for which the name “embryophore”’ may be suggested, and 
they lie in a cavity which may be regarded as the body-cavity 
of the ovicell. 

The zoccia can be identified, in a good and well orientated 
series of sections, by an examination of the proximal end of 
the colony, which, it will be remembered, retains throughout 
life the form assumed by the young Lichenopora. Hence 
it may be confidently stated that z is a section of the pri- 
mary individual of the colony, and that z? and z® represent 
respectively the second and third zowcia, as defined in the 
earlier part of this paper. 

z} possesses a polypide, whose tentacles, enclosed in their 
tentacle-sheath, are seen in the figure; z? is quite empty 
basally, but contains some embryos distally, as shown by the 
figure; z° contains a brown body basally, but it has no polypide, 
and it opens at the level of the section into the ovicell. Two 
of the zoccia contain brown bodies, that of A not being visible 
in the section figured. 

The fertile brown body is in the neighbourhood of z?, near 
whose upper end is developed the trumpet-shaped aperture of 


1 The secondary embryos are more developed in fig. 11 than they would 
have been in fig. 12. 


ON THE DEVELOPMENT OF LICHENOPORA VERRUCARIA. 95 


the ovicell (not seen in the section figured). There is no 
aperture in connection with z’, or indeed with any of the other 
zocecia. The colony has every appearance of being perfectly 
healthy. The sections of its basal part show that nearly 
every zoecium has a functional polypide. Only one of these 
polypides has any trace of a testis, which consists of a small 
mass of degenerating spermatozoa. 

The occurrence of an ordinary brown body in z’? makes it 
improbable, as will be shown later, that this was the fertile 
zocecilum ; and it is in the highest degree probable that the 
primary embryo of this colony was produced by 2%. This is 
indicated by the position of the apertnre of the ovicell and by 
that of the fertile brown body, as well as by the occurrence of 
embryos within the cavity of z2. The connexion between z°® 
and the ovicell probably implies that z*? had been occluded 
(cf. p. 89). 

My preparations include a practically continuous set of stages 
- between fig. 11 and the commencement of the development ; 
and my results are based on the examination of complete 
series of sections through some three hundred colonies, in 
eighty-nine of which no embryos can be detected, while in 
seven more, degeneration had set in after the commencement 
of the development. The large number of fertile colonies 
examined enables me, to a large extent, to eliminate abnorma- 
lities in the development. 

An examination of the cases in which normal embryonic 
development was not found will give some useful information. 
The negative result in at least twenty cases was probably due 
to the fact that the colonies were too young. A careful com- 
parison of my measurements of entire colonies and of sections 
shows that the commencement of the embryonic development 
is to be looked for in colonies which measure about *40— 
‘43 mm. in total length, and about ‘25 mm. in transverse 
diameter, i.e. in colonies which are at about the stage of 
fig. 2. 

In the next place I am able to confirm the suggestion I 
formerly made (6, p. 212) to the effect that the normal develop- 


96 SIDNEY F. HARMER. 


ment of the ovicell is dependent on the activity of the poly- 
pides. On decalcifying a number of colonies of Lichenopora, 
a considerable number of all ages and sizes will be found to 
be practically empty. In studying the embryonic development, 
time is simply wasted if preparations are made of colonies in 
which the polypides are not for the most part fully active and 
functional, or at least in which the polypides have not recently 
been functional. A recently degenerated polypide may be 
difficult to distinguish, in a preparation of an entire colony, 
from a functional polypide; whereas it is quite easy to recog- 
nise the small, compact, brown body which indicates that 
the histolysis of the polypide took place at some more distant 
period. 

The conditions under which the embryos are nourished 
appear to be very different in different Cyclostomes. In 
Crisia ramosa I have shown that the embryos are contained 
in a highly protoplasmic reticulum, which, it can hardly be 
doubted, serves for the transference of nutriment to the de- 
veloping larve. In C. eburnea the reticulum is reduced to 
a minimum, and the path of the nutrient substances is pro- 
bably somewhat different from that inC. ramosa. Analogous 
phenomena probably occur in the genus Lichenopora; and 
L. verrucaria may be compared with C. ebur nea, in which 
the reticulum is but slightly developed. Although I have not 
yet been able to obtain a supply of material for the proper 
examination of other species of Lichenopora, I have obtained 
a series of sections through a single fertile colony of L. hispida, 
Flem. The ovicell contains a comparatively small number 
of young secondary embryos, which are embedded in a large 
solid nucleated mass which probably corresponds, in func- 
tion at least, with the “ follicle” of C. ramosa (cf. 6, Pl. 
xxii, fig. 6). 

My examination of this single series of sections of L. 
hispida enables me to state that the embryonic development 
of that species probably differs to a very considerable extent 
from that of L. verrucaria. Another curious difference 
between the two species may here be pointed out. While in 


ON THE DEVELOPMENT OF LIOHENOPORA VERRUCARIA. 97 


L. verrucaria the brown bodies formed by the degeneration 
of the polypides are absorbed from time to time, so that a 
zocecium hardly ever contains more than one or at most two 
brown bodies, those of L. hispida (again judging from my 
single series of sections) accumulate at the basal end of the 
zocecium, a considerable part of which is tightly packed by a 
mass of brown bodies. 

It seems clear that considerable differences in detail occur in 
the embryonic development of different Cyclostomes; but I 
believe that it will be found that the phenomena are funda- 
mentally the same throughout the group. 

Whatever may be the character of the arrangement by which 
the embryos are directly nourished, there can be no doubt that 
it is the polypides which ultimately supply the material at the 
expense of which they grow. In a very large proportion of 
the cases in which no embryo could be discovered, the poly- 
pides were not functional in all or some of the zowcia. A 
similar result is obtained by the examination of the degenera- 
tion of the embryo in fertile colonies. Leaving out of account 
doubtful cases, I have fifteen colonies in which degeneration of 
the embryo has clearly occurred. In eleven of these cases all 
the polypides of the colony had degenerated, this being pro- 
bably the proximate cause of the degeneration of the embryo. 
In some cases new polypide buds were being developed, and 
the colony would obviously have survived, and would probably 
have developed a new embryo from one of the younger poly- 
pides. One of the other four cases probably points in the 
same direction. The fertile zocecium contains two brown 
bodies, a polypide, and an embryo which commenced to dege- 
nerate in the ‘ suspensor stage ” (cf. fig. 24, a normal embryo 
at this stage), The remains of the embryo are between the 
two brown bodies, and it is highly probable that the distal 
brown body represents the polypide which originally supported 
the suspensor. The abnormal degeneration of this polypide, 
at an earlier period, probably resulted in the degeneration of 
the embryo, which is still just recognisable, although a new 
polypide has grown up in the fertile zocecium. 

VOL. 39, PART 1.—NEW SER. G 


98 SIDNEY F. HARMER. 


A much larger number of degenerating embryos might have 
been obtained, if the colonies of which sections were to be cut 
had been taken at random, instead of having been chosen with 
some regard to the condition of their polypides. 

I am unable to suggest any definite cause for the degenera- 
tion in most cases. A Protozoon, a species of Folliculina, 
is commonly found with the basal end of its tube embedded in 
the Lichenopora, its free end projecting from the surface of 
the colony. These organisms can often be recognised in the 
sections of the colonies, and in many cases they are found in 
degenerating specimens. It is quite possible that the Foll1- 
culina may in these cases be the cause of degeneration ; but 
it is equally possible that it is the effect, and that the Protozoa 
settle down in parts of the colony which have lost their zocecia. 
They are found, not uncommonly, in parts of the colony in 
which their presence does not seem to have had any effect on 
the adjacent zocecia. 

Degeneration may start at any stage of the embryonic de- 
velopment. I have observed it most commonly in colonies in 
which the embryo had developed up to the “ suspensor stage ; ” 
but I have also found it commencing at later stages, as, for 
instance, at the beginning of embryonic fission, or even later. 

In describing the development of the structures found within 
the ovicell of Lichenopora verrucaria, it is desirable to 
keep the account of the first brood of larve entirely separate 
from that of later broods. The appearance of an ovicell con- 
taining larve belonging to the first brood has already been 
considered (fig. 11); and this is the condition which is usually 
found in colonies which have a diameter of 1 mm. or there- 
abouts. The origin of all these complicated struc- 
tures is to be looked for in colonies which consist 
of only three or four zoecia. 

For descriptive purposes it will be convenient to classify 
the embryos in a certain number of stages; and it will be 
seen that the form of the entire colony has a distinct relation 
to these stages, although wide variations occur in the period 
at which any particular stage of development is passed through. 


ON THE DEVELOPMENT OF LICHENOPORA VERRUCARIA. 99 


The following stages are naturally marked out by the phe- 
nomena of the development. 

Stage A.—Formation of the definitive egg (figs. 15—18). 
The entire colony in which this stage occurs measures about 
‘45 mm. in length, and about ‘25 in transverse diameter 
(between figs. 2 and 8). 

Stage B.—Formation of the follicle and suspensor (figs. 
19—22). The colony does not differ materially in size from 
those in which Stage A occurs. 

Stage C.—‘Suspensor stage” (figs. 10, 23, 24), the 
embryo being supported by a functional polypide. Colony in 
the condition of figs. 3 and4; usually about ‘50—-60 mm. long, 
and about °35—'55 mm. broad. 

Stage D.—The polypide has degenerated, and the embryo 
is supported by a brown body found near the distal end of the 
fertile zocecium (figs. 27, 28). Colony as in Stage C. 

Stage E.—Disappearance of the “ suspensor”’ and enlarge- 
ment of the embryo (figs. 29, 30). 

Stage F.—Commencement of embryonic fission (figs. 31— 
35). Diameter of colony about ‘60—:80 mm. 

Stage G.—Fully formed ovicell (figs. 7, 11, 12). 

The measurements given here and elsewhere, so far as they 
refer to sections, must be taken as approximate. The plane 
of the sections may not be such as to permit of a satisfactory 
measurement being taken at all, while in the case of an 
oblique series the whole length or breadth of a colony may 
not be seen in any one section ; and the necessity of estimating 
the length obviously introduces a source of error. The delicate 
calcareous lamina which forms the edge of the entire colony 
is usually not distinct in sections. Lastly, there is consider- 
able variation in the size of the colony and in the time at 
which development commences. 

In spite of these drawbacks, I believe that it will be useful 
to record the measurements of the colony, wherever possible, 
as a guide to future investigators. The measurements given 
in the sequel refer in all cases, unless the contrary is stated, 
to the sections, and not to the colony before decalcification, 


100 SIDNEY F. HARMER. 


During my study of the development of L. verrucaria, I 
have been much struck with the frequency with which the 
zocecia z” and z® form the starting point of the. ovicell. This 
result was first obtained by a study of sections; but, as has 
already been pointed out, it was afterwards amply verified by 
an investigation of the entire colonies. Some statistics will 
bring out the striking nature of this phenomenon. In 108 
colonies belonging to Stages B to G, or being fertile colonies 
in which degeneration of the embryo was taking place, I have 
been able to determine the fertile zocecium. In the cases of 
the older colonies (Stages E to G), it is easier to recognise the 
fertile zocecium when it happens to be z? or z? than when it is 
a younger zocecium ; and the older stages have consequently 
been omitted in the second column of the table. 


Table showing the Frequency of the Occurrence of 
z* or z° as the Fertile Zowcium. 


Stages B—G 
and degenerating Stages B—D. 
fertile colonies. 
No. of cases in which z? is fertile ‘ ‘ 44 j 99 
2 rs Z . ‘ - Si ‘ 15 
29 y) 2° or 23 ee) . . 8 ° 5 
55 R 5 Pizaee : : 15 ; 1 
“4 3 Piz us : ; 8 : 6 
re a a younger zocecium is 
fertile . : 3 6 ‘ 5 
- s two zoccia are fertile . 6 : 3 
“A a the fertile zocecium could 
be determined . 5 108 ‘ 75 
Total number of colonies investigated . : 175 : 95 


As it is by no means easy to determine the fertile zocecium 
in unfavorably orientated sections, there is obviously room 
for error in the above results; but as I have only entered 
cases in which I have felt confident that I could really deter- 
mine which zoccium was fertile, I believe that the table is 
substantially accurate. In estimating the bearing of these 
numbers it must be noticed that in a considerable proportion 
of the cases in which the fertile zocecium is entered as ?2? or 


ON THE DEVELOPMENT OF LICHENOPORA VERRUCARIA. 101 


?28, the doubt usually arises in consequence of the difficulty of 
distinguishing between z? and z® in the sections. Thus in 
most cases where a fertile zoceclum was marked ?2z?, it was 
certainly z° if it was not 2’. 

Another line in the table refers to cases in which the fertile 
zocecium was either z? or z®, but it was not certain which of 
these two was fertile. 

We thus obtain the result that in the colonies of all ages 
up to Stage G the fertile zocecium was either z? or z> in 
96 cases, while a younger zocecium was fertile in only 6 
cases in which a single embryo was present. The correspond- 
ing numbers for Stages B to D only are 67 and 5. 

Even if a liberal deduction is made from the 96 of the first 
column or the 67 of the second, on the ground of uncertainty 
or possible errors of observation, we are still entitled to con- 
sider that the fertile zocecium which produces the first brood 
of embryos in L. verrucaria is normally one of the two zocecia 
which are first formed by budding after the larva has meta- 
morphosed itself into the primary zocecium of the colony. Two 
of the six cases alluded to in the table as colonies in which 
two zocecia are fertile, have embryos in both z? and z3. In two 
more cases the fertile zocecia were probably z? and z?, In the 
fifth case z?, and in the sixth case z® was fertile in addition to 
a younger zocecium than either of these two. 

The table further brings out the fact that z? is more often 
fertile than z?. An examination of young colonies which have 
been stained and mounted whole, without decalcification, shows 
that z is really older than z’, although there is no evidence of 
this in the later stages. Thus a colony ‘82 mm. long by 
‘15 mm. broad had a mature polypide in z', and a small 
polypide-bud in z*. Another, °35 mm. by ‘24 mm., had poly- 
pides in z! and z?; while a third, -43 mm. by ‘27 mm., was 
developing for the first time a polypide-bud in 2’, in addition 
to having functional polypides in the older zoccia. 

z” is thus clearly older than z’, and we accordingly have the 
following striking result. ‘There is a strong tendency in the 
development of Lichenopora verrucaria for the ovicell to 


102 SIDNEY F. HARMER. 


owe its origin to the oldest zocecium which is formed asa bud 
and does not result from the metamorphosis of the larva; or, 
in other words, the ovicell tends to be formed from the first 
blastozoite of the colony. I have found no case in which the 
ovicell owes its origin to z'; but it must not therefore be con- 
cluded that the primary individual differs from the others in 
being devoid of generative organs. z! commonly has a well- 
developed testis, while I have on several occasions observed 
structures in z! which I consider to be egg-cells. 

It will be seen from the figures of the entire colonies that the 
determination of z®? and z? is easy enough in sections in many 
cases in which the orientation is good. It is more difficult 
when, as commonly happens, the disc from which the colony 
originates is injured or lost when the Lichenopora is removed 
from the seaweed. The determination of the fertile zocecium 
is greatly facilitated by noticing the position of the tentacles of 
the polypides. 

The woodcut on p. 78 shows that z! may be considered to 
have an inner (axial) and an outer side. During the retracted 
condition of the polypide the tentacles lie close to the inner 
side of the zocecium, while the cecum of the stomach is situated 
more externally. The orientation of the polypide of z’ is like 
that of the polypide of z'; but the polypide of z* faces in the 
opposite direction, its tentacles lying, like those of z!, close to 
the inner border of the zoccium. ‘This appears to be true of 
all the polypides developed in any zocecium. After a given 
polypide has degenerated into the condition of a brown body, 
the newly formed polypide is found to have the same orienta- 
tion of its tentacles as its predecessor had. The younger 
zocecia have a corresponding arrangement. In any colony the 
polypides are normally orientated during retraction in such a 
way that the tentacles of a polypide are nearer to the centre of 
the circle enclosed by the growing margin of the colony than 
is the stomach. This point is, however, not brought out by 
any of the figures which illustrate this paper. It may be noted 
that the number of the tentacles is constantly eight. 

A further point which deserves special notice is the character 


ON THE DEVELOPMENT OF LICHENOPORA VERRUCARIA, 103 


of the “ brown bodies.” Barrois (2 and 3) has shown that the 
brown body formed by the degenerating parts of the larva 
during metamorphosis has disappeared completely by the time 
that the first polypide is mature. He has not, however, de- 
scribed the way in which this disappearance takes place. 
Ostroumoff (16, p. 185) states that the cecum of the primary 
polypide is formed ‘“ at the cost of the brown body;” but he 
does not enter into further details. 

I can confirm Barrois’ statements with regard to the dis- 
appearance of the degenerated remains of the larva, since I 
have found that the primary zocecium, when very young, pos- 
sesses a mature polypide, but has no trace of a brown body. 
I have found this condition lasting till the stage is reached 
when z” has a mature polypide, and z’ possesses either a bud or 
amore or less mature polypide. At a later stage the primary 
zocecilum contains a brown body as well asa polypide. The 
smallest colonies in which I have found this condition mea- 
sured *33 X ‘16 mm. and *35 x ‘21 mm. respectively. This 
implies that the functional polypide is in these cases not the 
original tenant of the zocecium, although in some colonies of 
the same or even a larger size I have found no brown body. 
The degeneration of a polypide, in fact, results in the forma- 
tion of a “brown body,” which in the younger stages of the 
colony nearly always remains as a distinct structure in the 
zocecium. In the older stages, as we shall have occasion to 
see, the brown bodies commonly disappear from the zocecia. 
When present they are easy to recognise, as they appear as 
bright yellow bodies even in sections which have been stained 
with Grenacher’s hematoxylin. A part of the cecum of the 
stomach has the same colour in sections. 

I do not think that the primary polypide of z! develops a 
testis, although that organ is commonly present as an append- 
age of the polypide of z' in cases where the corresponding 
zocecium possesses a brown body, 


104 SIDNEY F. HARMER. 


Stage A.—Formation of the Definitive Egg. 


Figs. 17 and 18 have already been described in my pre- 
liminary note (8). Fig. 17 represents a colony consisting of 
three fully formed zocecia only. The section is parallel to the 
surface of attachment, in what may be called a horizontal 
plane. The examination of the complete series of sections 
showed that both z! and z? have a brown body, a polypide, and 
a testis. The level of the section figured is not such as to 
show the characteristic orientation of z? (see p. 102); z> has a 
brown body, a polypide, and a conspicuous cell (seen more 
highly magnified in fig. 18), which obviously resembles an 
egg. A few nuclei to the right of the egg probably represent 
the commencement of the investments of the embryo. Sper- 
matozoa were found floating freely in the body-cavity of z! and 
z’, but none were found in the fertile zocecium in this parti- 
cular case. 

I have not observed the process of fertilisation: with any 
degree of certainty, but I believe that it must occur at about 
this stage. The fertile zooecium may or may not have a testis, 
while testes are commonly present in other zocecia, particularly 
during the early stages of embryonic development. In some 
cases, and particularly in old colonies in which new broods of 
embryos are developing, the testes are enormous, being some- 
times as much as *30 mm. long. It can hardly be doubted 
from these facts that fertilisation does occur at some period. 
Considerable masses of ripe spermatozoa are found in the body- 
cavity of many of the zocecia; and it is probable that they pass 
thence, in some way, to the fertile zocecium. The young sper- 
matozoa are always in groups of four, as in Crisia (6). 

The correct identification of the large cell shown in fig. 18 
is clearly of the first importance, and I have thought it 
desirable to figure two other cases of the same kind (figs. 15 
and 16). 

Fig. 15 shows an egg attached toa polypide-bud. z!, z?, and 
z° contained polypides, but none of the other zocecia were old 
enough to possess more than buds. The egg seems to belong 


ON THE DEVELOPMENT OF LICHENOPORA VERRUCARIA. 105 


to the primary polypide-bud of a zocecium which is younger 
than z* and z°, since no brown body could be detected in it. 
A precisely similar egg occurred in another polypide bud of the 
same age in the same colony. 

Fig. 16 shows part of the polypide of z? of a rather long and 
narrow colony, measuring ‘48 mm. in length. The first three 
zocecia are fully formed; and each has a brown body, a poly- 
pide, and a testis. z! has either one or two eggs, and z? has 
the two eggs shown in the figure. I have observed in all six 
cases in which eggs were developed in z!, and in one of these 
the egg was found in the recently formed brown body. It 
appears to me that z! may occasionally give rise to true eggs ; 
but I have observed no single case in which an embryo is 
formed in that zoecium. It is probable that the eggs shown 
in fig. 16 would (one or both) have developed, and that the 
zocecium (z°) which contains them would have become fertile. 

In another case the zocecia of a comparatively old colony 
had formerly contained brown bodies and polypides. All the 
latter had, however, degenerated (probably simultaneously), 
forming a series of recent brown bodies, in which the remains 
of testes can be detected. Each zocecium thus contains an old 
brown body, and a younger, half-formed brown body; and, as 
is commonly noticed in such cases, new polypide-buds have 
been developed to take the place of the old polypides. One of 
these, already recognisable as a young polypide, was found to 
possess an egg similar to that shown in fig. 15. The colony 
having undergone a complete process of degeneration, it is not 
surprising to find that this egg is developed in a zocecium 
which is younger than either z? or z°. 

It is obvious from the foregoing statements that the first 
appearance of the eggs differs in different cases. Hither one 
or two eggs may be present, and they may occur with or with- 
out a testis. In one case I have found them in all the first 
three zocecia. More commonly I have detected an egg in only 
one zoecium iu acolony. The diameter of the egg is about 
14 (average of four measurements). ‘This is not very differ- 
ent from the measurement (17°6 ) which I have recorded (6) 


106 SIDNEY F. HARMER. 


for the egg of Crisia, and the correspondence is even closer 
if we take the largest Lichenopora egg (16) of which I have 
a record. 

In a single case I have found an egg-like cell, 9°6 u in dia- 
meter, at the growing margin of acolony. This cell corresponds 
closely in appearance and position with the eggs which I have 
described in the growing-points of Crisia. I do not, however, 
feel by any means sure that this condition really represents a 
normal phase of the development. Some of my other prepara- 
tions suggest that the egg is more probably differentiated in 
situ from the outer layer of a young polypide bud. Moreover 
the brown body with which the developing embryo is associ- 
ated implies the loss of a polypide in the fertile zocecium. One 
might therefore expect, a priori, that the egg would make its 
appearance in a bud formed to replace a pre-existing polypide, 
and not in a part of the growing edge where no polypides have 
yet become mature. 

I have found cells which I regard as eggs in thirty-eight 
colonies. In. most cases I have been unable to detect more 
than a single egg in a colony, although in some cases two or 
even three eggs may occur in a single zoecium. The egg- 
bearing zocecilum is commonly 2? or z®, but in several colonies 
the egg occurred in a younger zocecium. In one or two of the 
latter cases either z? or z> was already preoccupied with an 
embryo of its own. ‘The observations I have noted down with 
regard to the occurrence of these cells point strongly in the 
direction of their being eggs. They do not occur promiscuously 
in all the zocecia. Should a colony have been successful in 
developing an embryo no further trace is ordinarily seen of eggs. 
In young colonies which have no embryo these cells are of 
common occurrence, and in very young colonies they are found 
in just those zoccia which might be expected to produce them. 
At the same time the results are not perfectly consistent, and I 
can best interpret them by assuming that there is considerable 
variation with regard to the first origin of the eggs. This 
appears to me to be particularly the case with young colonies 
in which the development has not been quite normal, whether 


ON THE DEVELOPMENT OF LICHENOPORA VERRUCARIA. 107 


in consequence of the degeneration of the polypides or from 
other causes. 


Stage B.—Formation of the Follicle, &c. 


Although the embryo and its investments can be recognised 
with certainty at this stage, it is by no means easy to make out 
all the details. 

Fig. 19 shows a part of the polypide, with the embryo, of a 
colony which possessed functional polypides in 2}, z’, and 2’, 
and had two other zoccia which had reached the stage at which 
the polypide-bud becomes obvious. The fertile zocecium is 
either z? or z>; it possesses a brown body and a small testis, 
neither of which is shown in the section figured. The body- 
cavity contains some ripe spermatozoa, some of which are seen 
in the immediate neighbourhood of the embryo. The latter 
consists of several nuclei embedded in a mass of protoplasm 
which lies in a definite cavity. This is bounded by a layer of 
cells which may be termed the follicle. 

Fig. 22 is from z* in an older colony which had four old and 
one half-grown polypide, and two buds. No brown bodies 
could be distinguished in this case—a very unusual phenome- 
non in colonies of this age, although quite common in older 
colonies. The four mature polypides are, however, conspicu- 
ously old, and their alimentary canals have solid contents which 
are probably the missing brown bodies. This question will be 
considered later. The investments of the embryo are rather 
more definite than in fig. 19, and it is more difficult to make 
out with certainty the limit between the follicle and the em- 
bryo. At least five of the central group of nuclei clearly belong 
to the latter. 

Fig. 21 shows a transverse section of the follicle and embryo 
at a stage between the last two figures. The colony measured 
‘40 by ‘21 mm., and possessed three polypides only, of which 
either z? or z> was fertile. The follicle is somewhat thicker 
than in the cases previously considered, and possesses more 
than one layer of nuclei. A brown body is present, as usual, 
in the fertile zocecium. 


108 SIDNEY F. HARMER. 


Fig. 20 is from a colony which possessed three polypides only, 
each of them being accompanied by a brown body. The fertile 
zocecium is almost certainly z?, but if not itis z’. The structures 
connected with the embryo show a distinct advance in develop- 
ment. The follicle is more fully formed, and there is also pre- 
sent an elongated group of nuclei which is easily recognised as 
the “suspensor” of the next stage. 

My notes show a perfectly consistent result with regard to 
the age of the colony at which this stage is passed through. 
This is brought out by the following table: 


Embryos in Stage B. 


No. of colonies 


Colonies possessing three polypides only 7 5 “ 16 

+f r four or five polypides_ . : : 3 
Colony larger, the fertile zocecium being younger than z? or 23 1 
Sections not good and result uncertain : ; : 5 
Total number of colonies examined . : : 25 


It is clear that Stage B is most commonly passed through in 
colonies which contain three fully developed polypides, with 
perhaps one or two young buds in younger zoecia. The mea- 
surements of the colonies are remarkably uniform at this 
stage, the total length of the colony (in section) ranging from 
‘40 to ‘48 mm. 

It may be noted that although most of the above colonies 
were in the three-polypide stage, it does not follow that any 
colony containing that number of polypides will also contain 
an embryo in Stage B. We have already seen that a consider- 
able number of these colonies contain eggs. In some cases, 
moreover, it appears that the polypides may be well formed in 
z'—z3 before any of these zocecia undergo a histolysis of their 
polypide, or at least before the polypides of z? and 2° experi- 
ence that fate. Now since the occurrence of a brown body is 
all but universal in all fertile zocecia, of whatever age, and 
probably occurs in all fertile zocecia which are developing 
normally, we can hardly expect to find an embryo in any z? or 
z in which the first polypide of the zocecium has not had time 


ON THE DEVELOPMENT OF LICHENOPORA VERRUCARIA. 109 


to degenerate and to be replaced by a new one. The brown 
body which is found in connection with the developing embryo 
is alluded to in this paper as the fertile brown body. 

The exact origin of the cells forming the embryonic invest- 
ments could not be made out, but I feel confident that they 
are not formed as a modification of an ordinary polypide-bud. 
They appear to me to be differentiated in situ from the cells 
of the funicular tissue which surround the egg, aided probably 
by the cells which form the capsule of the fertile brown body. 
On two occasions I have, however, found young colonies in 
which the appearances could be interpreted by supposing that 
a polypide-bud was concerned in the production of the struc- 
tures connected with the embryo. I have come to the conclu- 
sion that such a view would be erroneous ; and I regard the 
colonies in question as really belonging to Stage D, although 
they have reached that stage rather precocicusly. 


Stage C.—‘‘Suspensor Stage,’ with Brown Body and 
Polypide. 


I have ventured to call this stage by a name taken from the 
development of the embryo in a flowering plant, in conse- 
quence of the resemblance of a cord of cells supporting the 
embryo to the structure known to botanists as the suspensor. 
By this term I refer to the inner cord of cords (figs. 23, 24) 
which is contained in an investment continuous with the cells 
which immediately surround the embryo. ‘These cells, which 
enclose the embryo, may be termed the “follicle ” (fig. 24). 

The whole set of structures which are developed in imme- 
diate relation with the embryo form the “‘embryophore,” a 
structure which has been seen at a later stage of development 
in fig. 11 (cf. p. 94). 

The general appearance of the embryo, with its embryonic 
membranes, at this stage, may be gathered from fig. 10, repre- 
senting part of the fertile polypide in a colony which is about 
‘67 mm. long. The colony possesses in all about nine brown 
bodies, one of which is accompanied by a polypide-bud, and 


110 SIDNEY F. HARMER. 


the rest are in zocecia which have a functional polypide: that 
is, about nine zocecia are fully formed, and of these the third 
(probably) is fertile. The fertile polypide, like some of the 
others, has a testis, from which the embryophore hangs down 
freely into the body-cavity. The fertile brown body occurs in 
the immediate neighbourhood of the embryo. This stage is 
clearly not very different from fig. 20, which has already been 
described. 

Fig. 23 is from a somewhat younger colony (measuring ‘53 
by ‘32 mm.). Five zoccia are here fully formed, with brown 
bodies and polypides, and of these z? is fertile. The outline 
of the fertile brown body is projected from other sections on to 
the figure. In this case the fertile zocecium has no testis, 
although that organ is present in some of the other zoccia. 
The suspensor is, as is frequently the case, curved in such 
a way as to keep the follicle in close contact with the brown 
body; but this curved condition may be merely the position 
assumed during the retraction of the polypide. The suspensor 
obviously consists of more than a single row of cells. Its 
investing cells are intimately connected with a layer which 
forms a kind of capsule round the fertile brown body. The 
follicle consists of a protoplasmic mass containing two more or 
less distinct layers of nuclei, covered externally by a few more 
flattened nuclei, which probably belong to an external flat 
epithelial layer. 

Fig. 24 (from a colony ‘48 by ‘45 mm.) shows the embryo 
and its accessories with unusual clearness. The fertile zocecium 
is 2°. The section suggests that the suspensor is really a tube, 
and that the two rows of nuclei are on opposite walls of its 
lumen. This is borne out by other sections, and particularly 
by transverse sections. The lumen can sometimes be made 
out; but it probably does not always exist as anything more 
than the finest passage. 

Fig. 25 represents an unusual condition, in which two 
embryos are developing in the same follicle. The colony was 
about ‘56 mm. long, and z? was fertile. I have found two per- 
fectly similar cases belonging to the next stage (Stage D), and 


ON THE DEVELOPMENT OF LICHENOPORA VERRUCARIA. 111 


it is quite possible that the phenomenon is not very uncommon. 
After the commencement of embryonic fission, at any rate, it 
would be quite impossible to say whether one embryo or two 
were originally present. It will be remembered that two eggs 
are sometimes borne by a single polypide (fig. 16). It may 
fairly be inferred that under certain circumstances both eggs 
may develop. 

Some idea of the general appearance of the embryo at 
Stage C may be obtained by the inspection of figs. 23, 24, but 
it is often by no means easy to make out the structure of the 
embryo, or indeed to be sure how many nuclei belong to it, and 
how many may have been derived from intrusive follicle-cells. 
I have previously pointed out (6, p. 215) that the follicle- 
cells in Crisia do push their way in among the blastomeres ; 
and I have obtained a good many preparations of Lichenopora 
in which this process appears to be taking place. I can say 
no more with regard to the structure of the embryo than that 
it is a mass of protoplasm containing nuclei; and I am unable 
to say whether, at this stage or even later, these nuclei are 
arranged with any relation to embryonic layers. 

Whatever doubt may exist with regard to the exact details 
in Stages A and B, there cannot be the slightest doubt that the 
“suspensor stage ” is a perfectly normal phase of the develop- 
ment. I have examined some thirty-eight colonies in which 
the suspensor was present in the condition described above. 
With hardly any exception, I have found this stage in colonies 
which measure ‘40—'68 mm. in length (in sections) and ‘30— 
*55 mm.in breadth. The fertile zoeecium may or may not have 
a testis, and I cannot lay down any general rule on the subject. 

In two or three observed cases more than one zocecium was 
fertile. The only other variation which need be recorded, 
although it might equally well have been considered under the 
next stage, is the occasional occurrence of a young polypide bud 
in a zocecium which has reached the normal suspensor stage, 
and has recently lost its polypide by degeneration. 

In some cases this bud appears simultaneously with the 
development, in other zocecia, of buds which are clearly formed 


112 SIDNEY F. HARMER. 


in consequence of the degeneration of the polypides. This is 
a case of a phenomenon repeatedly noticed, in which an 
epidemic of degeneration attacks all the polypides of a colony 
simultaneously. In a large number of cases this probably has 
no permanently injurious effect on the colony, as new polypide 
buds are developed in all the zowcia, even while many of the 
parts of the degenerating polypides are still distinctly recog- 
nisable. Iam convinced that the simultaneous degeneration 
of the polypides is often responsible for the degeneration of the 
embryo; and in some of the above-mentioned cases in which a 
young bud occurs in the fertile zocecium, the embryo would 
probably have ultimately become involved in the brown body 
This is borne out by examination of a colony, ‘51 mm. long by 
‘37 mm. broad, in which all the polypides had moderately 
recently degenerated. Polypide-buds are being developed to 
replace them, and an advanced bud is present in the fertile 
zoecium. ‘This bud possesses an egg, although eggs could not 
be found in any of the other zocecia. The inference is that the 
egg was destined to replace the (degenerating) embryo present 
in the same zocecium. 

Should the buds become functional polypides with sufficient 
rapidity, I see no reason why the embryo should not survive. 
It is possible, on the contrary, that the occurrence of a young 
polypide-bud together with a degenerating polypide in the 
fertile zowcium (Stage C) always implies that the embryo has 
already commenced to degenerate. I am induced by the study 
of one series of sections to think that this is not necessarily the 
case. The fertile zocecium contains a polypide which is clearly 
young. The brown body is large; near the suspensor it is 
distinctly old, while at the opposite end it is as distinctly 
newly formed. ‘The embryo and its accessories are normal in 
character. This case suggests a fusion of the original fertile 
brown body with a new one formed by the degeneration of the 
fertile polypide. A new polypide has, however, been de- 
veloped in the same zoccium, and the health of the embryo 
does not seem to have been in any way affected by these 
changes. 


ON THE DEVELOPMENT OF LICHENOPORA VERRUCARIA, 118 


It is impossible to avoid being struck with the very remark- 
able similarity between the stage which has just been described 
in Lichenopora and a phase in the development of Pluma- 
tella, as described by Kraepelin (11; see especially pl. i, 
figs. 65—68). The figures given by that observer might almost 
do duty for Lichenopora, except that the ‘‘suspensor”’ is 
rather longer and the “ follicle” less complicated than in that 
genus. The “suspensor’’ of the fresh-water form is, more- 
over, attached to the body-wall instead of being borne by a 
polypide or by a brown body. 

Kraepelin and Braem (4) have investigated the origin of 
this structure in Plumatella, and they both arrive at the 
conclusion that the structure which supports and contains the 
egg is a polypide-bud. Braem, indeed, dissents from the 
statement of Metschnikoff (13), by whom the structure was 
first described, to the effect that it is an ordinary bud; and 
points out certain differences between it and a bud which is 
destined to grow into a polypide. But the embryophore in 
Phylactolemata would appear to be a two-layered bud, and 
the part which corresponds to my “suspensor” is simply the 
inner or ectodermic layer of that bud. 

Braem further calls attention to the occurrence, in the 
morula stage, of certain nuclei which are smaller than those of 
the segmentation spheres. These may possibly have some 
relation with the smaller nuclei which I have myself noticed 
in Lichenopora during Stage C. 

The earliest changes undergone by the fertilised ovum of 
Plumatella, according to Kraepelin’s account, appear to be 
extremely similar to those which occur in Cyclostomes. In 
Plumatella, as in Lichenopora, a time arrives when the 
embryo is a sharply marked spheroidal mass supported by the 
end of the “‘ suspensor.” 

From this stage the development of the Phylactolemata is 
very different from that of the Cyclostomata. In the former 
group the embryo enlarges, becomes two-layered and acquires a 
central cavity which is completely closed onall sides. Oneor more 
polypide-buds are invaginated from the two-layered body-wall 

VOL. 389, PART 1.—NEW SER. H 


114 SIDNEY F. HARMER. 


into the central body-cavity, and the embryo finally escapes as 
a larva provided with a larger or smaller number of polypides. 
In Lichenopora, on the contrary, the morula enlarges while 
retaining an embryonic, undifferentiated character in its cells. 
It then undergoes repeated fission, and the definitive secondary 
embryos produced in this way then acquire a two-layered 
character which is strikingly suggestive of the Phylactolema- 
tous embryo. 

In spite of the extraordinary resemblance between the “sus- 
pensor stage” of Lichenopora and that of Plumatella, I 
have not been able to convince myself that the result is arrived 
at, in the former, in the way described by Kraepelin for the 
latter. The structure of the embryophore in Lichenopora 
would indeed be most easily explained by supposing the 
‘* suspensor ” to represent the ectodermic part of a bud, but I 
have not been able to obtain any evidence that this is the case. 
There are, however, instances among marine Polyzoa where a 
new polypide-bud is formed in the immediate neighbourhood of 
the brown body (cf. 7, p. 140), instead of in close connection 
with the body-wall. It is possible that the Lichenopora 
suspensor may belong to this little-understood class of polypide 
buds, but I must emphasise the fact that I have no internal 
evidence tending to prove that it is a bud. 

Were it not for the very marked similarity of the early stages 
of the two groups there would be little reason for anticipating 
an agreement in detail between Cyclostomata and Phylacto- 
lemata, particularly as the details of Lichenopora are so 
different from those of Crisia, another Cyclostome. 

Kraepelin (10) has, moreover, urged a series of reasons for 
believing that the Ctenostomata, among the marine Polyzoa, 
are the nearest allies of the Phylactolemata. As I have made 
no special study of the latter group, I have arrived at no con- 
clusions of my own with regard to their affinities, but the 
details of their early developmental phenomena appear to me 
to be sufficiently similar to those of Cyclostomes to make it 
worth while to call special attention to the fact. 

One other suggestion may be made with regard to the 


ON THE DEVELOPMENT OF LICHENOPORA VERRUCARIA. 115 


Phylactolemata. In spite of the admirable researches of 
Braem (4) and of Kraepelin (11), the mode of the first origin 
of the statoblasts cannot be said to be beyond doubt. Although 
in the absence of any observations of my own on this point I 
do not wish to commit myself to this view, it appears to me not 
impossible that some connexion exists between the formation 
of the statoblasts and-the process of embryonic fission. The 
natural view of older writers that the statoblasts are winter 
eggs may hereafter prove to have had more truth in it than 
has recently been supposed. Verworn (21), indeed, has at- 
tempted to resuscitate the older view, although his results 
have not commended themselves to later inquirers. Braem’s 
description and figures in particular lend some colour to the 
suggestion that the formation of the statoblasts may be a modi- 
fied form of embryonic fission. For fear of being misunder- 
stood I must, however, repeat my statement that I do not con- 
sider that I have sufficient grounds either for accepting or for 
rejecting this suggestion. 

We may now return to the consideration of Lichenopora, 
and we pass to— 


Stage D.—Suspensor Stage, with Definitive Fertile Brown 
Body. 

The normal occurrence of this stage is supported in the 
clearest possible manner by my sections. 

Fig. 28 represents a typical case. The colony measures 
‘43 by ‘37 mm., and it contains six brown bodies, one of which 
is fertile, while each of the others is accompanied by a poly- 
pide. No other zoccium is sufficiently old to have a polypide, 
and the colony is one in which this stage of development has been 
reached comparatively early. zis fertile; it contains a brown 
body, from which the embryophore hangs down into the zoe- 
cium. The brown body is not directly connected with the 
orifice of the zoccium, although it is now situated near the 
upper end of the body-cavity. 

Fig. 27 (Pl. 9) is probably a somewhat later stage, in 
which the brown body has come close up to the layer of body- 


116 SIDNEY F. HARMER. 


wall which is invaginated at the orifice of the zoecium. The 
colony is larger than the preceding specimen, measuring 69 mm. 
by ‘51 mm., and possessing at least twelve brown bodies. The 
fertile zocecium is here neither z? nor z’, but a younger indivi- 
dual of the colony. The figure shows that the suspensor and 
the follicle are still easily recognised. The brown body has, 
however, acquired a much more definite capsule of cells than it 
had in the previous stage. This capsule is not so fully formed 
in the somewhat earlier condition shown in fig. 28; but in 
fig. 27 it has become extremely definite. Its protoplasm and 
its numerous nuclei take up hematoxylin with an avidity that 
shows that the tissue is actively growing, and it is indeed this 
tissue which appears to be, somewhat later, closely concerned 
with the development of the trumpet-shaped aperture of the 
ovicell. The embryo is distinctly larger than it was in Stage C. 

Fig. 26 represents the embryo of another colony (in Stage D), 
which was slightly younger than that from which fig. 28 was 
taken, as it had only five brown bodies and four polypides ; z? 
was fertile. The embryo in this case possesses a peripheral 
layer of nuclei, which surround a central group of three nuclei. 
I am unable to say whether there is any morphological differ- 
ence between the central group and the others. 

It might be supposed, a priori, that Stage D originates by 
the degeneration of the fertile polypide in Stage C. I have 
distinct evidence that this is the case. The polypide is seen to 
begin to degenerate, and the brown body already present 
begins to become confluent with it. The old brown body and 
the degenerating polypide become surrounded by a common 
capsule of cells. The fertile brown body of Stage D hence 
differs from that of Stage C, inasmuch as it contains the 
remains of what was the fertile polypide in the earlier stage, 
in addition to the original brown body. Should development 
proceed normally, no further polypide is developed in the 
fertile zocecium, and the brown body moves to its upper 
end, and eventually comes into close relation with what was 
formerly its orifice. 

The size of the colony during Stage D is in most cases from 


ON THE DEVELOPMENT OF LICHENOPORA VERRUCARIA. 117 


*50 to ‘70 mm. in length, and from ‘35 to ‘55 mm. in breadth, 
although these limits may be passed in either direction. The 
average of eight cases in which I was able to estimate the 
length of the colony (in sections) was ‘53 mm., and the average 
breadth (twenty-two cases) was ‘44mm. The entire colony 
(fig. 4) might well have been in this stage, although still 
older colonies are commonly in Stage D, as may be shown by 
counting in the sections the number of their fully formed 
zocecia. 

The distinction between Stages C and D is obvious enough 
as regards the embryo and its connected structures. But the 
change from one stage to the other implies no more than the 
degeneration of a polypide and its alteration into the condition 
of a brown body. We should not, therefore, expect to be able 
to distinguish by external features alone between colonies in 
the two stages. This is all the more difficult, since the period 
in the growth of a colony which corresponds with a particular 
stage in the embryonic development varies within wide limits. 


Stage E.—Disappearance of the Suspensor and Enlarge- 
ment of the Embryo. 


This stage immediately precedes the beginning of embryonic 
fission. Fig. 29 represents an embryo found in a colony mea- 
suring ‘56 mm. in breadth. The fertile zocecium was almost 
certainly z?. The embryo has unmistakably increased in size 
compared with the embryos in the previous stage. While that 
of fig. 27 measures 28 in greatest length, that of fig. 29 
measures 70 y, which is exactly two and a half times as much. 
Other changes have also taken place. The suspensor has 
ceased to be a definite structure, and the embryo has moved 
towards the brown body, so as to take the place originally 
occupied by the suspensor. The follicle is still distinct in this 
series of sections, but the section figured only cuts its edge. 
The first indication of the commencement of this stage is given 
by the longitudinal elongation of the embryo, which then 
begins to extend beyond the original limits of the follicle into 
the stalk connecting the latter with the brown bedy. We 


118 SIDNEY F. HARMER. 


have seen that the suspensor is potentially a tube, and it 
appears to me that the cells of the suspensor lose the regularity 
of their arrangement, and that the embryo probably passes 
into the midst of the altered cells of the suspensor, perhaps 
by widening out the original lumen. There are, however, 
some reasons for thinking that the suspensor may be pushed to 
one side during the elongation of the embryo, and I cannot 
pronounce definitely on this point. In any case the suspensor 
ceases to be recognisable shortly after the close of Stage D, 
and the embryo invariably elongates so as to approach the 
fertile brown body with its distal or upper end. Towards the 
end of Stage E the embryo may have completely left its original 
follicle, and may lie entirely in the place of the original 
suspensor. 

I have found but few colonies (not more than seven or eight) 
in this stage ; but those few formed a series completely bridg- 
ing over the interval between Stages D and F. The length of 
four of these embryos was 30, 53, 70 (fig. 29), 72 u respectively. 
The average transverse diameter of the colonies measured in 
Stage E (seven cases) was ‘562 mm., the extremes being °48 
(‘77 long) and ‘62 mm. 

The embryophore shown in fig. 29 occupies a position in 
the zocecium exactly like that of the corresponding structure 
in fig. 27; that is to say, the cells surrounding the brown 
body are in contact with an invagination of the body-wall at 
the orifice of the fertile zoecium. The invagination is, how- 
ever, considerably deeper in the zocecium from which fig. 29 
is taken than in that shown in fig. 27. A thin layer of cells, 
not present in the earlier stage, now stretches across the mouth 
of the invagination from one edge of the rim of the zocecium 
to the opposite edge. 

I regard this layer as the remains of the calcareous film 
which occludes the mouth of the fertile zocecium (cf. pp. 86, 
87). The smallest entire colony in which occlusion was just 
commencing measures ‘77 mm. in transverse diameter, whereas 
the largest colony in Stage E (measured from sections) was 
‘62 mm. broad. This difference is quite unimportant, partly 


ON THE DEVELOPMENT OF LICHENOPORA VERRUCARIA. 119 


because of the variation in the size of colonies at the same stage 
of embryonic development, and partly because, in comparing 
the measurements of entire colonies with those of sections, 
something must be deducted from the former series of measure- 
ments. The calcareous lamina which borders the entire colony 
is usually not very apparent in sections, or may become some- 
what crumpled. In correlating the entire colonies with the 
sections it will be well to notice that sections of colonies in 
Stage E give no evidence of an aperture belonging to the 
ovicell. 

Special attention may here be directed to the fact that a 
brown body may often be seen in the occluded zocecium of an 
entire colony. That brown body is of course the fertile brown 
body; and the fact that it may be visible in a dry colony 
when other zocecia are not seen to possess a similar structure 
is of course due to the constant occurrence of the fertile brown 
body near the original orifice. If brown bodies are present in 
the other zoccia, they are usually placed at the end of the 
cecum of the stomach, and are consequently deeply placed in 
the colony, in a position where they cannot be seen. 

The fertile zocecium to which fig. 29 refers has yet one 
further point of interest. Atashort distance below the embryo 
its cavity is completely divided by a septum which passes trans- 
versely to the long axis of the zoecium. This septum, which 
is commonly noticed in, and is indeed probably a normal 
feature of, later stages as well, is probably calcareous during 
life. It cuts off the proximal, empty end of the zocecium from 
the distal end which contains the embryo. 


Stage F.—Commencement of Embryonic Fission. 


This stage is by far the most important of all, as demon- 
strating the fundamental identity of the developmental pro- 
cesses in Crisia and Lichenopora. Fortunately there is no 
stage in the latter in which my evidence is clearer than in this 
one. I have obtained some twenty-seven series, any one of 
which by itself would have been almost sufficient to prove the 
occurrence of embryonic fission. 


120 SIDNEY F. HARMER. 


The youngest figured (fig. 31) shows an embryo at the very 
beginning of this stage. The figure is taken from a colony 
which measures ‘83 mm. in transverse diameter. The diameter 
of the embryo is 45, but it must be remembered, in comparing 
it with fig. 29, that it is cut transversely to its long axis. The 
embryo is now provided with a clear outer layer of nuclei. The 
preparation for fission is indicated by the tendency of the centrally 
placed nuclei to arrange themselves in groups or in regular 
series, and by the appearance of the small slits or vacuoles 
marked S in the figure. 

Fig. 35 is a longitudinal section through an older embryo, 
in which fission has definitely begun. The brown body (not 
cut in this section) is unaltered, and the embryophore still 
hangs down freely into the fertile zocecium. The embryo is 
no longer a rounded mass, but it consists of an irregular series - 
of pieces, which show clear evidence of being engaged in fission. 
The embryonic protoplasm and nuclei take up hematoxylin or 
borax carmine with great readiness; and there can be no doubt 
from their appearance that they are engaged in active growth. 
The pieces into which the embryo divides at this stage have no 
very regular arrangement; but, as will be seen from figs. 33—35, 
the embryonic fragments are at first arranged irregularly round 
a more or less definite central space. The appearance of 
figs. 31 and 83 suggests that a kind of vacuolation of the 
central part of the primary embryo takes place, and that the 
embryonic cells thereby become dissociated into a series of 
peripherally situated groups. In tracing an embryo, such as 
that shown in fig. 34, through a series of sections, the parts of 
the primary embryo, where not completely separated from one 
another, are seen to be connected in the most irregular manner. 
Sometimes they are united with one another laterally ; some- 
times an embryo is united with its vis-a-vis by a diagonal 
connection passing across the centre of the whole mass; some- 
times a piece of embryonic tissue is prolonged into several 
finger-like processes, not unlike those of the primary embryo 
of Crisia (6, pl. xxiii, fig. 11). There is, in fact, no regularity 
or definite method in the breaking up of the primary embryo 


ON THE DEVELOPMENT OF LICHENOPORA VERRUCARIA. 121 


of Lichenopora; and every primary embryo examined at 
this stage differs in these respects from every other primary 
embryo. 

The main fact is, however, the same. The embryo at this 
stage has ceased to be a coherent structure,and has resolved itself 
into a number of complicated lobes of embryonic tissue, some 
of which have actually become separate from their neighbeurs, 
and some are ready to become separate at a moment’s notice. 

Even now a differentiation of two kinds of embryonic cells 
is not necessarily apparent. In fig. 34 there is some indication 
of the occurrence of inner cells, which are probably destined 
to give rise to the inner layer of the secondary embryo. Other 
preparations do not show any such clear differentiation. 

The general features of this stage are very characteristic. 
The fertile brown body is invariably present, and it is situated 
near the upper end of the fertile zoecium. The embryophore 
is shaped something like the lower (closed) end of a test-tube, 
and it still hangs down freely into the body-cavity of the 
zocecium. ‘The embryo or its parts usually lie in spaces which 
probably appear in the surrounding cells of the embryophore, 
by a process of vacuolation of the protoplasm. The aperture 
of the ovicell is beginning to develop. 

The present is a convenient place to raise the question of 
the function of the suspensor. It might have been supposed 
that it was a tube carrying spermatozoa to the egg, if it had 
not been for the fact that the suspensor is formed after the 
development of the embryo has commenced. The time at which 
one would expect fertilisation to take place is the stage before 
the embryonic investments are completed, when the egg is 
hardly separated from the fluid of the body-cavity. At this 
time spermatozoa are commonly found in the neighbourhood of 
the egg or of the embryo in its early stages (cf. fig. 19). 

The function of the suspensor appears to me to be probably 
connected with the nutrition of theembryo. During Stages 
C and D the suspensor remains quiescent; but it probably 
contributes to the formation of the mass of protoplasm in which 
the secondary embryos are supported from Stage F onwards. 


122 SIDNEY F. HARMER. 


It may be noted that the embryophore may now contain 
ege-like cells (fig. 34, a), which are similar to the giant-cells 
that I have described in Crisia. I have never found these 
cells showing the slightest evidence of being really eggs; 
whereas the dividing primary embryo is quite unmistakable 
at this stage. I therefore attribute a purely subsidiary part to 
these cells, and I do not believe that they have any direct 
share in the development of embryos. I have not found giant- 
cells during the earlier stages of the development. 

The aperture of the ovicell commences to develop at this 
stage, as is distinctly shown by the sections. We have seen 
that the fertile brown body of earlier stages is surrounded by 
a deeply staining mass of cells which may be in immediate 
contact with an invagination of the body-wall (fig. 27). In 
this stage the condition closely resembles a stage which I have 
described in Crisia (6, pl. xxiii, fig. 12), except that the 
latter has no brown body. ‘The tubular aperture of the ovicell 
of both Crisia and Lichenopora is closely connected at its 
base with the deeply staining mass of cells or its derivatives. 
In the latter genus these cells grow towards the distal surface 
of the zoccium which contains them, just as in Crisia they 
grow towards the distal surface of the ovicell. A cavity next 
appears in the mass of cells, immediately above (distal to) the 
brown body. This cavity, which is at first merely a series of 
vacuoles, is usually perfectly definite near the brown body, and 
may extend as a space, which appears crescentic in a longitu- 
dinal section, halfway round the brown body. Tracing this 
space upwards, it becomes less definite, and it extends into the 
base of the tubular aperture of the ovicell. A transverse sec- 
tion of the ‘‘ aperture ” at this stage usually shows an external 
thin, body-wall which surrounds the part of the body-cavity 
which extends into the tube. The centre of the transverse 
section is occupied by a solid mass of cells, whose diameter is 
about half of that of the entire tube. Lower down the solid 
mass becomes excavated by a lumen which becomes continuous 
with the cavity which occurs near the brown body. Below 
the latter the cavity passes into an irregular space, traversed 


ON THE DEVELOPMENT OF LICHENOPORA VERRUCARIA. 128 


by strands of nucleated protoplasm, which thus divides the 
space into smaller cavities in which the derivatives of the 
primary embryo lie. 

I believe that the processes which take place at this stage 
are practically like those which occur in Crisia. The whole 
mass of the embryophore becomes highly vacuolated, and is 
thereby transformed into a series of more or less definite spaces 
in which the secondary embryos will lie. These spaces become 
more definite near the aperture of the ovicell. In the next 
stage I have often observed mature embryos in the tubular 
aperture, obviously on their way out to the exterior. 

Very few of the colonies which I have observed in this stage 
have a smaller transverse diameter (in sections) than ‘60 mm. 
The extreme measurements I have noted are ‘45 mm. and 
1:04 mm. The average of the measurements of the transverse 
diameter of twenty-four colonies in this stage is ‘747 mm. 
There is great probability that Stage F corresponds with a 
condition of the entire colony in which the fertile zocecium is 
occluded, and the very beginning of the formation of the roof 
of the ovicell is taking place. 


Stage G.—Ovicell well developed. 


The Production of the First Brood of Secondary 
Embryos is at its Height. 


Fig. 11, which shows a characteristic horizontal section of 
a colony in this stage, has already been described (p. 94). 
The fertile brown body still remains compact and conspicuous ; 
and it occurs in the immediate neighbourhood of the trumpet- 
shaped aperture of the ovicell. During this stage it is, how- 
ever, no longer found in the fertile zocecium, which in the 
section drawn (fig. 11, z?) is seen to contain some secondary 
embryos at its upper end. ‘The cavity of this zoccium has 
become continuous with that of the ovicell in the manner pre- 
viously described (p. 87), and the brown body has passed out 
into the cavity of the ovicell proper. The brown body still 
forms a kind of centre from which the lobes of the embryo- 


124, SIDNEY F. HARMER. 


phore start ; and its position at the base of the aperture of the 
ovicell materially aids in the discovery of that structure in the 
series of sections. 

The secondary embryos do not lie freely in the cavity of the 
ovicell, but are still contained in a protoplasmic reticulum, 
which is the modified embryophore. The latter has largely 
increased in size with the commencement of embryonic fission, 
and has become branched, its lobes extending round the zoecia, 
which pass through the cavity of the ovicell (cf. fig. 6). The 
general appearance of a young embryophore in this stage in 
an entire colony is shown in fig. 12. 

The most striking difference between the development at 
this stage and that of Crisia consists in the absence of the 
primary embryo of the latter, and it often requires close exami- 
nation to prove that fission is still proceeding in Stage G. 
Such, however, is undoubtedly the case, although there is no 
coherent primary embryo left at this period. 

Fig. 37 illustrates the manner in which the number of second- 
ary embryos is increased during Stage G. An embryo similar 
to the smallest ones seen in fig. 11 is dividing transversely. 
The external and internal layers, which are so characteristic a 
feature of the Cyclostome embryo, are now well marked, and 
the inner layers of the two halves of the dividing embryo are 
completely separated from one another. The longest diameter 
of the embryo figured is 56 pn. 

In some colonies it is easy to demonstrate the occurrence of 
this process, and there are many cases where its recent occur- 
rence may be inferred from the fact of two small embryos lying 
close together in a position which suggests that they have 
recently separated from one another. It is obvious that if the 
embryo shown in fig. 37 had been cut transversely to its longest 
diameter it would have been by no means easy to be sure that 
fission was taking place. Another difficulty in proving the 
occurrence of this process is due to the fact that the normal 
development of a secondary embryo which is not going to 
divide further may, in section, appear very similar to the divid- 
ing embryos. The young definitive secondary embryo, which 


ON THE DEVELOPMENT OF LICHENOPORA VERRUCARIA. 125 


is destined to form a single larva, is a small ovoid mass of cells 
consisting of a well-defined external layer and an equally 
clearly defined inner layer. One wall of the embryo becomes 
invaginated to form the larval “sucker ”’ (cf. 6, pl. xxiv, figs. 
22,23). This process is seen to be taking place in some of 
the embryos shown in fig. 11. 

It is not always easy to distinguish between cases in which 
the sucker is being invaginated and cases of embryonic fission. 
One difference may, however, be pointed out between the two 
processes. The invagination of the sucker never results in 
the division of the inner layer of cells into two separate por- 
tions, nor does this layer even appear divided in a section of 
the embryo. Fission, on the contrary, results in the complete 
separation of the inner layer into two halves, one of which 
belongs to each of the products of division. In some cases the 
ovicell contains larger and more irregular masses of embryonic 
tissue. These probably give rise to more than two secondary 
embryos. 

I have satisfied myself of the normal occurrence of the 
process of fission, as described above, in a number of colonies ; 
and I believe that the process continues, in healthy colonies, 
for a considerable time. I have not made preparations of 
colonies of all the larger sizes; and I am not able to say what 
is the upper limit of size reached during this stage. The 
largest colony among my sections, in which the embryonic 
structures are in Stage G, measures 1°60 mm. in diameter, 
and the smallest measures ‘80 mm. I have, however, found 
mature larve in somewhat abnormal colonies which measure 
only ‘56 mm. in diameter. The average diameter of fifteen 
colonies measured is ‘93 mm. 

Although the fertile brown body passes into the cavity of 
the ovicell during this stage, it is still possible to obtain satis- 
factory evidence of the identity of the fertile zocecium in well- 
orientated horizontal sections. If an ordinary zoccium loses 
its polypide without developing a new bud to take its place, 
the previous existence of the polypide is shown by the occur- 
rence of a brown body in the zocecium. There is no way by 


126 SIDNEY F. HARMER. 


which the brown body can be got rid of except through the 
agency of a new polypide. In the fertile zocecium, on the 
contrary, the brown body passes upwards to the neighbourhood 
of the orifice (beginning of Stage D), and the basal end of 
the zocecium is left completely empty. In examining the base 
of colonies in the later stages of embryonic development, one 
of the oldest zocecia is usually found to be empty; and it may 
safely be inferred, in most cases, that this was the fertile 
zocecium. z! commonly possesses a functional polypide even 
during Stage G; and either z? or z3, as the case may be, is 
usually in the same condition. The other one of these two is, 
however, usually empty. In some cases the lower half of the 
fertile zocecium is seen to be cut off from the upper by a 
transverse septum, as described in Stage E (see p. 108). It is 
probable that this is the normal arrangement, and that the 
function of the septum is to restrict the embryos to that part 
of the fertile zocecium which is on a level with the rest of the 
ovicell. 

The active production of secondary embryos during this 
stage seems to have a well-marked effect on two sets of struc- 
tures in the rest of the colony; namely, on the testes and on 
the brown bodies. 

Colonies which are in the earlier stages (A—C) have testes 
in some of their zocecia in the great majority of cases. There 
can be no doubt that the colony of L. verrucaria is ordinarily 
hermaphrodite ; and, as we have seen, a testis may or may not 
be present, with the embryo, in the fertile zocecium. Prouho 
(18) has recorded some extremely interesting observations on 
the succession of the polypides in Alcyonidium duplex; 
in which a polypide which produces spermatozoa degenerates 
into the condition of a brown body, and is succeeded, as a 
normal part of the life-history, by a polypide which produces 
an ovary. It is quite possible that phenomena of an analogous 
nature may occur in Lichenopora, although the details are 
obviously different. If continuous observations of the zocecia 
of a young living colony could be made from day to day, it 
might be possible to show that the apparent irregularity in the 


ON THE DEVELOPMENT OF LICHENOPORA VERRUCARIA. 127 


production of testes did really follow some definite law. The 
only general statements that I can make are to the effect that 
testes may be present during any of Stages A—G; but that 
while they are normally present in the younger stages they 
are much less common in the older ones. Thus of nineteen 
colonies in Stage F, in which I have noted this particular 
point, seven had well-developed testes, while twelve had none 
at all, or only slight remains. In some cases in which no 
functional testes occur in the colony, the former presence of 
these bodies can be proved by the presence of their degenerat- 
ing remains in the recently formed brown bodies. Of ten 
colonies in Stage G, in which I have noted the presence or 
absence of testes, only one had well-developed male organs; 
while in nine none were present, or at most the last remains 
of testes which had previously been active. 

This seems to indicate that the energies of the colony are 
given as much as possible to the production of embryos during 
Stage G, and the development of spermatozoa is meanwhile 
suppressed. 

It has already been pointed out that the healthy development 
of the embryos is dependent on the functional activity of the 
polypides. Were these to degenerate to any great extent 
during Stage G, it is probable that the development of the 
embryos would be interfered with. This is borne out by the 
examination of various colonies in which degeneration is 
actually taking place, and even more strikingly by the character 
of certain colonies in which the embryonic development has 
clearly been modified by conditions of insufficient nutrition. 
These are colonies in which all or most of the polypides have 
degenerated. The ovicell should be in Stage G, and it has 
actually succeeded in producing perhaps three to six mature 
embryos. But instead of finding the ovicell crammed with 
embryos of all ages intermediate between these and the youngest 
secondary embryos, the ovicell contains nothing in addition to 
the mature embryos except a very few small fragments of 
embryonic tissue. It may be concluded that these are not 
colonies in which the production of secondary embryos is 


128 SIDNEY F. HARMER. 


approaching an end, from the fact that the ovicell is small and 
the embryophore is compact and unbranched. 

Some colonies in Stage G which have no polypides, or only 
one or two polypides, appear quite healthy. It is probable 
that in these cases the production of embryos would have soon 
been injuriously affected, as a single polypide is probably 
insufficient for the nutrition of a whole ovicell full of embryos. 
This probability is increased when it is noticed, as is commonly 
the case, that many of the remaining polypides have quite 
recently degenerated ; showing that the colony has had a full 
complement of polypides up to a very recent period. 

The interdependence of the development of the embryos and 
the nutritive conditions of the whole colony is, however, no- 
where more clearly shown than in the very remarkable fact 
that a large proportion of the older colonies have no brown 
bodies at all. A brown body is of course present in a zocecium 
which has lost its polypide altogether, without developing a 
bud to replace it. But most of the zoccia in a healthy colony 
—during Stage G, for instance—contain a polypide without 
any brown body. In the earlier stages, where the mass of 
embryonic tissue present in the colony is very small, polypides 
can degenerate, and new buds can develop to take their place 
without materially affecting the health of the colony. And as 
a matter of fact we do find that most young colonies contain 
polypides, degenerating polypides, fully formed brown bodies, 
and young polypide-buds, which are developing to take the 
place of the old polypides. 

Were any extensive degeneration of the polypides to take 
place in any colony containing numerous embryos, it is 
probable that the colony could no longer bear the strain of 
producing its embryos. In a large proportion of the colonies 
in Stage G, there is no trace of the complete degeneration of 
polypides, there are no brown bodies, and there are no polypide 
buds, either in new zoccia or in old ones. Even the increase 
of the size of the colony, by the formation of new zocecia, thus 
appears to be retarded during the height of the formation of 
secondary embryos. 


ON THE DEVELOPMENT OF LICHENOPORA VERRUCARIA. 129 


I have formerly brought forward certain evidence (7) tending 
to show that the formation of brown bodies is, to a large extent, 
a kind of excretory process, and it would be surprising if 
Lichenopora were altogether exempt from the necessity of 
forming brown bodies, whatever their function may be. The ° 
explanation of the apparent anomaly of the absence of these 
structures appears to me to be that parts of the alimentary 
canal degenerate during the functional activity of the poly- 
pide. A newly formed polypide has the inner border of the 
epithelial cells of its alimentary canal clear and sharply defined. 
Parts of the alimentary canal of old polypides are in the same 
condition, but certain parts of the cecum of the stomach are 
in a very different state. The cecum is coloured during life 
with the pigment which is so characteristic of parts of the 
alimentary canal in Polyzoa. In sections these parts contain 
a very small quantity of protoplasm, and consist principally of 
yellowish granules which are very similar to the granules which 
largely compose the brown bodies—that is to say, there is no 
very great difference between the latter and those parts of the 
alimentary canal which are pigmented during life. I believe 
that this implies that excretory substaces accumulate in the 
epithelium of the alimentary canal as the polypide grows older. 
In the case of the younger colonies the entire polypide usually 
atrophies in course of time ; but in the case of older colonies, 
in which it is of the utmost importance that every zocecium 
should be contributing its share towards the nutrition of the 
ovicell, the excretory granules are discharged into the stomach 
continuously during the life of the polypide. 

The facts are that the alimentary canal of these poly pides 
usually contains solid substances which resemble fragments of 
brown bodies—that certain parts of the epithelium have no 
definite inner limit, but pass off into a cloud of granules which 
extend into the alimentary canal, and that the polypides are 
usually distinctly larger than those of young colonies. It is 
not always easy to get polypides (in sections) which are well 
orientated for the purpose of measurement ; but the measure- 
ments I have been able to obtain support the impression one 

VOL, 39, PART 1.—NEW SER. I 


180 SIDNEY F. HARMER. 


receives from the inspection of colonies of different ages, that 
polypides in old zocecia, which have no brown bodies, are dis- 
tinctly larger than those commonly met with in younger colo- 
nies in which brown bodies are present in the same zoccia. 
‘ The significance of the larger size of these older polypides is, 
as I take it, that they have gone on growing after the time 
when degeneration would have taken place in a younger colony. 
The difference in the size of the polypides, at different ages of 
the colony, becomes very apparent in comparing the size of 
the brown bodies formed from their degeneration. The 
diameter of the fertile brown body in young colonies is about 
‘05 mm.; that of a fertile brown body belonging to a later 
generation (fig. 36), and formed in colonies in which large 
polypides are present, may be as much as ‘09 or ‘10 mm. 

It is not always easy to decide, in a given case, whether the 
appearance of the alimentary canal implies the passage of a 
regular brown body into the canal or the partial degeneration 
of part of the epithelium. I have no doubt that the old brown 
bodies may be removed by passing into the alimentary canal of 
the newly formed polypide, as in Flustra and some other 
Polyzoa. Although one or even two brown bodies may occur 
in a single zoecium of Lichenopora verrucaria, I do not 
think I have ever found more than two; and in most cases 
where a polypide has recently degenerated, one does not find 
an older brown body as well, even though the age of the colony 
makes it certain that there has been an older brown body. In 
some of these cases it is probable that the old brown body 
fuses with the degenerating polypide; but in others it may be 
that the old brown body has passed into the alimentary canal 
before the degeneration of the polypide again takes place. 
Some of my sections indicate the occurrence of this process, 
and the brown body seems to become attached to the wall of 
the stomach, and finally to pass into it. Nuclear structures 
and other parts of the brown body may in these cases occur 
freely in the lumen of the alimentary canal. Even without 
actual evidence of this kind it would almost be necessary to 
assume that this takes place; for in no other way do I see any 


ON THE DEVELOPMENT OF LICHENOPORA VERRUCARIA. 131 


explanation of the total absence of brown bodies in old zocecia. 
In L. hispida the alimentary canal probably does not absorb 
the brown body, since a considerable number of old brown 
bodies occur in the proximal parts of the zoccia (cf. p. 97), 

The old polypides are not merely different in size from the 
young ones, but their several parts may be very differently 
arranged, during the retracted condition at least. The polypide 
of young colonies is like that of many other Polyzoa,—e. g. 
that of Alecyonidium as figured by Prouho (18, pl. xxiii, 
fig. 3). The short pharynx lies in the same line as the ten- 
tacles, and from its lower end a long cesophagus passes upwards 
(parallel with the tentacles) to open into the stomach (ef. fig.19). 
The intestine and rectum continue the direction of the ceso- 
phagus, while the caecum hangs down into the body-cavity 
parallel with the cesophagus. The apex of the cecum is formed 
of more protoplasmic cells, but its sides contain the yellow 
granules which have already been mentioned. In the poly- 
pides of old colonies the whole of the cecum is much swollen. 
Its apex has not altered its place, as is shown by the position 
of the group of protoplasmic cells. The upper end of the 
stomach or the region of the intestine is, however, commonly 
bent, so that the intestine and rectum may be directed down- 
wards (away from the orifice of the zocecium). The walls of 
the cecum are at the same time greatly swollen and very 
granular. The lumen may be in places quite obliterated; and 
in any case, parts of the wall have no sharp inner boundary. 
The epithelium here shades off quite gradually into the granular 
contents of the stomach. This might be ascribed to defective 
preservation, were it not for the fact that other parts of the 
same sections are well preserved. But since, in many cases, 
no polypide-buds are being formed, and since none of the 
zocecia are losing their polypides, I believe that this is an 
arrangement by which none of the zocecia are rendered in- 
efficient for nutritive purposes during the time at which there 
is the greatest strain on the energies of the colony. 

The phenomenon which has just been described appears to 
be characteristic of old colonies (Stage G and later stages), but 


132 SIDNEY F. HARMER. 


it is by no means restricted to these. Brown bodies are absent 
in a considerable number of colonies in Stage F, and rarely in 
earlier stages, sometimes occurring even in quite young 
colonies. 

Whatever the fate of other brown bodies, the fertile brown 
body always occurs in a normal colony. I have not examined 
many colonies larger than 15 mm. in diameter, and I cannot 
say how long this structure remains compact. In one colony 
I found a fertile brown body which had almost divided into 
two, and I have also noticed a brown body of the second 
generation of embryos divided into a considerable number of 
fragments which are contained in the embryophore. In all 
the colonies I have examined in Stage G the fertile brown body 
is still compact. 


The Occurrence of Second Broods of Embryos. 


My observations on this part of the development are very 
incomplete, but I have noticed one or two facts of interest. 
There is no doubt that new broods do make their appearance 
in old colonies, and this may possibly indicate a second year’s 
growth (cf. p. 89). That we are not here dealing with the 
first brood is shown by the large size of the colony, and by the 
fact that it possesses a large ovicell, which, however, contains 
no embryos. 

I have not traced the first origin of the new brood of embryos. 
Fig. 36 is from a colony 2°2 mm. in diameter. No other pri- 
mary embryo was found. Most of the polypides had enormous 
testes, and a few had egg-cells in a corresponding position 
near the apex of the cecum. The embryo is carried by an 
embryophore supported by a fertile brown body, and corre- 
sponds in general with Stage D or E of the first brood. Ihave 
not been able to obtain a stage intermediate between the 
occurrence of eggs at the lower end of a polypide and the 
section figured. It does not appear to me probable that a 
normal suspensor is formed. The cells which surround the 
embryo are exceedingly loose and scanty, and the part of the 
embryophore between the brown body and the embryo is short 


ON THE DEVELOPMENT OF LICHENOPORA VERRUCARIA. 183 


and has but few nuclei. It seems to be growing down from 
the brown body to enclose the embryo. The very large size 
of the brown body is at once apparent, and judging by the 
condition of the other polypides of the colony it may be inferred 
that it has been formed by the degeneration of a large polypide 
(cf. p. 130). The investment of the brown body is well deve- 
loped, highly protoplasmic, and full of nuclei. 

Another old colony in which I have found the same pro- 
cesses taking place measured 3°4 mm. Its ovicell was well 
developed, and the polypides and testes were large. The ordi- 
nary zocecia were not accompanied by brown bodies, and no 
eggs were found. In this case I discovered no less than eleven 
primary embryvus, each of which was contained in an embryo- 
phore supported by a large fertile brown body. ‘The evidence 
given by these cases is quite concordant. ‘The brown body is 
as characteristic a feature of the second generation as it was 
of the first, and its presence indicates that the primary embryo 
is developed in relation with a polypide. Moreover the space 
in which the embryophore of fig. 36 lies is clearly a zocecium. 
One of the eleven embryos of the larger colony has already 
given rise to a good many secondary embryos, and its brown 
body is broken up into fragments. The other ten have a com- 
pact brown body, and are in a condition which exactly corre- 
sponds with Stage F of the first brood. In all these primary 
embryos, fission has commenced and is going on vigorously. In 
one case the embryophore is beginning to grow out of the 
space in which it was produced into the general cavity of the 
ovicell. 

The later development of the primary embryos of the second 
brood is thus identical with that of the first brood, but it is 
probable that the details of the early development are some- 
what different. One of the noteworthy features of the case 
is the large number of zocecia which may become fertile in a 
single colony. 

I have unfortunately no other observations on this interesting 
part of the life-history of L. verrucaria. From the great 
clearness of the embryo in all the cases observed, it is quite 


134 SIDNEY F. HARMER. 


possible that the earlier processes of the development would be 
more easily made out in the second brood than in the first 
brood. 


General Conclusions. 


The only genus with which Lichenopora can be compared, 
as regards the embryonic development, is Crisia. In no 
other case have the early processes been described at all. 

In Crisia certain zocecia take on the character of ovicells. 
A polypide-bud makes its appearance in the young ovicell 
while the latter is at the growing-point of the branch, and is 
not externally marked out in any way as an ovicell. The bud 
becomes connected with an egg, and develops a tentacle-sheath 
which shows it to be a modified polypide. The greater part of 
the bud, however, forms what may now, on the analogy of 
Lichenopora, be termed an embryophore ; and in this struc- 
ture the embryo develops and undergoes its fission. There is 
no fertile brown body; and the ovicell is recognisable exter- 
nally as a structure which looks like a zocecium that has been 
inflated. 

While the ovicell of Crisia is thus clearly a modified 
zocecium, which develops a polypide-bud in the ordinary way 
at the growing-point, the morphology of the ovicell of Lichen- 
Opora is more obscure. There is no doubt that the embryo is 
formed in a zocecium; but it is formed in a fully developed 
zocecium which probably is normally tenanted by two succes- 
sive polypides. Each of these becomes a brown body in turn, 
the second brown body fusing with the first to form the charac- 
teristic fertile brown body. The exact origin of the embryo- 
phore is not as clear as it was in Crisia; but there is nothing 
to show that it is morphologically a bud. The later history is, 
however, not dissimilar. The embryophore becomes vacuolated, 
and in the vacuoles so formed the secondary embryos are later 
found. Certain differences in the details of the embryonic 
fission have already been pointed out. The fertile zocecium 
becomes occluded, and its cavity then becomes continuous with 
a series of alveoli or interzocecial spaces, which occur between 


ON THE DEVELOPMENT OF LICHENOPORA VERRUOARIA. 135 


the upper ends of the zowcia. The ovicell is increased in size 
by the addition of fresh alveoli. 

What, then, is the morphology of the ovicell? Are we to 
regard it as formed by the fusion of a number of zoccia, repre- 
sented by the constituent alveoli? Or, is the ovicell a highly 
modified zocecium, or two zocecia in cases where two zocecia are 
originally fertile? Or, lastly, is no direct comparison possible 
between Crisia and Lichenopora in respect of their 
ovicells? The last question may probably be dismissed with 
the reply that the fundamental process of embryonic fission, 
and the character of the embryophore, are so similar in the 
two cases that the ovicells must be regarded in general as 
homologous structures. An objection to considering the alveoli 
to be suppressed zocecia is to be found in the fact that the 
zocecia reach the basal lamina of the colony, and the alveoli 
do not. This has previously been correctly pointed out by 
Smitt (20). The alveoli do not, moreover, develop a polypide 
bud at any period. 

We must not lose sight of the fact that Crisia and Licheno- 
pora are two very widely separated genera. It would not, 
indeed, be easy to choose any other pair of recent Cyclostomes 
which would be less nearly related toone another. More light 
will probably be thrown on the morphology of the ovicell of 
Lichenopora by the examination of genera which are less 
different from Crisia. 

I have obviously been fortunate in being able to study the 
ovicells of L. verrucaria. Waters has more than once 
(24, 26) alluded to the importance of examining the ovicells 
of this genus, and has stated that in some species “large 
numbers of specimens may be examined without any ovicell 
being found.” Whether this is due to the absence of the 
ovicells in these cases, or merely to the difficulty of recognising 
them, I have no means of deciding. But it must be expressly 
noted that Waters states (24, p. 261), in describing L. nove- 
zelandiz, Busk, that ‘‘ where there is no ovicell the zocecial 
tubes run into the centre, the central depression forming an 
inverted cone without cancelli;”’ and a similar statement is 


136 SIDNEY F. HARMER. 


made on the next page for L. grignonensis, Busk. Waters 
unfortunately does not figure any of these specimens with a 
depressed centre and without cancelli; but these features 
certainly point to the conclusion that no ovicell is present. 
Waters’ statements are in strong contrast with my own results 
on L, verrucaria, in which the ovicell is potentially present, 
as shown by the alterations in the fertile zoccium in most 
colonies which consist of no more than four or five, or even 
three zocecia ; and it is actually present and recognisable from 
the outside in all colonies which have reached a slightly later 
Stage. 

It would no doubt be going too far to assume that what is 
true for L. verrucaria necessarily holds for other species of 
Lichenopora. I am not at present able to make an inde- 
pendent examination of other species ; and although the ovicells 
of some of them have been described by Waters (24, 26) and 
others, the descriptions refer almost entirely to the external 
form, and I can find practically no points of comparison which 
assist in forming a conclusion as to the morphology of the 
ovicell. Waters in one place (25, p. 277) alludes to a feature 
which may have some relation with some of the phenomena 
described above in L. verrucaria. Speaking of Hornera 
(“Idmonea”’) fissurata, Busk, Waters states that he could 
not discover any opening to the ovicell, but that one of the 
lateral zocecia was much larger than the others, and it appeared 
that this change was “connected with the functions of the 
ovicell.” In each of the ovicells subsequently examined by 
Mr. Kirkpatrick one or two zowcia in proximity to the ovicell 
were enlarged and altered in direction. It would be interest- 
ing to know whether the enlarged zocecium in this case had 
any resemblance to the fertile zocecium in L. verrucaria. 

The ovicell seems to be practically identical in all species of 
Lichenopora in which it has been described. It always 
forms an inflated area between the upstanding zocecia, and it 
is usually known to occupy the centre of the colony. In some 
species it is extremely difficult to recognise the ovicell at all. 
I cannot help suspecting that the ovicell normally develops at 


hy 


ON THE DEVELOPMENT OF LICHENOPORA VERRUCARIA. 137 


an early stage in most, if not all, of the species of Licheno- 
pora, and that the presence of cancelli between the rows of 
zocecia is an indication that the ovicell is developing, even 
though no other evidence is obtained of its presence. 

There are, however, other genera of Cyclostomata in which 
the ovicell is of a simpler character than in Lichenopora. 
Thus in Discotubigera lineata, MacG., as figured by 
Waters (26, pl. xv, fig. 5), the ovicell is a lobed structure 
which extends over a comparatively small area of the discoidal 
colony near its margin. It opens by a single aperture to the 
exterior. A similar form of ovicell occurs in the genus Diasto- 
pora. It appears to me that these ovicells resemble those of 
Liche nopora reduced to simpler terms, and they are, perhaps, 
to some extent intermediate between the two genera which I 
have specially studied. Taking into account the origin of the 
ovicell in Lichenopora from an ordinary zocecium, the fact 
that the ovicell of Crisia is demonstrably a modified zocecium, 
and the existence of intermediates between these two extremes, 
I venture to suggest that that of Lichenopora is also to be 
regarded as formed by the modification of a zocecium. The 
fertile zocecium is prevented by the position which it occupies 
near the centre of the full-grown colony from expanding at its 
basal end. Expansion of some kind is necessary, in order to 
provide room for the swarm of secondary embryos ; andif there 
is anything in this suggestion, it may be supposed that the 
significance of the process is that the fertile zocecium has 
become swollen at its upper end. Instead of growing out as 
a free appendage of the colony, the swollen part has applied 
itself to the upper surfaces of the zocecia, and has filled up all 
the spaces intervening between the zocecia. The result is the 
same as if the ovicell had become amceboid, its pseudopodia 
extending between the zocecia, embracing them, and anasto- 
mosing on the further side of each. In the cases where two 
or more zocecia become fertile, the ovicell may be regarded as 
being composed of as many original zoovcia. 

The actual formation of the ovicell, by the formation of the 
alveoli and their subsequent fusion with one another, is of 


= 


188 SIDNEY F. HARMER. 


course somewhat different from this hypothetical process; but 
I claim to have established the facts (1) that the processes 
which precede the formation of the ovicell take place in a 
fertile zocecium, and (2) that the development of the cal- 
careous roof of the ovicell starts from the region of the orifice 
of the fertile zocecium. 

It remains to be seen whether the ovicells of such genera 
as Hornera, Diastopora, Tubulipora, and Idmonea fit 
in with this idea ; and I hope to have an opportunity of making 
a more extended study of this question. 

Next to the demonstration of the occurrence of embryonic 
fission, a process which had been previously proved for Crisia, 
the most striking fact which I have succeeded in establishing 
is the restriction of the production of an embryo to one or two 
of the oldest zocecia in the normal development. 

Even the second brood of embryos is really dependent on 
this function of the fertile zocecium, or zoccia, of the first 
generation ; for they are nourished in the ovicell which was 
developed in connection with the primary fertile zoccia. In 
the great majority of colonial animals a large number or all 
of the individuals of the colony become fertile. The phe- 
nomena which occur in Lichenopora are perhaps due to the 
fact that the colony in this genus is to be regarded as an 
individual of a higher order. The discoidal form and the close 
association of the zocecia have produced the result that the 
colony behaves in this respect as if it were a single individual. 
The production of fertile eggs is thus limited to one or two 
individuals, and in fact to one or two of the zoccia which are 
first formed as buds from the primary individual of the colony. 
In the case of Crisia, where the association of the different 
zocecia is much less intimate, a considerable number of the 
individuals of a colony may become fertile. Here, however, 
there is something analogous in the behaviour of the inter- 
nodes, each of which in the great majority of species normally 
produces no more than a single ovicell. 


ON THE DEVELOPMENT OF LICHENOPORA VERRUCARIA. 139 


REFERENCES. 


. Attman, G. J.—‘‘A Monograph of the Fresh-water Polyzoa,” 4to, Ray 


Society, 1856. 


. Barrots, J.—‘‘ Recherches sur |’Embryologie des Bryozoaires,” 4to, 


Lille, 1877. 


. Barros, J—‘ Mém. sur la Métamorphose de quelques Bryozoaires,” 


‘Ann. Sci. Nat. Zool.,’ ser. 7, vol. i, 1886, No. 1. 


. Bram, F.—* Unt. iib. d. Bryozven des stissen Wassers,” Leuckart and 


Chun’s ‘ Bibl. Zool.,’ Heft 6, 1890. 


. Harmer, 8. F.—“On the British Species of Crisia,” ‘Quart. Journ. 


Micr. Sci.,’ vol. xxxii, 1891, p. 127. 


. Harmer, 8S. F.—“On the Occurrence of Embryonic Fission in Cyclo- 


stomatous Polyzoa,” ibid., vol. xxxiv, 1893, p. 199. 


. Harmer, 8. F.—‘‘ On the Nature of the Excretory Processes in Marine 


Polyzoa,” ibid., vol. xxxiii, 1892, p. 123. 


. Harmer, 8. F.—“ Preliminary Note on Embryonic Fission in Licheno- 


pora,” ‘Proc. Roy. Soc.,’ vol. lvii, 1895, p. 188. ‘On the Develop- 
ment of Lichenopora verrucaria, Fabr.,” ibid., vol. lix, p. 73. 


. Hincxs, T.—‘ British Marine Polyzoa,’ 8vo, London, 1880. 
. Krazpetin, K.-—“ Die Deutschen Siisswasser-Bryozoen,” part J, ‘ Abh. 


Ver. Hamburg,’ vol. x, 1887. 


. Krarrewin, K.—Ibid., part 2. ‘ Abh. Ver. Hamburg,’ vol, xii, 1893. 
. Levinsen, G. M. R.—“ Mosdyr,” from ‘ Zoologia Danica,” 4to, Copen- 


hagen, 1894. 


. Merscunixorr, E.— Beit. zur Entwick. ein. niederen Thiere,” § Bull. 


Ac. St. Pétersbourg,’ vol. xv, 1871, p. 502. 


. Ostroumorr, A.—‘ Remarques relatives aux recherches de Mr. Vigelius 


sur des Bryozoaires,” ‘ Zool, Anz.,’ vill, 1885, p. 290. 


. Ostroumorr, A.— Contribution a1l’étude Zool. et Morphol. des Bryozo- 


aires du Golfe de Sébastopol,’ ‘Arch. Slaves de Biol.,’ vol. ii, 1886, 
p. 8. 


. Ostroumorr, A.— Zur Entwicklungsgeschichte d. Cyclostomen See- 


bryozoen,” ‘ Mitt. Zool. Stat. Neapel,’ vol. vii, 1887, p. 177. 


. Percens, E.— Unt. an Seebryozoen,” ‘Zool. Anz.,’ xii, 1889, p. 504. 
. Prouno, H.—* Cont. 4 )’Histoire des Bryozoaires,” ‘Arch. Zool. Exp. 


et Gén.,” ser. 2, vol. x, 1892, p. 557. 


. Riptey, 8. O.— Polyzoa, Coelenterata, and Sponges of Franz-Joseph 


Land,” ‘ Ann. Mag. Nat. Hist.,’ ser. 5, vol. vii, 1881, p. 442. 


140 SIDNEY F. HARMER. 


20. Surrt, F. A.—“ Krit. forteckning ofver Skandinaviens Hafs-Bryozoer,” 
‘Ofv. K. Vet.-Ak. Forhandl.,’ 1866, p. 395. 

21. Verworn, M.— Beit. z. Kenntnis d. Siisswasserbryozoen,” ‘Zeit. f. 
wiss. Zool.,’ xlvi, 1888, p. 99. 

22. VicELius, W. J.—*‘ Cont. ala Morphologie des Bryozoaires Ectoproctes,” 
‘Tijdschr. d. Ned. Dierk. Ver.’ (2), i, Afl. 3 and 4, 1886. 

23. VicrLtius, W. J.—‘ Zur Morphologie der marinen Bryozoen,” ‘ Zool. 
Anz.,’ x, 1887, p. 237. 

24, Waters, A. W.— Bryozoa from New South Wales, North Australia, 
&c. Part 3,” ‘Ann. Mag. Nat. Hist.,’ ser. 5, vol. xx, 1887, p. 253. 

25. Waters, A. W.—‘*On some Ovicells of Cyclostomatous Bryozoa,” 
‘Journ. Linn. Soc. Zool.,’ vol. xx, 1890, p. 275. 


26. Waters, A. W.—“On the Ovicells of some Lichenopore,” ibid., p. 280. 


DESCRIPTION OF PLATES 7—10, 


Illustrating Mr. Sidney F. Harmer’s paper “ On the Develop- 
ment of Lichenopora verrucaria, Fabr.” 


All the figures refer to Lichenopora verrucaria, Fabr. Slight varia- 
tions in the position of the camera lucida are responsible for the fact that they 
are not absolutely drawn to scale. All the figures on Plate 7, and figs. 12, 
13, on Plate 8, were drawn with Zeiss A. Figs. 11, 14, and 17 were drawn 
with Zeiss C, and the others with Zeiss F. 


PLATE 7. 


Fic. 1.—Two views of a very young left-handed colony (diameter ‘18 mm.). 
1, 2, and 8 are the first three zowcia. 

Fic. 2.—Somewhat older left-handed colony (diameter ‘30 mm.) seen from 
above. 

Fic. 3.—Left-handed colony with four zoccia (1, 2, 3, 4). (Diameter 
‘48 mm.; extreme length, ‘64 mm.) 

Fic. 4.—Right-handed colony with seven complete zoecia. The proximal 
lip of the colony is beginning to grow towards the disc (broken), and the 
commencement of two new zoccia can be seen in the space between the first 
three zoccia (1, 2, and 3). (Diameter °59 mm.) 

Fic. 5.—Colony in which the roof of the ovicell is developing. The colony 
is probably left-handed, with occluded z*. The aperture of the ovicell is not 


ON THE DEVELOPMENT OF LICHENOPORA VERRUCARIA. 141 


yet completely formed. The proximal edge of the basal lamina has nearly 
reached the disc. (Diameter, 1:18 mm.) 


Fic. 6.—Thick radial section of part of an old colony, showing the relations 
of the ovicell to the zocecia. (Diameter of the right-hand zoccium, outside 
the ovicell, = ‘192 mm.) 


Fie. 7.—A fully developed discoidal colony (diameter 1°92 mm.). The ovi- 
cell is obviously composed of a number of alveoli. The colony is probably 
left-handed, with occluded z?, some indication of which can probably be seen 
just beneath the single aperture of the ovicell. The zocecium to the left of 
the latter and above it is probably z?. 


Fic. 8.—Two zoecia from a very old colony, with secondary thickening. 
The porous roof of the ovicell is seen at the bottom of a set of honeycomb- 
like spaces, the so-called “ cancelli;”? some of these are covering the bases of 
the zocecia, and in the left-hand zocecium these are entirely or partially roofed. 
Two of the blister-like swellings thus formed have part of their roof still 
uncalcified. (Diameter of zoccia ‘14 mm.) 


Fig. 9.—From another old colony. The alveoli are still distinctly visible, 
and the cancelli are commencing in the interalveolar grooves, or at the base 
of the zocecia. The left-hand zocecium is closed by a porous calcareous cap. 
(Diameter of zocecia ‘13 mm.) 


PLATE 8. 


(N.B.—The measurements given below [except for figs. 12, 13] were made 
from the sections.) 


Fic. 10.—Part of a longitudinal section through the fertile zocecium (? z3) 
of a colony, ‘67 mm. in length, with about nine fully-formed zoccia. The 
elongated structure to the right of the fertile brown body is the embryophore. 


Fic. 11.—Section (diameter ‘80 mm.) of a colony with fully formed ovicell. 
The section is parallel to the flat surface of the plano-convex colony, and cuts 
the convex side near its upper surface (cf. fig. 6). The only zocecia seen in 
the section are thus the more centrally placed ones. z? was presumably the 
fertile zocecium, as it is empty basally, and contains secondary embryos, at its 
upper end. The fertile brown body has passed into the ovicell. z? also con- 
tains secondary embryos, but has a brown body basally, and was probably 
occluded without being fertile. The secondary embryos are contained in the 
“ embryophore,”’ whose lobes extend, in the cavity of the ovicell, between the 
zocecia. The zocecium A contained a brown body lower down. 


Fig. 12.—Entire colony (diameter 1°41 mm.), stained and mounted in 
Canada balsam. The shaded part shows how much of the ovicell is com- 
pletely roofed. The alveoli which are still unroofed are not specially indicated. 
The ovicell has an aperture, at the base of which is the fertile brown body. 


142 SIDNEY F. HARMER. 


The embryophore is younger than in fig. 11. Parts of the zocwcia which are 
seen either through other zocecia or through the roof of the ovicell are indi- 
cated by dotted lines. The colony is left-handed. z* is occluded (and was 
obviously fertile), and would hardly have been visible at all in a dry 
preparation. 

Fic. 13.—Occlusion of the fertile zocecium ina left-handed colony (diameter 
‘88 mm.). The roofing of the ovicell has only just commenced, as a calcareous 
film starting from the outer border of z?, and sweeping inwards in such a way 
as to leave z? open to the cavity of the ovicell. Although the figure is from a 
dry mount, the remains of the fertile brown body (not shown in the figure) 
are visible through the opening into z?. 

Fic. 14.—Part of the growing edge of an adult colony, more highly magni- 
fied, showing the way in which new zoccia are formed, and representing 
various stages in the development of the alveoli. See also pp. 81, 84. 


(Figs. 15—35, and Fig. 37, refer to the development of the first brood 
of secondary embryos. 


PLATE 9. 


Fie. 15.—A polypide-bud with an egg. One other bud in the same colony 
had an egg in a similar position. 

Fic. 16.—From a colony, ‘48 mm. long, consisting of three zoecia. One 
or two egg-cells were discovered in z}, as well as in the zocecium (z*) drawn. 

Fie. 17.—Horizontal section of a colony, ‘43 mm. long and ‘25 mm. broad, 
consisting of three complete zoccia only. 

Fic. 18.—Part of the left-hand zocecium (? z°) of the last figure, more 
highly magnified. The egg measures 12°8 » (= ‘0128 mm.). 

Fie. 19.—From a colony consisting of three mature zocecia. Development 
of the egg has commenced, and the follicle is becoming definite. The heads 
of mature spermatozoa are seen in the body-cavity. 


Fie. 20.—From a colony consisting of three mature zocecia. The follicle 
is more developed, and the suspensor has become a definite structure. 
(Diameter of the embryo, 11:2 p.) 


Fic. 21.—From a colony, ‘40 mm. long and ‘21 mm. broad, consisting of 
three mature zocecia. ‘The follicle is cut transversely, and the embryo is in 
about the same stage as that of Fig. 19. 

Fic. 22.—From a colony consisting of four mature and one half-grown 
zocecium. ‘The suspensor is not yet clearly formed. 

Fic. 23.—From a colony, ‘53 mm. long and ‘32 mm. broad, consisting of 
five mature zocecia. The figure is combined from three sections of the 
ocecium, the fertile brown body being projected on to the rest of the drawing. 


ON THE DEVELOPMENT OF LICHENOPORA VERRUOARIA. 143 


The suspensor and the cells which surround it are closely connected with a 
mass of cells, a, which form part of the investment of the brown body. 


Fic. 24.—Embryophore from a colony ‘48 mm. long and *45 mm. broad. 
(Length of suspensor, 41°6 » ; transverse diameter of embryo, 20°8 yu.) 


Fic. 25.—Showing the abnormal development of two embryos within the 
same embryophore. From a colony ‘56 mm. long, consisting of five mature 
zocecia. 


Fic, 26.—Embryo (Stage D) from a colony °56 mm. long and ‘37 mm. 
broad, consisting of five zocecia. The outer layer of nuclei is unusually 
distinct. 


Fic. 27.—From a colony ‘69 mm. long and *51 mm. broad, consisting of 
about twelve mature zoccia. The fertile polypide has degenerated, and the 
brown body has approached the orifice of the zoccium. The embryo is 
28 » long, and the length of the entire embryophore with the brown body is 
93 p. 


PLATE 10. 


Fic. 28.—From a younger colony, ‘43 mm. long and ‘37 mm. broad, con- 
sisting of six mature zocecia. The brown body has not yet reached the orifice 
of the zocecium, and development is consequently not quite so advanced as in 
Fig. 27. 


Fic. 29.—From a considerably older colony, *56 mm. broad. The upper 
end of the embryophore had met the orifice, as in Fig. 27, and evidence was 
obtained that the zocecium was occluded by a membrane stretching across the 
orifice. Length of the embryo 70 yu. 


Fie. 30.—(Stage H.) Obliquely longitudinal section through the embryo- 
phore of a colony ‘62 mm. broad. The upper end of the figure was turned 
towards the brown body. Diameter of the embryo 30°4 nu. 


Fic. 31.—Embryo from a colony *83 mm. broad. The occurrence of the 
split-like spaces s, and the arrangement of the inner mass of nuclei indicate 
that embryonic fission is about to commence. Longest diameter of embryo 
45°6 p. 

Fies. 32 and 33 are consecutive transverse sections of the larger portion 
of an embryo at the beginning of its fission. From a colony ‘93 mm. broad, 
containing two fertile zowcia. The longest diameter of the embryo is ‘68 p. 
A smaller piece of the embryonic tissue, not shown in the figures, is already 
completely separated from the portion drawn. 


Fic. 34,—Transverse section of the older embryophore from a colony 
(86 mm. broad) containing two fertile zocwcia. Embryonic fission is actively 
proceeding. The diameter of the giant-cell a is 17°6 p, and that of its 
nucleus is 9°6 p. 


144 SIDNEY F. HARMER. 


Fic. 35.—Longitudinal section of the embryophore, with actively dividing 
primary embryo. The brown body is not seen in this section, but would have 
come at the upper end of the figure. From a colony ‘53 mm. long. 


Fic. 36.—Longitudinal section of a young embryophore belonging to a 
second or later generation. The brown body is conspicuously larger than that 
of the corresponding stage (Fig. 28) of the first generation. The embryophore 
lies in a space which is clearly a zoecium. (Diameter of the colony 2°20 mm. 
Length of the embryo plus the embryophore 167°2 ». Longest diameter of 
the brown body 83°6 w; of the embryo, 86 zp.) 


Fie. 37.—Showing the fission of a young secondary embryo, belong- 
ing to the first brood. From a colony ‘93 mm. long and ‘88 mm. broad, 
similar to that-shown in Fig. 11. The longest measurement of the dividing 
embryo is 56 p. 


SEP <5 1896 


ON NAUTILUS AND SOME OTHER ORGANISMS. 145 


Letters from New Guinea on Nautilus and some 
other Organisms. 


By 


Arthur Willey, D.Sc. 


Ratu (New Brirary), German New GUINEA, 
via SINGAPORE ; 
April 5th, 1895. 

Dear Proresson Lankester,—Although I have not seen 
any mature ova of the pearly Nautilus since my last letter to 
you, yet I have made some further observations on the adult, 
which may be of interest. 

1. Ectoparasites.—In my previous note! I accidentally 
omitted to mention the occurrence of numerous Copepod 
parasites in the mantle chamber of the Nautilus. They are 
present in nearly every individual that I have examined, and 
are found attached to the branchix, the internal surface of the 
funnel, and in other regions of the pallial chamber. 

The parasites are a species of the genus Caligus,’? and 
possess the characteristic semilunar suckers on the first pair 
of antenne. ; 

When Nautili are placed in jars the Caligids emerge in 
large numbers from the mantle chamber, and swim about 
actively in the water, usually in close proximity to the sides 
of the vessels, from which they can be removed only with 
some difficulty, owing to the great adhering power of the above- 
mentioned suckers. 

2. Movements.—I send you a photograph of a Nautilus 
in its ordinary swimming attitude,—such a figure, I believe, 

1 “Natural Science,’ May, 1894. 
2 C. nautili, pending a more detailed description. 

VoL. 39, PART 2.—NEW SER. K. 


146 ARTHUR WILLEY. 


not having been previously published.1. From the photograph 
it will be seen that in swimming the animal elevates itself to 
such an extent that the eyes are raised above the free margin 
of the mouth of the shell. As to its movements on the sur- 
face of the water, there is very little to add to the excellent 
account given by the late Professor Moseley in his ‘ Notes by 
a Naturalist on H.M.S. Challenger.’ 

I can emphatically confirm Moseley when he says that “ it 
is probably a mistake to suppose that it ever comes to the 
surface voluntarily to swim about.’ Nevertheless a Nautilus 
is not necessarily ruined by being brought up from the bottom 
in a fish basket. If liberated within a reasonable time it is 
capable of returning to its natural habitat. I have proved this 
experimentally. 

When freshly captured Nautili are placed in aquaria they 
rise to the surface and sink to the bottom with the greatest 
facility. The rising to the surface, according to my observa- 
tions, is effected solely by the muscular activity of the animal, 
and is in no way dependent on any physical modification of 
the gas in the air-chambers. ‘The presence of the latter 
renders the shell extremely light and buoyant, so that it is, 
under normal circumstances, completely under the control of 
the muscles of the animal. I say ‘under normal circum- 
stances” advisedly, because there is one thing which the 
Nautilus cannot do, namely, turn upside down. It is neces- 
sary to insist on this, because Moseley gives a translation of 
the account given by Rumphius at the beginning of last 
century, which, I regret to say, so far as my experience has 
gone, is very wide of the mark. Always remembering that 
Rumphius was the first to demonstrate the cephalopodous 
character of the pearly Nautilus, the suspicion is forced upon 
me that he derived his account of its habits from the narrative 
of an ill-informed and imaginative Malay fisherman. Natives 
are so often remarkably well informed about the habits of 
animals, that the above supposition can by no means be taken 
as conveying a reproach to the old master. 


1 This photograph is reproduced as Fig. 24 at the end of this article. 


ON NAUTILUS AND SOME OTHER ORGANISMS. 147 


“On the bottom,” says Rumphius, according to Moseley’s 
translation, “the animal creeps with the other side! upper- 
most, with the head and tentacles on the bottom, and makes 
tolerably fast progress.” The only comment I can make on 
this statement is that it is inconceivable. I wish'I had the 
work of Rumphius to refer to, in order to find out how he 
succeeded in seeing the Nautilus at all on the bottom. He 
goes on to say, “The animals remain mostly at the bottom, 
creeping sometimes into hoop-nets set for fish and lobster-pots ; 
but after a storm, when the weather becomes calm, they are 
to be seen floating in troops on the surface of the water... . 
The floating, however, does not last long, for, drawing in all 
their tentacles, the animals turn their boats over and go down 
again to the bottom.” 

The Nautilus can no more turn its boat over than a suc- 
cessful balloon ascent can end by the cage coming down 
uppermost. Anyone can convince himself of this by placing 
an empty Nautilus shell in water. A remarkably small weight 
is sufficient to sink such an empty shell; and when the living 
animal retracts itself and ceases all muscular action, thereby 
converting itself, as it were, into a dead weight, it is heavy 
enough to sink several shells in addition to its own. 

There is a slight error in Moseley’s account of the move- 
ments of the Nautilus, which may as well be corrected. He 
says, ‘On either side of the base of the membranous opercu- 
lum-like headfold . . . the fold of the mantle closing the 
gill-cavity was to be seen rising and falling, with a regular 
pulsating motion, as the animal in breathing took in the 
water, to be expelled by the siphon.” It is not a fold of the 
mantle which is thus seen to pulsate, but the posterior free 
membrane-like expansion of the funnel on either side. 

Besides observing the movements of the Nautilus inthe narrow 
limits of jars and buckets, I have also placed them in the sea in 
shallow water, and will briefly describe one such experiment. 

On March 16th a Kanaka brought me six Nautili. All of 
them sank to the bottom of the buckets except one, which 


1 That is, the ventral side. 


148 ARTHUR WILLEY. 


swam about on the top all the time. About an hour after- 
wards I took the latter out in a canoe to observe its move- 
ments in one to three fathoms of water. On placing it in the 
sea it remained at first on the surface, swimming actively 
backwards, i.e. with the shell directed forwards. It fre- 
quently swam to the bottom and back again to the surface— 
often also swimming in the middle stratum of water. I have 
never found any necessity for framing an elaborate theory as 
to the rising and sinking of the shell. 

3. The Position of the Spadix or Hectocotylus.— 
I was surprised to find that the spadix in the male developed 
variably on the right and left sides of the cephalic system. 
Out of thirty-seven males which were examined ad hoe, 
twenty-three had the spadix on the left side, and fourteen had 
it on the right side. 

4. Pallial Arteries.—I will say a few words about the 
pallial arteries, in order to refer to the arterial blood-supply of 
the siphuncular pedicle. Anatomical evidence seems to point 
to the conclusion that the latter structure has more of a 
vestigial than of an actively functional physiological import- 
ance. mbryology will show whether or not it has any original 
relation to the primitive shell-gland. 

The so-called lesser aorta of Owen, arising from the anterior 
slightly incurved margin of the heart, divides immediately, as is 
known, into two branches. These are respectively the anterior 
and posterior pallial arteries. 

The anterior pallial artery bends inwards and somewhat 
downwards to the middle line, and then runs forwards below 
the skin on the surface of the renal sacs. At the point where 
it turns forwards a small intestinal branch is given off. Arrived 
at the anterior limit of the renal sacs, the anterior pallial 
artery passes into the substance of the mantle, and runs 
towards the free margin of the mantle, shortly before reaching 
which it bifurcates into two main marginal arteries, from which 
numerous radial arteries are given off (see Fig. 1). 

Exactly at the point where the anterior pallial artery passes 
into the substance of the mantle a pair of branches, right and 


ON NAUTILUS AND SOME OTHER ORGANISMS. 149 


left, arise from it almost at right angles to it. 


These may be 
called the branchio-osphradial branches of the anterior pallial 


A ee 7706 
Nee %Q 
__M.a 
(ge SS, cay A erage es, WIS ees apa. 
i WE ng 
= \ | 
/ __.boa 
\ i — he 
4 SS ae 
KO. ‘| KO. 
Zen. Ry ee es ; i + Bees yey 
Se 
h bee 


Fic. 1—N. pompilius, 


9. View of ventral surface of pallial and 
visceral regions, to show the course of the pallial arteries. A Funnel. 
Free edge of mantle. 


7. a. Radial paljial arteries. 
artery. a.p.a. Anterior pallial artery. 


a. g. Region of nidamental gland. 
b. 0. a. Branchio-osphradial artery. s. m. Shell-muscle. 7. 0. Region of 
renal organs. /. Heart. p.p.a. Posterior pallial artery. g. Gonad. 
Intestinal branch of anterior pallial artery. 


Mm. e. 
m. a. Marginal pallial 


1. Qe 


150 ARTHUR WILLEY. 


artery, since among their minor ramifications they send up 
branches to the tips of the branchiz supplying the integument 
of the latter, and also a small branch into each of the osphradia 
(i.e. into the osphradia of Lankester and Bourne, and into 
those which I described in my last letter). In the female they 
also supply the nidamental gland. 

The posterior pallial artery runs backwards over the ventral 
surface of the heart, leaves the pericardium through the orifice 
described by Owen, and then passes onwards below the skin to 
the left of the genital gland, between the latter and the gizzard. 
Upon reaching approximately the middle point of the posterior 
rounded surface of the body it, too, passes into the integument, 
and immediately divides into two main branches, right and 
left, which supply the dorsal and posterior regions of the 
mantle, including the siphuncular pedicle. A variable number 
of small secondary or tertiary branches go up into the siphun- 
cular pedicle, but there is one branch which is essentially the 
siphuncular artery (see Figs. 2 and 3). But even the definite 
siphuncular artery is not constant in its origin, but arises now 
from the right and now from the left of the two main branches 
of the posterior pallial artery. The siphuncular artery is 
therefore a minor ramification of the posterior pallial artery. 

Owen, followed by Vrolik, described the latter as passing 
in toto into the siphuncle. Keferstein, whose figure is more 
accurate in this respect, says that it passes ‘‘ nach hinten zur 
Korperhaut und besonders zum Sipho.” 

Injection indicates that the arteries which supply the 
siphuncle are homodynamous with the other ramifications of 
the posterior pallial artery. 

5. Pallial Veins.—It is not now necessary to go into de- 
tails with regard to the pallial veins, but I will point out how 
they may be seen to great advantage. 

When a Nautilus becomes moribund it usually rises to the 
surface, owing to an abundant production of gas in the interior 
of the body. If it is allowed to die, and is then removed from 
the shell, the veins are found to be injected with gas of some 
sort, and the finest ramifications of the veins, in the mantle 


ON NAUTILUS AND SOME OTHER ORGANISMS. 151 


at least, are displayed with a clearness which could be hardly 
attained by artificial injection. 


Fic. 2.—N. pompilius, ¢. View of dorsal surface of pallio-visceral 
region, to show the principal ramifications of the posterior pallial artery. 
m.e. Free mantle edge. ¢. Region of crop.  s.p. Siphuncular pedicle. 
1. Region of liver. s.a. Siphuncular artery arising from the right of the 
two main branches of posterior pallial artery. p.p.a. Posterior pallial 
artery. giz. Region of gizzard. g. Region of gonad. 7@ Region of 
intestine. 

N.B.—In front of siphuncular artery are seen two small branches which 
bend backwards and enter the siphuncular pedicle on its dorsal aspect. 
The pedicle is here represented turned forwards to expose its ventral 


surface. 

The mantle is simply riddled by these veins in a manner 
which defies one’s powers of draughtsmanship. The veins are 
collected into two main trunks, which le on either side of the 
anterior pallial artery, and proceed backwards to open into the 
afferent branchial vessels. At the sides of the mantle there 
are also a number of lateral pallial veins, which open into a 
large sinus situated over the shell muscle. 


152 ARTHUR WILLEY. 


6. Miscellaneous.—In New Hanover and New Ireland 
the Nautilus shell is called “ Togol,”’ and is used as a decoration 
on the outriggers of the canoes and on the houses, and also as 
a drinking vessel and for baling out canoes. Occasionally the 


Fic. 8.—N. pompilius, ¢. Similar view as in Fig. B, to show the 
siphuncular artery arising from the left of the two main branches of the 
posterior pallial artery. Letters as in Fig. B. 


shell of Nautilus umbilicatus is obtained, and the new- 
comer then acquires the useful information that that is the 
male of the Nautilus! This seems to be the idea of the 
Kanakas as well as of the outlying white men. 

In N ew Hanover, where primitive institutions still largely 
flourish undisturbed (including cannibalism), a curious custom 
prevails, of which I became aware while engaged in collecting 
finger-prints. The natives employ the cheiromantic creases 
for the purpose of creating factitious social groups or families, 
within the limits of which marriage is tabooed. They trace a 
fancied resemblance between the creases of the hand and the 


ON NAUTILUS AND SOME OTHER ORGANISMS. 153 


form of the wings of a given bird, and name accordingly. All 
those whose hand-creases are referable to the same bird con- 
stitute one family or “ta-uk,’’ apart from any question of 
‘blood relationship. Thus a man of the “kanai ta-uk,” or 
family of the sea-gull, cannot marry a woman of that “ ta-uk.” 

It is also tabu to eat bird-flesh, but they all eat pork—exactly 
the reverse of what occurs in New Britain, at least as far as 
the men are concerned. 

Probably most of the fighting i1 New Hanover is due to the 
complicated relations of the sexes. “ Plenty fight belong 
Mary.” 

Finally, I may conclude these brief notes by adding that if 
a Kanaka kills another Kanaka of lis own ta-uk he does not 
eat him, according to my information.—Yours very truly, 

Anrruur WILLEY. 


Ratum, New Britatn, 
via Singapore ; 
June 4th, 1895. 

Dear Proressor Lankester,—The following are a few zoo- 
logical observations, relating chiefly to the fauna of Blanche 
Bay, which I have incidentally made during the last few 
months. 

1. Polyclades.—The Polyclade fauna is rich both in species 
and individuals. Most of those that I have hitherto had any 
opportunity of observing have been taken close to the shore 
of the island of Rakaiya (or Raluan), in Blanche Bay, from 
the lower surfaces or in the crevices of volcanic stones. 

I have, however, on several occasions also obtained them 
from the reef opposite Ralum. Although a fact of no absolute 
value proceeding from so limited an experience, it is neverthe- 
less worth mentioning that the Polyclades fromthe reef off Ralum 
have all belonged to the subdivision Cotylea, and included a 
very fine Thysanozoon, attaining a length of 66 mm.; whereas 


154 ARTHUR WILLEY. 


those from the stones off Rakaiya, which I have collected in 
much larger numbers, have all been representatives of the 
subdivision Acotylea. 

Several of the latter have laid their egg-plates in my jars. 
There are, among others, two well-marked species of Stylochus, 
both characterised by their uniformly pitted or granular ap- 
pearance, but in one case of an opaque dark ashy colour, and 
in the other of a light brown sandy complexion. As I cannot 
match these with any descriptions in the literature at my com- 
mand, it will do no great harm to refer to the former as S. 
cinereus and to the latter as S. arenosus. 

Stylochus cinereus has a length of 24 to 27 mm., and a 
width of 10 mm. ; tentacles brownish yellow, covered with eyes, 
and 3°5 mm. from anterior margin of body; cerebral eyes ex- 
tending over proximal portion of outgoing nerves; anterior 
marginal eyes extending backwards to level of tentacles; 
margin of body cloudy, but light and unpigmented. 

S. arenosus has a length of 41 to 45 mm., and a width of 
16 mm.; tentacles covered with eyes except the tip, which is 
orange-coloured; tentacles 555 mm. from anterior margin ; 
cerebral and marginal eyesas in 8. cinereus; margin of body 
nearly colourless and translucent; about twelve pairs of in- 
testinal diverticula. 

It is difficult to obtain uninjured examples of S. cinereus, 
as it tears with the greatest readiness while being detached 
from the stones on which it lives. Lacerated examples pre- 
senting a ragged appearance thrive well in aquaria. The 
tendency to laceration seems to be a characteristic of the 
species as much as any anatomical feature. ‘There is a species 
of Cryptocelis (marginal eyes all round body) common here, 
which practises deliberate autotomy, this being one of its 
most characteristic peculiarities. 

S. cinereus laid several irregular patches of eggs on April 
27th. ‘The eggs were disposed in no particular order in a 
somewhat granular gelatinous matrix, each ovum being sur- 
rounded by its own membrane, and measuring 12 to 18, in 
diameter. Before the commencement of segmentation the 


ON NAUTILUS AND SOME OTHER ORGANISMS. 155 


ova of this species pass through a pronounced ameeboid stage 
(Figs. 4 and 5), subsequently becoming roundagain. It is also 
a favorable species for observing amceboid movements of the 
polar bodies. The movements are sluggish but unmistakable, 
and sometimes the polar bodies become widely separated from 
one another (Fig. 6). 


Fig. 4. Wie, fh. 


Fic. 6. 


Fies. 4 and 5.—Ova of Stylochus cinereus before commencement of 
segmentation, to show the ameeboid phase. Zeiss, 3 c, cam. luc. 

Fic. 6.—Ovum of S. cinereus preparing for the first division, to show 
the ameeboid character of the polar bodies. Zeiss, 3 c, cam. luc. 


Stylochus arenosus laid one irregularly contoured plate 
of eggs on May 6th. The ova were arranged in distinct rows, 
each ovum surrounded by its own proper membrane, and 
measuring 9 to 10 n. 

The egg-discs of two species of Planocera, occurring like 
the above species in shallow water off Rakaiya, offer a strong 
contrast to the egg-plates of Stylochus. I am under the 
necessity of naming these, at least provisionally, on the same 
principle as the above-described species of Stylochus. 

Planocera discus (mihi) attains a length of 45°5 mm., 
with a width of 28mm. The dorsal surface is coloured with 
delicate yellowish-brown (umber) reticulated markings, on 
which are scattered black nodal spots; interspaces whitish ; 


156 ARTHUR WILLEY. 


margin of body of the delicate brown ground colour inter- 
rupted by narrow white streaks; mid-dorsal region pitted 
with numerous minute black spots; tentacles brown with 
rose-yellow tip, 13 mm. from anterior margin ; large tentacular 
eyes at base of tentacles, more numerous than the smaller 
cerebral eyes; latter numbering about twenty-two to twenty-five 
on each side; female genital aperture about same distance 
from posterior end as the tentacles are from the anterior end ; 
pharyngeal bursz, seen from below, dense white; intestinal 
diverticula six or seven pairs. 

The margin of the body is always sinuous or skirt-like, and 
when at rest the animal is capable of assuming a nearly circular 
form. Both this and the following species can be handled with 
impunity, neither of them evincing the slightest disposition to 
laceration. 

Planocera discoides (mihi) is, for a Polyclade, an object 
of great beauty. I obtained two specimens of it from the 
bottom of a volcanic stone, on the top of which corals and 
sponges were growing off the south-west shore of Rakaiya, 
in some two or three feet of water at low tide. 

It reaches the length of 75 mm., with a width of 36 mm., 
always with a sinuous margin. Like the preceding species, it 
can assume an almost circular form. The body is remarkably 
transparent ; intestinal rami, seen from above, light brown, 
moniliform, anastomosing ; interspaces beset with numerous 
minute rubiginous spots; larger dark brown nodal spots 
scattered about dorsal surface ; margin of body light, pellucid ; 
tentacles pellucid, 22 mm. from anterior margin; female 
aperture same distance from posterior margin; large dense 
white shell-gland between the two genital orifices; eyes about 
twice as numerous as in P. discus; cerebrum in a clear pel- 
lucid area, which is deeply indented or excavated ventrally ; 
seven or eight pairs of intestinal diverticula, which present a 
dull greyish-white colour from below. 

P. discus laid a circular dise of egg-capsules on April 30th. 
The disc had the appearance of consisting of a series of con- 
centric circles, but closer examination showed that the rows 


ON NAUTILUS AND SOME OTHER ORGANISMS. Sd 


were arranged spirally, the spirals being here and there in- 
terrupted (Fig. 7). The egg-disc measured 15 mm. in diameter. 


Fic. 7.—Enlarged view of the egg-dise of Planocera discus. Only a 
few of the egg-capsules have been inserted in the drawing. 


Each row consists of numerous capsules packed closely to- 
gether in what amounts to a common gelatinous tube. The 
spirals can be unwound. In the outermost spiral were a few 
irregularly dispersed capsules. The normal capsules contained 
each from eight to eleven ova, and the latter had no other 
membrane round them individually. 

On May 4th the same individual, which had been kept 
isolated all the time, laid another much smaller disc of egg- 
capsules. I left these undisturbed, and in the course of time 
many of the ova developed into ciliated embryos with a pair of 
large primary eyes. The embryos swim about actively in the 
capsules, three or four to seven or eight in each capsule, sur- 
rounded by the fragments of those eggs which had not de- 
veloped. Many of the embryos had developed abnormally, 
owing probably to the fact that the water in which they were 
kept had not been changed frequently enough. Before seg- 
mentation commences, some of the ova in each capsule present 


158 ARTHUR WILLEY. 


an appearance of approaching fragmentation, and it is probable 
that this is a regular occurrence. Here also the polar bodies 
execute amceboid movements. 

In the case of P. discoides I found two egg-discs on 
May 27th on the same stone on which the adults were living. 
To the unaided eye they were not to be distinguished from 
those of P. discus, but microscopic examination showed 
that the gelatinous matrix! in which the egg-capsules were 


{> 


Fic. 8.—Four egg-capsules from an egg-disc of Planocera discoides 
each capsule contains four ova. Zeiss, 3 a, cam. luc. 


arranged in concentric rows was more continuous, and not 
divisible into concentrically disposed tubes; and, above all, that 
in each egg-capsule there were only four ova (Fig. 8). 

Of the four ova in each capsule, as a rule, only two develop 
into ciliated embryos; frequently, however, three, and rarely 
only one. I have in no instance found four ciliated embryos 
in a capsule. Those ova which do not develop undergo frag- 
mentation. I think the particles of the fragmented ova must 
be dissolved in the fluid contained in the capsules, and not 
absorbed by the remaining ova in the solid form. When two 
or three embryos are revolving in a capsule, there is usually no 
trace whatever of the previous existence of other ova in the 
same capsule ; and when there is only one embryo in a capsule 
it is no larger than other embryos. 

It is chiefly before the embryos begin to revolve that the 


1 A large number of amceboid bodies may be observed in the matrix in 
which the egg-capsules are embedded in this species. 


ON NAUTILUS AND SOME OTHER ORGANISMS. ~~ 159 


actual evidence of fragmentation is to be obtained (Figs. 9— 
11). Only rarely are the rounded particles derived from the 


Fie. 9. Iie, IO): 


CART 
i 
, 


4 


i 


sab 


2 
Lee 
AT 


rear 


Fics. 9, 10, and 11.—Planocera discoides. Kgg-capsules. In Fig. 9 
one ovum has undergone fragmentation; in Fig. 10 two ova have undergone 
fragmentation; in Fig. 11 the remains of the fragmented particles are seen 
lying against an embryo which had already acquired cilia, but had not com- 
menced to rotate. Zeiss, 3 c, cam. lue. 


fragmentation of the original ova, to be seen floating about 
in those capsules in which the embryos are revolving. 

2. Cirripathes anguina.—From the reef off Ralum I 
obtained a fine specimen of this unbranched Antipatharian, 
upwards of 9 feet in length. It reached up close to the surface 


160 ARTHUR WILLEY. 


at low water. The natives here call it “a pada ur-a-ta,” or 
simply “a ur-a-ta,’ which means “ the bones of the sea.” 

3. Nudibranchiata.—The Nudibranch molluscs, particu- 
larly the Doridide, are represented here by a great wealth of 
species of all sizes up to 60 mm. by 45 mm., many of them 
being brilliantly coloured. Their spiral egg-bands may be 
obtained without difficulty. Several species, including two 
Molids, have laid their eggs in my dishes. 

On the reef off Ralum, and in Blanche Bay too, there are 
to be found large numbers of very long spirally coiled ribbons, 
so thickly encrusted with sand that they appear to consist of 
nothing but sand. The natives call these ‘‘a pipia,” which 
means “‘ earth” or ‘“‘ ground.’ Some are tough and elastic, 
others brittle. ‘They are the egg-bands either of species of 
Doridide, or of other Opisthobranchiate Mollusca, but I have 
not yet been able to identify them. 

4. Onchidium.—One of the commonest Mollusca in the 
shallow water off the south-west shore of Rakaiya is a species 
of Onchidium. It is often found in the very heart of large 
friable stones, approximately at low-water mark, but some- 
times further out. It occurs in other parts of Blanche Bay as 
well. What service the dorsal eyes can be to it is not easy to 
imagine. Semper supposed it was to enable the Onchidium to 
perceive and escape from what he says is its chief enemy, 
Periophthalmus. But although I have seen a small species 
of Periophthalmus at Rakaiya, on the muddy shore of the so- 
called lagoon, it is difficult to accept Semper’s view. 

The papille which carry the eyes are obviously homologous 
with the retractile branched respiratory papille in the posterior 
dorsal region, and every transition can be observed between 
them. 

When kept in confinement in jars, Onchidium asserts its 
pulmonate qualities by often creeping out of the jars for long 
distances. 

5. Larve of Polygordius and Squilla.—It is worth 
recording the presence of larve of Polygordius in the “ Auf- 
trieb”” in Blanche Bay. In the narrow strait which separates 


ON NAUTILUS AND SOME OTHER ORGANISMS. 161 


Rakaiya from the mainland I have taken one larva of Poly- 
gordius in an advanced stage of metamorphosis, when the long 
body has grown out behind the trochophore. On the same 
occasion the “ Auftrieb ” contained also larvee of Squilla. 

6. Phosphorescence.—The astonishing phosphorescence 
which is to be observed when fishing with the tow-net at night 
in Blanche Bay is in large measure due to Copepods. and 
Ostracods. 

7. Ascidians.—The Ascidiau fauna here is richer than I 
expected to find it, since, as Herdman has pointed out, the 
tropics are not the metropolis of the Ascidians. 

The Didemnide, which form one of the most difficult groups 
of compound Ascidians from a systematic point of view, 
are strongly represented by red, white, yellow, and green 
varieties. 

Botryllus also occurs, though notcommon. I only know at 
present of two species—one very thin and white, attached to 
the lower surface of corals; the other thicker and purple, 
found growing on a T'ridacna shell. 

In regard to the compound Ascidians, I have been struck by 
the apparent absence of any member of the Polyclinide. 

The simple Ascidians are represented by numerous species, 
one of which especially has peculiarities of such a nature that 
I will describe it at some length, reserving an account of the 
other Ascidians for a future occasion. 

8. Styeloides eviscerans, n. sp.—In 1885, Sluiter 
described, in the ‘ Natuurkundig Tijdschrift voor Ned-Indie,’ 
Bd. xlv, a simple Ascidian, under the name of Styeloides 
abranchiata, n. gen. et sp., in which the branchial sac and 
intestine were absent. Such was the condition of the animal 
in other respects that Sluiter was led to suppose that this must 
be the normal state of things, and founded the new genus 
accordingly. 

In his extremely useful ‘‘ Revised Classification of the 
Tunicata” (‘Journ. Linn. Soc. Zovol.,’ vol. xxiii), Herdman, 
commerting on this species of Sluiter’s, says, ‘This is such 
an exceptional and remarkable case that I cannot help sus- 

VOL. 39, PART 2.—NEW SER. L 


162 ARTHUR WILLEY. 


pecting that the single specimen examined by Sluiter was 


merely an individual abnormality.” 
I venture to hope that I have found the solution to this 


enigma in the species about to be described. 


ae 
eS 
LG 


Ne 


Fie. 12.—Group of seven individuals of Styeloides eviscerans, n. sp. 
represented as lying attached to the surface of a fragment of stone. In the 


large individual to the right the digestive tract is indicated in process of 
extrusion through the atrial aperture. a. Anus with frilled margin. 7. In- 
h, Foreign organisms attached 


testine. 47.s. Branchial sac. e. Endostyle. 


to test. 


ON NAUTILUS AND SOME OTHER ORGANISMS. 163 


The species is not common, but I have obtained three or 
four examples of it from the lower surface of stones off the 
south-west shore of Rakaiya, in one half to one fathom of 
water. 

The accompanying sketch (Fig. 12) represents a fine and 
typical group of individuals of this species. The members of 
the group are so intimately connected together by the mutual 
fusion of their tests that one would at once suppose that they 
had arisen from a parent stock by budding. Such, however, 
is not the case, since by making incisions it is found that it is 
only a fusion of test-substance, and not a true organic union. 
Moreover sometimes isolated individuals are to be found, as in 
Fig. 18. 


~ 
eas 


- Fre. 13.—Styeloides eviscerans. Outline of isolated specimen. a. 
Atrial siphon. ¢. Processes of test. 

The general colour of the animal or group of animals is a 
characteristic dull reddish brown, the colour being more pro- 


164 ARTHUR WILLEY. 


nounced in the neighbourhood of the apertures, while the lips 
of the latter are a pure dark red, interrupted by four light streaks 
which indicate the quadripartite character both of the buccal 
and atrial orifices. 

As seen in the figures, the individuals are not always attached 
to the rock by the same side, but sometimes by the ventral 
side, sometimes by the right, and sometimes again by the left 
side (Fig. 12). 

The total length of the group represented in Fig. 12 was 
91°5 mm., and the greatest breadth of the group 31 mm. 
The animal of which an outline is given in Fig. 13 measured 
51°5 mm. in length, and the atrial opening was 20°5 mm. 
removed from the buccal aperture. By its external appearance 
alone it is an extremely well-marked species. The surface of 
the coriaceous test is in some places wrinkled and in other 
places smooth. 

The most remarkable peculiarity of the new species, how- 
ever, is the faculty which it possesses of evisceration. 

After I had had them for a short time in a vessel where 
everything was fresh and in good condition, I suddenly dis- 
covered a number of digestive tracts lying at the bottom. On 
then inspecting the Ascidians, of which there were several 
species present at the time, I found that they were all living 
and in a bealthy condition. 

Eventually I actually observed the process of evisceration 
taking place (cf. large individual to the right in Fig. 12). It 
takes a rather long time before the process is completed. Itis 
effected by violent periodic contractions of the atrial siphon. 
After it is over the animal presents a perfectly normal and 
healthy appearance, opens and closes its siphons, and is suscep- 
tible to irritation and to the influence of cocaine. So constant 
is this ejection of the digestive tract that if it is desired to pre- 
serve specimens intact, they must be placed in alcohol imme- 
diately after capture. 

The dissection of an individual which has ejected its bran- 
chial sac and intestine discovers no laceration whatever; and 
undoubtedly, in ignorance of the habit of evisceration, one 


ON NAUTILUS AND SOME OTHER ORGANISMS. 165 


would be tempted to suppose, as Sluiter did, that the absence 
of an alimentary canal was the normal condition. 

In adopting Sluiter’s generic name the diagnosis must of 
course be amended. 

Before evisceration takes place the branchial sac is found to 
have the usual vascular connections with the mantle, but the 
endostylar area seems to have a very loose attachment to the 
mantle, and can be readily detached. When the branchial sac 
is ejected the dorsal tubercle (Figs. 14 and 15) and peripharyn- 
geal groove are left behind, and there is a corresponding 
triangular excision in the wall of the ejected branchial sac. 


Fie. 14. Iie, INK, 


Fic. 14.—Styeloides eviscerans. Dorsal tubercle. yp. . Peripharyn- 
geal band. d. ¢. Dorsal tubercle. g. Ganglion. 
Fie. 15.—Dorsal tubercle of another individual of 8. eviscerans. 


The dorsal lamina is a simple undulating or crumpled mem- 
brane, and there are four folds of the wall of the branchial sac 
on each side. 

The genital saccules have the characteristic subeylindrical 
form, and occur on both sides attached to the mantle. Curious 
bodies called endocarps, whose nature I do not understand, also 
occur on the inner surface of the mantle as in Sluiter’s 
species. 


166 ARTHUR WILLEY. 


The latter is probably a distinct species from the one I have 
described, although there are many features common to them 
both, particularly the external form and mode of attachment. 

I am a little puzzled to understand what Sluiter says about 
the endostyle, and am inclined to think there must be some 
mistake about it, as there is no trace of a typical endostylar 
epithelium in the section figured by him. I have even 
observed a line or ridge in the ventral surface of the mantle 
corresponding very closely to his Taf. viii, fig. 2, but this possi- 
bly represents the former line of contact between the endo- 
style and the mantle. 

There are naturally a great many more questions to be 
answered in connection with this remarkable Ascidian, but I 
have probably said enough to show that its property of evis- 
ceration is its most distinguishing peculiarity, and thus to 
afford an explanation for an otherwise inexplicable anomaly. 

Yours very truly, 
Artour WILLeEY. 


Raum, GerMAN New GUINEA, 
via SINGAPORE ; 
September 24th, 1895. 


Dear Professor LANKESTER, 


1. Significance of the Siphunele in Nautilus 
pompilius. 

Being desirous of obtaining, if possible, experimental evi- 
dence as to the physiological significance of the siphuncle 
in the pearly Nautilus, I have made several successful attempts 
to cut the siphuncle without otherwise injuring the animal. 
The evidence supplied by the experiment cannot be regarded 
as conclusive, on account of the altered conditions of depth and 
temperature to which the Nautilus is exposed by being brought 
up to the surface, but it may be well to consider what the 
results indicate. 


ON NAUTILUS AND SOME OTHER ORGANISMS. G7 


At first I sawed through the shell into one of the chambers, 
and then cut the siphuncle. This method has the disad- 
vantages of injuriously affecting the efficiency of the chambers, 
and of causing a more or less considerable loss of blood to the 
animal. The latter will, however, live in confinement about 
as long as untouched individuals. 

A young Nautilus operated upon in this way on June 26th 
was placed in the sea in shallow water, for its movements to 
be watched. It sank slowly to the bottom, and then for a long 
time made active revolving motions about the vertical axis, 
but scarcely made any progressive movements. 

On another occasion (July 10th), after several trials, I found 
that the best way of performing the operation is to saw through 
the shell in the neighbourhood of the posterior portion of the 
body of the animal, over the cardiac region, and not to tamper 
with the chambers. If the shell be held mouth downwards, 
this point lies approximately in the same vertical and trans- 
verse plane with the points where the free margin of the 
mouth of the shell merges into the umbilicus. When a large 
enough hole has been made in the shell to admit the scissors, 
the shell being still held upside down, the ventral visceral 
portion of the body usually detaches itself from the shell, or 
can be readily caused to do so, and, sinking inwards, exposes 
the root of the siphuncle, which can then be severed. On 
returning the shell to its normal position the body immediately 
resumes its normal intimate contact with the wall of the cavity 
in which it lives, and the pressure so exerted prevents any 
extensive loss of blood. Under these conditions the operation 
does not, as a rule, appear to affect the vitality of the animal 
in any degree. 

A Nautilus! which was treated in this way on July 10th, 
on being placed in the sea swam about very vigorously for 
some time in the middle stratum of water, but most of the 
time at a little distance from the bottom. On September 13th 
I operated on four more individuals taken in Talli Bay, on 


1 It should perhaps be mentioned that in this particular individual I acci- 
dentally cut into the last chamber, and plugged the opening with wax. 


168 ARTHUR WILLEY. 


the north coast of the Gazelle peninsula. One of them showed 
a tendency to sink to the bottom, which it always performed 
very gradually. In this one I had accidentally punctured the 
mantle over the heart. The others remained floating and 
swimming about on the surface during the whole time of 
observation. They did not go far in one steady direction, but 
tended to go in circles, as in fact did another one whose 
siphuncle was uncut. If one of the individuals floating at the 
surface was forced down to the bottom with a hand-net, it 
would slowly rise to the surface again. This also often happens 
with a Nautilus that has not been operated on. 

The results indicated by the above experiments, which, it 
may be added, are worth repeating, may be summarised as 
follows: 

The cutting of the siphuncle (a) does not temporarily affect 
the vitality of the animal; (3) does not prevent it from making 
movements of translation ;! (y) does not prevent it from floating 
at the surface; (6) does not prevent it from sinking to the 
bottom. 

It still remains to be ascertained whether a Nautilus whose 
siphuncle has been cut, having sunk to the bottom of the sea 
in shallow water, will undertake a journey to the surface. 
My experiment of July 10th would seem to indicate that this 
" might be expected to occur. 

The above experiments do not appear to oppose the view 
which I expressed in a former communication—that the si- 
phuncle of Nautilus pompilius is, in some measure, of the 
nature of a vestigial structure. 

It might indeed be legitimate to suppose, on the principle of 
the correlation of organs, that in the Nautiloidea the course 
of evolution has led to a reduction of the siphuncle pari 
passu with an increase in the efficiency of the chambers as 
hydrostatic organs. 

1 Tn speaking above of progressive movements I mean, of conrse, in the 
usual backward direction. 


ON NAUTILUS AND SOME OTHER ORGANISMS. 169 


2. Some Features in the Arterial System of N. 
pompilius, as determined by Injection. 


(1) Cireulus Pallialis.—After successful injections a sin- 
gular feature in the circulatory system is to be observed. The 
marginal pallial artery, which I described and figured in a 
former note, is found to pass on each side, dorso-laterally, into 
a branch of the dorsal aorta, so that a complete arterial circuit 
is produced. 

I have even partially injected the marginal pallial arteries 
from the dorsal aorta itself, but the injection fluid did not 
proceed very far in this centripetal direction, owing no 
doubt to the resistance it met with from the action of the 
heart. 

For this remarkable arterial circuit, produced by the con- 
fluence of the marginal pallial arteries, which arise ultimately 
as branches from the so-called “lesser aorta,” with a pair of 
branches! from the great aorta, I propose the above name of 
“circulus pallialis.”” It is illustrated in the accompanying 
sketches (Figs. 16 and 17). 

In Fig. 16 the posterior convex extremity of the body is 
supposed to be somewhat tilted up, in order to show the whole 
outline of the septum-producing area of the mantle. This out- 
line is very distinct in fresh specimens, and the region of the 
mantle enclosed by it is distinguished from the surrounding 
portions of the mantle by its greater thickness and opacity. 
. As already stated, this is the portion of the mantle which 
manufactures the septa, and it has an abundant arterial supply 
through the ramifications of the two main branches of the 
posterior pallial artery. These ramifications may be grouped 
together as the pallio-septal arteries; and it is surprising 
to see how rigidly they are confined within the septal 
contour. 

It will be noticed that the latter makes on each side a 


1 Yor reasons which will presently appear, these may be called the pallio- 
nuchal arteries. 


] 


70 ARTHUR WILLEY. 


posep a? ll mn 


Fie. 16.—N. pompilius, ¢. Dorso-posterior aspect of visceral region, to 
illustrate the circulus pallialis and the septal contour. x. m. Nuchal mem- 
brane. x.a. Nuchal artery. m.e. Free mantle-edge. p.z.a. Pallio- 
nuchal artery. m.p.a. Marginal pallial artery. co/. Columellar or shell 
muscle. p.c.a@. Posterior columellar artery. p. p. a. Posterior proventricu- 
lar artery. s.c. Septal contour. s. Siphuncle. p. s. a. Pallio-septal arteries. 
s.a. Siphuncular artery. 7. Liver. 7. Intestine. 7. Testis. post. p. a. 
Posterior pallial artery. g. Gizzard. 

N.B.—The dorsal aorta and its branches are indicated by dotted lines. 
They show dimly through the skin when injected. 


ON NAUTILUS AND SOME OTHER ORGANISMS. tat 


a 
-- ~ 


~ 
4 \ 
/ \ 


‘ \ 
\ 


pT TTT 
———————— \ 


Fic. 17.—N. pompilius, ¢. View of nuchal region, to further illustrate 
the circulus pallialis. The dorsal free mantle-edge is reflected and a median 
incision made. 4%. Hood. coz. Conecavity at base of hood, in which the 
nuchal membrane (z. m.) lies. f. Dorso-posterior portion of funnel. x. a. 
Nuchal artery. co/. Columellar muscle. m. p. a. Marginal pallial artery. 
p.m. a. Pallio-nuchal arteries. d.a. Dorsal aorta. m.e. Free mantle-edge. 
c. e. Cut edges of mantle and body-wall. 

N.B.—The dorsal aorta shapes its course in this region in accordance with 
the state of repletion of the erop. 


172 ARTHUR WILLEY. 


symmetrical figure with the outline of the great shell-muscles! 
where the latter abut on the shell. 

In Fig. 17 the union of the marginal pallial artery with the 
left pallio-nuchal artery is represented from the inner surface 
of the mantle. Here it is seen that the two arteries unite in 
the dorso-lateral angle where the mantle and the funnel-flap 
fuse with the body-wall; and furthermore, that from the same 
point an artery is given off which passes forwards and gives 
off branches to the nuchal membrane. The latter structure 
was accurately described by Owen as a “semilunar ridge” 
lying in the concavity at the base of the hood, and applied to 
the involute convexity of the shell. 

Owen thought it might serve to prevent the shell from 
encroaching too much upon the hood ‘‘in the act of creeping.” 
We now know that the animal does not creep on its hood with 
reversed shell. 

The nuchal membrane would seem to be responsible for the 
dense black colour of the involute portion of the shell, and 
possibly also exerts a lubricating influence. Keferstein calls 
it the “ Nackenlappen.” 

(2) Genital Arteries (Fig. 18).—Nolessthan three arteries 
arise directly from the heart which, to my knowledge, have 
hitherto escaped attention. ‘They are (a) the artery of the 
genital duct or gonaducal artery ; (0) the artery of the genital 
gland or genital artery; (c) the artery of the pear-shaped 
gland. 

The accompanying sketch (Fig. 18) obviates the necessity of 
a detailed description. The genital artery is submedian, and 
the main trunk lies on the dorsal side of the genital gland. 

Both the gonaducal artery and the artery of the pear-shaped 
gland give off a branch which passes into the perigonadial 
membrane, and this apparently trifling fact, combined with the 
‘subsymmetrical relations of the gonaduct and the pear-shaped 
gland, may indicate that the latter is the metamorphosed 
‘genital duct of the left side, and not, as I believe has been 

1 For purposes of nomenclature it will be found convenient to speak of the 
great shell-muscles as the columellar muscles. 


ON NAUTILUS AND SOME OTHER ORGANISMS. 173 


LON 


WAN 7 i 
es ie Z\ \ ASKEW SS aces 
ee to 


_— 
-— 
a> 


Fig. 18.—N. pompilius, ¢. Genital arteries from below. r. Rectum. 
r. a. Rectal artery. (N.B.—The rectal arteries are very variable.) aut. p. a. 
Anterior pallial artery. 7. a. Intestinal artery. (N.B.—This artery usually 
passes to the right of the rectum, as shown in this figure; but in one 
instance I have observed it to pass down to the left of the rectum.) 47. v. 
Branchial veins. post. p. a. Posterior pallial artery. p.s.g. Pear-shaped 
gland with its artery. a’. Branches of the preceding artery and of the gona- 
ducal artery, which supply the superjacent perigonoidal membrane. gez. a. 
Genital artery and its branches. ¢. Testis. ¢.0. Aperture of testis. p. v.0. 
Orifice of communication between the pericardial and visceral portions of 
body-cavity, through which the posterior pallial artery passes. goz. a. Gona- 
ducal artery. v.s. Vesicula seminalis. (N.B.—This structure, tlie testis and 
pear-shaped gland are closely united to the heart by a membrane.) J. v. 
Needhamian vesicle or spermatophore sac. #. Dotted line to indicate where 
the pallial duplicature merges into the body-wall ventrally. 


174 ARTHUR WILLEY. 


suggested, the morphological equivalent of an entire left genital 
apparatus. 

In the female the ramifications of the genital artery pass 
up on to the surface of the individual ova, and form a kind of 
capillary system, the finer branches following, but not always 
confined to, the reticular markings formed by the ridges of the 
follicular membrane which project into the yolk (Figs. 4 
and 5). 

The meshes formed on the surface of the ova by the inter- 
section of the follicular ridges or plications are much wider in 
submature ova than in the less mature, and the ridges would 
presumably be found to flatten out in completely ripe eggs, 
although it has not been my good fortune hitherto to have 
found any such. At the animal pole of the egg the ridges are 
absent, and those which lie at the margin of this area form in- 
complete meshes as described by K 6lliker in the ovarian ova 
of other Cephalopods (Fig. 19). 


Fie. 19.—Fresh ovarian ovum of N. pompilius, to show the reticular 
markings produced by the plications of the follicle. j. a. Clear polar area, in 
the centre of which lies the germinal tract. 


The clear polar area of the ovum has usually a sub- 
triangular shape, and from each of the corners of the 
triangle what may be called a line of weakness occurs in the 
follicular wall, bound on either side by incomplete meshes 
(Fig. 19). 

The arteries which traverse the surface of the ova give off 
minute branches which pass inwards, as it were, into the 
depths of the follicular ridges; and these deep-lying vessels 
anastomose with one another, while the superficial branches 
appear, as a rule, not to form anastomoses. It may be added 


ON NAUTILUS AND SOME OTHER ORGANISMS. 175 


that the impression of anastomoses is much more readily 
conveyed by examination with a hand-lens than it is by the 
use of the compound microscope. 


Fic. 20.—Ovary removed from body, and seen from dorsal aspect. On the 
right of the figure the follicular meshes of a submature and a half-mature 
ovum are partially inserted to show difference in size. o. a. Neck of ovarian 
sac, which bears the aperture. o. Ova. gez.a. Genital artery. Only the 
more superficial branches are indicated. 


The germinal tract appears in the centre of the clear polar 
area as a faint whitish spot, and is turned towards the ventral 
aspect of the ovary (Fig. 20). The older ovarian ova are ren- 
dered shapeless by mutual pressure, with, however, a roughly 
oval outline. In this condition an ovum may measure 15:5 
mm. in length, with a breadth of 11°55 mm. When the pres- 
sure is released by slitting open the ovary the ova round up, 
and those which are submature have an average diameter of 
some 10 mm. The yolk is viscous and glutinous, and pos- 
sesses a translucent brownish tinge. ‘The nearly ripe ova 
rupture with the utmost facility. 

From a consideration of the size and relative states of 
maturity of the ova, it might be expected that they are laid 
singly. Every month, from December to September inclusive, 
I have been able to obtain over-ripe males (with spermato- 
phores in the dorsal buccal recess and in the Needhamian 
vesicle) and submature females. Once in July I obtained a 
male with a discharged spermatophore capsule in the buccal 
recess ; in fact, I have found this more than once. 

From these facts, and from the fact which I have previously 


176 ARTHUR WILEY. 


mentioned of the relative scarcity of the females in compari- 
son with the males, I draw three provisional conclusions: 
firstly, that during the process of reproduction (fecundation 


Fic. 21.—Female genital organs of N. pompilius, seen from below, to 
show the direction in which the polar areas of the ova lie. yp. a. Polar areas of 
ovarian ova. o. a. Aperture of ovarian sac. g. Uterine portion of oviduct. 
v. Vaginal portion of oviduct. 


and spawning) the females live in retirement; secondly, that 
the females practise what the Germans call ‘ Brutpflege;” 
and lastly (what I regard as almost certain), that repreducues 
takes place all the year round. 

(3) Cephalic Arteries (Figs. 22 and 23).—For a descrip- 
tion of these arteries it will suffice to refer to the explanation 
of the figures. I will, however, call attention to the varia- 
bility of the right and left anterior proventricular arteries ; 
the latter was absent from the individual represented in fig. 7. 
The two main trunks into which the dorsal aorta divides be- 
hind the brain may be called the right and left innominate 
arteries. 

It is a singular fact that the great median buccal artery 
always springs from the right innominate artery. The con- 
stancy of this origin would seem to indicate that it is poten- 
tially a paired structure. 


ON NAUTILUS AND SOME OTHER ORGANISMS, iar 


Fic. 22.—N. pompilius, g. Dissection, from above, of the cephalic 
region to show the cephalic arteries. An incision has been made through the 
nuchal membrane, the hood, and the buccal membrane. The brain-capsule has 
been opened, and the median portion of the mantle behind the nuchal region 
has been removed. 0.c. Buccal cone. ent. inf. Dorso-lateral inner row 
of tentacles (the superior labial processes of Owen). 7. a. Labial arteries 
supplying the buccal membrane and fringe. s. m. a. Superior mandibular 
artery. s.7. Superior retractor muscles of the jaws (Owen). 4%. a. Buccal 
artery. 7@. m.d. Paired inferior mandibular arteries. 2. m. Nuchal mem- 
brane. a.c. a. Anterior columellar artery. a. p. a. Anterior proventricular 
artery. f. Dorso-posterior portion of funnel. 2. e. Free mantle-edges 
c. e. Cut edge of mantle. p.z.a. Pallio-nuchal artery. d. a. Dorsal aorta. 
cer. Brain with cerebral arteries. 2%. Cut edge of hood. 

N.B.—Apart from the cerebral arteries, all the anteriorly directed branches 
of the innominate arteries pass below the cerebral capsule. 


VOL. 39, PART 2,—NEW SER. M 


178 ARTHUR WILLEY. 


Attention may also be drawn to the two arterioles which 
arise from the base of the right proventricular artery and 
supply the wall of the dorsal aorta. 


Fic. 23.—N. pompilius, g. Sketch to show the arteries which arise 
from the innominate arteries after the brain has been cut through. 4% a. 
Right tentacular artery, which gives off branches to all the tentacles of its 
side. The innermost or mesiad branch (o/f a.) supplies the bipartite 
laminated organ of van der Hoeven, which possibly represents a pair of 
inferior labial processes, and may have an olfactory function. fa. Infun- 
dibular artery, which passes through the cartilage into the funnel. The 
latter also receives minor branches from the columellar arteries. oc. a. Arteries 
to the eye. ped. a. Pedal artery, a convenient name for the common trunk 
from which the infundibular and tentacular arteries arise. opt. Optie gan- 
glion. cer. Brain cut across and drawn aside. a.c. a. Anterior columellar 
artery. a. p. a. Right anterior proventricular artery. Compare its origin in 
Fig. 7, where it arises from the aorta. d. a. Dorsal aorta. /. a. p. a. Left 
anterior proventricular artery, usually but not invariably present; it some- 
times arises from the left innominate artery, as indicated by the dotted lines ; 
and sometimes, as in this example, from the dorsal aorta. zz. Innominate 
arteries. d.a. Buccal artery. 7. m. a. Inferior mandibular arteries. 


ON NAUTILUS AND SOME OTHER ORGANISMS, 179 


8. Further Remarks, 


(1) Capillaries.—As far as I have been able to observe, 
there seems to be a true capillary system in the free portion 
of the mantle, that is to say, in the pallial duplicature. I 
have previously described the pallial arteries and veins. There 
may be another system of capillaries in the funnel, which has 


Fie. 24.—Photograph of a living specimen of Nautilus pompilius, 
taken by A. Willey at Ralum in 1895. 


an astonishingly rich vascular supply, and at whose base two 
large veins, the infundibular veins, may be observed to 
pass into the vena cava. But I have observed no veins in con- 
nection with the genital organs, and I gathered from micro- 


180 ARTHUR WILLEY. 


scopic examination that the follicular arteries on the surface 
of the ova possessed free openings. 

(2) Blood.—I will here only mention that the blood is a 
syrupy fluid with a pronounced blue tint, which becomes very 
dark on exposure to the air. The corpuscles comprise amceboid 
aud fusiform cells, the latter being somewhat Gregarina-like 
in appearance. 

Yours very truly, 
Artuur WILLEY. 


THE BRAIN OF A FQTAL OkNITHORHYNCHUS. 181 


The Brain of a Fetal Ornithorhynchus.' 
Part I.—The Fore-brain. 
By 


G&. Elliot Smith, Wi.D., Cia. i., 
Demonstrator of Anatomy, University of Sydney, N.S.W. 


With Plate 11. 


Tue foetus, with whose brain this paper is concerned, was 
received already preserved in alcohol from the Australian 
Museum of Sydney by Professor J. 'T. Wilson, who has 
described in full the external appearances and measurements.” 
It is sufficient here to mention that its extreme length, mea- 
sured along the dorsal contour, is 80 mm. | 

The head was split sagittally some distance from the middle 
line, and the larger piece, after being embedded in celloidin, 
was cut in a complete series of coronal sections, each 50 1 
thick. These were stained with haematoxylin and decolourised 
with alcohol acidulated with picric acid, which afforded a very 
useful counter-stain. By means of the picric stain the nerve- 
fibres were clearly demonstrated, so that it was easy to make 
out their general distribution sufficiently accurately to compare 

1 This paper forms part of a thesis dealing with the ‘ Anatomy and Histology 
of the Cerebrum of the Non-placental Mammal,” which was awarded the 
University Medal when presented to the examiners for the M.D. degree at 
the University of Sydney, March, 1895. ‘The account of the gross anatomy 
is now being published elsewhere. ‘The account of the histology of the adult 
brain will be published shortly. 

2 « Description (with figures) of a Young Specimen of Ornithorhynchus 
anatinus,” ‘Proceedings of Linnean Society N.S.W.,’ vol. ix, part 4, 2nd 


series) p. 682. 


182 G. ELLIOT SMITH. 


with Weigert- and Golgi-stained specimens of the adult brain. 
For the opportunity of examining this valuable series T am 
deeply indebted to the generosity of Professor Wilson, whose 
kindly interest and valuable advice I gratefully acknowledge. 

In recording the disposition of parts in a single stage of a 
developing brain, the account, in the absence of other stages 
to show the origin and destiny of the various regions, must be 
almost purely descriptive. This paper, therefore, will indicate 
the apparent differences in the specimen under discussion from 
the ordinarily accepted account, rather than attempt to ex- 
press an opinion upon any of the morphological questions 
raised. Many interesting and important facts are elucidated, 
however, by a comparison with the adult brain, a description of 
whose histology forms the bulk of the thesis of which this 
paper constitutes a part. 

A comparison with some valuable serial sections of foetal 
Perameles and Macropus in various stages of development 
has thrown a considerable amount of light upon many other- 
wise obscure points. These specimens were placed at my 
disposal by Professor Wilson and Mr. J. P. Hill. I have also 
examined some imperfect series of brains of foetal Phalan- 
gista. Several foetal Echidne, which were lately received, 
are unfortunately not ready for examination, but will be 
described subsequently. . 

The literature of this subject is very scanty and incomplete. 
Before the Linnean Society of New South Wales, Dr. C. J. 
Martin and Mr. J. P. Hill recently described the condition of 
the neural plate in the early embryo of Platypus,! and Selenka 
has described early stages of the Opossum brain.? C. L. 
Herrick has published brief notes on, and figured sections of, 
the foetal brain of Didelphys.? Osborn,‘ in his well-known 
paper upon the corpus callosum, described the hippocampal 


‘«On a Platypus Embryo from the Intra-uterine Keg,” * Proceedings 
Linnean Society of New South Wales,’ vol. ix, part 4, 2nd series, p. 736. 

2 * Das Opossum,’ Wiesbaden, 1887. 

> * Journal of Comparative Neurology,’ 1894. 


4 *Morphologisches Jahrbuch,’ Bd. xii, 


THE BRAIN OF A FQITAL ORNITHORHYNCHUS. 183 


region of the cerebrum of pouch specimens of Macropus of 
various ages. These brief notes are all that I know referring 
to the development of the brain of the non-placental mammal. 

In connection with Osborn’s work, one cannot refrain from 
expressing regret that he was so biassed by his previous work 
upon the Amphibian and Sauropsidan cerebrum (Part I of 
his paper) as to misrepresent the condition which obtains in 
the young kangaroo. I am well acquainted with the decep- 
tive appearance which is presented by the brain of the foetal 
Macropus, especially when coloured with ordinary nuclear 
stains—an appearance which lent itself so unfortunately to 
Osborn’s purpose ; but the important position occupied by the 
marsupial in his series should have demanded a more critical 
examination than his paper indicates. For, as the connecting 
link between the whole sub-mammalia—to which his argument 
mainly applied—and the mammalia—to which alone can the 
distinguishing names be applied a priori—M acropus formed 
the link upon which the cohesion of the whole chain of his 
argument depended. By his erroneous interpretation of the 
condition in Macropus his argument lost all its cogency ; for 
in showing, as he clearly did, the homology of that commis- 
sural band, which in the non-placental mammal is undoubtedly 
fornix commissure, with the dorsal commissure of Sauropsida 
and Amphibia, he prepared the way for the statement that the 
corpus callosum—as distinct from the fornix commissure— 
exists only in Placentalia. 

In describing the foetal brain there is, unfortunately, no 
satisfactory account of the adult brain to which one can refer. 
The accounts of Meckel and Zuckerkandl are quite valueless, 
from the fact that their descriptions were based upon very 
badly preserved material. Owen’s description and Garner’s 
notes are reliable as far as they go, but they are very brief and 
general, and deal only with the macroscopic anatomy, as also 
does Turner’s paper. The only attempt at a systematic 
account of the histology—that of Hill—is so clearly biassed 
by its writer’s previous work, that it gives a very erroneous 
and altogether misleading idea of the structures it is supposed 


184 G. ELLIOT SMITH. 


to describe. In this aceount, therefore, such brief references 
will be made to the histology of the adult brain as are neces- 
sary to a proper understanding of the foetal structures. 

Forming as it does the link between the Mammalia and the 
lower Vertebrates, Ornithorhynchus occupies a position of 
supreme importance and interest to the morphologist. This 
importance is enhanced in the case of the brain, because this 
organ presents a number of very significant transitional fea- 
tures, several of which shed an important light upon the 
morphology.of the cerebrum in the whole Vertebrate series. 
In addition there are a number of purely “ individual”? fea- 
tures—notably the enormous development of the trigeminal 
nerve and sensory tract—which give added interest to the 
study. Although in the histological differentiation of its parts, 
and in their general arrangement, the cerebrum of Platypus 
clearly conforms to the mammalian type, yet numerous fea- 
tures—such as the arrangement of the hippocampus, the dis- 
position of the structures in the lamina terminalis, and the 
“ precommissural area”’—indicate its close Sauropsidan aflini- 
ties. It is of interest to note that the resemblances to the 
lower brain are even more marked in the foetus under con- 
sideration than in the adult. 


General Account of the Fore-brain. 


At the outset one is immediately confronted with the difficul- 
ties of the vexed question of nomenclature. With the increasing 
knowledge of brain anatomy it is perhaps only natural that 
more exact divisions and distinguishing terms, more in accord- 
ance with the advanced ideas, should be introduced. But 
when familiar terms are used in an altogether unfamiliar sense, 
or applied to entirely different parts of the brain in order to 
give place to new terms no more appropriate than those they 
supersede, the enormous intrinsic difficulty of this subject 
becomes seriously increased, and an undesirable element of 
confusion is introduced. ‘The accurate system of nomenclature 
introduced by His loses much of its value by a confusing 
application of Huxley’s terms ‘ prosencephalon” and “ thala- 


THE BRAIN OF A FATAL ORNITHORHYNCHUS. 185 


mencephalon ” in quite a different sense from that in which they 
are universally used, which must prevent the general acceptance 
of his suggestions. Although His’ system of nomenclature is 
otherwise very convenient for descriptive purposes, the mor- 
phological value of his divisions is seriously open to question, 
especially the somewhat arbitrary limits to his “ Endhirn.” 
The nomenclature of v. Kupffer seriously adds to the confusion. 
One cannot but admire the exactness of Burt Wilder’s definitive 
terms, but they are so uncouth and bizarre that the English 
student has to learn a new language before he can master the 
simplest description. In this paper the most unambiguous 
terms will be selected from the different systems, the alterna- 
tive names being indicated in brackets; but as far as possible 
Huxley’s well-known nomenclature will be employed. 

Hach cerebral hemisphere consists of an ovate mass, attached 
to the lateral aspect of the front end of the neural tube. The 
maximum length of each hemisphere, including the bulbus 
olfactoril, is 5°15 mm., and at its mid-point its height and 
depth are about half that measurement. The diencephalon 
(with the median part of the “ Endhirn”), whose extent is 
schematically represented by the shaded area in fig. 1, is of 
the same depth as the posterior part of the hemisphere, but is 
much shorter. Posteriorly the hemisphere overlaps the me- 
sencephalon (mes.) to a slight extent. 

In a median sagittal section through the fore-brain, which 
is represented somewhat diagrammatically in fig. 2, the optic 
nerve is seen as a prominent object (opt. n.) in the floor, so 
that it forms a convenient starting-point. Immediately behind 
the optic nerve the floor sinks to form the recessus infun- 
dibuli (7. i.) of His (recessus postopticus of Burckhardt), 
from the posterior extremity of which the tubular hypophysis 
(hyp.) extends at first downwards, and then bends forwards at 
its extremity. ‘The anterior wall of this extremity contains a 
number of neuroblasts and a distinct layer of nerve-fibres, 
which extends upwards on to the floor of the third ventricle. 
The bent extremity of the tubular hypophysis is closely sur- 
rounded by the large glandular (buccal) part of the hypo- 


186 G. ELLIOT SMITH. 


physis (pit.), which extends forwards for a considerable dis- 
tance. It consists of a dense mass of highly convoluted 
tubules (fig. 11, pié.), lined with cubical epithelium. Imme- 
diately in front of the optic chiasma the floor again becomes 
depressed (fig. 2, 7. 0.), and on either side of the middle line 
forms a distinct recess (fig. 11, 7. 0.). A comparison with the 
early stage of the Perameles brain shows that this recess is 
the remains of the optic diverticulum, and therefore cor- 
responds to His’s recessus opticus (preopticus—Burck- 
hardt). According to His, the recessus opticus corresponds 
to the anterior extremity of the limiting furrow (Grenz- 
furche) between his alar (Fligelplatte) and basal (Grund- 
platte) laminze, to which Reichert has given the name “ sulcus 
Monroi,” without, however, recognising its important mor- 
phological significance. 

Immediately in front of the recessus opticus the median 
wall takes a sudden bend into the vertical direction, and at the 
same time becomes enormously thickened to form a large mass 
rhomboidal in section (fig. 2, 2. inf.), which His calls the 
“lamina terminalis” and Burckhardt the “lamina 
infra-neuroporica.” According to His it is formed by the 
meeting and fusion in the middle line of his Fliigelplatten, 
but according to v. Kupffer and Burckhardt it forms part of 
the floor (Bodenplatte). The researches of His seem to 
indicate clearly enough that it corresponds to part of the frontal 
suture (Schlussnaht), but more convincing evidence is required 
before it can be granted that it is formed purely from the 
Fligel-, aud not also from the Grund-, platte. This question 
will be again referred to later on. In the dorsal part of the 
lamina infra-neuroporica a large rounded mass of fibres will be 
seen (a. c.) near the posterior surface. This is the anterior 
commissure. In all Vertebrates this phylogenetically very 
ancient and primitive connecting link between the two hemi- 
spheres crosses the middle line in the same position in the 
lamina infra-neuroporica, so that it forms an easily recognisable 
basis for comparing the perplexing regions which surround it, 
Immediately in front and above it there is a small scattered 


THE BRAIN OF A F@TAL ORNITHORUYNCHUS. 187 


bundle, consisting as yet of very few fibres, which also cross 
the middle line in the same lamina. This is the first rudiment 
of the fornix-commissure (fig. 3, f. c.). It will be noticed 
that it lies below the foramen of Monro and in the floor of 
a diverticulum of the third ventricle (figs. 2 and 3), which is 
often known as the ventriculus communis (Osborn). In 
the adult Ornithorhynchus brain the fornix commissure 
occupies a very different position, i.e. in front and above the 
foramen of Monro, and entirely dorsal to the third ventricle. 
This altered position of the commissure in adult Mammalia 
and many Sauropsida probably depends mainly upon the growth 
of the lamina terminalis, especially that part lying between 
the anterior and fornix commissures; but partly also on the 
backward growth of the hemisphere, which is accompanied by 
a corresponding growth of its dorsal commissure. In many 
Amphibia, where the lamina terminalis does not grow to the 
same extent as it does in the higher animals, the fornix 
commissure maintaius into adult life a position exactly cor- 
responding to that met with in the foetal Platypus. In the 
arrangement of their commissures the Sauropsida are inter- 
mediate between the Amphibia and Prototheria.! 

Above the lamina terminalis the median wall suddenly be- 
comes thin again and forms a thin plate, which appears to 
spring from the anterior edge of the lamina terminalis. This 
thin wall (fig, 3, 7. ”.), which is distinguished by von Kupffer 
as the lobus olfactorius impar, bounds a little diverti- 
culum (fig. 8*), which Burckhardt calls the recessus neuro- 
poricus. The same region is distinguished by His as the 
angulus terminalis, and is regarded as the dorsal limit of 
the frontal suture line, and the last point to lose its connec- 
tion with the ectoderm. Immediately dorsal to it there is a 


1 Further and more detailed examinations of the cerebrum of the adult 
Ornithorhynchus and a number of reptiles since this paper was written 
have demonstrated a much closer resemblance between adult and foetus than 
this paper indicates, as well as the marked similarity of both to the higher 
Reptilian condition. A more detailed account of the region of the commis- 
sures will be published in a short time. 


188 G. ELLIOT SMITH. 


small transverse layer (/. sup.), which Burckhardt calls the 
lamina supra-neuroporica, and describes as ‘ ein kurzer 
ependymatoscr Abschnitt, . . . der erste Abschnitt der Scheitel- 
platte.” All authorities now seem to agree in regarding this 
as the anterior extremity of the roof (Deckplatte). This 
must be admitted if the angulus terminalis is really the dorsal 
extremity of the “ vordere Schlussnaht.” The lamina supra- 
neuroporica takes a sudden bend backwards (fig. 3) to form a 
horizontal band, which gives origin in many lowly Vertebrates 
to the plexus inferiores, and in higher animals to the 
plexus laterales as well, or exclusively. In the specimen 
under consideration, however, although the plexus laterales do 
not actually spring from this lamina, they are formed from the 
caudal prolongations of its lateral parts (vide fig. 15) on 
either side of the paraphysis (par.). Immediately dorsal to 
this lamina the anterior wall of the third ventricle is bulged 
out to form a large sac (fig. 2, par.), which constitutes the 
paraphysis of Selenka, of which a fuller description will be 
given below, 

The whole extent of the roof (actual) of the third ventricle 
as far back as the superior commissure (fig. 2, s.c.) is deeply 
invaginated to form a complex choroidal fold (fig. 2, ch. 3; also 
figs. 9—15), which later on forms the “ diaplexus” of Wilder. 
It is evident, therefore, that no Zirbelpolster of Edinger (par- 
encephalon, prepinealis Zwischenhirndach) exists in this brain, 
since the velum or origin of the median plexus extends right 
back to the superior commissure. The latter is a very small 
band of transverse fibres connecting the two ganglia habenule. 
Immediately behind the superior commissure is the epiphysis 
cerebri, which consists of a small solid clump of cells (fig. 2, 
ep.) with deeply stained nuclei, lying behind a small recessus 
pinealis and upon the posterior commissure (p.c¢.). The latter 
is a broad band of transverse fibres, and is new generally re- 
cognised as the limit of the diencephalon. 

The constricted anterior pole of each hemisphere is distin- 
guished as an olfactory bulb by the application of the olfac- 
tory ganglion to its periphery (figs. 18 and 4, olf. g., J.g.). 


THE BRAIN OF A F@TAL ORNITHORHYNCHUS. 189 


On tracing the cerebrum caudally in a series of sections, the 
basal aspect of the olfactory bulb will be found to be con- 
tinuous with two regions which already present distinct histo- 
logical features, although there is little superficial indication 
of any such division. These two regions are the tuberculum 
olfactorium (figs. 6 and 7, ¢.0.)—a relatively iarge basal 
area next to the middle line—and the pyriform lobe, whose 
surface is covered by the already well-developed ‘ external 
olfactory radiation” (e.0.7r.) of Edinger. The tuberculum 
olfactorium only extends back as far as the lamina terminalis, 
but the pyriform lobe extends the whole length of the hemi- 
sphere (figs. 6—13, pyr. and e.o.r.). In the adult the pyri- 
form lobe is limited laterally by a well-defined fissure (fissura 
rhinalis or ectorhinalis of Turner)—figs. 8 and 14, f.r.,— 
and is separated from the tuberculum olfactorium by a deep 
fissure (fig. 8, f.er.)—fissura endorhinalis of Turner. In 
the foetus there is no sign whatever of the fissura ectorhinalis, 
although a shallow endorhinal groove (fig. 10, f. er.) is already 
to be made out. ‘This appears to correspond to what His has 
wrongly called the rhinal fissure. If this is the case, he has 
excluded from his ‘ lobus olfactorius” the pyriform lobe, 
which does not become superficially distinguished from the 
pallium (Turner) until late in ontogeny. In Notoryctes, 
even in the adult brain, no rhinal fissure appears to be present, 
and in Perameles nasuta it is very shallow and imperfect. 
The rhinal fissure appears late, not only in ontogeny, but also 
in phylogeny. It appears, therefore—and in this I am sup- 
ported by His’s figures—that the so-called rhinal fissure of the 
early human foetus cannot be the ectorhinal—as His believed 
—but the endorhinal fissure. 

The tuberculum olfactorium is not confined to the basal 
aspect of the cerebrum, but crosses the ventro-mesial angle to 
become continuous with the ‘‘precommissural area of the 
mesial cortex ” (figs. 6 and 7, p.a.). In the foetal brain there 
is no line of demarcation between these two regions, since they 
present exactly similar histological features, but in the adult 
(fig. 8 *) there is a clearly defined boundary line. The “ pre- 


190 G. ELLIOT SMITH, 


commissural area” extends from the mesial aspect of -the 
olfactory bulb in front to the lamina terminalis behind (vide 
figs. 2 and 4, p. a.). The tuberculum olfactorium (locus per- 
foratus anticus) and the precommissural area (gyrus subcallosus 
of Zuckerkandl) correspond to the “ posterior olfactory lobule” 
of His. The whole of the mesial wall of the cerebrum 
dorsalto the precommissural area forms part of the 
rudimentary hippocampus (figs. 2, 6, and 7, hip.). It will 
be seen, therefore, that the hippocampus extends as far forwards 
as the olfactory bulb (the characteristic histological structure 
may be recognised at a distance of 150 u behind the olfactory 
ganglion). The part of the region in immediate relation to 
the precommissural area is the rudimentary fascia dentata, 
the dorsal part being the “ Anlage” of the cornu Ammonis. 
As it is traced backwards the hippocampus is found to pass to 
the dorsal side of the foramen of Monro (fig. 15) and choroid 
fissure, the ependymal choroid fold being attached to the 
ventral margin of the ‘‘ Anlage” of the fascia dentata. The 
hippocampal rudiment is co-extensive with the fissura choroidea, 
and forms its dorsal boundary (figs. 9 to 13). In the greater 
part of its extent this hippocampal “ Anlage ” is bulged into 
the lateral ventricle, giving rise to a shallow trough (the 
“ Bogenfurche” of Schmidt, fissura arcuata). To this 
trough the name “hippocampal furrow” (“ Ammons- 
furche ” of Mihalkovics) may very appropriately be applied. 
It has an exact morphological significance, being strictly con- 
fined to the region which is to become hippocampus. It is 
clear, therefore, from the relations of the anterior extremity 
of the hippocampus, that the true Bogenfurche cannot form 
the ‘ fissura prima” of His. 

The position occupied by the hippocampus in this foetus— 
entirely dorsal to the foramen of Monro and fissura choroidea 
—is typical of its true and essential relations in the whole 
Vertebrate series. The recognition of the homologous region 
in the sub-mammalian brain constitutes the key to the whole 
question of the nature of the commissures, because all the 
commissural fibres arising in this area must belong to the 


THE BRAIN OF A F@TAL ORNITHORHYNCHUS. 191 


hippocampal or fornix commissure. In the adult Platypus 
there is no difficulty whatever in recognising the hippocampus, 
since its histological features, whether studied by aniline blue- 
black staining or by the methods of Weigert or Golgi, exactly 
resemble those of the higher mammal (figs. 8 and 14). Upon 
this fact largely depends the extreme morphological import- 
ance of the cerebrum of Ornithorhynchus. For while the 
topographical relationship of the Prototherian hippocampus 
resembles that of the Sauropsida, it at the same time presents 
quite as high a histological differentiation as is found in the 
Eutheria—a fact which renders its identification quite certain 
and easy. By comparison it is easy to identify the correspond- 
ing region in the foetal brain, where the simplicity of its 
histological structure (fig. 15) and the absence of the com- 
plicated foldings and inrolling of the adult structure vividly 
recall the Sauropsidan hippocampus, and at a glance convinces 
one of the homology. In all adult Mammalia the pallium 
takes some share in the formation of the mesial cortex (figs. 8 
and 14, p.), but in the foetus the hippocampus extends right 
up to the supero-mesial angle, so as to exclude the pallium from 
any share in the formation of the mesial wall. Similarly in 
the reptile the hippocampus forms the greater part of the 
mesial wall, and extends forwards above the commissures tc- 
wards the anterior pole of the cerebrum, as Edinger and 
Meyer have described. Not only does it form the mesial wall, 
but it extends on to the dorsum of the cerebrum. Herrick, 
by attempting to draw too close a comparison with the 
Eutherian brain (where this region is greatly disturbed by the 
development of a corpus callosum), would limit the hippo- 
campus to the posterior extremity of the hemisphere. This 
is a gratuitous assumption, utterly unsupported by any evidence. 

The whole of the dorsal and lateral aspects of the surface 
of the hemisphere, from the hippocampus above to the pyriform 
below, constitutes the pallium (Turner). Both in its onto- 
genetic and phylogenetic history it continues to increase in 
extent after the more ancestral hippocampal and pyriform 
regions have reached their full development, both in size and 


192 G. ELLIOT SMITH. 


differentiation of structure. Thus on the one hand it en- 
croaches upon the mesial aspect of the cerebrum, and has a 
marked influence in the production of the complicated in- 
rolling of the hippocampus; on the other hand, it tends to 
overlap the pyriform lobe, producing the rhinal (ectorhinal) 
fissure (fig. 14, fr.). 

The corpus striatum (fig. 10, ¢. s.) is a large mass, which is 
closely related ventrally to the tuberculum olfactorium and 
pyriform lobes. Below the foramen of Monro it is closely 
connected (fig. 10) with the lamina terminalis, and behind the 
foramen it is connected with the “ Anlage” of the optic 
thalamus, which lies internal and entirely dorsal toit. The 
internal capsule is a very insignificant tract, consisting of a 
few scattered fibres (fig. 11, ¢.2.), which are only to be found 
in one region. Even in the adult (vide fig. 8) it forms a very 
scattered and relatively small system, which is indicative of 
the intermediate position of Platypus between the Sauropsida, 
which have no proper internal capsule (Meyer), and the mar- 
supials, where it is a well-formed fibre tract. 

Bebind the foramen of Monro the pyriform lobe (figs. 13 
and 14, pyr.) is directly continuous with the corpus striatum 
(c. s.), the connecting bridge being the rudiment of the nucleus 
amygdalz (nx. a.), which in the foetus gives rise to an enormous 
tenia semicircularis (¢.s.), or stria terminalis, 

The foramen of Monro forms a rounded aperture situated 
upon the lateral wall of the anterior extremity (Endhirn of 
His) of the median cavity of the fore-brain, about midway 
between the roof and the floor (fig. 2, f.m.). To be more 
exact, it opens into the lateral ventricle from a small recess 
of the third ventricle’ above the lamina infra-neuroporica, 
which is bounded anteriorly by the angulus terminalis (fig. 
3, 7. n.) and lamina supra-neuroporica (/. swp.), and dorsally by 
the horizontal extension backwards of that lamina. Into the 
dorsal aspect of this recess (fig. 15, a.) and in front of the 
foramen of Monro, the paraphysis opens (par.). The foramen 
of Monro is bounded dorsally in the whole of its extent by a 


99 ¢¢ 53 ¢¢ 


1 © Ventriculus communis, ventricle of telencephalon, aula.” 


THE BRAIN OF A FQITAL ORNITHORHYNCHUS. 193 


large mass of neuroblasts (fig. LO, ¢h.), the rudimentary optic 
thalamus. As this mass is traced caudally (figs. 11, 12, and 
13) it gradually increases in size, forming a large bulging of 
the Fltiigelplatte in the posterior part of the diencephalon. 
Where it forms the dorsal boundary of the foramen of Monro 
(fig. 11) this thalamic rudiment consists of three distinct 
parts—a mesial ependymal layer, an intermediate neuroblastic 
mass, and a lateral nerve-fibre layer. ‘The connections of the 
latter cannot be accurately traced, but it appears to be the 
tenia thalami. 

As this thalamic rudiment is traced forwards above the 
foramen of Monro the neuroblastic mass gradually diminishes 
in size, until at the extreme anterior limit of the foramen of 
Monro (fig. 15) it disappears altogether, only the ependymal 
and nerve-fibre layers being left. The nerve-fibre layer also 
soon disappears, leaving the ependymal layer only. 

The point at which the thalamic rudiment ceases corre- 
sponds exactly to the anterior limit of the foramen of Monro, 
so that the latter is separated from the roof (Deckplatte) by 
the whole width of the rudimentary optic thalamus. 

The foramen of Monro is further bounded above by the 
epithelial fold which forms the rudiment of the choroid plexus 
of the lateral ventricle, and which is attached to the ventral 
border of the fascia dentata (fig. 15, f. d.). 

Below, the foramen is bounded by the lamina infra-neuro- 
porica (figs. 2 and 10), which here becomes continuous with 
the corpus striatum (vide fig. 10). Posteriorly the foramen is 
bounded by the junction of the thalamic rudiment with the 
corpus striatum (fig. 11). Anteriorly the boundary is formed 
(vide the right side of fig. 15) by the fusion of the lamina 
terminalis with the ventral border of the fascia dentata. If, 
therefore, we leave out of consideration the choroid plexus of 
the lateral ventricle, it is evident that the foramen of Monro 
has no immediate relation to the Deckplatte in this brain. 
Before discussing the relation of the hemisphere to the Deck- 
platte the paraphysis must be described. 

VOL. 89, PART 2.—NEW SERIES. N 


194 G. ELLIOT SMITH. 


The Paraphysis. 


In describing the structures met with in a median sagittal 
section it was mentioned that the dorsal part of the (actual) 
anterior wall of the median cavity of the fore-brain was bulged 
out to form a large sac. The corresponding structure is well 
seen in the early embryo of Perameles (fig. 16, par.). Ven- 
trally it opens into the cavity of the “telencephalon” of His 
by a narrow neck. Posteriorly its lumen becomes suddenly 
restricted by two large bulgings from the lateral walls into 
the cavity. These bulgings are the anterior extremities of the 
optic thalami (fig. 17). On comparing the appearance of the 
Perameles brain with that of Ornithorhynchus (fig. 15) 
the general resemblance is at once apparent, the only essential 
difference being the presence of a large complicated median 
fold of the roof of the paraphysis in Platypus. ‘This is the 
anterior extremity of the ‘‘ Anlage” of the plexus choroideus 
medius—(ch. 3), the ‘‘diaplexus”’ of Wilder. In the early 
foetal brain of Perameles the median choroidal fold of the 
third ventricle is not yet developed, although that of the 
lateral ventricle (fig. 16) is well formed, so that the velum 
and roof of the paraphysis form a simple transverse band. In 
the Platypus embryo, however (fig. 2), a well-developed cho- 
roidal fold extends from the superior commissure to the 
lamina from which the lateral plexus arises, completely inva- 
ginating the paraphysis (figs. 7, 9, and 15) in the middle line. 
In Platypus the transition from optic thalamus to paraphysis 
is a very gradual one, so that in examining a series of coronal 
sections the lateral walis of the diverticulum would seem to be 
merely the forward continuation of the ependymal layer of the 
Fliigelplatten (fig. 15). This structure has attracted a con- 
siderable amount of attention since it was first noticed, and 
numerous names have been applied to it, and hypotheses 


1 «Vir die vordere Epiphyse, dem Adergeflechtknoten Goettes 
den Burckhardt in friiheren Arbeiten zuerst Plexus choroideus su- 
perior, spater Conarium genannt hatte, braucht er neuerdings mit Kupffer 
die von Selenka herrithrende Bezeichnung Parapliysis.’’—FRonizp, 


THE BRAIN OF A’ F@TAL ORNITHORHYNCHUS. 195 


advanced to explain it. As yet, however, its significance is 
veiled in mystery. In this brain it is enclosed in a sheath of 
dura mater formed by the splitting of the falx cerebri (fig. 7, 
fe.), and it is closely surrounded by a number of very large 
vessels. From the close relation of the organ to these large 
vessels, it has been suggested that it may have some nutritive 
function before the “ diaplexus” is developed (Gage). 


The Relation of the Roof (Deckplatte) to the 
Third Ventricle and Foramen of Monro. 


In describing the foramen of Monro, it has already been 
seen that the Deckplatte (“‘Scheitelplatte” of Burckhardt) is 
separated from its dorsal aspect by the whole width of the 
thalamic rudiment. The foramen opens into the small dorsal 
recess (fig. 15, d.) of the third ventricle which is bounded an- 
teriorly by the angulus terminalis and the lamina supra-neuro- 
porica. The latter structure is considered to form part of the 
Deckplatte, and the continuation backwards of its horizontal 
part in this specimen (fig. 3) gives rise to the choroid plexus 
of the lateral ventricle (fig. 15). In fact, it is generally con- 
sidered that the lateral plexus is a derivative of the Deckplatte 
(Minot). IRf this be so, it must be admitted that the foramen 
of Monro and the hemisphere are in immediate relation to the 
Deckplatte. Ifthis be so, the different parts of the morpho- 
logical roof will be seen to be widely separated in the region 
of the foramen of Monro (fig. 10). Above there is the un- 
doubted roof of the ‘‘aula;” whereas laterally, between the 
thalamic rudiment, and the hippocampal rudiment, is the 
lateral plexus whose epithelial covering is supposed to bea 
derivative of the Deckplatte (Minot). If, however, the hemi- 
sphere, as is generally supposed, extends up to the Deckplatte, 
how does the optic thalamus come to occupy the position 
which is roughly represented in fig. 10, ¢h.? Kven admitting 
that the choroid plexus of the hemisphere is a derivative of the 
roof which has extended backwards, invading, as it were, the 
mesial hemisphere wall, ouly two alternatives appear to remain 


196 G. ELLIOT SMITH. 


to explain the position of the thalamus in this specimen. 
Hither the hemisphere brain develops from the lateral aspect 
of the brain-tube on the ventral side of the anterior extremity 
of the thalamus, or the latter in its growth invades the Deck- 
platte. All the appearances, both in this brain and in those 
of all the Perameles and Macropus specimens, are against 
the latter hypothesis. It would appear that the hemisphere 
develops on the ventral side of the anterior extremity of the 
thalamus.) The lateral walls of the paraphysis in this speci- 
men look deceptively like anterior extensions of the Fliigel- 
platten. 

Taken as a whole, however, the general disposition of parts 
in the brain under consideration can be harmonised with the 
account of His. On tracing the pars subthalamica (Forel) or 
hypothalamus (Waldeyer) forwards in the series of sections, it 
will be found to become continuous with a large mass lying on 
the mesial side of the corpus striatum and on the ventral side 
of the optic thalamus. This mass (fig. 11, *), which forms 
the lateral wall of the ventral part of the third ventricle, on 
being traced forwards will be found (fig. 10) to become con- 
tinuous with the lamina infra-neuroporica. What is this mass? 
According to the hypotheses of Burckhardt and v. Kupffer it 
must be considered simply as the anterior extremity of the 
Grundplatte ; whereas, according to His, it must be the anterior 
end of the Fitigelplatte which has become bent towards the 
base of the brain. In fig. 11 the fibres of the external capsule 
(c. e.) will be seen extending into the mass to cross the middle 
line (fig. 10) in the lamina infra-neuroporica, so that the mass 
in question is divided into two parts—a ventral, which in the 
adult is easily recognised as the tuberculum olfactorium ; and 
a dorsal, which in the adult is corpus striatum. In the adult 
the anterior extremity of the optic thalamus fuses with the 
lamina infra-neuroporica (‘septum pellucidum”) above the 
anterior commissure, so that the combined mass appears to 
rest upon the anterior commissure. It is almost impossible 


1 Without earlier developmental stages no opinion can be expressed con- 
cerning these questions (25th April, 1896). 


THE BRAIN OF A FQ@ITAL ORNITHORHYNCHUS. 197 


(fig. 8) to draw any line of demarcation between corpus 
striatum and tuberculum olfactorium. This being the case, 
this region (fig. 11, *) may doubtfully be homologised with 
the corpus striatum, as His has done. It will be noted from 
the figures that the transition from undoubted hypothalamus 
to this doubtful region takes place at the optic chiasma—the 
termination of His’s ‘‘ Grenzfurche.” If, therefore, it is granted 
that this doubtful area (*) is corpus striatum, it follows that 
the hemisphere brain extends down as far as the optic recess, 
as His has described in the case of the human feetus. 

Having now given a general account of the disposition of 
parts, there are certain regions which need further consideration. 


The Olfactory Bulb. 


This part of the cerebrum, whose structure is roughly re- 
presented in transverse section in fig. 18, consists of four 
distinct layers, as Vignal! has pointed out. Immediately sur- 
rounding the ventricle there is a thick mass of small cells with 
deeply stained nuclei. This is the ependymal layer (e). 
Surrounding it there is a stratum of about the same thickness, 
but much lighter in appearance—the layer of nerve-cells 
(6). Examined with the high power, this is found to consist 
of amass of well-formed neuroblasts (fig. 19). Posteriorly this 
layer is found to be continuous with the essentially nervous part 
of the cerebral cortex. It is probable that in the adult the 
neuroblasts which compose this layer mostly develop fibres of 
the pars olfactoria of the anterior commissure. The most 
superficial layer is imperfect, consisting of two distinct gan- 
glionic masses, each composed of neuroblasts like the second 
layer. The larger mass (olf. g.) covers the whole inner surface 
of the bulb, and extends on to the ventral and slightly on to 
the dorsal aspects of the bulb. In this region the true olfactory 
nerves terminate. Hence I have distinguished it as the 
“olfactory” ganglion. On the outer aspect of the bulb there 
is a second smaller mass (J. g.), presenting’ exactly similar 
histological features. In this mass the fibres derived from 

1 Quoted in Minot’s ‘Human Embryology.’ 


198 G. ELLIOT SMITH. 


Jacobson’s organ terminate. Hence I have distinguished it as 
the “ganglion of Jacobson’s organ.” The further 
significance of this arrangement I have discussed in the 
‘ Anatomischer Anzeiger.’ 

Separating both ganglionic masses from the second or nerve- 
cell layer there is a distinct layer of nerve-fibres (y), which 
form the root bundles of the “ olfactory [internal and external | 
radiations” of Edinger. It corresponds to the molecular layer 
or “ Randschleier” of the cortex cerebri. 

The surface regions of the hemisphere are already differen- 
tiated into two histologically distinct districts, (1) a basal 
region (comprising the tuberculum olfactorium and the pre- 
commissural area), which presents an amorphous and un- 
differentiated structure; and (ii) a dorso-lateral region—the 
pallium,—which already exhibits characteristically specialised 
features. Interposed between these regions on either side 
there are two districts—the hippocampus and pyriform lobe— 
which present intermediate or transitional features. These 
divisions have a further and much wider significance. The 
undifferentiated basal regions—including not only the pre- 
commissural area and tuberculum olfactorium, but also the 
prosencephalic part of the olfactory bulb, the corpus striatum 
and lamina terminalis (“septum lucidum ’”’)—are the ancient 
parts of the brain phylogenetically, and develop early in 
ontogeny. The transition regions appear next in phylogeny— 
the hippocampus and then the pyriform lobe. The highly 
differentiated pallium appears last, and continues to increase 
in size and in histological specialisation long after the other 
regions have reached the acme of their evolution, or are even 
in their decline. Thus in the Prototheria the hippocampus 
and pyriform have already reached the height of their develop- 
ment, whereas the pallium progressively increases in size and 
complexity in the Hutherian series up to Homo, in whom the 
lowest stratum—the olfactory—the tuberculum olfactorium 
and precommissural area are disappearing. 


THE BRAIN OF A FQTAL ORNITHORHYNCHUS. 199 


The Pallium. 


Five distinct layers are recognisable in the pallium (fig. 20). 
These are arranged in the following order—proceeding from 
the surface, and comparing with the layers of the olfactory 
bulb. 


Pallium (fig. 20). Olfactory bulb (fig. 18). 


suence 1. The olfactory ganglion 

(an added layer). 
erlbesmolecule player 7) x carunsen -sntiosede: 2. The nerve-fibre layer (y). 
2, The superficial nerve-cell or “mantle” 

layer (e) 

3. The middle nerve-fibre layer (@) 
4. The deep nerve-cell layer (c) 
Depth evependymal layer {(@) cues ade acacieeent' ee ependymal layer (e). 


pan 


...8. The layer of nerve-cells (6). 


The superficial or molecular layer of the pallium forms 
a clear peripheral band containing only scattered cells and 
nerve-fibres, It is the rudiment of the molecular layer of the 
adult cortex, and also appears to correspond to the ‘ Rand- 
schleir.” In the pyriform lobe and hippocampus (fig. 10) it 
becomes greatly thickened. In the basal regions it does not 
exist as a distinct layer. 

The superficial nerve-cell layer (e)—the mantle layer 
—consists of a very regular stratum of densely packed neuro- 
blasts, from which the most important constituents of the 
cortex develop. 

The middle nerve-fibre layer (d) contains a large 
number of nerve-fibres and scattered neuroblasts. It appears 
to be the ‘ Anlage” of the corona radiata, although in the 
region of the corpus striatum it is separated from the corona 
proper by a large mass of cells. 

The deep nerve-cell layer (c)—the Rolando-cell layer of 
Ludwig Lowe—is more or less intimately joined to the epen- 
dymal layer (a), but a slight interval (b) where the cells are 
not so closely packed indicates the line of separation. The 
destiny of this layer is uncertain, 


200 G. ELLIOT SMITH. 


“The Precommissural Area of the Median Cortex.” 


The general relations of this region have already been 
described. In a scheme (fig. 5) which indicates the hypo- 
thetical arrangement of parts in the anterior wall of the 
primitive brain-tube it will be seen that the olfactory bulbs 
(olf. g), situated at each lateral angle, are separated by a 
broad plate of nervous tissue. With the development of 
the hemispheres this plate becomes bent forwards on either 
side of the middle line. The lateral part, which now forms a 
sagittally placed wall (p.a.), is the “ precommissural area,” 
and the median band connecting the two “ aree ” is the lamina 
terminalis (/.¢.), which later develops into the septum lucidum. 
His calls the “ precommissural area” part of the “ olfactory 
lobe.” In fact, as the homologue of Zuckerkandl’s “ gyrus 
subcallosus,” it must form part of His’s “ posterior olfactory 
lobule.” On morphological and histological grounds, however, 
the precommissural area cannot be separated from the rudi- 
ment of the “septum lucidum” (compare figs. 7 and 9), both 
having equal claims to be considered as parts of the “ olfactory 
lobe ” or rhinencephalon. 


The Hippocampus. 


As the pallium (fig. 15, p.) is traced round the supero-mesial 
angle of the hemisphere the cellular elements will be found to 
undergo a striking rearrangement. The molecular or super- 
ficial layer (fig. 15) will be found to become very much broader 
than it is in the pallium, and, by comparison with later stages 
of Perameles, is readily recognised as the homologue not 
only of the molecular layer, but also of the stratum lacunosum 
and stratum radiatum of the adult hippocampus. The cells 
of the superficial nerve-cell layer lose their densely packed 
arrangement, and the cells of the second, third, and fourth 
layers become directly continuous with a column of somewhat 
scattered cells, which produce the pyramidal and polymorphous 
cells of the hippocampus. As this cell column approaches the 
margin of the hemisphere the cells become more densely packed 


THE BRAIN OF. A F@TAL ORNITHORHYNCHUS. 201 


again to form the rudimentary fascia dentata (fig. 15, f. d.). 
Since, in the hippocampus, the cells of the fourth layer of the 
pallium separate from the ependymal layer, the latter appears 
as a much more sharply defined layer (fig. 15). Between the 
ependymal layer and the irregular column of neuroblasts there 
is a relatively wide nerve-fibre layer—the alveus. This is the 
homologue of and is continuous with the fifth layer of the 
pallium. 

This simplicity of hippocampal formation presents a close 
resemblance to the corresponding region in the reptile brain. 
The anterior extremity of the hippocampus in all non- placental 
mammals maintains this primitive simplicity of structure into 
adult life (fig. 8). The much-disputed questions of homology 
will be discussed later on in dealing with the histology of the 
adult organ. 

The fornix system is as yet imperfectly formed, and presents 
a reptilian-like simplicity of arrangement. Although the de- 
scending fibres of the fornix (figs. 7,9, 15, d.f.) are already 
well formed, there are only a few scattered commissural fibres 
in the lamina infra-neuroporica. 


The Pyritorm Lobe: 


As the pallium is traced into the pyriform lobe (figs. 6, 7, 
9—12) the molecular layer suddenly increases in thickness. 
This layer contains the fibres of the external olfactory radia- 
tion (e. 0. 7.).. The superficial nerve-cell layer suddenly 
becomes more scattered, and loses the regular appearance 
_ found in the pallium proper. This layer corresponds to the 
layer of “double pyramids” (Kolliker) found in the adult 
pyriform lobe (fig. 8). The other layers of the pallium are 
represented in the pyriform lobe merely by diffusely scattered 
nerve-cells. Behind the lamina terminalis the pyriform lobe 
is continuous mesially with a large mass of diffusely scattered 
nerve-cells—the nucleus amy gdale. 


The Corpus Striatum and Optic Thalamus. 


The corpus striatum (c.s.) forms a large mass projecting 
into the lateral aspect of the lateral ventricle, and lying in the 


202 G. ELLIOT SMITH. 


concavity of the external capsule (c.e.).Anteriorly it is 
directly continued into the tuberculum olfactorium (figs. 7 
and 8), which Ganser calls the “ cortex of the head of the 
corpus striatum.” Posteriorly it is closely connected with 
the pyriform lobe (figs. 12 


14) through the intermediation 
of the nucleus amygdale. Except in the one region imme- 
diately behind the foramen of Monro (fig. 11), there is no sign . 
of any division into caudate and lenticular nuclei by an in- 
ternal capsule. 

Histologically the rudiment of the corpus striatum presents 
numerous well-formed neuroblasts, which give rise to nerve- 
fibres, whose connections will be described immediately. 

The optic thalamus presents in all sections a large mass of 
neuroblasts, which varies in size and appearance in different 
sections (figs. 1O—13). The only other structure yet differ- 
entiated in the thalamic region is the ganglion habenule 
(fig. 13, G23). 

The great thalamic nucleus, which appears to cor- 
respond to the “ centre mediane de Luys ” (Lowe and Minot), 
is relatively very largely developed in this foetus, and, as may 
be seen from fig. 14, this precocious development is an indica- 
tion of the huge proportions of the adult optic thalamus, 
whose enormous growth in the lateral direction reduces 
the corpus striatum in this region to a very thin lamella 
(fiz. 14, ¢. 8.). 

Connecting the great thalamic nucleus and the corpus 
striatum there is a huge bundle of fibres (fig. 12) which stands 
out extremely clearly, since there is practically no internal 
capsule to obscure it. By means of this fibre tract, which 
apparently corresponds to Edinger’s ‘‘radiatio strio-thala- 
mica,”’! a very intimate connection is established between the 
corpus striatum and cortex cerebri (pallium) on the one hand, 
and the ‘great thalamic nucleus”? on the other. A large 
proportion of these fibres appear to arise from neuroblasts 


in the corpus striatum. Others appear to spring from the 
thalamus. 


1 Only in part. 


THE BRAIN OF.A F@TAL ORNITHORHYNCHUS. 203 


The following anatomical facts concerning the brain of 
Platypus appear to be associated with one another :—an enor- 
mous trigeminal nerve; a huge development of substantia gela- 
tinosa Rolandi in pons and medulla, and fillet in the mesence- 
phalon; a large and precocious development of the optic 
thalamus (where in the adult the great bulk of the fillet ter- 
minates), and an exceptionally abundant supply of fibres con- 
necting the latter with the corpus striatum and cortex cerebri. 
The important bearing of this upon the vexed question of the 
termination of the fillet (lemniscus) will be discussed later on 
when dealing with the adult. It may bestated here, however, 
that an examination of the Ornithorhynchus brain clearly 
shows that the great majority of the fillet fibres terminate by 
terminal arborisations around the large multipolar cells of the 
optic thalamus, which, in turn, are connected with the cortex 
cerebri and corpus striatum. 

The huge stria terminalis, which arises from the neuroblasts 
in the nucleus amygdalz, extends vertically upwards behind 
the “‘radiatio strio-thalamica,” and then takes a sudden bend 
into the horizontal plane (fig. 12, ¢.s.). The picric stain does 
not permit one to accurately trace its anterior connections. 

The account of the optic nerve and its central connections 

will be deferred until the rest of the brain is under considera- 

tion. Before then I hope to examine a series of foetal Echidne, 
and to be able to supplement these fragmentary and imperfect 
notes. 


Literature. 


The last few years have yielded an unusually rich and abun- 
dant contribution to the knowledge of the difficult subject of 
cerebral morphology; and in spite of the different interpreta- 
tions of the vast mass of data collected, there is, beneath all 
the conflicting statements, a remarkable degree of uniformity 
in the essential facts of cerebral development and phylogeny. 

No purpose would be served in a descriptive paper such as 
this in quoting all this bulky literature. In a paper entitled 
“ Die Homologien des Zwischenhirndaches und ihre Bedeutung 


204 G. ELLIOT SMITH, 


fir die Morphologie des Hirns bei niederen Vertebraten,” 


which appeared in the ‘ Anatomischer Anzeiger,’ Bd. ix, Nos. 
5 and 6, p. 152, Rud. Burckhardt began a discussion, which 
was vigorously carried on in the same journal by Studnicka 
and Rabl-Riickhard. Before the Anat. Gesellschaft W. His 
discussed the same subject. His paper was subsequently 
published in the ‘Arch. f, Anat. u. Entw. ; 1898, p. 157, “ Uber 
des frontale Ende des Gehirnrohres.” In the same number 
His published a paper on nomenclature, ‘ Vorschlage zur 
Eintheilung des Gehirns.” 

C. von Kupffer’s contribution to the argument is found in 
his ‘Studien zur vergl. Entwickelungsgeschichte des Kopfes 
des Kranioten,’ and also in the report of the Anat. Gesellschaft. 

An admirable summary of Burckhardt’s views is found in 
his paper entitled “‘ Der Bauplan des Wirbelthiergehirns” in 
‘Morpholog. Arbeiten, hrsg. v. G. Schwalbe, Bd. iv, 1894, 
p. 131. 

Critical digests of the literature will be found in Froriep’s 
paper, “‘ Entwick. des Kopfes” (‘ Ergebnisse der Anatomie und 
Entwickelungsgeschichte,’ hrsg. v. F. Merkel und R. Bonnet, 
Bd. iii, 1893); also the papers of Sorenson and Herrick in 
recent numbers of the ‘ Journal of Comparative Neurology.’ 

To Dr. L. Edinger I am indebted for a copy of his paper 
entitled “Die Faserung aus dem Stammganglion Corpus 
Striatum ” (‘ Verhdl. Anat. Gesellschaft,’ May, 1894), contain- 
ing an account of the “ radiatio strio-thalamica.” 


EXPLANATION OF PLATE 11. 


Illustrating Dr. G. Elliot Smith’s paper on ‘ The Brain of a 
Foetal Ornithorhynchus.” 


Fic. 1.—Diagrammatic representation of the brain of foetal Ornitho- 
rhynchus, reconstructed from serial sections X 6. The Zwischenhirn (d/.), 
which is shaded, is represented as though the cerebrum were trausparent. 
The oval cerebral hemisphere is seen to slightly overlap the mesencephalon 
(mes.). 6.0. Bulbus olfactorii. J. Jacobson’s organ. 2.J. Nerves from 


THE BRAIN OF A F@TAL ORNITHORHYNCHUS. 205 


Jacobson’s organ. n.olf. Olfactory nerves. opt.x. Optic nerve. 2.3, 
Oculo-motor nerve. G.g. Gasserian ganglion. 

Fie. 2.—Scheme of median sagittal section of the fore-brain, the hemi- 
sphere indicated by dotted lines. 6.0. Bulbus olfactorit. p.a. The “ pre- 
commissural area of the median cortex.” fdp. The rudiment of the hippo- 
campus (including the fascia dentata). p.c. Posterior commissure.  e.p. 
Epiphysis cerebri (pineal). s.c. Superior commissure. vel, Velum. ch.3. 
 Diaplexus”’ of Wilder. par. “ Paraphysis” of Selenka. 2 J/, Foramen of 
Monro. a.c. Commissura anterior, J/. ézf. Lamina infra-neuroporica., 7. 0. 
Recessus opticus of His (preeopticus—Burckhardt). opé.z. Optic nerve. 7.2. 
Recessus infundibuli of His (r. postopticus—Burckhardt). Ayp. Hypophysis 
cerebri. pit. Pituitary gland. 

Fic. 3.—Part of Fig. 2 enlarged.—/. sup. Lamina supra-neuroporica of 
Burekhardt. 1.2. Recessus neuroporicus (angulus terminalis, lobus olfac- 
toriusimpar). fc. Fornix commissure. 

Fig. 4.—Scheme of horizontal section through a mammalian brain, to show 
the relation of the ‘‘ precommissural area” (p.a.) to the lamina terminalis 
(2.¢.) and olfactory bulb (4.0.). Olf.y. Olfactory ganglion. c.s. Corpus 
striatum. ¢4. Diencephalon (His). 

Fie. 5.—Scheme of hypothetical arrangement of the terminal plate of the 
brain-tube, to show that the precommissural area and lamina terminalis are 
parts of one structure. Compare with Fig. 4. 


Fies. 6, 7, 9, 10, 11, 12, and 13 are a series of coronal sections of the 
fetal Ornithorhynechus brain, each x 12. The hippocampal rudiment 
(Aip.) may be seen extending through all the sections, lying in front above 
the precommissural area (p.a@.), which in turn is continuous with the tuber- 
culum olfactorium (¢.0.) and pyriform lobe (pyr.), on whose surface is the 
“external olfactory radiation” (Hdinger), e.0.7. yp. Pallium. c.e. External 
capsule. 


lig. 6.—Counting the sections from before backwards (each section being 
50 p thick), this is No. 24. It passes a short distance behind the olfactory 
bulb. 


Tig. 7.—(No. 35.) Just in front of the lamina terminalis. The falx cerebri 
(fx.) splits above to enclose the paraphysis (par.) and a number of large 
vessels. d./. Descending fornix fibres (Reichbundel). 


Fic. 8.—Section through the corresponding region in the adult brain. 
Weigert stain x 3. yp. Pallium. 4.f. Fissura hippocampi. f. d. Fascia 
dentata. d.f. Descending fornix fibres. p.a@. Precommissural area, ¢. 0. 
Tuberculum olfactorium. * Line of demarcation between p.a. aud Z. e. 
Jer. Fissura endorhinalis (Turner). pyr. Pyriform lobe. e. 0.7. External 
olfactory radiation. f.7. Fissura ectorhinalis. ¢.2. Internal capsule in 
corpus striatum (c.s.). ¢.7. Corona radiata. alv. Alveus. 


206 G. ELLIOT SMITH. 


Fie. 9.—(No. 38). Passing just in front of the foramen of Monro on the 
right side, and through it on the left. References as above. 


Fie. 10.—(No. 47.) Through foramen of Monro (f 4). a.ec. Anterior 
commissure. ¢h. Optic thalamus. Bf The “hippocampal furrow” 
(Bogenfurche). 

Fie. 11.—(No. 51.) Just behindthe foramenof Monro. /.f. Longitudinal 
furrow. 7.0. Recessus opticus. pdf. Pituitary gland. G.g. Gasserian gan- 
glion. c.¢. Internal capsule. 

Fie. 12.—(No. 80.) y.4. Ganglion habenule. z.¢4. Nucleus thalami 
(centre mediane de Luys). 7.s¢. ‘ Radiatio strio-thalamica” of Edinger. 
¢.s. Stria terminalis (running horizontally). G@.g. Gasserian ganglion. Viz, 
Fibres of fifth nerve proceeding from G.g. to the pons. 

Fic. 13.—(No. 89.) z..a. Nucleus amygdala, from which the large stria 
terminalis (¢.s.) arises and passes vertically upwards to cross on the dorsal 
aspect of the radiatio strio-thalamica (7. s¢.). pos. The “ tuberculum quinti” 
of the pons Varolii. 

Fic. 14.—Section through the adult Ornithorhynechus brain corre- 
sponding to Fig, 12 in the foetus. x 2. References as above. m.c. Middle 
(soft) commissure. p. swb¢. Pars subthalamica (hypothalamus of Waldeyer). 

Vie. 15.—Part of Fig. 9 enlarged. B/ Hippocampal furrow (Bogenfurche). 
ch. 3. Median choroidal invagination (diaplexus). Ap. Hippocampus. /. d. 
Fascia dentata. p. Pallium. c.s. Corpus striatum. d./. Descending fornix 
fibres. 2. xf. Lamina infra-neuroporica. 

Fic. 16.—Coronal section through the corresponding region of an early 
foetus of Perameles nasuta. c#./. Choroid plexus of the lateral ventricle. 

Fig. 17.—Section a short distance behind Fig. 16. ¢4. Optic thalamus. 
J. M. Situation of foramen of Monro. Bf Situation of Bogeufurche. 

Fic. 18.—Z.s. Olfactory bulb of foetal Platypus. x 25. olf.g. Olfactory 
ganglion. J.g. Olfactory ganglion into which the nerves from Jacobson’s 
organ are inserted. j. Nerve-fibre layer. d. Layer of nerve-cells. ¢. Hpen- 
dymal layer surrounding the ventricle. 

Vie. 19.—Neuroblasts (Zeiss D) from d, (Fig. 18). 

Vic. 20.—Supra-ventricular cortex cerebri (pallium) of feetal Platypus. 
x 96. a. Kpendymal layer. e. Deep nerve-cell layer (Rolando-cell layer of 
Lowe). d. Middle fibre layer. ¢. Superficial nerve-cell layer (mantle layer). 
J. Superficial fibre layer (molecular layer). 

Fies. 6, 7, 9, 10, 11, 12, 18, 16, 17, and 18 were drawn by means of a 
camera lucida. Fig. 15 was drawn from a microphotograph, for which I am 
indebted to Professor Wilson. 


ON ARHYNCHUS HEMIGNATHI. 207 


On Arhynchus hemignathi, a new Genus of 
Acanthocephala, 


By 


Arthur E. Shipley, 
Fellow and Tutor of Christ’s College, Cambridge, and University Lecturer in 
the Advanced Morphology of the Invertebrata, 


With Plate 12. 


ANATOMY. 

In the summer of 1894 I received from Mr. Perkins, of Jesus 
College, Oxford, seven small parasites which he had noticed ad- 
hering lightly to the skin around the anus of a species of bird, 
Hemignathus procerus, which he collected in the island of 
Kauai, one of the Sandwich Island group. Each of these para- 
sites was divided into three regions,—a head, a collar, and a 
trunk ; and, in fact, they have an almost ludicrous resemblance 
to a young Balanoglossus with one or two gill-slits (figs. 1, 11, 
and 111). On investigating their anatomy it at once became 
evident that the animals belonged to the group Acantho- 
cephala, and, further, that they differed from the other 
members of the group in the absence of what is perhaps their 
most characteristic organ,—from which, indeed, they take their 
name—the hooked proboscis or introvert. ‘The absence of so 
characteristic a structure, and the fact that the parasites were 
found outside the body, i.e. not as an endoparasite, but as an 
ectoparasite, lightly attached to the skin, made me think that 
perhaps the hooked introvert had been left behind in the in- 
testine of the host, and that the body of the parasite had passed 
out of the alimentary canal of the bird. However, careful 
inspection failed to reveal any trace of a scar or mark where 
the introvert might have been broken off; and although in 


208 ARTHUR E. SHIPLEY. 


the absence of hooks and introvert sheath, &c., the anterior 
part of the body which I have called the head is as unlike the 
typical introvert as possible, still in its relation to the lemnisci 
and to the ligament it occupies the position of that organ, 
and until we can get further information I think the best 
plan is to regard this part of the body as equivalent to the 
eversible part of more normal forms. 

The second of the three regions into which the body is 
externally divided is shorter than the head and smaller in 
diameter; it may be termed the collar. The third or poste- 
rior region, which may be called the trunk, is the longest and 
the most slender of the three; behind it tapers to a point 
where the orifice of the genital duct is situated, and this 
end of the animal is always a little turned up (figs. 1, 0, 11, 
and vit). The exterior of the collar and trunk are smooth 
or lightly wrinkled, but the head is covered with a number of 
small depressions or pits which give it a very characteristic 
appearance, and which are well seen in sections. The head is 
attached to the collar by a narrow neck, which is surrounded 
and concealed by the edge of the collar. It is obvious in 
sections (figs. v and x11). All the specimens were somewhat 
shrivelled and apparently distorted. The largest measured 
3°5 mm. in length, the smallest 2°> mm.; had they been fully 
distended they would probably have been 1 to 1:5 mm. longer. 
The body-cavity of the head is continuous with that of the 
neck, and the latter opens freely into the cavity of the trunk 
(fig. 13). The first-named space is by far the largest. The 
lumen of the collar region is reduced by the great thickness of 
the walls of this part of the body, and both here and in the 
trunk much of the internal space is occupied by the lemnisci 
and the reproductive organs. 

The skin is one of the most characteristic features of the 
Acanthocephala, and as far as I know is only paralleled by 
that of the Nematodes, but it possesses certain features not 
found in the last named group. The whole body is covered by 
a thin cuticle which does not vary much in thickness in the 
different regions of the body, and which is invaginated a short 


ON ARHYNCHUS HEMIGNATHI. 209 


distance into the genital pore. Beneath this is the true epi- 
dermis, or subcuticle as itis called; this has in my specimens 
the usual structure met with in the group so well described by 
Hamann, and consists of a matrix of a fibrillar nature, the 
fibrils being as a rule arranged radially, in which are embedded 
a certain number of ameeboid nuclei (figs. vi and x). This tissue 
is much thicker in the region of the collar than elsewhere, and 
it is thicker in the trunk than in the head. It is pierced in all 
directions by a series of tubes or lacunze which have no definite 
lining, but which seem to be mere splits in the fibrillar matrix. 
The lacune—except in the head—have a general circular direc- 
tion which is very well marked in the trunk region where each 
runs into a lateral longitudinal split (fig. x). They contain a 
small amount of coagulum, the remnant of the fluid which 
circulates in them ; during life this fluid, in other species, holds 
in suspension fat and coloured oil globules. If these are 
present in my species they must have been dissolved out in 
the processes which precede embedding. The circular lacunze 
of the trunk not only communicate with one another by means 
of the two longitudinal lateral lacunee (fig. xiv), but they open 
into one another by numerous small branches which have an 
oblique or longitudinal direction. In the head the lacunz 
have a general longitudinal course; they are not, however, 
straight, but twist in and out between the pits on the surface ; 
they anastomose freely (fig. iv). Thus in a transverse section 
of the head the lacunze appear as round holes more or less uni- 
formly arranged in the skin, and the same effect is produced 
by a longitudinal section of the trunk. 

In the collar region the subcuticular tissue is much thick- 
ened, and the lacunar system forms a single more or less 
definite ring which gives off numerous branching anastomosing 
twigs (fig. v). 

Although the above account attempts to give the general 
course of the lacunz in the skin, it should be mentioned 
that there is considerable irregularity in the arrangement, 
and one is almost inclined to believe that the canals do not 
remain permanent, but that they sometimes close up and 

VoL. 39, PART 2.—NEW SER. O 


210 ARTHUR E. SHIPLEY. 


new ones appear. As they have no lining of any kind, such a 
closing would leave no trace. 

As Schneider,! Hamann,” and Kaiser® have shown in the 
species investigated by them, the lacunar system of the intro- 
vert is completely shut off from that of the neck—if it be 
present—and of the trunk by a fold inwards of the cuticle 
which cuts the subcuticular tissue in two. I have not been 
able to find any such cuticular ring in the species in question, 
but the state of preservation of my specimens does not allow 
me to say definitely that it does not exist. 

The lemnisci are two elongated sac-like prolongations of 
the subcuticular tissue which are attached anteriorly to the 
skin at the junction of the head and collar. They extend back- 
wards to the extreme posterior end of the body, and are slightly 
bent so that a longitudinal section may cut them in two or 
three places (fig. x111). Histologically they are composed of 
the same substance as the subcuticle in direct continuity with 
which they arise, and they are traversed by a similar system of 
canals. Physiologically they seem, as Hamann suggests, to 
act as reservoirs for the fluid of the canal system of the intro- 
vert; when the fluid they contain is forced into the spaces of 
the introvert the latter is everted. It is withdrawn again into 
the body by special muscles. In most species the canal system 
of the lemnisci opens into that of the introvert in front of the 
cuticular ring, and is thus completely independent of that of 
the trunk. If we assume that the head of my species corre- 
sponds with the introvert of other forms which have lost its 
introvert sheath, the lemnisci open into the same region of 
the skin as they do in other Acanthocephala. 

The nuclei of tae subcuticle and of the lemnisci are very 
remarkable; they cc:.espond in structure with those described 
by Hamann in Neorhynchus claveceps, in which species 
according to this observer both the skin and the lemnisci retain 
in the adult their embryonic condition. As in Neorhynchus 

1 «Arch. Anat.,’ 1868, p. 584. 


2 «Die Nemathelminthen,’ Hefte 1 and 2, Jena, 1891 and 1895. 
3 * Bibl. Zool.,’ Heft 7, 1892, p. 1. 


ON ARHYNCHUS HEMIGNATHI. i | 


the number of nuclei is very small, some twelve to twenty seem 
to suffice for the whole of the subcuticle, and perhaps two to 
four for each lemniscus. The structure of the nucleus shows a 
most striking resemblance to an ameeba with rather short pseudo- 
podia (figs. x, x11, and xtv)._ No single nucleolus can be de- 
tected, but numerous chromatin particles are present, and 
in some a distinct vacuole can be observed. These nuclei are 
scattered about in a most irregular fashion ; not one may be 
seen in a number of consecutive sections, and then perhaps 
three or four may appear, and from their large size persist 
through several sections. The nuclei lie, as a rule, embedded 
in the substance of the subcuticle ; more rarely they are found 
in the lacunz. Although there is no proof, one is tempted to 
believe that the nuclei wander through the subcuticle and 
lemnisci in an amceboid manner, and that the small number of 
nuclei which are found in these tissues is compensated for 
partly by the large size of each, but more especially by their 
mobility. Similar amoeboid nuclei undoubtedly move about, 
fuse with one another, and undergo fission in the subcuticle of 
the larval forms of Neorhynchus claveceps. 

Within the subcuticle and completing the skin on the inner 
side, is a layer of circular muscles, and still more internally a 
layer of longitudinal muscles (figs. vi and xv). The muscles of 
these layers are but a single fibre thick, and they are not very 
uniformly present. The circular layer is most complete in 
the region of the trunk, and I have figured a section to show 
this (fig. x11). The longitudinal layer is even less definite, 
but scattered fibres can be detected here and there (figs. xiv 
and xv). Hach fibre appears to be spindle-shaped, and in the 
circular muscles has the striated portion only on its outer face, 
forming a thin band; the inner half of the fibre consists of 
vacuolated strands of protoplasm in which is a nucleus. The 
longitudinal layer of muscles alone is continued over the 
lemnisci (figs. 1x and xiv). These muscles are not covered 
on their inner side by any layer of epithelial ceils, neither 
does any such layer cover the ligament, but both tissues lie 
freely exposed to the fluid of the body-cavity. 


212 ARTHUR E. SHIPLEY. 


In the more typical Acanthocephala the anterior end of the 
body terminates in a hollow eversible portion provided with 
rows of hooks whose number and shape have a certain sys- 
tematic value. This introvert can be withdrawn, not into the 
general body cavity, but into the cavity of the introvert sheath, 
which is shut off from the general body cavity by a double 
(chinorhynchide) or a single (Neorhynchidz) wall. The ex- 
trusion of the introvert is believed to be effected by fluid being 
forced into its lacunze by the lemnisci. It is retracted by 
special muscles attached to the inside of its tip; besides these, 
other retractor muscles run from the outside of the introvert 
sheath, and these serve to retract the whole sheath and its 
contents into the trunk. The chief nerve ganglion lies as a 
rule on the posterior end of the introvert sheath, usually in 
the middle line, but in the Gigantorhynchidz it is placed to 
one side. From the posterior end of the introvert sheath, and 
having its origin between its two walls when they are present, 
the ligament runs backward, traversing the body cavity, and 
ending in the funnel-shaped internal opening of the oviduct in 
the female and in the vas deferens in the male. 

Owing to the absence of an introvert and its sheath, the 
relations of the ligament in the present species is somewhat 
altered. It takes its origin from the anterior end of the head, 
and at first seems to consist of a few strands of muscular 
fibres which arise from the muscles of the skin (fig. x1). All 
my specimens but one proved to be mature females, whose 
ovaries had broken up into the egg masses which are character- 
istic of the Acanthocephala. These egg masses consist of 
packets of a dozen or more cells of which the peripheral layer 
develop into ova at the cost of the central cells which serve 
them as a food supply (figs. v1, x1, and x11). These packets 
coexisted in my specimens with ova in various stages of 
development, some without any egg shell, whilst others were 
provided with a thick deeply-staining membrane. The whole 
lumen of the head was crowded with these ova. In the region 
of the collar the ova were confined by a thin-walled membrane, 
and in the trunk there were two such masses of ova, which, 


ON ARHYNCHUS HEMIGNATHI. 213 


however, seemed less mature than those lying in the head. 
Lying amongst the various organs in the body-cavity were 
a number of very finely granular masses, which I take to be 
the masses of spermatozoa (figs. vi and x). Of the complex 
system by means of which the ova leave the body, little could 
be made out beyond the fact that a well-marked funnel is 
present opening into the posterior end of the body-cavity 
of the trunk (fig. 1x). I failed, however, to find a second 
opening near the narrow end of the funnel such as occurs 
in other forms, but this may have been due to the poor state 
of preservation. 'The funnel leads into a duct which opens 
on the posterior end of the trunk. 

The testes are two in number, and lie one behind the other 
in the ligament, though owing to its looping both may appear 
in the same transverse section. The spermatozoa do not 
escape into the body of the male as the ova do into that of the 
female, but pass down a duct in the ligament which opens 
at the end of the body. Traces of accessory glands were seen, 
but the details were not clear. 

The brain lies on or in the ligament just behind its point 
of attachment to the skin of the head (figs. x1 and x11). 
Owing to the disruption of the ovaries in my female specimens 
the ligament could not be traced very far, but in the only 
male it reached from one end of the body to the other. ‘The 
brain consists of a few large ganglion cells with a clear homo- 
geneous cytoplasm and deeply-stained nuclei; the divisions 
between the cells were very sharp and straight (fig. xr). In 
the females this mass of cells lay in the ligament; in the male, 
on the other hand, it occupied the centre of the fibrous and 
muscular strands which compose that body (fig. xv). In the 
former I could trace no nerves leaving the brain, but in the 
male two nerves surrounded by muscles pass backward ; these 
obviously correspond with the retinacula of other forms, 


Classification. 
Until recently the group Acanthocephala included but one 
genus, Echinorhynchus, which comprised several hundred 


214 ARTHUR E. SHIPLEY. 


species. Recently, however, Hamann! has pointed out that 
these species present certain differences which enable him to 
divide the group into three families, each with a corresponding 
genus. ‘To these I venture to add a fourth, to include the 
remarkable form above described. This family may, I think, 
be called the Arhynchide, and the new genus Arhynchus, 
which name refers to the absence of the eversible introvert ; 
and, inasmuch as it is convenient in naming a parasite to have 
some indication of its host, I think the specific name may 
be hemignathi. 

If these terms be adopted, the classification of the Acan- 
thocephala will be as follows, the characteristics of each of 
the first three families being taken from Hamann’s papers. 


ACANTHOCEPHALA. 


I. Family EKcuinoruyncuipa. ‘The body is elongated and 
smooth. The introvert sheath has double walls, and the 
introvert is invaginated into it. The nerve ganglion is in the 
introvert sheath, mostly embedded in it and central in posi- 
tion. ‘The hook papille are only covered with chitin at their 
apex, and the hooks have a process below. 

Genus Echinorhynchus, with the characters of the 
family. 

The vast majority of Acanthocephala belong to this family ; 
a few may be mentioned. EH. proteus, found in many fishes 
aud varying in size with its host ; its larval forms inhabit the 
Amphipod Gammarus pulex, and are also found in the body- 
cavity of numerous fresh-water fishes. E.clavula occurs in 
many fishes and in the intestine of a species of Bufo. E. 
angustatus is found also in fishes, with its larval form in the 
Isopod Asellus aquaticus. EK. moniliformis is said to 
attain maturity in the human intestine; its usual host is a 
mouse, and its larval host the larva of a beetle, Blaps mucro- 
nata. KE. porrigens invests the intestine of the rorqual 
and H. strumosus that of a seal. There are many others. 


1 Loe. cit., and ‘ Zool. Anz.,’ Bd. xv, 1892, p. 195. 


ON ARHYNCHUS HEMIGNATHI. 215 


If. Family Gicanroruyncnipm. Large forms, whose body 
is ringed and flattened during life like that of a Tenia. 
The hooks are like those of a Tzenia, the hook-papilla being 
entirely covered with chitin. There are two root-like pro- 
cesses to each hook. The introvert is muscular, has no lumen 
and the introvert cannot be retracted into it, but the whole 
retracts into the body-cavity. The ganglion is excentrically 
placed to the side, behind the middle of the so-called sheath. 
The body-cavity is enclosed in a structureless membrane, and 
is traversed by membranes stretched transversely. The 
lemnisci are long, coiled, with a central lacuna. 

Genus Gigantorhynchus, with the characters of the 
family. 

Hamann includes three species in this family—G. echino- 
discus, G. tenioides, and G. spira; and points out that 
EK. gigas agrees with them in all points but that of the external 
annulation. The first of the above-named species occurs in the 
intestine of anteaters, and has been found in Myrmecophaga 
jubata and Cycloturus didactyla. G. tznioides 
has been found in a species of Cariama, Dicholophus 
cristatus; and G. spira lives in the king vulture, Sarco- 
rhampus papa. E. gigas in the adult stage occurs in the 
small intestine of swine, and its larval host is believed to be 
the grubs of Melolontha vulgaris and Cetonia aurata in 
Europe and of Lachnosterna arcuata in the United States.! 
It is recorded once from the human intestine. 

III. Family Neoruyncuipa#. Sexual maturity is reached 
in the larval state. The introvert sheath has a single wall. 
A few giant nuclei only are found in the subcuticle and in 
the lemnisci. ‘The circular muscles are very simply developed, 
and the longitudinal muscles only present in places. 

Genus Neorhynchus, with the characters of the family. 

This genus includes but two species, N. claveceps and 
N. agilis. They both present interesting cases of pzdo- 
genesis, the large embryonic nuclei of the young larva do 
not break up into numerous nuclei as they do in the com- 

1 C, W. Styles, ‘Zool. Anz.,’ Bd. xv, 1892, p. 52. 


216 ARTHUR E. SHIPLEY. 


moner species. N. agilis is found in Mugil auratus and M. 
cephalus; N. claveceps in the Carp, Cyprinus carpio, 
its larva form according to Villot! in the fat bodies of the 
Neuropterous insect Sialis niger; it has also been found in 
the alimentary canal of the leech Nephelis octocula, and 
specimens of the water-snail Limnzea have been articially in- 
fected with it. 

IV. Family Aruyncuip#. Short forms, with the body 
divided into three well-marked regions,—head, collar, and trunk. 
The head is pitted, the collar smooth, and the trunk wrinkled, 
not annulated—in spirit specimens. There is no eversible in- 
trovert, and no introvert sheath and no hooks. The sub-cuticle 
and the lemnisci have a few giant nuclei, and the lemnisci are 
long and coiled. 

Genus Arhynchus, with the characters of the family. 

This family in the length and curvature of its lemnisci re- 
sembles the Gigantorhynchide, and in the persistence of 
the embryonic condition of the nuclei in the sub-cuticle and 
the lemnisci, the Neorhynchide ; but in the shape of the body, 
its division into three well-marked regions, the absence of 
eversible introvert, introvert sheath, and hooks, it stands alone, 
though to some extent nearer to the Neorhynchide, in which 
the introvert is relatively small, the introvert sheath simple, 
and the number of hooks reduced, than to either of the other 
families. 

The single species Arhynchus hemignathi was found 
attached to the skin around the anus of a Sandwich Island 
bird, Hemignathus proceros. This bird is a member of a 
family Drepanidide, which is entirely confined to the Sand- 
wich Island group. Professor Newton tells me that it is 
probable that the “ food of Hemigunathus consists entirely 
of insects which it finds in or under the bark of trees;”’ hence 
it is probable that the second host of this parasite, if such 
exists, must be looked for amongst the Insecta. 


Tue ZootocicaL LABORATORY, CAMBRIDGE ; 
March, 1896. 


1 «Zool. Anz.,’ Bd. viii, 1885, p. 19. 


ON ARHYNCHUS HEMIGNATHI. 27 


DESCRIPTION OF PLATE 12, 


Illustrating Mr. A. E. Shipley’s paper on “ Arhynchus 
hemignathi, a new genus of Acanthocephala.” 


In some cases the names of the various structures are written on the figures, 
in others the following abbreviations have been adopted. czve. mus. The layer 
of circular muscles in the skin. gez.d. Genital duct. gez.p. The external 
opening of the duct. dace. The lacune in the skin. Jda/. lac. The large lateral 
lacune of the trunk. dem. The lemnisci. Zig. The ligament. dong. mus. The 
longitudinal muscles of the skin. mus. The muscles from which the ligament 
arises. uc. The amoeboid nuclei of the skin and the lemnisci. sperm. Coagu- 
lated masses of spermatozoa in the body-cavity of the female. 

Fies. 1, 0, and 111.—Three views of three different specimens of Arhyn- 
chushemignathi. Hach x 20. The division of the body into three regions 
is well marked. The details are shown in Fig. 1. Figs. 1 and 1 are rough 
sketches. 

Fig. 1v.—A transverse section through the head of a female, crowded with 
ova and egg-masses; the ligament is shown in section. x 40, 

Fig. v.—A transverse section through the same, just below the edge of the 
collar. In the centre is the neck, which fuses with the collar a few sections 
further back. The big circular canal of the collar is shown at dae. x 40. 

Fic. vi.—A transverse section through the trunk of the same. The upper- 
most lemniscus is cut in two places. The ovary is double, and shows egg- 
masses as well as eggs; some coagulated masses of spermatozoa are lying in 
the body-cavity. x 40. 

Fic. vit.—A surface view of the external opening of the genital duct. 
x 40. 

Fic. vill.—Some developing ova, highly magnified. 


Vie. 1x.—A transverse section through the trunk near the genital pore, 
taken from the same series as figs. Iv, V, aud vi. It shows part of the funnel- 
shaped internal opening of the genital duct, gex. d. x 40. 

Fic. x.—A transverse section from another specimen taken behind the 
opening of the genital duct. This shows the arrangement of the lacune and 
their communications with the lateral lacune, /aé. lac. 

Fic. x1.—A longitudinal section through the central part of the skin of the 
head, showing the origin of the ligament and the ganglion cells of the brain, 
lying in a mass of ova and egg-masses. 


218 ARTHUR E. SHIPLEY. 


Fite. x1.—A small portion of the skin in section, showing the single layer 
of circular muscle-fibres. x 40. 


tre. xui.—A median longitudinal section through a female. The whole 
body-cavity full of ova and egg-masses. The ligament is seen in the head, and 
the genital duct near its opening in the trunk. The left lemniscus, cut twice, 
is alone seen. X 80. 

Fic. xtv.—A transverse section through the trunk of a male, showing one 
of the testes. This section shows also the longitudinal muscles on the 
lemnisci and the large lateral lacune, Jat. lac. x 40. 

Fic. xv.—A transverse section through the head of a male, showing the 
brain in the ligament, and the longitudinal muscle-fibres very well developed. 
x 40. 


ZOOLOGICAL OBSERVATIONS IN THE SOUTH PACIFIC. 219 


Zoological Observations in the South Pacific. 
by 


Arthur Willey, D.Sc.Lond., 
Balfour Student of the University of Cambridge. 


With Plate 13. 


In the course of a brief stay in Sydney I had the opportunity 
of working up for publication some material obtained by me. 
My warmest thanks are due to Professor W. A. Haswell, who 
has generously placed space and appurtenances at my disposal 
in his laboratory at the University of Sydney. 


ip 


On a New Ampuioxus FROM THE LoutstapE ARCHIPELAGO 
(Asymmetron caudatum, n. sp.). 
(With Plate 13, figs. 1 to 4.) 


Last January, as a result of my dredging operations in the 
magnificent lagoon of the Deboyne groups of islands, of which 
Panaieti is the largest, I obtained, on two occasions, specimens 
of the family Branchiostomide (a single specimen only on each 
occasion), which proved to be a new species, nearly related to 
Asymmetron lucayanum, Andrews. 

‘They were dredged up from a depth of one fathom and seven 
fathoms respectively, living in clean coral sand. In the same 
habitat a species of Ophelia is very abundant, and to the 
casual gaze it bears a striking resemblance to the Amphioxus, 
with which it is associated. This Ophelia burrows in the sand 
with its pointed snout, ike Amphioxus, but more deliberately, 


220 ARTHUR WILLEY. 


and swims with exactly the same motions of the body as 
Amphioxus, which is not the case with most other Annelids, 
At first I thought I was obtaining a fine series of Amphioxus, 
but the great majority turned out to be Ophelids. 

The species now to be described, for which I propose the 
name Asymmetron caudatum, n. sp., adopting Kirkaldy’s 
system! of subdivision of the Branchiostomide, is perhaps 
interesting on account of its geographical distribution relative 
to other species of Amphioxus, rather than from the possession 
of novel features. 

That it should be entirely distinct from its relatives in 
Torres Straits, from whom it is removed by a distance of less 
than six hundred miles, and, on the other hand, closely allied 
to a species residing in the Bahamas, upwards of eight thousand 
miles away, on the other side of the American continent, is 
certainly a noteworthy fact in distribution. 

The specific differences between species of Amphioxus are 
frequently of apparently little moment, but may be of im- 
portance when taken in conjunction with geographical distri- 
bution. On this principle the Japanese Amphioxus recently 
described by Andrews?” should at least be regarded as a marked 
variety of rather than identical with Amphioxus Belcheri. 

As mentioned above, I only succeeded in obtaining two 
individuals of Asymmetron caudatum, because the unpro- 
pitious season of the year, when one is liable to be overtaken 
by westerly gales, rendered it imadvisable to delay for any 
considerable length of time in the lagoon. 

The larger specimen was an immature female, and measured 
29°5 mm.; the other was a mature or submature male, 
measuring 20 mm. in length. Other differences, particularly in 
respect of the caudal fin (cf. figs. 1 and 3), occurred between 
the two individuals ; and these differences, combined with the 
marked contrast in size, would seem to point to a sexual 


1 Kirkarpy, J. W., * A Revision of the Genera and Species of the Branchio- 
stomide,” ‘ Quart. Journ. Mic. Sci.,’ vol. xxxvii, 1895, pp. 8303—323. 

2 AnpRews, HE. A., “An Amphioxus from Japan,” ‘ Zool. Anz.,’ 1895. 
This should be named A. Belcheri, var. japonicus. 


ZOOLOGICAL OBSERVATIONS IN THE SOUTH PACIFIC. 221 


dimorphism, which, if found to be constant among a large 
number of individuals, would be a fact of some importance. 

My two specimens alone prove that an appreciable sexual 
dimorphism may occur, whether as a variation or as a fixed 
property. 

The following are ali the numerical facts that I have ascer- 
tained regarding the two specimens. 

1. Female. Length, 29°5 mm.; number of myotomes, 60; 
formula, 40—9—11; length of tail or urostyloid process be- 
hind myotomes, 3 mm.; distance of anus from root of tail, 
2°5 mm.; distance of atriopore from root of tail, 6°5 mm. 

2. Male. Length, 20 mm.; number of myotomes, 64; 
formula, 44—9—11; number of gonads, 30, disposed uni- 
laterally on right side. 

The characteristic generic features of Asymmetron, exhibited 
by the new species, are as follows.! 

The right metapleur is continuous with the ventral fin; the 
latter has no fin-chambers and no fin-rays; the right and left 
metapleura pass equally into the rostral fin in front ; an inter- 
tentacular membrane is present betweenthe ventral buccal cirri ; 
the buccal cirri are plain; and lastly, there is a pair of post- 
atrioporal, or, better, post-atrial caeca, which are, however, very 
short, ending abruptly between the posterior lips of the 
atriopore. 

The dorsal fin-spaces become very irregular posteriorly, and 
cease at about the 54th myotome. 

In addition to the length of the body and the number of 
myotomes, an important specific feature of Asymmetron 
caudatum is the rostral fin, which is marked off from the 
dorsal fin by a pronounced constriction of somewhat varying 
depth (cf. figs. 2 and 4). The rostral fin is indeed a striking 
feature in the fresh object, but is less marked after preservation. 

The caudal fin also varies considerably in the two individuals, 
its urostyloid portion being much more constricted off from 
the succeeding portion in the region of the myotomes in the 
male specimen than in the female (figs. 1 and 3). It is, how- 


' T have verified all the points mentioned by sections and preparations. 


222 ARTHUR WILLEY. 


ever, important to note that in any case there is a distinct 
caudal fin extending to the posterior extremity, and Kirkaldy’s 
opinion that there is no caudal fin in Asymmetron is somewhat 
misleading. 

From the above description it is evident that A. caudatum 
comes very near to A. lucayanum, but differs definitely and 
distinctly from it in size, number of myotomes,! and in the 
possession of a prominent rounded rostral fin, marked off above 
and below by a constriction. 


i 


On tHE NEPIONIC SHELL OF THE Recent NavtiLus. 
(With figs. 5 and 6 on Plate 13.) 


While conversing recently with Mr. Charles Hedley, 
Conchologist to the Australian Museum, Sydney, I had 
occasion to refer to the very large size of the ovarian ova 
of Nautilus pompilius, apparently larger than the ova of 
any other known Cephalopod; while the fluidity of the yolk 
would lead to the presumption that the ova, when deposited, 
would be provided with a firm covering, whether horny, 
coriaceous, or calcareous. 

Mr. Hedley then drew my attention to two young shells of 
N. pompilius in one of the museum cases, in each of which, 
at the same region in both, there was a very distinct discon- 
tinuity in the formation of the shell by the animal, and added 
that this line of demarcation probably represented the limit 
reached by the shell at the time of hatching from the egg. As 


1 The average myotome formula for A. lucayanum as given by Andrews 
is 44—9—13 = 66, and the average length 13 mm. The maximum length, 
according to Andrews, is 16mm. out of a large number of specimens. 
Kirkaldy gives 19 mm. as the maximum length of specimens sent to Prof. 
Lankester by Prof. Agassiz. 


ZOOLOGICAL OBSERVATIONS IN THE SOUTH PACIFIC, 223 


the size of this young shell behind the line of division between 
it and the later formed shell-substance seemed to correspond 
roughly with the dimensions of the ovarian ovum, as stated by 
me in a former communication,! I at once became interested 
in Hedley’s suggestion. 

The term “nepionic ” was invented by Hyatt in 1890, and 
embodied by Jackson in his important memoir on the ‘ Phy- 
logeny of the Pelecypoda,’? and has been adopted by Pilsbry 
in the ‘Manual of Conchology.’ As I understand it, it is 
intended to be applied to the larval stages, sensu stricto, of 
all animals. To quote Hyatt, “ This term should be confined 
to the designation of stages of growth which are immediately 
continuous with later stages, and parallel or referable in their 
origin to the adults of allied existing or fossil forms which are 
not so remote as those from which the embryonic stages are 
derived.” 

The nepionic period is succeeded by the adolescent or neo- 
logic stages, which “‘ are of as great importance for tracing the 
genealogy of small groups as are the [nepionic] characters in 
larger groups.” 

In the case of those terrestrial Gastropods which lay large 
hard-shelled eggs full of yolk, the development and growth of 
the shell take place within the egg until the close of the 
nepionic period, by which time the shell has in many in- 
stances attained a very respectable size. Thus the nepionic 
shell of the above-mentioned molluscs is that portion of the 
true shell (as opposed to the embryonic shell) which develops 
within the egg. 


1 “Letters on Nautilus, &c.,” see ante, p. 175. 

* Jackson, Kopert Tracy, “Phylogeny of the Pelecypoda, the Avi- 
culidse, and their Allies,” ‘ Memoirs Boston Soc. Nat. Hist.,’ vol. iv, No. 8, 
1890. 

Proressor Hyarr introduced the term “nepionic ” in a foot-note to the 
preceding memoir as a substitute for the term “silphologic” suggested by 
him ina previous paper (Hyarr, ALpuEvs, ‘ Values in Classification of 
the Stages of Growth and Decline, with Propositions for a New Nomen- 
clature,” ‘Proc. Bost. Soc. Nat. Hist.,’ vol. xxiii, Part 3, 1887). 

1 am indebted to Mr, Hedley for these references. 


224 ARTHUR WILLEY. 


The nepionic shell frequently stands in marked contrast to 
the post-nepionic or (with the large terrestrial Gastropods) 
post-natal shell, both by a definite line indicating a tem- 
porary discontinuity of growth, and by its different colour 
and minute sculpturing.! In fact, it is obvious that a 
definite break in the continuity of shell-formation is more 
likely to occur in those cases in which, after the completion of 
the nepionic shell, the young mollusc has to go through the 
operation of hatching, after which it recommences to form 
post-natal shell-substanee. This supposition is, I believe, in 
accordance with the facts. 

With regard to Nautilus, about whose reproduction we are 
so much in the dark, it is a distinct point gained to know, at 
least with respect to its shell, at what stage the animal 
hatches. 

The existence of the line of discontinuous growth, at a 
particular point, in young shells of N. pompilius, might not 
be regarded as conclusive in itself, but I have since found a 
precisely similar and constant discontinuity of growth in a 
number of shells of N. umbilicatus and N. macrom- 
phalus. 

In the two last-named species the nepionic shell differs 
from the post-nepionic shell by the presence in the former of 
a minute plication, and of a peculiar glossy sheen. In N. 
pompilius the nepionic shell also has a peculiar sheen, and 
presents a contrast to the post-nepionic shell in the colour 
of its markings, which are brick-red, those of the latter being 
crimson. 

Fig. 5 (Pl. 13) represents a drawing of the umbilical region 
of N. umbilicatus, Lam. On looking down into the um- 
bilicus at the first whorls of the shell, it will be seen that 
after the young shell had completed rather more than one 
entire whorl, there occurred a sharp and abrupt interruption 
in the growth of the whorls, and (in the actual specimen, 
although not shown in the figure) the character of the shell 


1 Compare Hepiry, Crartzs, “On the Structure and Affinities of Panda 
atomata,” Gray, ‘ Records Austral. Mus.,’ vol. ii, 1892, pls. v and vi. 


ZOOLOGICAL OBSERVATIONS IN THE SOUTH PACIFIC. 225 


added after this line,! is very distinct from the previously 
formed shell, as already mentioned. Sometimes there is a 
visible depression in the shell at the nepionic line. 

In the shells of all species of Nautilus there are sometimes 
to be seen one or several strongly marked lines indicating 
occasional, but not regularly periodical, interruptions in the 
growth of the shell,—periods of rest, in fact. These, however, 
are neither constant in position nor in their occurrence in 
different individuals. On the other hand, the nepionic line 
occurs in a constant position in all individuals of a species, 
although in N. pompilius it cannot be seen beyond a certain 
stage, owing to the overgrowth of the whorls. 

Altogether I have seen the nepionic line in eighteen shells 
of the three above-named species of Nautilus. 

Fig. 6 (Pl. 18) is an outline sketch of a young shell of 
N. pompilius? with still perforated umbilicus, to show the 
nepionic line marking off the prenatal from the post-natal 
shell. ‘The following are the measurements of this shell: 

(a) Length of complete shell (from the anterior free margin 
to a point directly opposed to it on the posterior convex surface 
of the shell), 32 mm.; greatest width of complete shell, 19 mm. 

(b) Length of nepionic portion of shell, 27 mm. (this varies 
slightly in different individuals from 25 to 27 mm.); greatest 
width of nepionic portion of shell, 16 mm. 

It may be observed that the above dimensions of the nepionic 
shell of N. pompilius are a good deal in excess of those of 
the ovarian ovum which I previously described ; but not more 
so relatively, I think, than is the size of other newly hatched 
Cephalopods to the original size of the freshly deposited ova or 
of the mature ovarian ova. 

Knowing the size and character of the nepionic shell of 
recent Nautilus, and the size and character of its submature 
ovarian ovum, I think there is ample justification for the con- 
clusion that the nepionic line marks the period at which the 
young individual hatches from the egg. 


1 T shall refer to this in future as the “ nepionic line.” 
* The sketch was taken from a specimen in the Australian Museum by 
kind permission of Mr. Charles Hedley. 


VOL. 39, PART 2.—NEW SER. P 


226 ARTHUR WILLEY. 


There is strong presumptive evidence that the animal of 
Nautilus at the time of hatching already possesses the main 
features of the adult, with the possible addition of a yolk-sac, 
This follows both from the size of the nepionic shell, which 
comprises a number of chambers, and from the consideration 
of a very small specimen of Nautilus, which I obtained in 
New Britain and mentioned in a former publication.!, I have 
unfortunately mislaid the shell of this specimen, but it had a 
perforated umbilicus, through which daylight could be seen, 
and was at least as small as the shell from which Pl. 13, fig. 6, 
was drawn. 

Jackson (loc. cit.) refers to Hyatt’s figures of a young 
Nautilus Koninckii, in which is indicated “a smooth ne- 
pionic period, succeeded by a fluted neologic stage.” Ihave . 
been unfortunately unable to refer to Hyatt’s well-known 
memoir on the “ Embryology of Fossil Cephalopoda ”’ (‘ Bulletin 
Mus. Compar. Zool. Cambridge, Mass.,’ vol. iii, 1872). 

In the first part of his “ Beitrage zur Entwickelungsge- 
schichte der fossilen Cephalopoden” (‘ Paleontographica,’ 
vol. xxvi, 1879-80) Branco figures a number of minute shells 
of Ammonites, which sometimes, when they measure even less 
than 1 mm. in diameter, exhibit a deep circular constriction ; 
and this not only in cases in which the adult shell is grooved, 
but also in many instances in which the latter is not grooved. 

If this groove were to be identified or compared with the 
nepionic line of recent Nautilus, which is itself sometimes 
depressed, as mentioned above, it would indicate that the ova 
of the Ammonites referred to were much smaller and poorer in 
yolk than those of tne recent Nautilus. This would be an 
interesting conclusion if it could be substantiated. 


1 * Natural Science,’ June, 1895. 


ZOOLOGICAL OBSERVATIONS IN THE SOUTH PACIFIC. 227 


IDOE 


On some VARIATIONS IN THE SHELL oF NAUTILUS, WITH 
Description or A New Variety (N. POMPILIUS, VAR. 
MoRreETONI, NOV. VAR.). 


(With figs. 7, 8, and 9 on Plate 13.) 


Of the obvious differences that exist between the shell of 
N. umbilicatus and that of N. pompilius the following 
may, for my present purpose, be particularly remembered. 

1. In N. umbilicatus (see Pl. 13, fig. 5)— 

a. The umbilicus is perforated, the shell being so loosely 
coiled that a small aperture, which can be looked through, is 
left in the centre of the whorls. 

(3. The umbilicus is large and open, so that all the whorls 
can be distinctly seen. This holds good also for N. macrom- 
phalus, but here the umbilicus is not perforated. 

y. The free margin of the shell does not merge into the 
umbilicus, but tends abruptly inwards and downwards, so as to 
form a prominent shoulder, nearly at right angles to the 
lateral free margin. It finally fuses laterally with the anterior 
convexity of the shell, which is pigmented black. 

2. In N. pompilius— : 

a. The umbilicus is imperforate (excepting very young 
individuals). 

. [3. The umbilicus is practically obliterated through a depo- 
sition of callus by the dorso-lateral angles of the mantle. 

y. The free margin of the shell merges directly into the 
umbilical region. 

The different characters tabulated above are usually very 
distinct, but they may be considerably weakened by variation. 

One variety, for instance, of N. pompilius will present all 
the characteristics of the species with the single exception 
that the free margin of the shell does not pass directly into 
the umbilicus, but forms a shoulder. 

In another variety the free margin of the shell will pass 
gradually into the umbilicus, but the latter will be perforated. 


228 ARTHUR WILLEY. 


Finally, in a third variety the deposition of callus will have 
been entirely left out and the whorls loosely coiled, so that we 
shall have the above-mentioned shoulder, a perforated um- 
bilicus, and visible whorls. 

The latter variety is not identical with the so-called N. 
stenomphalus, but is, in a sense, intermediate between the 
latter and N. umbilicatus. I obtained a single example 
from New Guinea waters, kindly given to me by the Hon. M. 
H. Moreton, of Government House, Samarai, British New 
Guinea. I have pleasure in naming it N. pompilius, var. 
Moretoni. 

In acollection of shells of N. pompilius it may be ob- 
served that the extent of deposition of callus over the um- 
bilicus is very unequal; and that while there is usually a thick 
prominent plug of white callus, sometimes the deposition of 
the latter has been so deficient as to leave a deep umbilical 
depression on cither side. 

A peculiar feature about the variations now being dealt with 
is that they may occur on one side of the shell only, the other 
side being more or less normal. 

Pl. 13, fig. 7, represents the umbilical region of a variety of 
N. pompilius, whose umbilicus was only partially perforated. 
On the right side of the shell (the latter being considered with 
the convex side downwards, as it is in the fresh condition) 
there was a deep hole, about a centimetre in depth; on the 
other side there was a deep umbilical depression, which, how- 
ever, was closed over with callus, with the exception of a 
minute pin-hole aperture—perhaps half a centimetre in depth. 
This latter could only be made out by close inspection, and 
at first sight the Jeft side of this shell appeared to be practi- 
cally normal. 

It is clear that if this variation were carried a little farther 
in the same direction, we should have a form possessing the 
essential features of N. stenomphalus, and there is no 
reason at present to regard the latter as anything more than a 
variety of N. pompilius with a persistent perforated um- 
bilicus. In a specimen of N. stenomphalus in the Austra- 


ZOOLOGICAL OBSERVATIONS IN THE SOUTH PACIFIC, 229 


lian Museum only a portion of the coil was dimly visible, and 
the free margin of the shell passed directly into the anterior 
border of the umbilicus without a prominent shoulder.! 

In Pl. 13, fig. 8, an interesting variation is shown in which 
the umbilicus was imperforate, yet nevertheless the free margin 
of the shell did not pass simply into the umbilicus, but bent 
round in such a manner as to form a prominent shoulder, 
behind which there was a deep umbilical depression. I have 
two examples of this variation from New Guinea, and in both 
cases the above-described shoulder was confined to the left side 
of the shell, the right side being normal. 

It will be found convenient to name these varieties. That 
first described above, with partially perforated umbilicus, may 
be called N. pompilius, var. perforatus, nov. var. 

The second variety, with the shoulder, may be called N. 
pompilius, v. marginalis, nov. var. 

Now we come to the third variety to be described, namely, 
N. pompilius, var. Moreton, nov. var. 

In this variety, which is a very well-marked one, there is a 
prominent shoulder as shown in Pl. 13, fig. 9; the umbilicus is 
somewhat widely perforated, and a considerable portion of the 
whorls can be distinctly seen. ‘This shell was symmetrical in 
every respect. At the first sight of it one might well hesitate 
whether to regard it as a loosely wound pompilius, or as a 
closely wound umbilicatus, but the existence of intermediate 
varieties shows, I think with certainty, that it belongs to the 
pompilius group. 

All the variations described above belong to the category of 
substantive variations, according to Bateson’s nomencla- 


* It may be added here that there also seems no reason for regarding 
N. scrobiculatus as other than a variety of N. umbilicatus. On this 
view the species of recent Nautilus would be reduced to three,—N. pom- 
pilius, N. umbilicatus, and N. macromphalus. 

2 All the varieties described in this paper were obtained in New Guinea, 
and with one exception from the natives. I have certain evidence of the 
living Nautilus occurring in New Guinea waters, but the natives know 
nothing of it, and it is not in a workable condition for the zoologist 


230 ARTHUR WILLEY. 


ture,! and it is important to notice that they can be arranged 
in three groups as follows : 

1. Single variations, in which the variation affects only 
a single character, such as the umbilical shoulder or the um- 
bilical aperture. 

2%. Collective variations, in which a whole group of cha- 
racters are affected, as in the variety Moretoni. 

3. Incomplete variations, as in the two examples of my 
variety marginalis, where the umbilical shoulder was con- 
fined to one side of the shell. 

Finally, it may be pointed out that the species and varieties 
of the recent Nautilus seem to fall naturally into the following 
scheme, which, however, does not aim at suggesting any 
particular direction of evolution. 


N, macromphalus. N. umbilicatus. 


N. umbilicatus, v. secrobiculatus. 
N. pompilius, v. Moretoni. 


N. pompilius, v. stenomphalus. 


N. pompilius, v. perforatus. 
N. pompilius, v. marginalis. 


Ve 
fr 


N. pompilius. 


University oF SYDNEY ; 
May 18th, 1896. 


1 Bareson, W., ‘Materials for the Study of Variation,’ Macmillan, 1894. 


ZOOLOGICAL OBSERVATIONS IN THE SOUTH PACIFIC. 231 


EXPLANATION OF PLATE 13, 
Illustrating Dr. Willey’s “ Zoological Observations.” 
Figs. 1—4 refer to Asymmetron caudatum, n. sp. See p. 219. 


Fic. 1.—Asymmetron caudatum, n. sp. Posterior extremity of 
female from right side. d. Dorsal. v. Ventral. From living object. 

Fic. 2.—A. caudafum. Anterior extremity of female from right side. 
From living object. 

Fic. 3.—A. caudatum. Posterior extremity of male from right side. 
From living object. : 

Fie. 4. A. caudatum. Anterior extremity of male from left side. 
From living object. 


Fig. 5.—N. umbilicatus. Drawing of umbilical region to show the 
nepionic line, which is situated a little beyond the first complete whorl of the 
shell. From right side. 


Fic. 6.—N. pompilius. Outline sketch of young shell with perforated 
umbilicus, to show nepionic line. From right side. 


Fic. 7.—N. pompilius, v. perforatus, nov. var. Umbilicus partially 
perforated. The dotted line is merely to indicate the limit of the black 
pigment on the anterior convexity of the shell. Right side. 


Fic. 8.—N. pompilius, v. marginalis, nov. var. Umbilical region of 
left side, to show the umbilical shoulder and depression. 


Vie. 9.—N. pompilius, v. Moretoni, nov. var. Umbilicalregion, showing 
shoulder, perforation, and whorls. From right side. 


N.B.—The drawings are placed in such a way that the convex surface of 
the shell would be directed upwards and the mouth of shell downwards, 
because it is thought that they would be more intelligible in this position ; 
but the right and left sides are named with the shell considered in the 
natural position, with the convex surface directed downwards and the month 
upwards. 


} ; pays atk me ee a4 . 4 iva ue: 


z : eu 

5 valde oaihs OE, ee eet) Woe ety 
7 ‘ 

a D Din a we) << " 

= it Cm <> * i ’ be 


: me : ; “ew ire Rey <a ry - er 
: earl eo adie big a « hoy ' * on ya) Py phan i 


. ahh ality Yr i 5 4 : ‘ * = 
7 oer oe Pir Hl ~ rth. rad v 
- ~ 
. : iy 
‘ “a: > = 
ae? CYA os j 
7 a : ° C Fe tour’ 


whe a a 
y : ; 'S O2 as eS 3 ali ate 
; see (he (he = 


te 


; “ a 6 * ¢ - 
: } a; a ae genes | 
, : PD j . ‘ a : 
; 3 - 
. « 7 (NIAMS 7h aay 
Ps = © 
> 2 


: ; es ty et SETS)) 7 
a a be , pe ‘SAD Gare J 
: ‘ih 


s ; eta Jc 


: "oe res 


.  . 
i 0 tLe Pr i 
th fi a | 
ee ee et, 
ne ae ae Hig Bex 
hal £ ; 7 7 _ Lhe 5 hae “eo s Ht 
fai ald | i Paci &. aie mit ies he val hy By: 


* - i a 
7 . 
= . q - ce 
> Sista 


CHLAMYDOMYXA MONTANA. 233 


Chlamydomyxa montana, n.sp., one of the 
Protozoa Gymnomyxa. 


By 


E. Ray Lankester, M.A., LL.D., F.R.S., 
Linacre Professor in the University of Oxford. 


—_—- 


With Plates 14 and 15. 


I vo not know of any naturalist who has seen the beautiful 
Chlamydomyxa labyrinthuloides of Archer in a state 
of “ expansion ”’ since he described it in this Journal twenty- 
one years ago. Mr. Thisleton Dyer, now Director of Kew 
Gardens, saw it in its active streaming condition when Archer 
exhibited it at one of the meetings of the Dublin Microscopical 
Club five-and-twenty years ago. And the other members of 
that admirable band of naturalists saw it and testify to the 
correctness of Archer’s description and figures. 

Some seven years after the publication of his paper Archer 
sent both to me and (through Prof. Percival Wright) to Mr. 
Patrick Geddes samples of Sphagnum with the cysts of 
Chlamydomyxa attached to the fronds of the moss. I kept 
the specimens which I received in conditions which I hoped 
would favour the rupture of the cysts aud the out-crawling 
of the Chlamydomyxa-network, but entirely failed to obtain 
such a result. 

Mr. Patrick Geddes made a careful study of the cysts received 
by him, and wrote a very interesting article on them, accom- 
panied by a coloured plate, in this Journal, vol. xxii, 1882, 
p. 30. Mr. Geddes especially dwelt upon the remarkable 
laminated formation of the cellulose cyst-wall, and oun the divi- 


234 HK. RAY LANKESTER. 


sion and multiplication of the encysted mass. He pointed out 
the importance of Chlamydomyxa for the theory of the vege- 
table cell-wall and its laminated structure, and he was led to 
compare the encysted condition of Chlamydomyxa with the 
encysted condition of such forms as Heematococcus and Glao- 
capsa. He also made important observations on the red- 
coloured oily substance formed in the cysts. 

I frequently searched for Chlamydomyxa in the years fol- _ 
lowing the publication of Archer’s paper whenever I found 
myself in a moorland country with Sphagnum bogs. In the 
neighbourhood of Lervik, on the Stavanger Fjord in Norway, 
I searched in vain, as also on Dartmoor. I was, however, 
rewarded for my coutinued efforts by at last finding 
Chlamydomyxa in abundance on August 22nd, 1886, on the 
surface of Sphagnum gathered in small ditches cut in the 
bog which occupies clearings in the pine-wood at Pontresina 
in the Enghadine; and on two subsequent visits to Switzer- 
land at the same time of year, once at Zermatt (1890) and 
later again in the Enghadine (1892), I found and studied 
Chlamydomyxa. 

On the first occasion when I found Chlamydomyxa, viz. in 
August, 1886, at Pontresina, I was fortunate in being able to 
observe and draw specimens which expanded their remarkable 
networks under the microscope, and I showed such active 
specimens in the Hotel Saratz to my friend Ernst Haeckel, 
who happened to arrive there for a short holiday, and to Sir 
Edward Fry—a judge not only of legal but of biological 
matters. My drawings then made are reproduced in Plates 
14 and 15, figs. 1—4. I was able to recognise the cysts 
which accompanied the naked Chlamydomyxa as belonging to 
the latter, and, indeed, should have recognised them from my 
previous familiarity with the encysted specimens sent to me by 
Archer. But I have never seen expanded Chlamydomyxa 
since that first occasion. At Zermatt and at the Maloja in 
subsequent years I found only the cysts, and no persuasion 
which I could offer was sufficient to induce the Chlamydomyxa 
to leave its cyst. The cysts, however, were always very abun- 


CHLAMYDOMYXA MONTANA. 235 


dant, and varied in shape and size. It is not at all improbable 
that the month of August is either too late or too soon for the 
best chance of observing Chlamydomyxa in the motile condi- 
tion, and I should recommend other observers in the same 
locality to try the latter part of June and commencement of 
July, as affording the best chance of observing Chlamydomyxa 
in the unencysted condition. 

The Chlamydomyxa which I studied in Switzerland, whilst 
agreeing in essential features with Archer’s genus, appears to 
me to be certainly a distinct species, for which I propose the 
name Chlamydomyxa montana, ‘The grounds for this 
distinction will be apparent from the following description. 

Mode of Occurrence.—The specimens of Sphagnum which 
yielded the Chlamydomyxa had a peculiar ruddy-brown appear- 
ance and a glistening surface, differing from ordinary healthy 
growths of the moss. They were old and in a state of incipient 
decay. I scraped the surface of the leaflets of the moss and 
teased them on a glass slide, so as to remove and spread as 
much as possible for observation the matters adhering to the 
surface. 

General Appearances.—I then found, when the preparation 
was examined with the microscope, numerous olive-brown disc- 
like or ovoid bodies from the 535 to the =, of an inch in 
diameter, which very scon showed a movement of the colour- 
less border by which each was surrounded. The colourless 
border seemed to open out and spread itself in the form of a 
network of threads. 

The Threads.—The gradual separation and spreading of 
these threads is a very curious phenomenon, and certainly 
gives the impression that the threads are pre-formed, closely 
packed together, and that they gradually separate and straighten 
out. Iam not prepared to assert that such is the case, but I 
think it very likely. A fully expanded specimen of fair average 
size 1s represented in Plate 14, figs. 1 and 2. It is represented 
in two consecutive phases of expansion, separated by about five 
minutes, as seen under a high power—No. 10 immersion of 
Hartuack. If one taps the cover-glass the whole branching 


236 E. RAY LANKESTER. 


aud outstretched system of threads suddenly contracts, and 
forms again a clear colourless “ border” to the pigmented 
disc. But immediately the outpushing of the threads recom- 
mences, and seems to proceed by a straightening and disen- 
tanglement of the individual threads, which grow in length 
as one watches them, apparently being extruded from the 
central mass. The commencement of the process of expansion 
of the threads is shown in a drawing of another specimen 
(Plate 15, fig. 4). 

In all cases the process of expansion is accompanied by the 
formation of “ vacuoles”? in the colourless border-substance 
(see figs. 1—4), which change their shape and position as 
the spreading goes on. But they have not the character of 
the ‘contractile’? vacuoles of Heliozoa, although Archer 
speaks of such “ contractile ” vacuoles (perhaps by inadvertent 
use of the term “ contractile”) as occurring in his C. labyrin- 
thuloides. Archer also observed frequently solid food par- 
ticles such as diatoms to be entanged in the substance of the 
colourless material. I never observed any food particles thus 
engulfed in C. montana. 

The threads are of extreme tenuity, of equal diameter 
throughout, and appear to me not to undergo any measurable 
change in dimension. I did not see them shorten and thicken, 
but they appeared to become flexed and gathered together 
when a stimulus was applied to them. Further, I never saw 
any thread either fuse with a neighbouring thread or divide 
into two. It appeared to me (but the observation is difficult) 
that when two threads come together they may be very closely 
apposed, but nevertheless retain their distinctness; and con- 
versely that, where a thread seems to divide into two longi- 
tudinally, the case is really one of the separation of two pre- 
existing threads. 

My general conclusion is that the threads do not really 
form either a dendritic branching figure or a network, but are 
merely apposed so as to form when less expanded, or, to speak 
more accurately, when less straightened, an apparent meshwork, 
aud when more straightened and separated from one another 


CHLAMYDOMYXA MONTANA, 237 


an apparent tree-like structure, the appearance in both cases 
being illusive. A gentle to-and-fro swinging movement of an 
outstretched thread is sometimes seen. Such movement may 
be due to currents in the surrounding liquid. 

As to the further interpretation of the nature of the threads 
I will say more below. 

The Oat-shaped Corpuscles.—Travelling upon the threads and 
closely packed against one another in those parts of the hyaline 
border-substance which are, so to speak, not unravelled, are 
very numerous oat-shaped or fusiform corpuscles. They are 
represented in figs. 1—4, and one is drawn on a greatly 
enlarged scale in fig. 5. These corpuscles are described by 
Archer as leading characteristics of his C. labyrinthuloides, 
and are compared by him with the nucleated spindle-shaped 
bodies which travel upon the threads of the Labyrinthula of 
Cienkowski. In Archer’s species of Chlamydomyxa, however, 
these corpuscles were homogeneous, showing neither an envelope 
of protoplasm nor a central nucieus. The same is true of the 
fusiform corpuscles of C. montana. They are structureless. 
It is important to note that they are very much smaller in 
C.montanathaninC.labyrinthuloides. Archer states that 
in the latter they are ;,4,5 inch long and half as broad, whereas 


0 
I find that in C. montana they are not more than the ;,1 


12000 
ofan inch or two microns in length. Nevertheless they are 
perfectly distinct and uniform both in size and shape. I found 
that they were stained more strongly than the threads by the 
addition of a solution of iodine ; and since neither by this nor by 
any other method could I detect anything indicating a central 
nucleus or nuclei in Chlamydomyxa (Archer failed equally), I 
think that these oat-shaped corpusclesin both species of Chlamy- 
domyxa should be regarded as nuclei, the particles of a frag- 
mented scattered nucleus. In encysted specimens of Chlamy- 
domyxa montana I entirely failed (either by direct exami- 
nation or by use of reagents) to detect any trace of these oat- 
shaped corpuscles. 

Movement of the Oat-shaped Corpuscles—The movement of 
the oat-shaped corpuscles is the most interesting and character- 


238 E. RAY LANKESTER. 


istic feature presented by Chlamydomyxa. It must be dis- 
tinguished altogether from the straightening and expanding 
movement of the mass of filaments; at the same time it is not 
manifested until the filaments have become—some at least of 
them—straightened and free. Then as such a filament separates 
itself, and as it were slowly pushes itself forth in a straight 
line, first one, then another, and finally many of the oat-shaped 
corpuscles are seen to advance along it. They move slowly in 
one direction as a rule, stopping sometimes after a consider- 
able advance, and then resuming movement. They do not all 
travel at the same rate on one filament. I saw on several occa- 
sions one corpuscle overtake another and glide over 
the back (so to speak) of its more slowly moving companion, 
and advance in front of it. Archer also witnessed this pheno- 
menon in his larger species. Further, the corpuscles do not all 
travel in one direction on one and the same filament. Some 
are advancing towards the free extremity of the filament, whilst 
others are travelling away from it. The corpuscles travelling 
in opposite directions meet and pass one another, or sometimes 
on meeting come to a standstill, then after a time the two 
separate from one another, reversing their previous direction 
of movement. These movements are not novel in themselves, 
but similar to the movements of the granules in the psendo- 
podium of an Actinospherium or in the threads of the vacuo- 
lated cell protoplasm of the Tradescantia hair. What is 
peculiar in the case of Chlamydomyxa is this, viz. that the 
moving corpuscles are oat-shaped bodies of definite and uniform 
size, and that there is no visible coating of streaming proto- 
plasm embedding both them and the filament upon which they 
move (as there is in the case of the pseudopodium of Actino- 
spheerium). 

I am inclined myself, from a careful consideration of my 
own observations on Chlamydomyxa and Actinospherium, 
and of Cienkowski’s observations upon Labyrinthula, to adopt 
the view that the filaments of Chlamydomyxa are inert pro- 
ducts of the metamorphosis of its protoplasm, which have a 
certain amount of durability, but can be rapidly absorbed by 


CHLAMYDOMYXA MONTANA. 239 


the protoplasm which also gives rise to them, as is the case 
with the axial fibres of the pseudopodia of Heliozoa.'! The 
oat-shaped corpuscles also are to be regarded as inert nuclear 
bodies, inert so far as motility is concerned. 

The movement of the oat-shaped corpuscles along the threads 
is, I believe, produced by an exceedingly delicate coating of 
hyaline protoplasm. Of the existence of such hyaline proto- 
plasm, which is neither filament nor corpuscle, we have 
evidence, firstly in the agglutination of the filaments when not 
extended, and secondly in the movement of contraction and 
expansion of the mass of filaments. I do not think that it is 
possible anywhere actually to see as an isolated substance this 
delicate ‘‘ varnish ” of hyaline protoplasm, but it seems to me 
reasonable to infer its existence. 

Thus I should regard the filaments of Chlamydomyxa as the 
delicate skeletal supports of a still more delicate streaming 
protoplasm, in which, as in coarser pseudopodial expansions, 
parallel currents of different rates or opposed in direction 
may arise. 

The Central Pigmented Corpuscles.—The centre of the disc or 
irregular body, which is the form taken by Chlamydomyxa 
montana when in the free motile state, consists, as shown 
in figs 1—4, of a number of rounded corpuscles closely 
pressed against one another, and of a yellow-brown colour 
with a tendency to a greenish tint. There is no doubt that a 
colourless protoplasmic substance invests each of these cor- 
puscles, and whilst holding the mass together is continuous 
with the colourless marginal substance or border of the disc. 
These corpuscles are of fairly uniform size (1, of an inch in 
diameter), and are of viscid consistency. I consider them as 
identical in character with the green vesicles described by 
Bourne as forming the bulk of the structure of his Pelomyxa 
viridis (see this Journal, vol. xxxii ,1891), and I entirely agree 
with him in separating them altogether from relationship to 


‘ I am inclined to think that such an elastic filament, one-sided in posi- 
tion, must be present also in all cilia and other forms of vibratile protoplasm. 


24.0 E, RAY LANKESTER, 


ordinary chlorophyll corpuscles such as those of the leaves of 
plants. 

I consider that these coloured vesicles, which differ from the 
conception of “corpuscles” or ‘ granules” only in their 
relatively greater fluidity, are identical with the colourless 
‘“‘ Glinzkorper” (Greef) or refringent corpuscles of Pelomyxa 
palustris, as I have stated in a note to Gould’s paper on the 
structure of that Protozoon, published in vol. xxxvi of this 
Journal, 1894; and they all, no doubt, have an important 
chemical function in relation to the surrounding protoplasm. 
The colouring matters which develop in such vesicles appear to 
have a special character, and without necessarily agreeing 
exactly with what is called ‘“ chlorophyll,” are related to it 
and to such substances as diatomin, which accompany chloro- 
phyll in other more highly developed organisms. The 
colouring matter of Professor Bourne’s Pelomyxa viridis 
is clearly not simple chlorophyll, or if such chlorophyll is 
there it is mixed with other pigmented bodies. ‘The pre- 
dominant yellow-brown colour of the vesicles (granules or 
corpuscles) of the disc of Chlamydomyxa montana is 
suggestive of diatomin. Very possibly it masks chlorophyll. 
This is probable, because Archer represents his Chlamy- 
domyxa labyrinthuloides as containing vesicles like those 
in question, but coloured bright green; and further, because 
after encystation my C. montana undergoes the most re- 
markable changes in colour, developing a brilliant grass-green 
tint (in some examples), whilst accompanying this chlorophyll- 
like colouring matter there are always found in the cysts of 
C. montana irregular or central droplets of a brilliant crimson 
oily fluid, soluble in ether (see Pl. 15, figs. 8 and 9). 

Encysted Condition.—The cysts of Chlamydomyxa mon- 
tana closely resemble those of C. labyrinthuloides, as 
described by Archer and Geddes. ‘The substance of which the 
eyst-walls are composed yields the blue colour characteristic of 
cellulose when treated with H,SO, and iodine. The cysts are 
further remarkable in two features, and unlike the capsules 
usually termed ‘‘ cysts” among Protozoa. ‘They have (a) a 


CHLAMYDOMYXA MONTANA. 241 


most definite laminated structure, consisting of a series of 
“ shells ” or complete coatings of cellulose substance, deposited 
one within the other, on the surface of the protoplasmic 
organism. The protoplasmic organism apparently shrinks 
after the formation of a first cyst, and having grown smaller 
forms a second deposit, and so on, in some cases to a ninth or 
tenth. In the second place, (b) the living organism enclosed 
in the cyst appears to be by no means quiescent, but to be 
undergoing important chemical changes, as shown by its 
remarkable change of colour (to a bright green) and its de- 
velopment of “droplets” of red-coloured oily material. 
The activity of the encysted Chlamydomyxa is further 
and very markedly demonstrated by its movement, 
change of form, and division into separate masses. 
The fact that the shells or coatings of cellulose are so freely 
and abundantly deposited by the living matter enables one to 
follow these changes in the encysted Chlamydomyxa with 
certainty. Thus in Pl. 15, fig. 6, we have evidence of division 
of an original mass of living Chlamydomyxa into three; in 
fig. 7, part of the cyst at first formed and marked a has been 
deserted, whilst the original outer cysts are much larger than 
those now occupied by the living material, which has divided 
into two masses. In fig. 9, three cysts occupied by green- 
coloured Chlamydomyxa which have given rise also to the red- 
coloured oil drops, are represented. These three cysts pro- 
bably originated from the division of one parent mass. The 
division of the encysted Chlamydomyxa has been described in 
considerable detail by Professor Patrick Geddes in his paper 
already referred to. 

In reference to differences between C. montana and C. 
labyrinthuloides, I may point out that I have not seen the 
body or mass of C. montana when it is in the dendriform 
streaming condition, to be partly lodged in the ruptured cyst, 
as Mr. Archer saw and figured that of C. labyrinthuloides. 
Nor have I seen green pigment (presumably chlorophyll) and 
red oil drops in that phase, as Mr. Archer describes for C. 
labyrinthuloides. The dendriform streaming specimens of 

VOL. 39, PART 2.—NEW SER. Q 


242 E. RAY LANKESTER, 


C. montana which I observed (in considerable numbers for 
several days) were always entirely free from any cyst-wall, and, 
moreover, their central granules or vesicles were of an uniform 
yellow-brown colour. This may be due to specific difference, 
but it would be interesting to know what appearance C. mon- 
tana presents early in the season,—for instance, in June. 

The very regular circular cyst drawn in fig. 10 is a fairly 
common form. As shown in the figure, the cyst contents are 
yellow in colour, almost bright yellow, with radiating struc- 
ture resembling yolk-columns. As many as eight concentric 
lamin were observed in the cyst drawn in fig. 10. 

Affinities of Chlamydomyxa.—I do not think that either the 
observations made by me on C. montana, or the progress of 
our knowledge of Protozoa since Archer described C. labyrin- 
thuloides twenty-one years ago, enables us definitely to assign 
to Chlamydomyxa its position in relation to other Protozoa 
Gymnomyxa. 

I cannot agree with Professor Geddes that it should be 
regarded as related to the Algz, since affinities have to be 
determined by a consideration of all the circumstances, and I 
cannot see how Chlamydomyxa would fit in with known Alge. 

I fully agree with Archer that the nearest ally of Chlamy- 
domyxa is the Labyrinthula of Cienkowski, and that any con- 
sideration of affinities must be based on this alliance. Archer 
was inclined to regard the “ threads”’ of Chlamydomyxa as 
protoplasmic, whilst Cienkowski regarded those of Labyrin- 
thula as of a horny nature. I do not think that there is any 
essential difference between the threads of Labyrinthula and of 
Chlamydomyxa. I regard them both as a“ formed material,” 
differing from streaming protoplasm, and comparable to the 
axial pseudopodial filaments of Heliozoa, The fusiform travel- 
ling nuclei of Chlamydomyxa differ very greatly from the oat- 
shaped nucleated corpuscles of Labyrinthula both in size and 
structure. But it seems to me a reasonable view that the con- 
dition of Chlamydomyxa is derived from that of Labyrinthula, 
aud that the fusiform nuclei of the former represent the oat- 
shaped corpuscles of the latter in a reduced condition. The 


CHLAMYDOMYXA MONTANA. 243 


fact that in C. montana as compared with C. labyrinthu- 
loides the fusiform nuclei are still further reduced, being but 
one third their size, tends to indicate a progressive reduction 
of these bodies. The protoplasm surrounding the nuclei of 
the oat-shaped corpuscles of Labyrinthula is represented in 
Chlamydomyxa by an exceedingly delicate and practically 
invisible layer, which is also extended over the threads, and is 
the seat of the movement which is rendered visible by the 
translation of the nuclei. 

The Protozoa which come nearest to Chlamydomyxa and 
Labyrinthula are certain of the Mycetozoa; but each of the 
two genera in question differs in its own way from the typical 
Mycetozoa, especially as to reproduction, and nothing would be 
gained by sinking them taxonomicaily in that assemblage. 

The most remarkable feature in which Chlamydomyxa differs 
from Labyrinthula and from all other Protozoa is its “ encysted 
phase.’’ The enclosure of the general protoplasm in a cyst- 
wall may be compared with the fruit-formation and other 
cyst-like productions of the Mycetozoa. But the physiological 
character of the cysts of Chlamydomyxa, the activities of the 
encysted organism, and the great relative duration and im- 
portance of the encysted phase are peculiar to Chlamydomyxa, 
and may be explained by the fact that this organism is an 
inhabitant of fresh water, and subjected to the vicissitudes of 
temperature and evaporation of the inhabited water which we 
know are frequently associated with special protective struc- 
tures and aberrant phases of growth and activity. 

A point in which both Chlamydomyxa and Labyrinthula 
agree with the Mycetozoa is their epiphytic habit. 

In his paper on Chlamydomyxa labyrinthuloides, 
Archer draws attention to a terricolous plasmodium (found in 
a flower-pot) of an unknown Mycetozoon described by Cien- 
kowski. It is not possible to decide from his (Cienkowski’s) 
description whether he had before him Archer’s Chlamy- 
domyxa, but it is most probable that he had not, since 
he does not describe the characteristic structures of that 
organism. Similarly the Biomyxa vagans described and 


244 E. RAY LANKESTER. 


figured by Leidy, in his large volume on Rhizopoda in the 
United States Geological Survey of the Territories, 1879, 
presents some points of agreement with Chlamydomyxa, and 
was discovered by Leidy upon Sphagnum. Yet inasmuch as 
neither the delicate filaments nor the fusiform nuclei, nor the 
central coloured vesicles (granules), nor cellulose laminated 
cysts are ascribed by Leidy to his Biomyxa, we must suppose 
that it indicates a distinct organism, not even closely related 
to Chlamydomyxa. 


OxrorD; July 8th, 1896. 


EXPLANATION OF PLATES 14 & 15, 


Illustrating Professor Ray Lankester’s memoir on “ Chlamy- 
domyxa montana, n. sp.” 


All the figures except Fig. 6 are drawn to the same scale. 


Fic. 1.—Chlamydomyxa montana, n. sp., in an unencysted expanded 
condition. 


Fic. 2.—The same specimen-after an interval of five minutes. Drawn at 


Pontresina from a living specimen as seen with objective No. 10 immersion 
of Hartnack, August 22nd, 1886. 


Fic. 3.—Another and larger specimen observed and drawn at the same 
time. 


Fic. 4.—A specimen which is beginning to expand its filaments. After 
five minutes the threads would have a much wider extension. 


Fic. 5.—A single fusiform corpuscle (nucleus) greatly magnified, showing 
its shape and its relation to the filament. The actual length of the corpuscle 
iS =5455 Inch. 

Fig. 6.—An encysted specimen of C. montana which has divided into 
three. Pontresina, August, 1886. 


Fic. 7.—An encysted specimen of C. montana which has divided into 
two. Pontresina. 

Fic. 8.—An encysted specimen of C. montana which has developed 
chlorophyll and a crimson oil drop. The cyst-wall is simple. Pontresina. 


Fic. 9.—Three similar encysted specimens. Pontresina. These and the 
last were brought alive to London and drawn there. 


Fic. 10.—A spherical cyst with eight lamine to the cyst-wall and a golden 
yellow pigment. This last specimen observed and drawn at the Maloja in~ 
1892. 


JAN 4 1897 


OONSTITUTION AND DEVELOPMENT OF TERMITES. 240 


The Constitution and Development af the Society 
of Termites: Observations on their Habits; 
with Appendices on the Parasitic Protozoa 
of Termitide, and on the Embiide. 


By 


Professor B. Grassi im collaboration with Dr. A. Sandias. 


With Plates 16—20. 


Tue memoir, of which Ihave made the following translation, was originally 
published in the ‘ Atti dell’ Accademia Gioenia di Scienze Naturali in Catania,’ 
ser. 4, vols. vi and vii (1893-4), but has become more generally known 
through the appearance of a separate edition in the former year. But though 
its value was immediately recognised, the work has not yet become familiar 
to an extent commensurate with the importance of its contributions to natural 
science. 

The object of the treatise, itself an expansion and completion of certain 
preliminary papers on Termitidee by Professor Grassi, is set forth in that 
author’s introduction, and requires no further explanation; but it is perhaps 
permissible to point out that, over and above the results obtained in the 
pursuit of that object, the memoir is a signal and instructive example of a 
class of work but too seldom resorted to—the union of morphological re- 
search with an inquiry of the most prolonged and persevering character into 
the habits and bionomics of the living form. In these latter respects it forms 
a worthy parallel to the work of such great pioneers in the investigation of 
the social systems of insects as Smeathman, Huber, and Réaumur. And it 
is for this reason that no study of an abstract which merely summarises the 
results obtained can prove as convincing or suggestive as that of the original. 

An excellent general summary of the various writings on Termitide, which 
includes references to such facts as have been observed since the publication 
of the present memoir, is to be found in vol. v of the ‘Cambridge Natural 
History,’ by Dr. Sharp. In view of its existence it has been thought un- 
necessary to make more than a very few additions to the original text; these 
will be found indicated in the usual way by square brackets. 

Wa ter F. H. Branprorp, M.A.Cantab., 
Lecturer on Entomology at the Royal Indian 
Engineering College, Cooper’s Hill. 
VOL. 389, PART 3 —NEW SER. R 


246 B. GRASSI AND A. SANDIAS. 


To Fritz MUtier, 
on his Jubilee. 
Catania, 1892. 


INTRODUCTION. 


Ir has been recognised by Darwin, in his immortal work on 
the ‘ Origin of Species,’ that the existence of insect commu- 
nities, composed of different castes (kings, queens, workers, 
soldiers, and the like), furnishes his opponents with a trenchant 
weapon. Now these communities actually consist not merely 
of fertile forms (kings and queens), but of sterile forms 
(workers, soldiers, &c.), distinguished by important modifica- 
tions of structure and marvellous instincts, none of which are 
to be found in their parents. The sterile progeny would there- 
fore appear to be uninfluenced by one of the prime factors in 
the struggle for existence, that of heredity, owing to the im- 
possibility of transmitting to the offspring such modifications 
of structure or instinct as may be gradually acquired; and 
Darwin himself admits that it required great confidence in his 
theory not to renounce it in the face of this objection. 

I have had to deal shortly with this question on another 
occasion with regard to bees, for which I have observed that 
the cause of their apparent deviation from the normal rule 
may possibly be found in the existence of workers capable of 
oviposition, which apply themselves to all the tasks of the 
colony, and possess the characteristics of true workers, with 
the single difference that they lay parthenogenetic male ova. 

The males which hatch from these eggs would then transmit 
to the female ova—laid, as we know, by the queen—the cha- 
racteristic properties of the worker.! 


1 Darwin, however, explains the phenomenon by a very ingenious com- 
parison. ‘ According to M. Verlot,’ he writes, “some varieties of the 
double annual stock, from having been long and carefully selected to the right 
degree, always produce a large proportion of seedlings bearing double and 
quite sterile flowers ; but they likewise yield some single and fertile plants. 
These latter, by which alone the variety can be propagated, may be compared 


CONSTITUTION AND DEVELOPMENT OF TERMITES. 247 


As it is probable that the same explanation is equally ap- 
plicable to other Hymenoptera, I put aside that order of 
insects to devote myself to another, of remote relationship— 
namely, the Corrodentia, of which the Termitide constitute 
the typical form. But I found myself here in a much less 
cultivated field than that of the Hymenoptera, and it took 
me several years before I was able to see my way. 

In the meantime a school has arisen under Weismann, 
which denies the heredity of acquired characteristics ; this has 
somewhat modified, and perhaps rendered more interesting 
the general questions which have inspired my particular re- 
searches, as I will explain. 

It has been known for many years that the same bee larva 
may develop either into a queen or a worker in accordance 
with the nutriment it receives. Nutrition, therefore, possesses 
a most remarkable influence on the bee’s generative organs, 
and on many other characters correlated with their greater or 
less development (e.g. the faculty of producing wax, or of 
collecting honey, which is possessed by the worker only, and 
not by the queen). This would indicate that the environment 
has a powerful and direct influence on the genitalia, and would 
therefore tend indirectly to show that the much-disputed in- 
heritance of acquired characteristics is a possibility. 
with the fertile male and female ants, and the double sterile plants with the 
neuters of the same community. As with the varieties of the stock, so with 
social insects, selection has been applied to the family, and not to the indi- 
vidual, for the sake of gaining a serviceable end.”  [‘Origin of Species,’ 
ed. 6, p. 230. The whole chapter should be consulted.] But it may be 
objected that the difference between the queen and worker forms is far more 
profound than that between the simple and double stocks. 

Biichner [* Aus dem Geisteslebens der Thiere,’ translated under the title 
‘Mind in Animals,’ London, 1881], on the other hand, thinks that the 
explanation of the phenomenon should be found partly in atavism, and partly— 
as far, that is, as regards the marvellous instincts—in the instruction which the 
young receive from the colony. That atavism certainly plays a part in instinct 
is demonstrated ad evidentiam by a fact discovered by myself. It is known 
that certain silkworms become pup and moths without spinning a cocoon. 
Now my experiments show that the offspring of such moths may spin perfectly 
constructed cocoons. 


248 B. GRASSI AND A. SANDIAS. 


But this argument in its favour would be greatly weakened 
if the circumstance indicated by me above—namely, the pos- 
sibility of hereditary transmission by means of oviparous 
workers—should have played the part which I suppose it to 
have done. 

I determined, therefore, to investigate the origin of the 
workers and soldiers in Termitide,—in forms, that is, whose 
phylogenetic source is certainly absolutely distinct from that 
of the social Hymenoptera.) Such an inquiry would at 
least throw light on my hypothesis with regard to the bee, 
and in any case should lead to results of some utility for the 
problem I have several times referred to, the heredity of 
acquired characteristics, and for the subordinate and related 
question of the direct influence of environment on the gene- 
rative organs. 

The recent researches of Van Beneden, Boveri, O. Hertwig, 
and others on the ovum and spermatozoon had afforded yet 
another problem for solution, which would be modified accord- 
ing as to whether there should or should not exist special eggs 
or spermatozoa for the workers and soldiers of Termitidee. 

In short, the theory of evolution, the disputed heredity of 
acquired characteristics, and lastly, the theory which postu- 
lates the existence in every somatic cell of elements derived 
from both parents, these alike have all furnished me with 
motives for regarding the elucidation of the origin of caste 
forms in the social Termitide as a problem of the highest 
interest. This, then, is the main object of the present memoir, 
and it has been arrived at only by dint of prolonged observation 
and preliminary experimentation, the results of which are 
fully related here as a necessary corollary. 

The development of my argument has, so to speak, com- 


1 It must be recollected that, though we are acquainted with forms which 
perhaps can be ascribed to the Blattide from the Silurian strata, the Termi- 
tide, according to Scudder (in Zittel’s ‘Handbuch der Palaontologie,’ ii, 772), 
are absent in all the paleozoic, and appear only for the first time in mesozoic 
strata. Nevertheless the phylogenetic independence of the Termitide and 
Hymenoptera is indisputable. 


CONSTITUTION AND DEVELOPMENT OF TERMITES. 249 


pelled me to discuss neoteinia! and the simplification and 
degeneration of forms, and therefore to express my present 
views on the Thysanura, which for a long time formed the 
special object of my studies. 


In these preliminary remarks I have prominently set forth 
the fundamental ideas which have directed my work, and now 
pass at once to the consideration of the Termites. 

As yet two species only of Termitide, Calotermes flavi- 
collis, Fabr., and Termes lucifugus, Rossi, have been 
found in Italy. The latter forms the subject of Lespés’ 
admirable monograph,? but Calotermes flavicollis is known 
to us only by a few accounts of merely systematic importance, 
and not even complete.® 

However, Fritz Miller has made important investigations* 


1 (The term neoteinia has been introduced by Camerano (‘ Bull. Soc. 
Ent. Ital.,’ 1885, pp. S9—94) to denote the persistence during adult life of 
part or all of the characteristics normally peculiar to the immature, growing, 
or larval stages (e.g. the persistence of gills in the axolotl). It therefore 
covers much the same ground as is denoted by the term pedogenesis, but 
appears, so far as can be gathered from Camerano’s paper, to include a some- 
what wider class of facts than those comprised under the latter term, which 
would fall under his definition of total as opposed to partial neoteinia. 
Neoteinia, or the persistence of larval characteristics, does not necessarily 
imply that anticipation in time of sexual maturity which is usually connoted 
with the use of the term pedogenesis,—which, moreover, is strictly applied 
to agamic reproductions.—W. F. H. B.] 

2 « Recherches sur l’organisation et les meurs du Termite lucifuge,” 
‘Ann. Sci. Nat.’ (4), v (1856), pp. 227—282, pls. v—vii. 

3 (Hagen, “ Monographie der Termiten,”’ ‘ Linn. Entom.,’ x, 1—144, 270— 
395; xii, 4—342. The reference to C. flavicollis, op. cit., xii, pp. 54— 
61, pl. i, fig. 12; pl. ii, fig. 15.] 

4 “Beitrage zur Kenntniss der Termiten,” ‘Jen. Zeitschr.,’ vii (1873). 
I. “Die Geschlechtstheile der Soldaten von Calotermes,”’ pp. 333—340, 
pls. xix, xx. II. “Die Wohnungen unserer Termiten,” pp. 341—358. 
III. “‘ Die Nymphen mit kurzen’ Fliigelscheiden ? (Hagen), ‘ Nymphes de la 
deuxitme forme’ (Lespés). Hin Sultan in seinem Harem,” pp. 451—463. 
Id., ix, 1875. IV. “Die Larven von Calotermes rugosus, Hag.,” 
pp. 242 —264, pls. x—xiii. 


250 B. GRASSI AND A. SANDIAS. 


on another species of Calotermes, and these, as will be seen 
later, are in entire agreement with our observations on the 
European species. 

Nevertheless, as I have already mentioned, the study of 
Termitide in general is as yet very incomplete ; and although 
books of travel certainly abound in notices of these interesting 
insects, they mostly deal with detached and imperfect obser- 
vations. 

I have repeated Lespés’ investigations on Termes luci- 
fugus with particular reference to the so-called ‘“ nymph of 
the second form” which that distinguished naturalist de- 
scribed. 

He attributed to it an entirely different significance from 
that which it really possesses, owing mainly to the circum- 
stance that his period of study was confined to about nine 
months—from November, 1855, to August, 1856. But how 
difficult the elucidation of the true function of this nymph of 
the second form has been the reader will gather from the 
pages of this memoir; and I may state at once that it has 
necessitated the removal of many hundred cubic metres of 
earth and the cutting up of hundreds of trees, a task that has 
had to be carried out little by little, and has consequently 
demanded no small amount of patience. 

I am indebted to a small but fortunate discovery for the 
most interesting observations on Calotermes flavicollis, 
which I have to relate in the present work. 

If from three to twenty Calotermes of different ages are 
placed in a glass tube three to eight centimetres long (Pl. 19, 
fig. 11), closed with a cork and kept warm—for example, in the 
waistcoat pocket, unless in summer-time, they continue to live 
and constitute a family, or better, an independent colony ; they 
rear a fresh king and queen,’ if orphaned, and they rear soldiers, 


1 [The statements made on the sexual forms of Termitide will be more 
clearly understood if anticipated by a summary and definition of the terms 
employed in the course of the work. Sexual or royal forms are of two 
kinds, true and neoteinic. The true royal forms are imagos, or 
perfect insects, which acquire a complete development of the wings and 


CONSTITUTION AND DEVELOPMENT OF TERMITES. 251 


&c., if in want of them. In short, after a certain time the tube 
will contain a complete little nest, if it did not oxiginally 
do so. However, the insects do not hesitate to bore through 
the cork, and, unless they are watched, one is surprised to find 
some day that they have all made their escape. But fre- 
quently they remain in the tube, even though they have made 
holes putting them into communication with the outer air 
(PL 19, fig. 11). 

Some of these little nests can be kept alive for several 
months, but many die off after a few weeks, though not until 
they have afforded sufficient opportunity for making numerous 
observations through the glass walls by aid of a lens. It is 
desirable to use tubes of various lengths and calibres, because 
certain points are better seen in a wide than in a narrow 
tube. 

In addition to the insects, the tubes are partly filled with 
fragments of wood, which should be neither too dry nor too 
moist. In the former case the insects gradually shrivel, con- 
tract, dry up and die; and in the latter case there is a deposit 
of water vapour on the inner walls, and they are evidently 
killed by over-dampness. Death ensues more or less rapidly 
according to the amount of water deposited, and is sometimes 
almost as sudden as if the insects were suffocated or chloro- 
formed. Slow death due to over-dampness may be accom- 
panied by distinct cedema or reddish discoloration of the 
darkening of the chitin concomitantly with the maturation of the gonads after 
the last ecdysis. They leave the nest by “swarming.” The neoteinic 
royal forms undergo a premature maturation of the gonads whilst in a late 
arval or nymphal instar, or that of an immature and pallid imago. This 
maturation is accompanied by an arrest of development of other parts of the 
body; the chitin does not darken normally, and the wings do not grow 
further. They do not leave the nest. The latter class is further subdivided 
into complementary and substitute royal forms. Complementary 
forms are not found in Calotermes, but are normal components of the Termes 
nest, in which they are the only reproductive individuals. Substitute 
forms are developed to supply the loss of the true royal forms in Calo- 
termes, or the complementary royal forms in Termes. ‘The three varieties 
of royal examples all comprise kings and queens,—that is, individuals of 
each sex. ] 


252 B. GRASSI AND A. SANDIAS. 


body ; the latter is accompanied by the presence of a bacterium 
which I have not investigated. 

If the wood is too damp the colony may generally be saved 
by the simple precaution of leaving the cork out of the tube ; 
after a few days the moisture diminishes, and the cork must 
then be re-inserted, or the opposite extreme of undue dryness 
is quickly reached. With time and patience an exact estimate 
can be formed of the amount of moisture necessary, and it 
can then be easily regulated. The cork must fit closely, and 
want of oxygen need not be feared. At one time I was 
accustomed to uncork the tubes three or four times a day, 
but I subsequently found this to be quite unnecessary; if the 
Calotermites require change of air, they are able to provide it 
for themselves by boring through the cork. 

When these tiny nests are examined the tube may conve- 
niently be laid flat and left quiet, but an occasional shake is 
sometimes useful] to rouse the inmates to activity. 

Unfortunately Termes lucifugus does not flourish in 
these tubes, but drags on a feeble existence for a week or ten 
days at most; I have made many but quite unsuccessful 
attempts to establish a moderately suitable environment for 
them therein. The species is all the more difficult to study 
because, in comparison with Calotermes, the members of a 
colony are smaller and more rapid in movement. 

Many as have been the details made out through these glass 
tubes in my laboratory, there are others which will be more 
successfully undertaken by anyone who has the opportunity of 
applying my method to the large exotic species. 

Considering that the termites of tropical countries are 
among the most injurious of insects, I must point out a 
practical result of my investigations, and one of the highest 
importance. Contrary to the belief of residents and the 
accounts of travellers, such as Major Casati,! a nest of termites 
cannot be destroyed merely by killing the king and queen. 
If there remain alive some eight, ten, or twenty examples, 
which include any undifferentiated larve, or larve of perfect 


1 («Ten Years in Equatorial Africa,’ i, p. 165.] 


CONSTITUTION AND DEVELOPMENT OF TERMITES. 250 


insects or nymphs, these eight, ten, or twenty will form a new 
colony, which slowly but surely will become as flourishing as 
the original stock. 

It follows, therefore, that the extirpation of these insects is 
excessively difficult in practice; and their entry into any 
situation where they can prove destructive must be prevented. 
This is the strategy to adopt against them, for when once they 
have gained admission there is no way of getting rid of 
them. 

The present work has no claim to be a complete monograph 
of the Termitidz, and merely furnishes materials for whomso- 
ever will undertake so large a task. It is confined, in short, 
to the discussion of certain fundamental points. 

Before concluding I must express my thanks to Dr. Calan- 
druccio, who has given me no little help in this research. 
Many observations, particularly those on Calotermes kept 
in tubes, have been made under my supervision by my pupil 
Dr. Sandias, and have afterwards been carefully verified by 
myself. 


Tur CoLony or CALOTERMES FLAVICOLLIS. 
1. Situation and Nest. 


As with all Termitide, the life of Calotermes is intimately 
connected with the vegetable kingdom, for it lives exclusively 
on woody matter. At Catania I have found it only in the 
stems or roots of living but partially decayed, and therefore, 
as a rule, old plants, and never in those of which the stem or 
tap-root measures less than one or two centimetres in thick- 
ness. But when once established in a stem or main root it 
can extend into decayed portions of the branches or side-roots, 
even if of lesser diameter. Though always absent in entirely 
sound plants, it can be found in partly decayed examples of 
many kinds,—of many, but not of all, for I have never met 
with it in lemon, orange, or Agave americana, &c. It is 
also relatively scarce in cactus (Opuntia), even though large 
portions of the plant are dead; but if present, it may infest 


254 B. GRASSI AND A. SANDIAS. 


the rotten phylloclades, which in fact are sometimes the first 
portions to be occupied. 

It is difficult to find a tree with any decayed portion which 
has not been attacked by Calotermes in Catania and the 
adjoining provinces, at least in the low country, for I have 
not searched at a higher elevation than Nicolosi. At Castro- 
giovanni the species appears to be entirely absent.’ 

As mentioned above, it inhabits still living plants. Should 
the plant die the Calotermites survive until it has become 
completely dry, when they perish—a fact that anyone can 
verify by examination of the vine-stocks annually turned out 
from old vineyards. If the dead trunk-does not dry up, as 
is the case in marshy situations, the insects continue to 
flourish. 

Lucas and some other writers state that they have found 
Calotermes flavicollisin buildings. I have never observed 
this at Catania, but Dr. Sandias has found flourishing colonies 
at Trapani in the woodwork of verandahs, doors and stair- 
ways, &c., ten years old; probably the wood harboured the 
insects before being worked up for domestic purposes. It 
should be observed that the climate of Trapani is very damp, 
so that wood probably dries less there than at Catania.? 

In order that a plant may harbour Calotermes, it must 
(necessarily), as I have stated, exhibit some amount of decay, 
because such decayed portions alone are occupied. To proceed 
to details, Calotermes never invades the healthy parts, but 
encroaches at most on their boundaries. If a partly decayed 
vine-stock is infested, it is usually easy to make out that the 
healthy tissues are respected; yet should they contain an 
internal channel of decay barely larger than the body of a 
Calotermite, the insects can enter and excavate a gallery, which 
will then present but a very thin lining of decayed matter, and 

1 [In a foot-note at the end of the original work the authors state that, since 
it was printed, Signor Giuseppe Corona has discovered the species at Castro- 
giovanni. | 

2 The Marchese Doria informs me that he has confirmed Dr. Sandias’s 
observation at Genoa, although the climate there is certainly not moister than 
that of Catania. 


CONSTITUTION AND DEVELOPMENT OF TERMITES. 255 


will therefore appear to run in the sound structures. But, so 
far as I have observed, this lining of decayed tissue is never 
absent. 

It is undeniable that the process of decay is hastened by 
these burrows, particularly because they allow the infiltration 
of rain-water ; but a detailed study of these phenomena would 
lead too far from the subject. 

Calotermes greatly prefers to live in the deepest parts of 
the dead wood,—that is, the parts nearest the living tissues, 
and consequently the most recently decayed. Portions which 
have been long dead are usually too dry for it, and are chosen 
as a tenement by ants. 

Fertile pairs, unaccompanied by eggs or with very few 
offspring—that is to say, nests in process of formation or 
recently formed,—are to be found by special search at the 
places where a plant has been pruned a year or two previously 
and has subsequently rotted, provided that the pruning has 
not followed on antecedent decay. ‘This can be easily observed 
in the fig or vine. Fresh nests, therefore, are usually estab- 
lished where the dead portion is still so limited in extent as 
not to afford space for a numerous colony, and these situations 
are rarely preoccupied by other insects, such as ants, beetles, 
or other Termitide, &c. As the colony gradually increases 
and requires more room, the decayed area spreads, but to all 
appearance independently of the insects. That this is the 
case is rendered probable by an examination of progressive - 
decay in a district where Calotermes is absent, such as 
Lombardy, where the winter is too severe for them. 

Turning once more to the formation of the fresh termitaria, 
I do not wish to deny the possibility of its taking place in 
trunks in which extensive decay already exists; but such an 
occurrence is unusual, either because the wood is preoccupied 
by termites or other insects, or because the royal pairs must 
traverse parts which are already dry and swarming with enemies 
(ants), in order to arrive at the required spot where the decay 
is most recent. 

The corks of the tubes are attacked probably because they 


256 B. GRASSI AND A. SANDIAS. 


retain a certain amount of moisture; for if the insects are 
shut up with nothing but particles of cork they all die in a 
few days. 

To sum up, two conditions are essential for the life and 
well-being of Calotermes—a suitable temperature and a 
suitable amount of moisture. But whereas increase of tem- 
perature is favorable, at least up to a certain point, the degree 
of humidity can vary only within very restricted limits. 

Owing to the severity of the winter, Calotermes is absent 
from Lombardy, Piedmont, and Venetia, but it is found in the 
province of Genoa; its northward distribution is not accurately 
known. At Catania it is much more sluggish in winter than 
in summer, the ova do not develop, and the larve and nymphs 
do not moult. The influence of moisture has been repeatedly 
referred to; it is correlated with the nature of the cuticular 
structure (chitinous layer), which except in a few regions, such 
as the mandibles, is very thin, even in the adults and royal 
pairs, in both of which it is brown in colour. The younger 
the individual, or the more recent its ecdysis, the thinner is 
the chitin. In general the rule holds good for Calotermes 
that white and semi-transparent examples have a thinner 
cuticle than those which are more opaque and inclined to 
yellow (very old soldiers or substitute-queens). Young or 
freshly moulted specimens, usually distinguishable by their 
greater transparency and whitish or rarely faintly yellow colour, 
always require more moisture than those which are older or 
have not cast their skin for some time, and possess a more or 
less evident yellow tint. The imagos, before or after the loss 
of their wings, can endure a greater degree of dryness; but 
even these generally die in a short time if removed from the 
wood and exposed to._the open air. 

These conditions of temperature and moisture, naturally 
with some variation in degree, especially of the latter factor, 
must hold good for all other species of Termitide, and are 
probably correlated with their limitation to warm countries, 
their much wider European distribution in epochs when the 
mean temperature was presumably higher, and with their 


CONSTITUTION AND DEVELOPMENT OF TERMITES. 257 


choice of evening or night, the interval after a shower or a wet 
day as a swarming-time (Casati).! 

The nest of the Sicilian Calotermes resembles those of its 
congeners, which show, as a rule, very little architectural 
skill. Strictly speaking, they do not build any nest, but con- 
tent themselves with burrows excavated in wood and never 
quitted. These burrows vary very much in dimensions, and 
the same gallery may be narrow in one part and very wide in 
others. In one of the enlargements, towards the heart of the 
nest, the royal couple, with a numerous surrounding, is usually 
to be found. There is never a true royal chamber, such as 
has been described for many other species. The galleries are 
very variable in direction, but the widest and longest are 
generally subparallel with the long axis of the stem. A large 
number of transverse passages is commonly found, and these 
are often too narrow to allow room for more than a single 
individual at a time. According to Fritz Miller,? the burrows 
of Calotermes possess an inner lining of excrement. This 
appears at first sight really to be the case when they are exca- 
vated in rotten and somewhat softened wood. But if they 
are examined in cork, or in wood which, though dead, is still 
hard, the absence of such a lining is easily established, and 
this fact once determined, one can satisfy oneself that burrows 
made in damp wood are also unlined. 

In common with other termites, Calotermes avoids the 
outer layers of the cortex (as may be readily observed in vine- 
stocks), and thus protects the nest from the direct action of 
the atmosphere or infiltration of water. 

As previously stated, the colony provides for an increase in 
numbers by penetrating deeper into the wood. This penetra- 
tion is also determined by the fact that the superficial wood 
becomes too dry. The abandoned portion becomes then com- 
monly occupied by ants, the pitiless enemies of Calotermes. 
The termitarium is separated from the ant-burrows, or from 
openings caused by wind-cracks or axe-wounds by means of 


1 (‘Ten Years in Equatorial Africa,’ i, p. 166.] 
2 (‘Jen. Zeitschr.,’ vii, p. 343.] 


258 B. GRASSI AND A. SANDIAS. 


barricades of masticated wood, or more commonly of excre- 
ment, which is cemented together by saliva, with or without an 
admixture of disgorged wood. This disgorgement is normal, 
and not exceptional, as Fritz Miller believes. 

The following facts and other details of interest may be 
easily observed through the glass tubes. 

If some fragments of wood and a few Calotermites are 
put into a tube and left uncorked, the insects will be found 
on the following day to have established themselves either in 
the whole of the space filled with wood, or in its lower por- 
tion, at the bottom of the tube. In either case the occupied 
area is delimited by means of disgorged matter deposited in 
the interstices between the particles of wood. As far as can 
be seen the boundary wall is complete, but it lies at different 
levels and not entirely in the same plane; its margin can 
easily be made out at the point of contact with the glass, and 
will be found on revolving the tube to form a complete circle, 
but with some irregularities of level, obviously due to a 
selection of the smallest interstices for cementing up without 
reference to their higher or lower elevation. 

[The process of construction can be watched in the tubes ; 
the insect regurgitates a pasty mass, which is spread out, if 
necessary, by the antenne, so as to form a rounded pellet on 
the glass of about a millimetre in diameter, and of the colour 
of rotten wood. | 

The boundary wall is undoubtedly designed for protection 
against direct contact with the atmosphere ; this is proved by 
its not being constructed if the tube_is closed with a cork 
instead of remaining open. 

Such a wall may occasionally be built even in a corked 
tube, should part of the nest contain any offensive substance. 
Thus if flour be added to a small tube-nest, after a few days 
the upper part containing the flour, which will have become 
mouldy, is found to be cut off in the manner described’from 
the lower portion, in which all the insects have collected. 

Sometimes the Calotermites,in a tube may be seen to cement 


1 [‘ Jen.®Zeitschr.,’ vii, p. 3438.] 


CONSTITUTION AND DEVELOPMENT OF TERMITES. 259 


together pieces of wood which are unduly loose; but at other 
times they do not do so, although much incommoded by their 
mobility. In certain cases they line the lumen of the tube 
with disgorged matter, so as to form a continuous layer, save 
for a few irregular patches, but they generally spare them- 
selves this task, or merely begin and immediately abandon it. 

It should be added that if the tube is too damp its inner 
surface may be covered with moisture, to which the insects 
adhere, particularly when young, and so die. This is guarded 
against, at least partially, by the lining of disgorged substance. 

When the supply of wood in the tube is very small, the 
Calotermites usually have recourse to the cork, in which they 
bore one or more galleries a little wider than their body ; they 
gnaw away and hollow out the cork, carrying the débris into 
the tube, which is soon filled up, as the cork-dust is allowed 
to remain as a loose mass. 

If a corked tube is surrounded with sawdust, some of it can 
generally be found in its interior, having been carried in 
through galleries in the cork but little wider than the insects’ 
bodies. 

[The corks of tubes kept in the waistcoat pocket are often 
channelled with burrows leading to the exterior, through which 
the insects escape and are lost, in the belief that they are 
enlarging their nest. Sometimes books near which the waist- 
coat was laid at night were attacked, and the paper was found 
to be gnawed in places the next morning. | 

Burrows made in corks may often be seen to possess a sort 
of rounded lid at their external opening; this is cut out of the 
superficial layer of cork, and fits tolerably closely, though it 
usually falls in a little. It may be hinged to the cork by a 
more or less wide attachment. 

[Calotermites may be frequently seen to carry excrement to 
the bottom of the tube, and to accumulate it there, mixed 
up with wood and cork-dust ; and sometimes they keep the eggs 
in the same place. | 

If the stopper is too narrow the space between it and the 
tube is gradually filled up with excrement. 


260 B. GRASSI AND A. SANDIAS. 


Calotermes flavicollis differs materially from Termes 
lucifugus in the fact that it confines itself to excavating 
burrows, and is thus merely a borer; whereas the latter 
species not only bores, but builds tunnels, in order to connect 
pieces of wood at a considerable distance apart. 


2. Number of Individuals in the Colony. 


In these observations a normal development of the colonies 
isassumed. For should the queen, for example, die before the 
community has reared a successor, several months pass during 
which oviposition is suspended and multiplication is conse- 
quently brought to a standstill. 

A colony of Calotermes rarely consists of more than a 
thousand members, and is relatively numerous when it contains 
five hundred. This is correlated with the fact that the queen 
of our species is very far from attaining the colossal dimen- 
sions which are well known to occur in the queen-termites of 
tropical countries. 

After fifteen months of common life the king and queen 
may be surrounded with fifteen or twenty young, after another 
year with about fifty, and in the two or three following years 
the population increases till it reaches a maximum at which it 
becomes nearly stationary. This is attained when the king 
and queen have reached the largest possible size. 

Eight or ten winged adults may depart from a two-year- 
old nest, and the number leaving in successive years increases 
concomitantly with the increase in the population. 

At the time of maximum oviposition a queen of three to 
four years old lays as a rule four, five, or six eggs a day. 


3. The Different Castes (Plate 16). 


By way of preliminary, it should be stated that for the sake 
of orderly arrangement a knowledge is assumed in this part 
of the work of certain conclusions, the demonstration of which 
is postponed to the succeeding chapter. 

A colony of Calotermites contains— 

1. Undifferentiated larve (fig. 1), capable of becoming 


CONSTITUTION AND DEVELOPMENT OF TERMITES. 261 


either soldier larvee, or larve of perfect insects,—that is, larve 
possessing the first indications or rudiments of wings. 

2. Larve of perfect insects (fig. 3), and nymphs (fig. 4) 
derived from them. 

3. Larve of soldiers (fig. 2), and soldiers (figs. 5 and 
16); the latter derived from the former, and those derived 
from undifferentiated larve, from the larve of perfect insects, 
or from young nymphs. 

4. Perfect insects derived from nymphs (fig. 6). 

5. A royal couple, properly so called, derived from perfect 
insects (figs. 7—11). 

6. Larve of substitute royal forms (fig. 14), and 
substitute forms (figs. 12, 13, and 25) themselves derived 
from the former; those in their turn having originated from 
undifferentiated larve with fourteen or fifteen antennal joints, 
from larvee of perfect insects, or from nymphs. 

To proceed to details: a colony contains newly hatched 
larvee about 1 mm. in length,! with ten antennal joints of 
which nine are pilose, and one not. The glabrous third joint 
is relatively very long, and presents indistinct traces of a 
tripartite division. 'These traces soon become more marked, 
and the separation of the distal joints appears constantly to be 
clearly defined before that between the middle and the proximal 
joints, and possibly may actually precede it. 

As a result we have larve which are a little over 2 mm. in 
length, and exhibit twelve distinct antennal joints, of which 
the 3rd, 4th, and 5th are short and bare. In succession the 
5th joint, as far as I have seen, becomes pilose and relatively 
longer; then the 4th and finally the 3rd exhibit the same 
phenomena. 

We thus obtain examples with twelve antennal joints, all 
pilose; and in the meantime the body increases in size, so that 
such specimens measure about 4 mm. in length. 

The newly born larve are all perfectly alike,—that is, they 

1 In Termitide the length of the individual as a rule has only a relative 
value, as it depends largely on the degree of moisture in the atmosphere, the 
food supply, &c. 

VOL. 39, PART 3.—NEW SER. S 


262 B. GRASSI AND A. SANDIAS. 


are undifferentiated; but when they have attained an 
average length of 2 mm. they are divisible into two groups, 
one with a large head, little narrowed anteriorly (Pl. 16, 
fig. 2), the other with a small head, more evidently narrowed 
towards the apex (PI. 16, fig. 1). 

The former have become soldier larve, and will ulti- 
mately become soldiers. The latter are still undifferentiated ; 
they reach alength of 4mm. and acquire twelve pilose antennal 
joints (vide supra); the character of the head either remains 
unchanged, and continues to be undifferentiated, or be- 
comes modified by an increase in size and in the width of the 
anterior portion, so that they become soldier larve. 

Small soldiers (Pl. 16, fig. 16) are to be found which are 
derived from these very larvee ; they are less than 5 mm. long, 
and possess twelve pilose antennal joints, with an ill-marked 
suture between the 4th and 5th. 

It may be well to observe at once that soldiers of me- 
dium size with thirteen or fourteen pilose antennal joints, 
and large soldiers (Pl. 16, fig. 5) with fifteen to seventeen 
pilose antennal joints also exist. Further, the antenne of 
many soldiers are evidently mutilated. 

Moreover I have often found soldiers with antenne of twelve, 
thirteen (Pl. 16, fig. 19), or fourteen all pilose joints, of which 
the third is in process of division into two, both pilose ; some- 
times the division is completed in one antenna and hardly 
indicated in the other. In short, all intermediate stages exist 
between small, medium-sized, and large soldiers, and the fore- 
going distinctions have therefore a relative value alone. 

For a considerable time I supposed that the soldier of 
medium size originated from the small soldier, and grew itself 
into a large soldier; but as I have not succeeded in proving 
my hypothesis, in spite of long-continued research, I can no 
longer consider it to be well founded. 

Turning to the undifferentiated or small-headed larve, we 
have already seen that they can acquire antenne of twelve 
entirely pilose joints, and a length of 4mm. Subsequently a 
thirteenth joint is added, and their length increases to 6 mm., 


CONSTITUTION ANI) DEVELOPMENT OF TERMITES. 2638 


when also the thirteenth joint becomes pilose ; then a fourteenth 
joint is developed, and the whole fourteen are always found to 
be pilose in examples of 7 mm. and upwards in length. 

Iixamples with fourteen joints may exhibit evident traces of 
wings, which indeed may be already present in individuals 
with thirteen pilose joints, and a very indistinct rudiment of a 
fourteenth glabrous joint at the base of the third. But, on the 
other hand, examples with fifteen joints may be found without 
trace of wings. 

Those which have acquired wing-rudiments may be called 
larve of perfect insects. 

Undifferentiated larve, with thirteen or fourteen joints, as 
well as larve of perfect insects with thirteen (the fourteenth 
very imperfect) or fourteen joints may undergo a direct trans- 
formation into soldier larve with thirteen- or fourteen-jointed 
antenne, and from these originate the soldiers of medium 
size with the same number of antennal joints.! 

Lastly, undifferentiated larve with fourteen joints may 
either become larve of perfect insects without increase in 
that number, or they may remain unchanged till it has been 
increased to fifteen, all of which are pilose; and the latter 
condition may also be reached without change by the larvee of 
perfect insects. 

Both the undifferentiated larve and those of perfect insects 
with fifteen joints may give rise, without increase in that 
number, to the same kinds of forms that have been indicated 
for the larvee of fourteen joints,—soldier-larve, larve of royal 
substitutes, and, provided that they are as yet undifferentiated, 
larvee of perfect insects; and the last may acquire a sixteenth 
joint without increase of the wing rudiments (PI. 16, fig. 3). 
As a rule, therefore, the development of the sixteenth joint 
is accompanied by an increase in the wing-rudiments, so that 
the larva of the perfect insect becomes a nymph. 

I have found no undifferentiated larva with more than fifteen 
antennal joints; but, as I have said, we may have sixteen 


1 | have been unable to decide whether the number of joints is capable of 
increase in the soldier larve. 


264 B. GRASSI AND A. SANDIAS. 


completely pilose joints in the larve and nymphs of perfect 
insects. During this stage both of the latter may undergo 
the same fate as in the preceding stages, becoming soldier 
larvee, larvee of substitute royal forms, or, if already larve of 
perfect insects, nymphs. 

Both the larvee of perfect insects or nymphs may pass with- 
out change to the stage in which the antennz possess seven- 
teen joints. During this stage both may become larve of 
soldiers or royal substitutes. If this does not happen the 
nymphs acquire larger wing rudiments, and the larve of perfect 
insects become nymphs (Pl. 16, fig. 4). 

Before proceeding further it is desirable to give a fuller 
definition of the term nymph. 

It may be applied to such examples as are 8 or 10 mm. in 
length, with sixteen or seventeen antennal joints, and with 
wing-rudiments easily distinguishable by the naked eye. The 
expression, it must be observed, is incapable of precise defini- 
tion, inasmuch as there are no characters in insects with an 
incomplete metamorphosis which distinguish a larva from a 
nymph, except the wings, which have already begun to develop 
in larve of a certain age. It might possibly be adopted as 
connoting the earliest rudiments of wings, but this presents 
the difficulty that they are not easily detected. I therefore 
follow the terminology of Lespés and Hagen, and conventionally 
indicate as nymphs those individuals which have the beginnings 
of the wings readily visible to the naked eye. 

From the account just given it follows clearly that large 
soldiers are derived from the large soldier larvze, just as the 
moderate-sized and small soldiers originate from larve of 
corresponding sizes, and that the large soldier larve have arisen 
in their turn from undifferentiated larve with fifteen antennal 
joints, or from larvee of perfect insects with antenne of fifteen, 
sixteen, or seventeen joints, or lastly also from nymphs. The 
latter, possessing sixteen or seventeen joints, may be trans- 
formed into soldier larve, and subsequently become large 
soldiers in which the wing-buds can be distinctly made out 
with the naked eye (Pl. 16, fig. 26). At a later period these 


CONSTITUTION AND DEVELOPMENT OF TERMITES. 265 


rudiments are reabsorbed until hardly a trace is discoverable 
with the microscope. This origin of the soldier from the 
nymph is certainly infrequent. 

Nymphs with seventeen antennal joints may become imagos 
without increase in that number, or with an antecedent addition 
of one or even two joints; and the perfect insects may thus 
exhibit antennz with from seventeen to nineteen pilose joints, 
a numerical difference which has no relation to sex. Nymphs 
with seventeen to nineteen joints may be also transformed to 
royal substitute larve, but I do not believe that those with 
eighteen or nineteen joints can become soldiers, as I have never 
found so large a number in any example of that caste. 

The change from the nymph to the imago is accompanied 
by the development of pigmented compound eyes, while the 
wing-rudiments, from being vertical and closely appressed to 
the sides of the body, become dorso-lateral at their origin, 
nearly horizontal, and divaricate at the apex.1 These speci- 
mens have at last reached the adult stage (imago), and (Pl. 16, 
fig. 6) possess fully developed wings; at first white, they 
gradually become black and capable of flight. 

It will be seen from this account that examples with a 
number of antennal joints varying from twelve to seventeen, 
and therefore of very different lengths, can be transformed to 
soldier larve, and consequently to soldiers; and that those in 
which the number of joints varies from fourteen to nineteen 
may become larvee of substitute royal forms. The latter larve 
are not easily separable from the others; but in those with 
from fifteen to seventeen antennal joints the pigmentation of 
the compound eyes is evident. But it is absent or very scanty 
in substitute larve with fourteen, and some with fifteen joints. 
Ocular pigment is sometimes present in nymphs with seventeen, 
and always in nymphs with eighteen or nineteen antennal joints 
before their metamorphosis into substitute royal larve. 

The larvee of perfect insects or young nymphs are customarily 
selected for development into substitute forms (Pl. 16, fig. 14), 


1 In these specimens the wings have really reached their full development, 
but are enclosed in a chitinous sheath. 


266 B. GRASSI AND A. SANDIAS. 


and the wing-rudiments of the resulting royal examples are 
therefore absent or very ill-developed. [When a nymph with 
more evident wing-rudiments is selected for the throne, one or 
more of these (usually that of the right fore-wing) is bitten off 
CPLIG, tig. TS) 

Ihave repeatedly stated that the royal forms may originate 
in two ways; in one case they are derived from perfect black 
examples, of which the wings are fully developed and become 
detached along a special line of fissure, so as to leave a short 
stump (the Schuppe or Squama of authors). These ex- 
amples constitute the true black or normal kings and queens. 
In the other case the royal forms are substitutional, and origi- 
nate from examples which have suffered an arrested develop- 
ment of the wings, and in which the compound eyes are 
usually but not invariably pigmented! (Pl. 16, figs. 17, 25). It 
is remarkable that the antenne are never found intact in any 
king or queen, whether true or substitute, however young it 
may be. And the majority of royal forms possess a different 
number of joints, varying from thirteen to six or even four on 
either side. 

A most striking feature of the Calotermite colony is the 
entire absence of workers, in which this species agrees fully 
with the American form studied by Fritz Miller. 

It may, therefore, be concluded that the kingdom of 
Calotermes is composed of three castes: that of the soldiers ; 
that of individuals which reproduce without becoming black 
imagos ; and lastly, that of forms which lay eggs after the 
acquisition of fully formed wings,—that is, after reaching the 
stage of the perfect insect. 

The phenomena related in this section may be recapitulated 
in the following synthesis. 

The normal development of Calotermes up to the 
perfect stage may undergo deviation at different 


1 Perfect insects which are still white may also become substitute forms. 
They do not darken, and the wings are torn off, rarely along the special line, 
but usually irregularly as in Termes lucifugus (q. v.). These observations 
were made after the present work was completed. 


CONSTITUTION AND DEVELOPMENT OF TERMITES. 267 


ages. This deviation may lead to the formation of 
substitute royal forms, or of soldiers, after passage 
through the respective larval stages. The soldier 
larva may originate by modification of examples with 
from twelve to seventeen antennal joints; and the 
larva of the royal substitute forms by modification 
of examples with from fourteen to nineteen antennal 
joints. Perfect insects which are still white may 
also become substitute forms. 

I can confirm the fact discovered by Lespés' in Termes, 
aud by Fritz Miller?in Calotermes, that the caste of soldiers 
is composed of examples of both sexes. 


A minute description of all the forms in the different stages 
of development would be of little interest, and I shall therefore 
confine myself to a few points which have a special bearing on 
the problems I have undertaken to solve. 

Beginning with the sense-organs: 

(1) All trace of eyes is wanting in newly hatched larvee. 
The soldiers possess compound eyes, unpigmented and not 
prominent. The time of their appearance has not been de- 
termined; they are present in a rudimentary condition in 
larve of 3—4 mm. in length, with twelve antennal joints—the 
third glabrous, but they are not clearly distinguishable except 
in sections. Ata later period they are evidently faceted, but 
remain destitute of pigment. The compound eyes acquire 
pigment, as previously mentioned, in nymphs about to become 
imagos and in most examples selected for the dignity of sub- 
stitute kings or queens. In the perfect insect the eyes become 
more prominent and abundantly pigmented, and between them 
is developed an ocellus devoid of any trace of pigment, which 
I therefore regard as rudimentary. 

(2) Sensory hairs (Tastborste), characterised by their 
shortness and their connection with the nervous system, are 
very abundant over the antenne and the whole of the mouth 


1 «Ann. Sci. Nat.’ (4), v, p. 244. 
? (Jen. Zeitschr.,’ vii, pp. 333—340. ] 


268 B. GRASSI AND A. SANDIAS. 


parts, which possess no other form of nerve-ending, such as 
cones (Kegel) or papille (Zapfen). These hairs are espe- 
cially numerous on the apex of both pairs of palpi and on the 
antenne ; on the latter they are most abundant on the apical 
half of the terminal joints. There appears to be no marked 
difference between the sensory hairs of the mouth parts and 
antenne, and experiment shows that the latter organs are 
constructed so as to remain functional even when deprived of 
a certain number of joints. 

(3) The tibize exhibit the peculiar sense-organs discovered 
by Fritz Miller,! evidently tympanal organs (PI. 19, fig. 10). 

This is shown by the presence of the usual terminal rods, of 
a characteristic tracheal branch, the lumen of which is not 
accurately cylindrical, and which opens at either extremity 
into the main tracheal trunk of the tibia, and lastly, of tym- 
panic membranes. 

Tactile hairs are present on various parts of the body, and 
the so-called abdominal appendices (cerci) also appear to be 
essentially tactile. These appendices are really identical with 
the caudal cerci of Thysanura, reduced to a short basilar piece 
and a long terminal joint, of which the apical extremity is 
glabrous. ‘The remainder of their surface is covered with very 
long, fine, and readily vibratile hairs, in addition to others, 
such as are scattered over the body. 

The description just given of them goes to show that they 
correspond in Calotermes with those found in other insects. 
All the sense-organs here described, with the aforesaid excep- 
tion of the eyes, are fully developed at the time of hatching. 
The visual structures are certainly more or less imperfect ex- 
cept in the adults, and become functional concomitantly with 
the wings. Their unimportance in other stages is evident from 
the fact that pigment may be either present or absent in the 
eyes of substitute royal forms, and that individuals without, 
or with more or less imperfect eyes, apply themselves equally 
well to the work of the colony. 

[Calotermites move their antennz freely, and employ them 


1 [‘ Jen. Zeitschr.,’ ix, p. 234, pl. xii, figs. 32, 34.] 


CONSTITUTION AND DEVELOPMENT OF TERMITES. 269 


just as a blind man does his stick, as Lespés observes. But 
they are not accustomed to use them in mutual caresses, like 
other social insects. If the antennz are cut off at the base 
the insect becomes inert, stands in a fixed attitude, almost dis- 
regards the difference between light and darkness, quivers (vide 
infra) at rare intervals, and then for a shorter period and less 
violently than usual. It is not successful in soliciting excre- 
ment (vide infra), or does not directly attempt to do so. All 
these inconveniences are only partially exhibited if the an- 
tenne are detached more or less remotely from their base.] 


Fritz Miiller’s observation! that the number of antennal 
joints in Calotermes rugosus is increased by the successive 
formation of new joints at the base of the third is well known, 
and has led to the division of the antenna into two components 
(base and flagellum), as in other Arthropoda. 

I have attempted to determine the origin of the new joints 
in Calotermes flavicollis, but have not been successful in 
obtaining a clear idea of the process. 

In deciding whether any given joint is the most recent, its 
smallness, its freedom from hairs, and the indistinctness of 
the line of demarcation between it and the parent joint must 
be taken into account. But it will be readily understood that 
all these criteria are apt to fail. 

In the present case, the one which appears most practicable 
(the presence or absence of hairs) may certainly lead us 
astray. In fact, antenne of 13, 14, 15, 16, and 17 joints, with 
the third and fourth joints glabrous, are to be found, as well 
as completely pilose antenne of twelve to sixteen joints (fig. 
22). The natural inference is that the third or fourth joint, 
which was pilose in the latter examples, has become denuded 
in the former (e.g. for an antenna of fourteen joints, the 3rd 
and 4th pilose, to become fifteen-jointed (fig. 21) with the 3rd 
and 4th hairless, one or other of the latter joints must neces- 
sarily have lost its clothing) ; this may have taken place in 
connection with ecdysis, as we shall see later. 

1 [* Jen. Zeitschr.,’ ix, pp. 246, 247.] 


270 B. GRASSI AND A. SANDIAS. 


The broad fact remains that individuals exist with from 
thirteen to seventeen antennal joints, with either the third or 
fourth (fig. 20) or both joints glabrous, or without a single 
glabrous joint. 

How are the new joints formed? I regard it as certain 
that the 13th arises by unequal division of the 3rd, at the 
base of which it appears as a bud. The 15th and 17th appear 
to arise from the 4th, and the remainder like the 13th (Pl. 16, 
figs. 20—22). But closer study is necessary before arriving 
at a conclusion. 

It is of fundamental importance to notice that the new 
joints do not take origin from a zone of embryonic (undiffer- 
entiated) tissue, but from already differentiated structures 
(hy podermis, nerves, &c.). 


Calotermes moults periodically, and it is quite untrue 
that the number of antennal joints increases coincidently with 
the moult. It is more probable that the latter takes place 
after the process of joint formation, and in this we may per- 
haps find the explanation of the just related facts respecting 
the loss of antennal hairs. 

The number of ecdyses cannot be specified ; I have found 
examples of all sizes in process of moulting, and can go no 
further than to fix it at not less than five. The adults and 
soldiers do not moult ; and the latter are derived from soldier 
larve by an ecdysis, so that while the exuvie are larval, the 
new instar is that of a fully-developed soldier. 

As with other insects, the operation is accomplished by 
means of an anterior medio-dorsal fissure, through which are 
drawn out first the head and thorax, and lastly the abdomen 
(Pl. 16, fig. 15). In rare cases an example has been seen to 
be assisted in the operation by his comrades. 

Two ecdyses merit special attention. The first is that 
undergone by the nymph, furnished with wings apparently 
small in size (really of full dimensions and elaborately folded 
up under the old ensheathing cuticle), and with genital appen- 
dices ; from this it emerges with wings of full amplitude, and 


CONSTITUTION AND DEVELOPMENT OF TERMITES. 271 


without appendices should it be a female. The other impor- 
tant ecdysis is that in which the larva of a substitute royal 
form loses its genital appendices if a female, and this leads 
us to a detailed consideration of the latter organs. 


The genital, often wrongly termed the anal appendices,! are 
homologous with those of the ninth abdominal sternite of 
Thysanura, and are attached in Termitide to what is appa- 
rently the eighth, but is really the ninth sternite, the first 
being fused with the metasternum. They possess hairs which 
do not differ from the ordinary hairs scattered over the body. 

They are present in all sexually immature examples, but 
in the males only when mature. Further, the sternites differ 
in the sexes ; in the female the true (not the apparent) seventh 
sternite is strongly developed and semicircular, and the true 
eighth and ninth are small and possess a median fissure. In 
the male the true seventh is rather small, as are the true eighth 
and ninth, in which the fissure is wanting. ‘There is no penis. 


I have but little to say about the internal organs. The 
tracheal system agrees with Fritz Miiller’s description,’ and I 
have likewise observed the stigmata, tracheal trunks, anas- 
tomoses, &c. Further, this species exhibits the blind tracheal 
branches figured by Miiller, which I take to be suppressed 
trunks or tracheal vesicles. 

The alimentary canal presents the following features. In 
newly born larve the teeth of the proventriculus are colour- 
less and quite soft,—that is, covered as yet with a delicate 
cuticle; and the Malpighian tubules are four in number. 
Four others appear at the epoch when the antennez possess 
eleven distinct joints, of which the third and fourth are 
glabrous, and the former shows traces of division into two; 
these new tubules require a certain time in which to attain 
the size of the original four, and I have found them to be still 


1 [The genital appendices must not be confused with the abdominal appen- 
dices or cerci, referred to on p. 268.—W. F. N. B. | 
2 [‘Jen. Zeitschr.,’ ix, pp. 257, 259, pl. xiii.] 


Die B. GRASSI AND A. SANDIAS. 


much the smaller when the number of antennal joints has 
reached twelve, with the fifth joint already pilose, or with the 
third and fourth alone bare. I may note here that the appear- 
ance of intestinal Protozoa (vide infra) coincides -with this 
last stadium, though Joenia is confined to the large-headed 
larvee, whereas Monocercomonas is common both to these 
and the small-headed forms. 

The development of the four secondary Malpighian tubules 
proceeds in such a way that each is placed midway between a 
pair of the primary tubules; that is to say, the latter are 
equidistant from each other, and the secondary tubules are 
intercalated at equal distances between them, so as to produce a 
series of alternate large (primary) and small (secondary) 
tubules. 

Their mode of development is shortly as follows: they 
spring from the proctodeum at its junction with the mid-gut, 
exactly at the same level of the primary series. I have been 
unable to detect any special layer of embryonic tissue destined 
to give origin to them, and they may therefore be regarded as 
a direct derivation from the proctodeal epithelium. 

When the antenne possess twelve joints, all pilose, the 
difference between the earlier and later Malpighian tubules 
has ceased to exist. 

The salivary glands are highly developed in all castes and 
at every stage of growth. There is a single pair, as well as a 
large salivary reservoir, such as Miller has described (Pl. 19, 
fig. 7). There is an unpaired external opening in connection 
with the labrum. 

[The supra-cesophageal ganglia are situated as in Thysanura, 
with the olfactory lobes anterior, and the fungiform bodies 
posterior as in Termes lucifugus (see figs. 27—383c, fung.). 
The latter are relatively well developed when compared with 
those of Embiide (fig. 34) or Thysanura; there are two 
on either side, or four in all, not well separated from 
each other. As in other insects, they are characterised by 


1 [*Jen. Zeitschr.,’ ix, pp. 256, 257, pl. xii, fig. 42.] 


CONSTITUTION AND DEVELOPMENT OF TERMITES. 273 
the possession of small cells, of which the nucleus stains 
more intensely than in the other nerve-cells. 

Should a characteristic feature of these organs be sought, 
by way of contrast with those of less intelligent insects, it 
will be found in the abundance of these particular small 
nervous cells. | 

The visceral nervous system is well developed, and resem- 
bles that of Blattidz (fronto-labral commissures ; frontal and 
stomato-gastrie ganglia, &c.). The salivary glands are supplied 
(only ?) by branches from the subcesophageal ganglion, coming 
off from those to the labium or lower lip. 

The ventral ganglionic chain possesses six abdominal ganglia, 
and thus agrees with Lespés’ description; the sixth is very 
large, and is in correspondence with the true seventh and 
eighth abdominal segments. 

At the level of the fourth abdominal ganglion—that is, in 
the true fifth segment—there open numerous unicellular glands 
of unknown function. The retro-cerebral gland, which exists 
in Termes lucifugus,' is absent. 


Certain features of the different castes and larval forms will 
now be described. 

[Newly born examples are semi-transparent and almost pure 
white. If the mouth parts are detached the apex of the 
mandibles and the inner lobe of the maxille can just be seen 
with the microscope to be tinged with yellow, a feature which 
cannot be made out by examining intact specimens even with 
a good lens. After a few days these parts acquire a more or 
less pronounced yellow colour, owing to the thickening of the 
investing cuticle, and at this time a yellow line, caused by the 
approximated tips of the mandibles and maxille, may be 
distinguished at the front of the head even with the naked eye, 
though better with a lens. 

This yellow line appears early in forms which develop a 
large head (soldier larve), and is delayed in those of which the 
head remains small. At the time of its appearance the animal 

1 [Vide Pl. 16, figs. 28—33, and description. ] 


274 B. GRASSI AND A. SANDIAS. 


is seen to adopt as food a material which is usually of the same 
dirty yellowish colour as that of more developed examples, and 
is apparent through the translucent abdomen, so that the 
general white colour of the body is blotched with yellow. The 
importance which attaches to this fact will be explained in the 
subsequent chapter. | 

The youngest soldier larve are therefore distinguishable by 
the greater size of the head and the lesser constriction of its 
anterior portion, as well as by the yellow line in this region; 
moreover the thorax and the abdomen, which possesses 
the aforesaid yellowish blotching, appear to be somewhat wider. 

The body of small soldiers (5 mm. in length) does not become 
fully yellow, with exception of the head, which is golden- 
yellow, and the mandibles, which are brown. Their head is 
subglobose, and at first sight these small soldiers greatly 
resemble those of Termes lucifugus, but are distinguishable 
by the indistinctly marked neck. The relative width of the 
head and pronotum varies in soldiers of different sizes, as is 
shown in the plate (Pl. 16, figs. 5, 16). 

The body of medium-sized or large soldiers is golden- 
yellow, much deeper anteriorly, while the mandibles are 
coloured as in the small soldiers. The head is rectangular, 
and longer than broad, subquadrate. At the time of differen- 
tiation all the soldiers, large and small alike, are white, as I 
have stated. 

True kings and queens (6'5 to 7 mm. in length) of recent 
development are black, except for the apical portions of the 
legs and antennze, which are light yellow, and the anterior 
three fourths of the pronotum, which are golden yellow. 
With increase in bulk, at the first moment of the assumption 
of the definitive habit, the white intersegmental linear spaces 
become evident, and are most conspicuous when the royal 
forms have reached the maximum dimensions of which they 
are capable (Pl. 16, figs. 7,11). I may note here that a length 
of 10 mm. for the king and of 14 mm. for the queen is the 
greatest that I have observed. 

These white lines, which are not as yet evident in royal 


CONSTITUTION AND DEVELOPMENT OF TERMITES. 275 


forms as much as two years old (figs. 8 and 9), correspond with 
the interspaces between the first and second, and in succession to 
between the sixth and seventh abdominal somites. The second 
line is usually the widest, but is sometimes similar to the 
others. They occur alike in both sexes, but as the king in- 
creases in bulk less than the queen, they are naturally less 
marked in the former than in the latter sex at the same age 
and never ultimately reach the same degree of accentuation. 

The substitute king and queen are distinguished not alone 
by the almost constant presence of pigment in the compound 
eyes, but also by their pale-yellow colour, which deepens with 
age to a golden tint. 

In these royal forms alike the abdomen becomes enlarged, 
and I have found' examples of both sexes almost equalling 
the largest true kings and queens in size (PI. 16, figs. 12, 13). 
With the natural dilation of the abdomen, certain interseg- 
mental spaces, namely, those which are indicated above for 
the true king and queen, appear as whitish lines. 


I conclude by mentioning that I have looked in vain for 
any difference between the legs of the various forms. The 
tarsi are normally four-jointed; and the apex of the tibia 
possesses three spines, toothed on one side, and common to 
all the legs (Pl. 16, fig. 24). A plantula is present in the 
true imagos, but is wanting in all the other members of the 
colony, including the substitute kings and queens. 

These observations probably hold good for all species of the 
genus Calotermes; I infer this from the many incomplete 
notes which are scattered in the pages of various writers. 


4. Relative Numbers of the Castes. 


A nest contains a single king and queen, neither of which 
can be normally wanting. They may be either true or sub- 
stitute forms. At times one of the pair, either king or queen, 
is a perfect and the other a substitute form; and in such 
cases we have the strange phenomenon of courtship between 
a black and a yellow example. 

The existence of but a single royal couple is an assured 


276 B. GRASSI AND A. SANDIAS. 


fact, but it is not uncommon for two or more nests of Calo- 
termes to co-exist without (at least for us) well-defined limits 
in the same stem, which therefore appears to contain more 
than a single royal pair. In connection with this condition are 
certain facts which prove the existence of the terrible jealousy 
which is so remarkably shown, e. g., between separate families 
of the hive-bee. But for the sake of convenience I shall defer 
them to a succeeding section on Habits. 

When a colony is deprived either of king or queen, or of 
both, it furnishes a certain number of substitute royal forms, 
of which only one, if a single true form is missing, or two, 
that is, a pair, if both are absent, is called to the throne. 

In every nest, forms of different ages are always to be 
found; and as a general rule small individuals are more, or 
not less, numerous than large. 

Soldiers are relatively scarce, not more than from two 
to four being found in nests of from eight to fifteen inhabit- 
ants. In large colonies they exist in the proportion of one to 
‘very fifteen or twenty examples. 


5. Seasonal Variations in the Colony. 


It must be recollected, as a general rule, that the develop- 
ment of Calotermes is arrested during the winter months, 
that is, from the middle of November to the middle of April; 
and this condition explains the absence of individuals in pro- 
cess of moulting during that period. 

In these months the nest contains eggs, sometimes a hun- 
dred or more, which are always in the gastrulation stage. 
Their development remains stationary during the greater part 
of May, to reeommence towards the end of the month, so that 
newly hatched young are to be found from about the 10th of 
June till the end of July,a few eggs, from five to ten, hatching 
every day. In July it is evident that the number of young 
larvee increases, so does that of the eggs diminish until 
their final disappearance. Towards the end of July it is 
difficult to find a nest which still contains eggs, proving 
that oviposition is suspended at this time. The same holds 


CONSTITUTION AND DEVELOPMENT OF TERMITES, 277 


good for the interval from November to the middle of May; 
while the opposite condition exists in the second half of 
-May, in June, September, and October. Consequently 
oviposition remains suspended during winter and a large 
part of the summer. At the end of October many eggs 
are still to be found in the gastrula stage, and it follows 
from the preceding statements that they remain thus during 
the winter; but at the same month a certain number can be 
found in an advanced stage of development. From this, and 
from the fact that many new-born larve are to be found in 
that month, I can confidently state that the eggs laid in 
September (and perhaps in the early part of October) develop 
immediately without need of hybernation.! 

At all seasons, except from about April to the middle of June, 
larvee, with the features of being newly hatched, are to be found 
in the nests of Calotermes. The larve which are born in 
the second half of October make no progress in development 
until the following April. 

Larve in stages of development which succeed those with 
the characters of recent hatching are to be found at all times 
of the year. 

Nymphs are to be found in every nest during nearly all the 
year, being absent at most only in August and September. 

The perfect insects develop from July to October, a few 

_stragglers appearing in the spring. Their swarming is quite 
different from that observed in bees, and they leave the nest a 
few days after they become black. [As development is not 
simultaneous, they swarm in small groups of at most thirty 
examples, and occasionally singly or in pairs. A colony may 
therefore swarm, so to speak, ten or twelve times in all, from 
July to October. ] 

This fact explains the existence in many nests of a few 
winged examples still at the end of the swarming season, 
that is in October. 

1 TI must admit that there is a certain gap in my observations in the 
summer months, when I am unable to reside at Catania owing to the exces- 
sive heat. 

VOL, 39, PART 3.—NEW SER, T 


278 B. GRASSI AND A. SANDIAS. 


[As a rule a certain number of both sexes become mature at 
the same time, and the males invariably take flight two or 
three hours later than the females.| Further details on the 
subject of swarming will be found in the later section on Habits. 

Soldiers and soldier larve are to be found the whole year 
round, but only small soldiers are present in nests under two 
years old. 


6. Duration of Development, of Life, &c. 


These matters are very difficult to ascertain, but the follow- 
ing facts are certain : 

1. Owing to the interruption of development from November 
to April, individuals in the same stadium may be of different 
ages. 

2. Eggs which pass the winter in the gastrula stage will 
give rise in the following summer at most to soldiers with 
15-jointed antenne, or to larve with similar antennze, and with 
or without very short wing-rudiments. These larve become 
perfect insects and swarm in a later summer. 

These conclusions are the result of minute examinations of 
numerous small or orphaned nests, &c., some of which I shall 
proceed to record. 

By searching in the situations which have been described 
as the points of origin of new nests, from August to April, 
some two to twenty perfect insects, with the wings reduced to 
stumps, may easily be found; the majority are grouped in 
pairs, male and female, each of which may be accompanied by 
a few eggs. These pairs are recently formed, and, if originally 
numerous, are subsequently reduced to one or two (see the 
succeeding section on Habits). 

A pair is established, say, in August, and remains to the 
end of the autumn with only fifteen or twenty eggs; twelve 
months later, at the end, that is, of the following autumn, it 
will be surrounded with fifteen, twenty, or at most thirty 
young of different ages, the most advanced being large soldiers 
with at most fifteen antennal joints, or larve with a similar 
number, all pilose, and with a very slight indication of wings, 


CONSTITUTION AND DEVELOPMENT OF TERMITES, 279 


I must regard certain nests which may be found at the end 
of October, as having arisen from pairs which had only begun 
to lay eggs in May of the year following the swarm-period. 
In these the number of inmates is less than in the nests pre- 
viously referred to, and no larva possesses more than fourteen 
antennal joints, the third being glabrous, and the constric- 
tions between the 3rd, 4th, and 5th respectively being ill- 
marked. The wing-rudiments of these larve are not yet 
discernible, and the soldiers of such nests do not possess more 
than twelve antennal joints, and are small. 

The hypothesis that such nests may have originated from 
royal pairs disclosed only during the preceding summer, 
instead of the previous year, and consequently only three or 
four months old, is quite untenable ; as is indicated by the 
fact that nests are never found to contain only newly hatched 
or little-grown larve at the end of autumn, as must of neces- 
sity be the case if young can be born in the year in which the 
nest is founded. In short, the fact remains that at the end of 
autumn certain nests are found to possess undeveloped eggs 
only, whereas others already contain small soldiers, and larvee 
with fourteen antennal joints: nests intermediate between 
these two classes are entirely wanting, though they should be 
present if the above hypothesis were correct, inasmuch as the 
swarm-period lasts from the end of July till that of October. 

In March, 1891, I observed an absence of nymphs in the 
examination of several nests orphaned two years previously 
(occasionally they were present in very small numbers). 
These nests contained ova and forms in all other stages. It 
therefore appears that destruction of the royal pair in spring 
results in the suppression of swarming during July and 
August of the second succeeding year, but in those months 
only, because examples in which the wings are only just 
indicated in March will have become imagines by September 
and October. 

Nests orphaned between February and June sometimes con- 
tain no eggs, and usually no new-born larve in the following 
winter. 


280 B. GRASSI AND A, SANDIAS. 


It is consequently evident that the development of substi- 
tute royal forms proceeds very slowly. 

The data furnished by orphaned and young nests, and by 
many other observations which I have made, indicate that 
perfect and fully winged examples are not obtained before 
August from eggs laid in July of the preceding year. In 
short, Calotermes passes part of two years in the larval 
and nymph stages before taking flight. The soldier may com- 
plete its development in the same year in which it is hatched. 

It is difficult to say anything about the duration of life. 
I must deny the existence of any particular season when the 
soldiers die off, as Lespés has claimed for Termes. The 
life of the king and queen may be estimated at four or five 
years at least. 


7. Situation of the Different Forms in the Nest. 

I have already mentioned that Calotermes does not 
possess a royal chamber. The king and queen, whether true 
or substitute, usually remain in close proximity in the heart 
of the nest, where the inmates are always most crowded. 
They readily change their situation. The eggs are mostly 
near them or a little way off, and are never heaped together. 

Larve, nymphs, and winged forms, if present, are irregu- 
larly commingled ; but larve newly hatched tend to cluster 
together. The soldiers also are irregularly scattered, but a 
few are often found in close attendance on the royal couple. 
The soldiers are generally the first to make their appearance 
when a nest is opened, 

Substitute forms in process of development occur separately, 
or in groups of two to four in different parts of the nest. 

It thus follows that the component members of a Calotermite 
colony have no special situations in the nest. 


8. Certain Habits. 


I propose to deal here only with those habits which could 
not appropriately be dealt with in Section 3. 

Calotermites work, feed, and rest indifferently by day or 
night. When resting, they remain motionless without adopt- 


CONSTITUTION AND DEVELOPMENT OF TERMITES. 281 


ing any special attitude. They are fond of darkness, and 
when kept in a tube occupy the portions furthest from the 
light. They certainly work more actively in the dark, but are 
capable of doing so even when exposed to daylight. 

Oviposition continues both by day and night. 

Swarming takes place in the morning, usually from 9 to 
12; and it must therefore be recognised that Calotermes 
no longer avoids the light when it has reached the perfect 
stage. 

Before swarming the winged insects collect habitually in a 
spot which careful observation shows to be in the neighbour- 
hood of an exit-hole, putting the nest into communication with 
the outer air. 

Swarming takes place through this hole. The insects issue 
by ones or twos, so that the twenty or thirty examples ready 
to take flight emerge in perhaps a quarter of an hour. 

Once outside the nest, they run upwards for a few metres if 
the locality admits of it, and then only do they take wing. In 
a room they evidently fly towards the light, and if a wind is 
blowing they follow its direction. Some soon become tired and 
settle on the trunks or branches of neighbouring trees; the 
majority become lost to sight, but many certainly end by 
alighting on trees. It is here that they group themselves into 
pairs, the males and females of which must frequently be 
derived from separate nests, for, as I have mentioned, the sexes 
swarm separately ; this acts as a safeguard by which Calo- 
termes habitually avoids in-breeding. 

The winged forms have not been observed to pursue cach 
other in the curious way which will be spoken of under 
Termes. 

I may describe more precisely the manner in which the 
males and females come together when settled on a tree. The 
winged forms habitually search for a decayed spot, and when 
found they dig it out, after losing their wings, in order to bury 
themselves ; it is in this act of excavation that the meeting and 
subsequent pairing take place. 

The wings may be shed, merely by striking against an 


282 B. GRASSI AND A. SANDIAS. 


obstacle, or by becoming damp and adhering to some spot, 
while the insect continues to move about. 

But if not favoured by chance the Calotermite rids itself 
of its wings, as the following observation shows. Four perfect 
insects, which had recently left the nest, were captured by 
hand after flying about a room for some time, and were put 
under a piece of rotten wood. They had hardly settled down 
before they began to strip off their wings by resting their tips 
against some projecting corner of the wood and then moving 
backwards a little, so that the wings buckled towards the base, 
broke, and dropped off. When rid of them they began to 
gnaw the wood, at first along, and then across the grain; each 
worked by himself and at some distance from his fellows. 
Subsequently several chance encounters took place between 
them; they threatened to bite each other, and then ran off in 
different directions. They were of the same sex. If they had 
been of different sexes they would certainly sooner or later 
have copulated. 


In the colony of Calotermes all members work for the 
common welfare. The soldiers serve for defence, but as a rule 
only when some important enemy has to be combated; at 
other times nymphs and possibly the older larve assume the 
task. 

Cremastogaster scutellaris, Ol., which is abundant 
here, is one of the most formidable enemies of Calotermes, 
near which it makes its own nests. This ant enters the termi- 
tarium to massacre, whereas its own nest is never invaded by the 
Calotermite soldiers. If some examples of Cremastogaster 
are put into a tube containing a Calotermite nest, the follow- 
ing phenomena can readily be followed. The soldiers place 
themselves with gaping mandibles, waiting for any enemy that 
may come within reach. They then snap their jaws rapidly, 
shearing off antenne and legs, tearing the abdomen, or even 
cutting the ants in two at the level of the abdominal petiole. 
The soldier’s mandibles are seen to act like extremely sharp 
shears. 


CONSTITUTION AND DEVELOPMENT OF TERMITES. 285 


The ants themselves attack the Calotermites indifferently, 
but habitually avoid the heads of the soldiers, only daring 
occasionally to attempt to lop off their mandibles. As a rule 
they attack the soldiers from behind by biting the abdomen, 
and to protect it the soldiers creep under pieces of wood so as 
to leave the head alone free. 

If a few Cremastogaster and a larger number of Calo- 
termites are put into a tube, peace is usually concluded after 
about an hour’s conflict, with a certain number of dead and 
wounded on both sides. The ants take up a position in one 
part of the nest and the Calotermites in another. 

[Besides the ants, the soldiers of Termes are terrible 
enemies, but being small they are easily cut in two through 
the thorax by the Calotermite soldiers. The workers of 
Termes are much less dangerous enemies than the soldiers of 
that species. One, put into a tube-nest of Calotermes with 
soldiers, was at once placed hors de combat by a nymph, 
which cut off part of the buccal apparatus. Then sundry large 
larvee and other nymphs hurried up, bit off its legs, and tore 
its abdomen until the viscera protruded. Thesoldiers took no 
part except towards the end of the struggle, when one gave it 
a bite. Similar observations have been made several times, 
and show, as we said before, that the soldiers purposely reserve 
themselves for more important foes. 

Another Termes worker was put into a Calotermes 
nest containing no soldiers. The inmates took flight, pro- 
bably terrified by the knowledge that they were unprotected 
by soldiers; and the Termes succeeded in throwing the 
nest into confusion, until after some lapse of time a nymph 
plucked up courage to bite its abdomen, and thus killed it. 

A substitute queen of Termes lucifugus was introduced 
into another tube containing a Calotermite nest. A soldier 
promptly despatched her by decapitation, and then only did 
the nymphs intervene to tear the body. During the rest of 
the day the soldier never moved from the spot where he had 
killed the queen. | 

Besides protecting the nest the soldiers fulfil other duties, 


284 B. GRASSI AND A. SANDIAS. 


such as that of carrying the young and eggs on their 
mandibles. These organs are useless for gnawing wood, and 
their possessors therefore remain idle for hours together, 
while the rest of the colony is in full activity. Except for 
these tasks of defence from great dangers and wood-gnawing 
and consequently excavation of galleries, it is, so to speak, an 
absolute rule that all labour necessary for the community can 
be undertaken by any of its members. 

Newly hatched larve can be seen carrying a fragment of 
wood heavier than themselves. All forms except the soldiers 
excavate galleries, provided that their mandibles are sufficiently 
strong, which naturally is not the case just after birth or after 
a moult. Both king and queen, whether true or substitute, 
gnaw up wood and transport excreta, eggs, or wood-meal. 

Oviposition appears to be a very laborious process; in one 
case the egg was not extruded until an hour, and the succeeding 
egg until half an hour after appearance at the vulva. Once I 
saw a soldier assist the queen by raising and gently stroking 
her abdomen, but as a rule she lays without assistance. 

Hatching is effected without need of any assistance from 
inmates of the colony. The chorion is tolerably thick, and 
the eggs can be kept in a watch-glass without drying up, and the 
process of hatching observed. Moulting also is accomplished 
as arule without assistance. 

If the eggs and young are exposed by opening a nest, it is 
striking to see how the other inhabitants disperse without 
paying them the least attention, in contrast to the behaviour 
of ants; and yet one must recognise that the colony is as 
deeply interested in their welfare as ants are in that of their 
own offspring, and that it scatters simply because it is panic- 
stricken. ‘This may be proved by shaking a small nest made 
of loose pieces of wood in a glass tube or jar. For a moment 
all the inmates are thrown into disorder by terror, but they 
quickly recover themselves, become persuaded that it was 
merely an earthquake, so to speak, and devote themselves to 
the restoration of order by carrying the eggs back to their 
place at the bottom of the jar, and removing the young on the 


CONSTITUTION AND DEVELOPMENT OF TERMITES. 285 


mandibles of the soldiers, or in the mouth-parts of their other 
fellow-inmates. 


It must be recognised that Calotermites are perfectly well 
aware of the presence or absence of the royal pair, and they 
start about providing fresh ones directly they are orphaned. 
The following observation is important in this connection. 
A nest was divided, say, into three smaller nests, in one of 
which the royal pair was retained; they were put into separate 
tubes, kept together and uncorked in the same waistcoat 
pocket. After three or four days larve of substitute royal 
forms were found in them, except in the one containing the 
royal pair. The colony in the tree from which the original 
nest was taken had not turned their attention to raising sub- 
stitute forms, though they were 30 or 40 centimetres distant 
from any king or queen; whereas those in the tubes, though in 
close proximity to a royal pair, had at once begun to provide 
substitutes. It follows that the existence of the royal couple 
is certainly not perceived by means of any odour which they 
emit; and this affords us another marked contrast with what 
is observed in the case of bees. 


The copulation of Termitidze has been extensively dis- 
cussed ; some writers believe it to be accomplished in the 
open air, but the more general opinion is that the presence of 
the king in the nest shows that it takes place there, and is 
repeated from time to time. I regarded the latter supposition 
as possibly correct directly I found that the king is invariably 
present in the nest, and that his bulk increases, though to a 
less extent, concomitantly with that of the queen. 

At last, on April 17th, 1891, about 11 a.m., I detected the 
king and queen in coitu in a glass jar containing a small 
Calotermes nest. The pair appeared to be about three years 
old, judging by their size. They stood end to end inastraight 
line with the tips of their abdomens applied to each other ; 
their attitude was normal, with the dorsum uppermost, except 
for a slight upward flexure of the apex of the abdomen. After 
some half-minute they separated, perhaps owing to the ex- 


286 B. GRASSI AND A. SANDIAS. 


posure to lght. The contiguous parts of their abdomen 
exhibited a white substance which, when the king and queen 
became detached from one another, remained adhering under 
the hinder extremity of the queen. 

It is therefore certain that connection takes place in the 
nest and is repeated at intervals; and all my observations 
satisfy me absolutely that it cannot be accomplished in the 
open air, and is practicable only after loss of the wings. 

[I must regard certain relations which were observed to 
subsist between two substitute forms still far from maturity, 
because they were only about a fortnight old, as amatory in 
nature. One stood still while the other gradually approached, 
and when sufficiently near brought its antenne into contact 
with those of its fellow; it then quickly retired for some dis- 
tance, and returned later to repeat the pastime. This took 
place at least four times. On the fifth occasion the one which 
had been standing still, made movements as if to detain its 
companion ; they then remained together, and very rapidly 
stroked various parts of each other’s body, especially the apex 
of the abdomen. Their position during the reciprocal palpation 
of this part was almost that assumed during the act of coitus. ] 


Several writers have mentioned the convulsive movements 
characteristic of Termitide. These movements or quiverings 
are easily observed in Calotermes, and may be repeated 
periodically at very short intervals, almost at the frequency 
of the pulse-rate. 

In the act of quivering the tarsi are held motionless, while 
the body is shaken forwards and backwards; there may bea 
simultaneous slight lateral or vertical oscillation. Sometimes 
an example may stop whilst running, in order to quiver one or 
more times. 

Occasionally these convulsive movements are repeated a few 
times only, and then stop altogether; but at other times they 
recur after a few seconds or at most a few minutes’ rest, and 
may thus be continued sometimes for hours with many similar 
intervals of rest. 


CONSTITUTION AND DEVELOPMENT OF TERMITES. 287 


In the intervals between successive convulsions the insect 
remains still or progresses for a short distance only. These 
movements are executed by all members of the colony ex- 
cept those newly hatched. 

I have satisfied myself by careful observation of the phe- 
nomena exhibited in tube-nests that these convulsions serve 
as a cry to summon help or give alarm, or as a lament; in 
short, as a mode of intercommunication. 

When the insects are suddenly annoyed or disturbed by 
any cause, such as a rough shake of the tube, its change from 
a vertical to a horizontal position, sudden illumination, or the 
prolonged effect of too bright a light, all the members of the 
colony begin to quiver, except those that are running briskly 
about in search of a better situation. Moribund examples 
sometimes perform these movements at intervals of a few 
minutes, in some cases for a couple of hours. 

[If a few Termes lucifugus are put into a tube contain- 
ing the usual little nest of Calotermites, some of the latter, 
evidently excited by the very rapid movements of the intruders, 
run off rapidly in the opposite direction, and stop at intervals 
to quiver with much more energy than usual. This phenome- 
non is exhibited alike by larve, nymphs, and soldiers, but the 
latter quickly turn back to face the supposed enemy. It seems 
evident in such a case that those which are first aware of the 
presence of the Termes quiver violently in order to alarm 
the population. | 

[Sometimes the convulsions of an insect in the neighbourhood 
of the cork are quickly followed by the exit therefrom of 
others which have been burrowing inside it. Such cases 
appear to prove the utility of the action as a mode of 
signalling. | 

Members of the same nest clearly recognise each other. 
[As a proof, a few examples are removed from the nest and 
returned after five or six hours. The population is not dis- 
turbed or alarmed, and does not scurry about at their re-entry. 
It is a possible objection that these specimens have immediately 
recognised the nest, and therefore create no disorder on their 


288 B. GRASSI AND A. SANDIAS. 


return; to meet it a new nest was provided, from which 
certain individuals were excluded. They were introduced a 
few hours after the fresh nest was put in order and quiet, but 
they caused no disturbance although it was unfamiliar to 
them. As a control experiment, a few strange Calotermites 
were put into the same nest; the inhabitants took fright at 
once, and scattered in different directions. But after a little 
time all became quiet, and no fighting was actually observed. 
We may therefore conclude that examples taken from different 
nests readily fraternise ; and this applies to the soldiers as 
well as to the larve and nymphs. ] 

[I should mention with respect to the soldiers that if too 
large a number is added to a nest, these supernumeraries, as 
they may be called, are found to be killed or eaten one by one 
during the next few days.] 

[The following observation was made on one occasion. Half 
a dozen strangers, including a substitute royal example, were 
added to one of the usual nests in a tube. The royal specimen 
showed signs of hesitation and remained in the same spot, 
merely turning round and round, and straightening its legs as 
much as possible, as though to raise itself above the level of 
its neighbours. When an inhabitant of the nest approached 
and touched it, it drew back at once; the other did the same, 
so that they separated as rapidly as if they were stung. Some 
hours later the new-comer became quiet, and several inmates 
of the nest approached it to caress the antenne, &c. I must 
mention that the nest was orphaned. ] 

[We have already stated that Calotermites furnish a certain 
number of substitute forms when deprived of the king and 
queen. Why then is only a single royal pair to be found ? 
One day one of three substitute forms, which inhabited a 
small nest, was seen rapidly to pursue another with gaping 
jaws, with the evident intention of killing it, in which it was 
unsuccessful. Next day the nest contained only two such 
forms ; very probably the one which was pursued the previous 
day had been killed.] 

[Two true royal forms, a king and a queen, were introduced 


CONSTITUTION AND DEVELOPMENT OF TERMITES. 289 


into one of the usual nests, constructed the day before and 
containing only nymphs. One of the pair began at first to 
advance hesitatingly, while the other stood perfectly still ; but 
shortly one after the other gradually moved forwards. 

The nymphs then began to scatter in different directions in 
the endeavour to keep as far away as possible from the royal 
pair; these gradually retreated to the bottom of the tube, 
where they were left alone. But now and again a nymph 
approached with open jaws, and savagely bit one of the pair 
in the head or thorax, getting bitten with equal fury in 
return, and consequently retreating. 

Next day the royal couple was still left unattended, but was 
no longer actually disturbed. Then a second similar pair was 
introduced ; at once the nymphs all became greatly excited, 
attacked them and promptly reduced them both to helpless- 
ness by biting off their legs. The second pair was alive two 
days later, and the queen of the first pair was seen to attack 
the second king with open jaws; he retreated by dragging 
himself along the tube, being unable to run through the loss 
of all his legs. A day later he was dead, and the first queen 
was observed nibbling the stumps of his legs ; on the following 
day the queen of the second pair was also found dead. 
Several repetitions of this experiment were made, and always 
gave similar results. Whenever two or three supernumerary 
royal pairs are put into a nest, a single pair is all that can be 
found at the end of a few days.] 

I infer from all these facts that Calotermites exhibit those 
phenomena of jealousy and hostility which are so well known 
in bees ; however, their manifestation is less rapid. 

The observations just recorded, and many other facts 
omitted for the sake of brevity, lead us to the fundamental 
conclusion that the colony of Calotermes tolerates 
neither supernumerary royal examples nor super- 
numerary soldiers. Both one and the other are 
slaughtered. On the one hand, then, the colony can 
provide itself with royal forms or soldiers when they 
are required (vide infra); on the other hand, it rids 


290 B. GRASSI AND A. SANDIAS. 


itself of them when they are over-abundant. These 
facts imply the possession by Termites of a faculty 
which may be termed a sense of proportion or nume- 
rical sense. 

By way of conclusion, I must add that though I have had 
Calotermites under observation for several years, I may still 
have failed to detect a great part of their marvellous instincts. 
This is owing to two circumstances: 1, they are often slug- 
gish ; 2, when a nest is opened the population is thrown into 
such a state of astonishment that it usually does nothing but 
run away and give signals of alarm. And almost all the 
observations here recorded are due to the method of employ- 
ing tube-nests. 


Tue CoLony oF TERMES LUCIFUGUS. 
1. Situation and Nest. 


Several writers, particularly Lespés, have published observa- 
tions on this subject ; but I shall summarise merely my own 
investigations. 

Here in Sicily Termes lucifugus usually inhabits plants, 
rarely furniture or the wooden beams of buildings. It is most 
common in plants of which the stem or tap-root measures at 
least three quarters of a centimetre in diameter; but when it 
has once entered a stem or large branch it will pass on into 
the very smallest twigs and roots. It mines irregular galleries, 
and often avails itself of old beetle-burrows (Bostrychus, 
&c.); Calotermes does the same thing. Like that species, 
Termes leaves the outer layers untouched, so that a trunk 
may be completely mined out and yet appear sound, while the 
hand can easily be thrust into it by breaking through the thin 
intact superficial layer. 

Owing to the extreme tenuity of this layer in the smallest 
roots invaded, the galleries may appear at first sight to be 
tunnelled directly in the ground ; but I have never been able 
definitely to establish the existence of such a mode of con- 
struction. 


CONSTITUTION AND DEVELOPMENT OF TERMITES. 291 


I have never found Termes in the orange, lemon, or vine; 
but, unlike Calotermes, it is very common in cactus. It 
continues to live and flourish in perfectly dead and dry wood, 
even when employed in the construction of roofs, doors, 
furniture, casks, &c. 

It is characteristic that whereas Calotermes confines itself 
to the original host plant, Termes successively attacks fresh 
plants or timber, and may thus pass from the furniture of a 
house to a tree, or the reverse, or from one piece of furniture 
to another. The following facts are related to this peculiarity : 

A few workers of Termes, which are probably explorers, 
are sometimes found to make a spontaneous appearance in the 
open air and in broad daylight. 

On other occasions Termites travel by availing themselves 
of natural cracks, e. g. in lava, or by hollowing out small dry 
rootlets underground, or in reeds by means of the tubular 
lumen. Frequently they have recourse to galleries, which are 
quite distinct from the burrows made in wood, because they 
are fabricated or built up, so to say, by the insects themselves. 
They are therefore not merely miners, but also builders. 
These galleries are of two kinds, tubular or D-shaped in section. 
Galleries are usually constructed in tubular form in the absence 
of any suitable base on which to build them; if such be present 
the gutter-shaped gallery (semicircular in section) is resorted 
to, but even in this case that part of the base which is enclosed 
between the walls of the tunnel is incompletely cemented over. 
Galleries of the latter kind are made by preference in the angle 
of junction of two walls (Pl. 17, fig. 40). 

Gutter-shaped galleries may reach the considerable length 
of eight or more metres. In the choir of the principal church 
of Pedara I have seen such a tunnel leading from the stalls to 
a crack in the wall adjoining the wooden ceiling, in which the 
crack disappeared. The insects travelled between the choir 
and the ceiling by means of the tunnel and then of the crack, 
which they made use of apparently without modification. In 
such cases it is sometimes difficult to distinguish the track of 
the insects right up to the point where the gallery leaves off. 


292 B. GRASSI AND A. SANDIAS, 


Similar galleries also occur in the spacious ceiling of the 
Benedictine church at Catania. 

In the building of the Botanical Garden at Catania the 
Termites have invaded the benches of the school, the book- 
cases and window-frames, &c. Here I found two gutter-shaped 
tunnels over 40 cm. in length on the wall of a room; they 
began from the timber invaded by the insects, and ran along 
the surface of the wall, to finish at a point where no fissure 
was discoverable. 

The direction of the gutter-shaped galleries may vary as 
required ; they are usually vertical or oblique, less often hori- 
zontal, and they may branch in various ways. The tubular 
galleries are mostly short, rarely exceeding 5 cm., and are 
narrower than an ordinary pencil. They serve to connect two 
gutter-shaped galleries or two portions of a nest; and a 
gutter-shaped gallery may become tubular for part of its 
course. 

On one occasion I found a much flattened tubular gallery 
about 15 mm. in width, 4 mm. in depth, and 5 cm. in length, 
and somewhat irregular. This fragile structure was suspended 
from the ceiling, and contained a certain number of Termites; 
there were some apertures at its free extremity (Pl. 17, fig. 38). 
At other times I have found much shorter flattened galleries 
hanging from the ceiling (Pl. 17, fig. 39; Pl. 18, fig. 15). 
The purpose of these structures escapes me, but, recollecting 
the excrescences built on plants by certain tropical Termites, 
I suspect them to be a rude attempt at a nest. A similar 
explanation may perhaps be advanced for the galleries found 
at the Botanical Garden of Catania, which terminated in a 
free extremity. 

Both forms of gallery usually have a diameter from 2 to 6 mm. 
Their lumen varies at different points, and is generally large 
enough to allow several insects to pass at once; the internal 
surface is tolerably smooth, while the outer is irregular and 
rugged. Their colour is variable, but is usually of a chalky- 
grey tint. They are composed of fecal and disgorged matter, 
and of triturated wood. When connected with a plaster cor- 


CONSTITUTION AND DEVELOPMENT OF TERMITES. 293 


nice they contain distinct scattered white specks, which are 
fragments of plaster. 

These Termite galleries are invariably very light, porous, 
and friable. When the Termites meet with large empty spaces 
while in process of enlarging their nest, they may fill them up 
by building a complicated labyrinth, as many ants are known 
to do (Pl. 18, figs. 14, 16); and they readily adapt such pro- 
jecting pieces of wood as the space may contain to what they 
are building by covering them over with the materials of 
construction, or cementing them suitably together. 

Excellent specimens of both classes of gallery can be obtained 
by putting Termites into a glass jar half full of broken-up 
cactus-phylloclades, and closed with a cork or merely witha 
sheet of paper (Pl. 18, fig. 17). 

If many such nests are formed, one or more can generally be 
kept alive for six, eight, or more months. I have published 
elsewhere an account of one of these nests, which I repeat 
textually. 

“For eight months I have kept a colony of Termes luci- 
fugus without king or queen in a jar half full of crushed-up 
phylloclades of Opuntia, and closed with a sheet of paper tied 
over the mouth instead of with a bung. ‘The jar holds three 
litres, and its mouth is wide and polished. 

“ At the beginning of April the Termites were seen to have 
settled in the bottom layer only of the rubbish, and the 
remainder, some 7 cm. in thickness, was quite uninhabited. 
It was not till the 20th of May that a few specimens appeared 
init. Some days later a semicircular gutter-shaped tunnel was 
found adhering to the walls of the empty part of the jar (the nest 
material occupying barely a half). This tunnel put the rub- 
bish into communication with the paper cover, which presented 
a small aperture large enough to admit the body of a Termite 
at the extremity of the tunnel. 

“ All kinds of forms in the colony (larve, nymphs, soldiers, 
workers, and winged adults) went backwards and forwards by 
this gallery, which in its greater part would only allow room 
for a single individual at a time, but which widened here and 

VOL. 89, PART 3.—NEW SER. U 


294, B. GRASSI AND A. SANDIAS. 


there so that two could pass simultaneously. This primary 
tunnel, as it may be called, was the only one constructed during 
five days, but it was made to bifurcate by the addition of a 
lateral branch. This came off at an acute angle from near the 
middle of the primary gallery, and ran upwards to terminate 
on the margin of the vessel, where the polished surface began. 
During the next few days the uninhabited part of the rubbish 
was tunnelled with numerous burrows, opening on its upper 
surface by several small holes, some of which were continued 
upwards by tubular chimney-like galleries of different heights 
(the largest measured 5 cm.), vertical or oblique in direction, 
and varying in width, usually just capable of being traversed 
by a single Termite. The free ends of these chimneys were 
sometimes closed, sometimes open, and in that case an inmate 
of the colony would not infrequently peep out as from a watch- 
tower (Pl. 18, fig. 13). 

“ Swarming took place on June Ist. On the following days 
the Termites did very little work, to all appearance. They 
lengthened a chimney and built another gallery like the first, 
but not connected with it and not reaching the lip of the 
vessel ; and lastly, they made an incomplete extension of the 
primary tunnel by carrying it along the junction of the lip and 
the paper. The latter exhibited two fresh holes at this point. 

“One day I destroyed the newly formed section of the 
primary gallery by taking off the paper, which I purposely 
put back so that it did not accurately cover the margin of the 
vessel, but left a gap; next day the primary gallery was con- 
tinued horizontally outwards for about 1 cm. from the margin 
of the vessel ; this new portion did not run on the surface of 
the glass, and, like the chimneys, was tubular instead of being 
semicircular (Pl. 18, fig. 17). A day later it was dismantled 
and destroyed by the Termites; this did not surprise me, 
as I had already observed them make and unmake portions 
of galleries. 

**On June 20th the tops of the chimneys and openings were 
closed up, the galleries were unoccupied, and the whole of the 
colony had once more retired to the bottom of the vessel.” 


CONSTITUTION AND DEVELOPMENT OF TERMITES. 295 


This colony was unfortunately killed at the end of June by 
over-dryness. 

The structures built before swarming were certainly made 
chiefly for the purpose of facilitating that procedure. But 
similar erections, with the exception of the chimneys, can be 
obtained at a time remote from the swarm-period, or after it 
has elapsed, as in the case just described, where the Termites 
evidently found the vessel unsuitable, and attempted to 
abandon it, but were unsuccessful, and therefore all died. 

I state that they endeavoured to quit the jar because on 
other occasions I have observed a general migration from a 
vessel which had been inhabited some time, the contents of 
which were found on examination to be unsuited to their 
welfare (too damp, mouldy, over-dry, &c.). 

Termes does not line the galleries hollowed out of wood 
with excrement. The wood-meal produced by burrowing 
varies in colour with the material from which it is derived. 
Burrows are very easily made in certain cactus-phylloclades 
which keep their white colour after death, but become very soft ; 
the inner surface of these burrows and the dust removed are 
both white. 

Termites customarily select the softer parts of the wood, but 
when these are all destroyed they attack the hard parts, and 
thus form spacious chambers, openings, &c. 

Temperature and humidity are as important for Termes as 
for Calotermes; and as far as is known at present the 
geographical distribution of these forms coincides.' But it 
must be remembered that.the former can flourish at a some- 
what lower mean temperature than the latter, so at least I 
imagine, for the following reasons : 

Oviposition commences as early as the beginning of May. 
In October and November forms in intermediate stages between 
the nymph and the larva with the earliest indication of wings 


1 (Termes, however, extends farther north, occurring in France at 
Toulouse, Bordeaux, Rochefort, and La Rochelle, and in Italy in Tuscany 
and Venetia, the present writer having lately found a winged example at 
Venice.—W. F. H. B.] 


296 B. GRASSI AND A. SANDIAS. 


are very scarce; they begin to appear in December, and 
become abundant in January and February. The nymphs 
accomplish the imaginal transformation by April or May. 
But the fact that examples with the characteristics of those 
recently born are to be met with in spring, though the nest 
may have contained no eggs since the month of September, 
shows that the development of the earliest stages is arrested 
during the winter, as in Calotermes. Termes, it must be 
noted, requires a less degree of moisture than Calotermes, 
and can therefore live with comfort in dry and seasoned timber, 
and in desiccated portions of trees, &c. 

During the warmer months they bury themselves deeper and 
deeper in dead roots, so that their nests appear to be depleted, 
and it becomes difficult to procure complementary royal forms, 
eggs, new-born larve, &c., without digging to a great depth. 

Termes and Calotermes often share the same tree, but 
the former habitually confines itself to the dead and drier, the 
latter to the moister parts. But it will readily be understood 
that there is no sharp demarcation between these two regions, 
and therefore none between the two colonies. 


2. Number of Individuals in the Colony. 


It is practically impossible to make any accurate estimate 
of the limits of a nest of Termes, as will be seen later. How- 
ever, a single tree, which certainly does not harbour more 
than one nest, will contain at times as much as a litre of Ter- 
mites. Asa rule, the offspring of one nest extends to several 
trees, and the population of a single colony may therefore be 
reckoned as upwards of two litres—that is to say, very many 
thousand individuals. 


3. The Different Castes (Plate 17). 


The society of Termes lucifugus differs widely from that 
of Calotermes, or of such other Termitidze as have hitherto 
been adequately studied. 

Its characteristic feature is the invariable absence of a true 


CONSTITUTION AND DEVELOPMENT OF TERMITES. 297 


royal pair,—that is, a pair derived from winged imagos, which 
have lost their wings (except the stumps). This statement of 
mine will appear bold, but I have examined thousands of nests 
during a period of about seven years, and am in a position to 
make it without the least fear of contradiction. Small colonies, 
founded by a true royal pair, are to be obtained only by arti- 
ficially enclosing winged Termes in glass jars partly filled 
with wood. Nothing of the kind is ever found in nature. On 
one single occasion I lighted on a true royal pair, though 
without eggs, in January,—that is, about six months after 
swarming. 

For the present we may leave out of sight these artificial 
nests in glass jars to consider those found under natural 
conditions. 

The principal differences between the Termes colony and 
that of Calotermes are as follows: 

1. Termes possesses the caste of workers, which is wanting 
in Calotermes. 

2. On the other hand, Termes has no true royal pair, but 
its place is supplied by a large number of sexually mature 
individuals, which I term complementary royal forms 
(Pl. 17, figs. 16, 17, 21). These complementary forms have 
the characters of larve just about to become nymphs—that 
is, with the wing-rudiments relatively shorter than in the 
nymph. Their length differs a little, however, in different 
examples, and occasionally they are entirely wanting. 

3. Orphaned nests—that is, nests from which the comple- 
mentary forms have been abstracted—contain numerous substi- 
tute royal forms. These may resemble the complementary 
forms, but their wings are frequently entirely wanting (Pl. 17, 
fig. 15), or else developed as in the nymph (id., fig. 23). 
Sometimes they have the characters of an imago which has 
become brown in a few places only, and has the wings mu- 
tilated (id., fig. 24). 

The ordinary nest of Termes may evidently be compared 
with the orphaned nest of Calotermes, with this difference, 
that the former is much richer in royal examples, which 


298 B. GRASSI AND A. SANDIAS. 


usually possess some trace of the wings, whereas these are 
entirely absent in most of the substitute forms of Calo- 
termes. 

The orphaned nest of Termes has a still closer resemblance 
to that of Calotermes, for the royal forms, as I have said, 
are frequently destitute of wing-rudiments. 

In short, in the nest of Termes, as in the orphaned nest of 
Calotermes, individuals of which the wings have never been 
fully developed are invariably raised to the throne. 

There is a further important distinction: in the Calo- 
termes nest the king is always to be found beside the queen, 
whereas in numbers of Termes nests examined I have only 
twice found a single king associated with troops of queens. 
These kings were observed in the hot season,—that is, at the 
time when substitute queens are most difficult to find, because 
then the insects habitually bury themselves deep in the ground. 
I cannot doubt that I might have found many others if I had 
been able to continue my investigations in August and 
September. 

The complementary or substitute king is certainly present 
in the nest about the time when the queens of either kind 
reach maturity, and he disappears after pairing. The colony 
must therefore rear fresh kings every year, which become 
mature in August and September, fertilise the queens, and die. 
By way of confirming this inference, I may say, in addition to 
the facts just related, that recently orphaned nests contain as 
many examples in process of becoming substitute kings as 
those about to become substitute queens. Complementary 
kings in process of development can be found from the middle 
of March onwards, but are always very rare in non-orphaned 
nests, while developing complementary queens are entirely 
wanting. 

Finally, I may add that all possibility of parthenogenesis is 
excluded, as will be seen farther on, by the constant presence 
of abundant spermatozoa in the spermathece of the substitute 
queens. 


CONSTITUTION AND DEVELOPMENT OF TERMITES. 299 


A nest of Termes contains !— 


I. Very young larve, the head of which is alike in 
those of equal length (PI. 17, fig. 1). 

They include forms from the smallest (scarcely 1 mm. long, 
with antenne of eleven joints, the third bare, the rest pilose) 
(Pl. 17, fig. 1) to those a little over 2 mm. in length (with 
antenne of twelve joints, the third bare, the rest pilose, some- 
times with the fourth bare in one antenna only). Larve 
between these two groups are intermediate in length, and have 
either eleven pilose antennal joints or twelve, the third and 
fourth bare and the rest pilose. There are four Malpighian 
tubules in the smaller larve (Pl. 18, fig. 11), eight (four large 
and four small) in the larger (id., fig. 12). I may add at once 
that the smallest larve possess no parasitic Protozoa. 


II. Examples 2°25 to 3°75 mm. long, with twelve 
entirely hairy,or thirteen antennal joints; some with 
large (Pl. 17, fig. 2), others with small heads (id., fig. 3). 

Those with large heads may be regarded as young 
workers, and may become either adult workers or 
soldiers. 

The smallest members of this group still possess four small 
[secondary] Malpighian tubules, and may be free from Pro- 
tozoa (always from Trichonympha). 


III. Examples with fourteen antennal joints, 3°75 
to 4 mm. in length, with or without very short 
wing-rudiments. 

They fall into the following category : 

A. Forms without a trace of wings, with the head 
relatively large, the abdomen stout and rather short, and 
the colour of the body less conspicuously white (Pl. 17, figs. 
5, 78). These are more or less immature workers, capable of 
becoming adults, or of transforming into soldiers. They are 
derived from either the large- or small-headed forms of the 
preceding stage (11). 

1 As before, I must assume a knowledge in this section of certain experi- 
ments which will be described later. 


300 B. GRASSI AND A. SANDITAS. 


B. Soldiers with all antennal joints pilose. De- 
rived from large-headed forms of the preceding stage (11). 

c. Examples without trace of wings and with a 
relatively small head (Pl. 17, fig. 4). Derived from small- 
headed forms of the preceding stage (11). 

p. Examples with very short wing-buds (Pl. 17, 
fig. 6). Also derived from small-headed forms of the preceding 
stage (11). 


IV. Individuals 4 to 6 mm. or more in length, with 
fifteen or sixteen antennal joints. Some possess very 
short or partly developed wing-buds, and belong to 
groups c and p of Stage 11 (Pl. 17, fig. 9). Others have no 
trace of wings, and these may be of three kinds: a”, more 
or less youthful workers; 3B’, soldiers; c*, larve of 
royal forms, complemental or substitutional, with- 
out sign of wings,and with the head small. a? are 
derived from forms a or c of Stage 111; B* from forms 4, 
possibly also from forms B, c, and p of Stage 111; c?, lastly, 
from form c of that stage. 


V. Individuals with seventeen or eighteen anten- 
nal joints, incapable of flight, and infertile, or at 
least far from maturity. 

These may be of five kinds: a°, soldiers (with not more than 
seventeen antennal joints, all pilose) (Pl. 17, fig. 14); B%, 
adult workers (circa 5 mm. in length) (id., fig. 13); c’, 
“nymphs ofthe first form’! (7—8 mm. long), with long 
wing-pads, the genital organs little developed (id., fig. 10) ; 
pb’, “nymphs of the second form”? (4—8 mm. long, 
with more or less short wing-pads, the genital organs well 
developed) (id., figs. 19,20, 22); n3, larve of complementary 
or substitute royal forms, without trace of wings (7— 
9 mm. in length). As will be seen later, the forms p® are also 
larve of royal substitutes. Form a* may originate from a? of 
Stage rv, and perhaps from 8’ as well ; B® is derived from form 

1 [Lespés, ‘ Ann. Sci. Nat.’ (4), v, pp. 248—251, pl. v, fig. 6.] 
2 [Id., pp. 251—254, pl. v, fig. 7.] 


CONSTITUTION AND DEVELOPMENT OF TERMITES. 301 


A? of Stage 1v; c® and p® from the forms with wing-buds of 
Stage tv; and lastly, u* always from form c? of Stage rv. 

To this fifth group belong further certain individuals 
which I regarded as abnormalities before I ‘was acquainted 
with the facts I have recorded about the soldiers of Calo- 
termes. 

They are nymph-soldiers, or nymphs with the buccal ap- 
paratus of soldiers. They probably lose their wings and 
become simple soldiers, as do those of Calotermes.! 


VI. Perfect insects—that is, with fully developed 
wings and capable of flight (Pl. 17, fig. 11). The number 
of antennal joints remains, as in the nymphs, at seventeen or 
eighteen, but they are always entirely pilose. These specimens 
are distinguished by a general piceous colour, save for the 
mouth parts, tarsi, and apices of the tibie, which are testa- 
ceous. They originate from the nymphs of the first form. 


VII. Complementary (Pl. 17, figs. 16—18, 21) or sub- 
stitute (ad., figs. 15,22, 25) royal forms, sexually ma- 
ture, or nearly so, but in the guise of the larva or 
nymph, or else resembling an imago, partly infuscate 
and with the wings torn. 

The latter spring from immature, not fully darkened perfect 
insects ; the others, in the larva or nymph form, arise from 
the royal larve 2n® of group V, or from nymphs of the second 
form, or finally from nymphs of the first form (fig. 23). 
The royal forms derived from nymphs of the first form or 
from not fully darkened perfect insects are all substitute, 
never complemental forms. 

The wing-rudiments may be wanting, or more or less short. 
The body may be of the whitish yellow of old paper, or may be 
more or less extensively blotched with brown. The abdomen 
is much inflated, so that these forms are generally recognisable 
at a glance. 

They exhibit a further characteristic in the possession of 


‘ It now seems to me probable that they should be compared with the 
egg-hatching workers of the honey-bee.—G. B. Grassi, October, 1896. 


302 B. GRASSI AND A. SANDIAS. 


longer abdominal hairs than those of the nymphs and perfect 
insects; these hairs are generally transverse in direction, 
whereas they point obliquely backwards in the nymphs and 
imagos. 

The number of antennal joints may increase in the adult 
workers, larvee of complementary or substitute forms, and 
perhaps in the soldiers, to the maximum figure stated above. 


The general law laid down for the castes of Calotermes is 
equally applicable to Termes, with the introduction of a slight 
modification, due solely to the existence of the workers, which 
are wanting in Calotermes. 

It will stand thus:—The regular development of 
Termes up to the perfect insect may undergo a devi- 
ation at various periods of life, which leads to the 
formationof workers, of complementary or substitute 
royal forms, or of soldiers; the last passing through 
the stadium of the young worker. The deviation in 
question may take place at various periods. 


As the anatomical structure of Termes is in general agree- 
ment with that of Calotermes, I shall restrict myself to a 
few very brief remarks thereon. 

Turning to the sense-organs, it is noticeable that the soldiers 
of this genus do not possess compound eyes, nor do the larve 
which become sexually mature, until the wing-rudiments begin 
to appear. Consequently the complementary and substitute 
royal forms, which have no trace of wings, are equally desti- 
tute of compound eyes. But these are present and pigmented 
in all the other complementary and substitute forms. The 
pigment is also acquired by the nymph of the first form when 
just about to change to the perfect insect. 

In the nymph of the second form there is a distinct structure 
in the neighbourhood of each compound eye, which I interpret 
as a rudimentary pigmentless ocellus. This may also be easily 
seen in nymphs ready for the final change and in the perfect 
insect; but I have been unable definitely to find it in sub- 
stitute forms (I have not looked for it in sections). 


CONSTITUTION AND DEVELOPMENT OF TERMITES. 3808 


The well-known law of Fritz Miller as to the increase in the 
number of antennal joints is certainly inaccurate for Termes 
lucifugus. 

The 11-jointed antenna acquires its 12th joint by subequal 
division of the 38rd, so that it cannot be said whether the 
resulting 3rd or 4th is the younger. When the number of 
joints exceeds twelve it is an indisputable rule that the third 
is the youngest if the number is odd, and the fourth if the 
number is equal. And when the number of joints is odd no 
fresh joint develops until its predecessor has become pilose ; 
for I have never found an example with an odd number of 
joints which were all pilose, while it is easy to find examples 
with an equal number of joints which are or are not all pilose 
(Pl. 17, figs. 28—37). 


I shall now describe the substitute and complementary royal 
forms somewhat fully. 

They rarely exceed 11 mm. in length, and those without 
trace of wings probably never attain a length of more than 
6 or 7mm. The only two kings I have found measured about 
7 mm. in length. The shape both of the head and thorax in 
the examples possessing wing-rudiments is exactly similar to 
that of the nymph, and therefore of the perfect insect. On the 
other hand, those which have no sign of wings find their 
parallel in the small-headed larve; their pronotum has a 
characteristic shape, which will be better understood from the 
figure than from a laborious description, and which distin- 
guishes them from the workers without a shadow of un- 
certainty (Pl. 17, figs. 26, 27). 

The colour of the body, as before mentioned, is generally of 
the pale yellow tone of old paper, sometimes with an aureous 
tinge. The head is dull aureous-yellow, with brown com- 
pound eyes, which are wanting, as I have said, in the com- 
pletely wingless forms. Many old examples exhibit large 
areas which appear brown or sepia-coloured to the naked eye, 
but are seen under the microscope to be yellow sprinkled with 
minute black spots. These areas are as follows: 


304 B. GRASSI AND A. SANDIAS. 


1. The thoracic and abdominal terga (fig. 17). The latter, 
however, exhibit an immaculate longitudinal median vitta and 
a similar rounded spot on either side. The vitta may or may 
not be evident on the metathoracic, and is wanting on the 
mesothoracic tergite; it is present on the pronotum, where it 
forms a cruciform mark with a similar transverse vitta towards 
the apex. As a rule the side margins of the pronotum are 
also immaculate. 

2. The lateral limb of the abdominal sternites; but a few 
black spots may also be seen to exist over the median portion. 

3. The thoracic pleurze (side-pieces). 

4. The basal portion of the legs. 

It is important to notice that completely yellow examples 
may at times have the wing-rudiments more developed than in 
those with brown markings, although the greater abdominal 
development, or their known history, may show the latter to 
be the older. 

The wing-pads are dirty white and exhibit abundant trachez, 
arranged as in the fully formed wing. 

Hairs are distributed everywhere except on the interseg- 
mental spaces. As I have already said, those on the abdomen 
are longer and transversely directed (figs. 17 and 18), and thus 
distinguish these forms from those in which the wings are 
fully developed (fig. 12). The genital appendices are always 
absent in the female. 

The abdominal hairs and black maculation furnish 
characteristic points of difference from the fully 
winged forms. The latter are found to become black by 
uniform darkening,—that is, without first presenting the black 
spots peculiar to the complementary or substitute examples. 

The number of antennal joints in this latter group of royal 
forms is variable, and may differ on each side of the head ; it 
may be 14, 15, 16, or 17; but the last number is exceptional. 
In no case are the antenne intact, as the examination of the 
last joint as well as the third and fourth shows (PI. 17, fig. 29). 
I have found queens without indication of wings and with 
sixteen-jointed antenne, which were obviously truncated at the 


/ 


CONSTITUTION AND DEVELOPMENT OF TERMITES. 305 


apex ; and the condition of the third and fourth joints being 
pilose, led me to infer that eighteen would have been present 
if they were intact. But this is never the case in these forms. 

There remain for description those substitute forms which 
are derived from perfect insects which have not become black, 
and have the wings torn off (Pl. 17, fig. 24). They measure 
about 6 mm. in length, and possess the customary long out- 
standing hairs. The dark compound eyes are conspicuous, 
and the ocelli can also be made out. The antenne are 
curtailed as usual, and the wings are rarely torn off exactly 
along the hind margin of the squama, but so as to leave an 
additional portion of varying, usually small size, the laceration 
following a very irregular course, as if the wing had been 
gnawed off. 

The body is generally of a yellow colour, and is not spotted 
with black, but the margin of the pronotum, especially the 
posterior, and the hind margins of the meso- and meta-notum 
are of a uniform brown, even when seen through the micro- 
scope. 

Occasionally the head and the entire meso- and meta-notum 
are brown ; frequently, also, the thoracic pleurz and the outer 
face of the basal portion of the legs. In some examples the 
apex of the abdomen is brownish. In some the wings are of a 
uniform dirty white, but in many others the squama, the costal 
margin, and perhaps part of the torn edge are brown. 

The genital appendices are present in the male, but are 
wanting in the female, as in other complementary and sub- 
stitute queens. 


The stages of growth of substitute or complemental forms 
with longer or shorter wing-buds are important, and require 
notice. They are to be found by selection of the examples 
with seventeen or eighteen antennal joints and rudiments of 
the wings. According to Lespés’ classical researches, these 
examples are of two kinds, with the wing outgrowths respec- 
tively strongly and feebly developed. The former (fig. 10) are 
his “nymphs of the first form;” the latter (figs. 19 and 20), 


306 B. GRASSI AND A. SANDIAS. 


b 


his ““nymphs of the second form,” are further characterised 
by the bulkier and more ovoid abdomen. The eyes of the 
latter group begin to become pigmented and prominent, and 
their ocelli are visible. Their antennz are intact, and the 
hairs resemble those of the nymphs of the first form. 

These nymphs of the second form become complementary 
royal forms by a moult in which they acquire the characteristic 
direction of the abdominal hairs and, if of female sex, lose the 
female genital appendices. 

They exhibit a marked development of the genital organs 
which will be subsequently described. 


In the examples, previously mentioned, of which the head 
begins to enlarge, much the same development of the man- 
dibles and maxille takes place as in Calotermes flavi- 
collis. 


With respect to the general colour of the inmates of the 
colony, I should add that the workers are normally dirty white 
or yellowish, and the soldiers more distinctly yellow ; freshly 
moulted or very small specimens, and most undifferentiated 
forms, or those destined for sexual maturity, are pure white. 


The legs are alike in all the forms; the anterior tibiz 
possess three, the others two apical spines. 


Adult and fully winged examples exhibit the well-known 
sexual differences of the seventh, eighth, and ninth abdominal 
sternites, viz.—1. The seventh (the apparent sixth) is strongly 
developed and semicircular (with the rounded edge posterior in 
the female), very short in the male. 2. The eighth is reduced 
to two lateral lobes in the female, and is small and entire in 
the male. 38. The ninth nearly resembles the eighth. A 
similar disposition is found in the mature substitute and com- 
plementary forms. As in Calotermes, the ecdyses are rather 
numerous in Termes lucifugus, and do not bear the sup- 
posed relation to the increase in the number of antennal joints. 


CONSTITUTION AND DEVELOPMENT OF TERMITES. 307 


4. Relative Numbers of the Castes. 


It is impossible to indicate the relative numbers of the dif- 
ferent forms with any degree of certainty, owing to the exces- 
sive difficulty of fixing the limits of a colony. But, as before, 
the soldiers of Termes are far less abundant than the other 
forms. The workers occur in enormous numbers; the young 
and larve are also very numerous, the “nymphs of the first 
form ” relatively much less so. 

Examples in process of development into royal forms are 
common only in trees which, though densely populated, 
contain no or very few royal forms. Nevertheless a certain 
number of male “ nymphs of the second form” can easily be 
found from March to June, especially in trees containing com- 
plementary or substitute queens. 

Complementary forms occur in 6 or 8 per cent. only of trees 
invaded by Termes, but in these there may be upwards of a 
thousand, though the number usually oscillates between fifty 
and two hundred. 

Substitute royal forms, varying in number from about ten to 
two hundred, are confined to those nests which have been par- 
tially or completely orphaned, either by destruction of the com- 
plementary kings and queens, or by cutting down a tree and 
removing it a kilometre or so away from its original situation. 

In the nests which I have mentioned as having been ob- 
tained in glass jars, a single true royal pair was found. 


5. Seasonal Variations in the Colony. 


The colony differs very much at different times of the year. 
Eggs are present in May, June, and July, and, in all proba- 
bility, in August and September as well. 

The youngest larve are never present in April and May. 

Nymphs of the first form are not met with in June and 
July. Forms with tolerably well-marked wing-buds, and with 
fourteen to sixteen antennal joints, are absent or very scarce 
in October and November; they increase in number in 
December, to become abundant by January or February. 


308 B. GRASSI AND A. SANDIAS. 


Winged imagos occur from the beginning of April to the 
middle of June (a few stragglers being exceptionally found as 
late as September). 

The remaining forms are present, as a rule, all the year 
round, except the kings, or incipient royal forms of either sex. 

The latter have already been repeatedly referred to, and 
must receive further notice in the following chapter. 

Lastly, I must not omit to mention that nymphs of the first 
form are absent in some years from certain nests, which have 
probably been orphaned at a previous period. 


6. Duration of Development, Life, &c. 


The eggs hatch fifteen or twenty days after they are laid. 

Many observations lead me to conclude that the very young 
larvee found in the winter do not develop farther than the 
nymph of the first form in the following summer, and there- 
fore must certainly live through a second winter before acquir- 
ing wings; e.g. larvee born in October, 1889, will not have the 
wings fully developed until April, 1891. 

Thus, too, the complete or almost complete absence in 
October and November of examples with 14—16-jointed an- 
tennz and distinct wing-buds compels me to believe that those 
hatched in May have become nymphs already, or else that 
they do not yet possess fourteen antennal joints. The latter 
hypothesis is correct; for if the former were, a much larger 
number of nymphs ought to be found in particular nests than 
is actually the case. 

A small colony obtained in a glass jar was furnished with 
a number of fully winged individuals in the early part of May ; 
on the 20th of December it contained, beside the other in- 
habitants, five workers with 12—14-jointed antenne. 

Several orphaned nests were placed in large glass jars in 
January ; next October they still contained small specimens 
with 13—14-jointed antennz, while no substitute forms had 
developed. 

These further facts agree sufficiently with the hypothesis 
which I regard as correct. 


CONSTITUTION AND DEVELOPMENT OF TERMITES. 309 


Tn conclusion, a lapse of eighteen to twenty mouths may be 
estimated to take place between the times of hatching and of 
reaching the perfect state. The workers and soldiers probably 
require a much shorter time, and both can apparently be 
developed by the autumn from eggs laid in May of the same 
year. 

The duration of life of single specimens cannot easily be 
estimated. The soldiers and workers certainly do not die off 
about the middle of June, as Lespés pretends. The king, 
whether complementary or substitute, does not live more than 
a couple of months after reaching sexual maturity, whereas 
the corresponding queens will live for several years. 


7. Situation of Different Forms in the Nest. 


The fertile or nearly fertile royal forms usually live in a very 
remote part of the nest, often collected together in a deep root 
or in the heart of a large trunk. 

The only two mature kings I have found accompanied the 
queens; the latter may be surrounded with numerous new-born 
larvee and clumps of 30 or 40 to 100 eggs. 

Eggs and larve can be carried long distances, as is shown 
by the occasional presence of numerous young larve and some- 
times of eggs as well in trees destitute of queens. 

The larvee, nymphs, winged forms (if present), soldiers, and 
workers all occur mingled together, both in the midst of the 
queens and elsewhere. Recently invaded situations contain, 
as a rule, chiefly workers and soldiers; and isolated workers, 
which one is tempted to look on as explorers or pioneers, may 
sometimes be found under stones, or in reeds, &c. Certain 
parts of the nest often contain principally workers and soldiers ; 
others larvee with wing-rudiments and nymphs. 

[I would add further that when a nest is opened certain parts 
will contain nothing but soldiers in the neighbourhood of the 
egg-clumps. It is a mistake to regard this as a normal state 
of things; the disturbance and noise of opening the nest have 
put all the other inhabitants to flight, while the valiant soldiers 
alone remain to protect the eggs. | 

VOL. 39, PART 3,—NEW SER. a8 


310 B. GRASSI AND A. SANDIAS. 


VIII. Certain Habits. 


The swarming will be first described. 

A swarm was observed on May 14th, 1891. A host of 
Termites issued in groups of two to five from a single crack in 
a tree-trunk, rose on the wing for about three metres, and 
then followed the direction of the wind. Many fell on the 
neighbouring plants, and others were seized by ants as they 
emerged. The swarm began at 10 a.m., and lasted for more 
than two hours. 

This incomplete observation was made by the laboratory 
servant, who chanced to see a similar occurrence on May 24th, 
1891. In the second case the winged forms were seen to issue 
from several holes, at which a few soldiers and workers also 
appeared. The imagos were all destroyed, at first by two 
lizards, which remained on watch near the nest and devoured 
them directly they appeared. When the lizards were driven 
off the insects were seen to come out in groups of six or eight, 
run a certain distance, spread their wings, and then take 
flight. At first they rose a certain height, and then followed 
the direction of the wind (a light scirocco). A few struck 
against branches and fell to earth, but they quickly got up 
again. This swarm lasted from about 9.30 to 11 a.m. When 
it was over the nest was opened, and found to contain many 
imagos ready for swarming about half a metre from the exit- 
holes. These were all females, as were those (twenty-six) 
collected during the swarm. 

Sinilar swarms were observed on various other occasions. 
One of special interest took place in the laboratory from the 
nest in a glass jar, which I have previously described, men- 
tioning its occurrence on June Ist. The details are now given. 

Towards the end of April I examined the nest at the bottom 
of the jar, and observed the appearance of perfect insects, 
some brown, others still white, but becoming brown in the 
course of afew days. After May 20th, suspecting that swarm- 
ing would take place soon, I covered the vessel with a wider 
and taller glass bell-jar. Both were placed on a pane of glass, 


CONSTITUTION AND DEVELOPMENT OF TERMITES. 311 


and the jar was not completely closed, because the paper did 
not accurately fit its mouth. 

On June Ist, about 9 a.m., I found a number of winged forms 
with a few soldiers and workers on the part of the sheet of 
glass which lay between the two jars; the poor little animals 
had taken advantage of the spaces left by the ill-fitting paper 
to emerge, and were then vainly trying to escape from the 
bell-jar. Nothing was disturbed, and twenty-four hours later 
no change had taken place, except that the winged forms were 
more numerous, and some were moribund. 

I then decided to liberate them, and removing the bell-jar I 
put them by means of a quill pen into a receptacle containing 
suitable materials for a nest—but without success, for they all 
died in less than forty-eight hours. No more winged examples 
could be detected through the walls of the glass jar; it 
was replaced unopened on the pane of glass, and once more 
covered with the bell-jar. Twenty-four hours later a large 
number of larve and several soldiers were again found on the 
pane. They were then returned to the nest by means of a 
quill, and did not escape again. 

The observation here recorded was made some years ago, 
and though imperfect is not wanting in significance. 

Another swarm was observed on a May morning of the 
present year (1892). There happened to be a small level 
place in front of the holes from which the perfect insects 
issued, and on this they ran to and fro before taking wing. 
They were accompanied by soldiers and workers which had 
emerged into daylight, evidently to protect the swarm. 

Several writers have recorded that black Termites, when 
ready to fly, perform certain movements which can be best 
followed by putting the insects on a sheet of paper; these are 
the so-called love passages, often described, and especially by 
Fritz Miller. 

They are not exhibited by Calotermes, but may easily be 
observed in Termes lucifugus. 

The majority of examples dispose themselves after the loss 
of their wings in pairs, one behind the other. More exactly, 


ole B. GRASSI AND A. SANDIAS. 


the one in front attempts to run away from the other, which 
pursues it and palpates the extremity of its abdomen, and 
sometimes the sides as well. In some cases the pair is com- 
posed of a wingless individual in front and a winged one 
behind ; or a male in front and a female behind; or the 
opposite; or both may be of the same sex, whether male or 
female. 

If a few workers are put among the imagos, one of the latter 
may often be seen to pursue a worker in the same manner. 
And occasionally three examples, instead of merely a pair, 
may be seen, one following the other. 

I believe that the meaning of these supposed amorous dis- 
plays is entirely different from that usually assigned to 
them, and that the pursuer wishes to solicit the dejecta of 
the one pursued; this will be explained in the following 
chapter. 

I have said that the adults lose their wings (the persistence 
of the squama being understood) (Pl. 17, fig. 12), and I must 
now explain more minutely how this occurs. Suppose that the 
wings of a specimen are accidentally allowed to touch the moist 
walls of the glass jar, they stick to it, and readily break off as 
soon as the owner tries to run away. 

[The insects perform various movements on their own account 
in order to tear off the wings. I have seen one raise and 
lower them, and at the same time put the hind leg over them 
so as to hold them down to the surface on which it was stand- 
ing. Another example got rid of them by violent fluttering ; 
and a third, which had only one wing left, tried to tear it off 
at first by forcible flapping, and then succeeded by holding it 
firmly with one of the hind legs. ] 


It often happens that the imagos lose their wings while still 
in the nest, but they nevertheless abandon it, as do those of 
Calotermes. 

In fact, the perfect insect has an imperious craving to quit 
the nest in which it develops. 

Winged examples artificially enclosed in a corked tube 


CONSTITUTION AND DEVELOPMENT OF TERMITES. 313 


quickly gnaw through the cork and escape. This happens even 
if the wings have fallen off. ; 


It will be recollected that the two sexes swarm at different 
times. This is demonstrated by— 

1. The fact that all members of a swarm are found to be 
of the same sex. 

2. The presence in many nests, late in or at the end of the 
swarm-period, of black winged forms, which are all males or 
all females only (by a rare exception a single male may be 
found to every twenty or thirty females), whereas nests in which 
the imagos are still white contain a male to every two or three 
females. 

The convulsive movements spoken of under Calotermes 
are exhibited alike by Termes, and are common to all mem- 
bers of the colony except the newly born, and have the same 
significance. Moreover the soldier is able simultaneously to 
produce a special crepitus (creaking), which arises whenever 
the head is held horizontally during the act of quivering by 
friction between the hind margin of the occiput and the ante- 
rior margin of the pronotum. But whenever the head is held 
in the normal position during the act—that is, somewhat de- 
flexed—no perceptible sound is produced, owing to the absence 
of friction. 

The soldiers, therefore, possess two distinct modes of commu- 
nication ; and itis noticeable that those of Calotermes always 
hold the head obliquely deflexed when quivering and produce 
no sound, 

I may add that this characteristic crepitus may be heard at 
very short intervals by applying the ear to a trunk containing 
a nest of Termites. This proves that the quivering motions 
are a constant feature in normal and undisturbed nests, in 
which they are therefore not employed to give indications of 
alarm or distress; and I conclude that, besides these signifi- 
cations, the convulsive movements must also have the value of 
ordinary speech ; that they constitute, in short, a means of in- 
tercommunication. The same conclusion holds good for Calo- 


314 B. GRASSI AND A. SANDIAS. 


termes; and I imagine that the quivering of both species 
produces a sound which is perceptible to the tympanal organ 
of the tibia, but is inaudible to the human ear. 

Termites, moreover, may communicate by means of the 
antenne. Thus, if a few are placed on a table, they usually 
arrange themselves in single files, which circle round the 
objects standing thereon ; and, in such a case, if two Termites 
moving in opposite directions chance to meet, they recipro- 
cally touch their antennz and then continue each on its own 
course (vide also previous statements). 

Tasks necessary for the common welfare, with a few excep- 
tions, are undertaken by all the inmates of the colony ; but the 
soldiers are unable to gnaw wood, owing to the great elonga- 
tion of their mandibles. 

Substitute or complemental royal forms have never been 
seen to prepare wood-meal, or to transport it, or ova,&c. Yet 
all these duties are carried out by the perfect insects with fully 
developed wings, before or after they have been shed. Newly- 
born larve may easily be found carrying about wood powder. 

[The soldiers serve for defence, like those of Calotermes. 
Two soldiers, one a Termite, the other a Calotermite, were put 
together in a small glass vessel. They accidentally came into 
contact and began to fight. The Termite, having the advan- 
tage of great quickness in movement, whereas Calotermes is 
sluggish, bit off some of its enemy’s legs, and was proceeding 
to further hostilities, when the other seized an opportune 
moment and cut its head off. On other occasions the soldiers 
of Calotermes tore the abdomen of the soldiers of Termes 
to pieces. 

The soldiers’ mandibles may appropriately be likened to a 
powerful pair of shears. Termite soldiers become formidable 
when put into one of the customary little nests of Calo- 
termes deprived of soldiers, rapidly cutting off the antennz 
of numerous examples, and biting them in various places. 
But if they are few in number, the Calotermites eventually 
reduce them to helplessness by shearing off their mandibles, 
and then pursue, tear, and kill them. 


CONSTITUTION AND DEVELOPMENT OF TERMITES. 315 


A fight is invariably provoked by putting soldiers of both 
species into a jar; if large larve or nymphs of Calotermes 
are introduced into a colony of Termes they are usually left 
alone by the soldiers, which probably fear them ; whereas Calo- 
termites (see p. 283) always kill any workers or nymphs of 
Termes which have been added to their nest. Termes 
soldiers fraternise, and do not fall out, even when taken from 
different nests. 

If triturated wood, soldiers, workers, and young of Termes 
are put together in a jar, the soldiers are soon seen to post 
themselves on the top of the rubbish, evidently on guard. 

Sometimes inmates of the same nest (soldiers and workers, 
or the latter inter se) come to blows, and wound each other 
ferociously in the thorax or abdomen, and do not stop unless 
others interfere to separate them. These internecine battles 
can be provoked, e.g. by overturning the contents of a nest, 
and are perhaps due to each individual imagining that his 
neighbour is the cause of the disturbance. 

Termites shun the light, and prefer to collect in the darkest 
parts of a vessel. ] 


(To be continued.) 


316 B. GRASSI AND A. SANDIAS. 


EXPLANATION OF PLATES 16—20, 


Illustrating Professor B. Grassi’s and Dr. A. Sandias’s paper 
on “The Constitution and Development of the Society of 
Termites: Observations on their Habits ; with Appendices 
on the Parasitic Protozoa of Termitide, and on the 
Embiide.”’ 


The first number after the explanation of each figure indicates the ocular, 
the second the objective of the microscope employed. Kor. = Koritska 
microscope, withthe tube in. Hart. = Hartnack microscope, with the tube in. 
T. = Termes lucifugus. C. = Calotermes flavicollis. 

Instead of the expression that a given individual possesses, e.g. seventeen, 
antenual joints, the abbreviation ‘‘ with seventeen joints” is employed. 


PLATE 16. 


Calotermes flavicollis. 


Fic. 1.—Small-headed larva, with twelve joints. The third, fourth, and 
fifth indistinct ; the former not pilose; the fifth with short hairs (distinguish- 
able with a higher amplification only). 


Fie. 2.—Large-headed larva, with thirteen joints. The third short, and 
not pilose. 


Fie. 3.—Small-headed larva, with sixteen joints. The third and fourth 
scarcely indicated, and not pilose. Wing-rudiments distinctly present, but 
very short, and visible only with higher amplification. 


Fic. 4.—Nymph, with seventeen joints. The third pilose, the fourth not. 
Fic. 5.—Large soldier. 

Fig. 6.—Perfect insect, with fully developed wings. 

Fie, 7.—True queen, in the fourth year of maturity. 

Fic. 8.—Outline of a true king or queen, in the second year of maturity. 
Fic. 9.—Outline of a true queen, in the third year of maturity. 

Fic. 10.—Outline of a true king, in the fifth year of maturity. 


Fic. 11.—Outline of the abdomen of a true queen, in the fifth year of 
maturity (drawn approximately to the same scale as Fig. 10). 


Fie. 12.—Young substitute queen. 


Fie. 13,—Outline of the abdomen of a substitute queen, in the third year 
of maturity (drawn to the same scale as Fig. 11). 


CONSTITUTION AND DEVELOPMENT OF TERMITES. 317 


Fic. 14.—Developing royal substitute form, with fifteen joints; the third 
the shortest, and not pilose. 


Fic. 15.—Exuvie of a developing substitute form. 
Fie. 16.—Outline of the body of a small soldier (the antenne are omitted), 
Fie. 17.—Outlines of the bind angle of (4) the mesonotum and (a) the 


metanotum of a substitute queen, with very slight traces of the wing- 
outgrowths. 3, 4, Kor. 


Fic. 18.—Outline of the right half of the meso- and metanotum of 
developing substitute form. The anterior wing-rudiment is torn off. 


Fic. 19.—Antenna of a small soldier. 3, 4, Kor. 

Fie. 20.—Base of a fifteen-jointed antenna. 3, 4, Kor. 

Fic. 21.—The same in a different stage, but still fifteen-jointed. 

Fic. 22.—The same after development of a sixteenth joint. 

Fie. 23.— Ovary and oviduct of a soldier. 3, 4, Hart. 

Fic, 24.—Posterior leg of a young nymph. 

Fic. 25.—Old substitute queen, with wing-rudiments. 

Fic. 26.—ialf of the thoracic terga, exhibiting wing-rudiments, in a young 
soldier. 3, 5, Hart. 

Termes lucifugus. 


Fic. 27.—Brain of a worker, by transmitted light. 3, 5, Hart. 

Fies. 28 to 33.—Series of horizontal sections of the brain of a young 
nymph, to exhibit the fungiform bodies (interpreted as psychic centres). They 
are represented by the darker and more closely dotted portions (which possess 
small, deeply staining nuclei). 

Fig. 28 represents the most superficial section, and is followed by the 
others in numerical order. 

Fig. 34 represents, by way of comparison, a similar section through the 
brain of Embia, at the point where the fungiform bodies attain the maximum 
dimensions. gl. Retro-cerebral gland. [This gland of unknown function 
exists (only ?) in the nymph of the first form, the perfect insect, and the 
soldier. It eliminates a transparent secretion, which can be spirted out for 
some distance. ] 


PLATE 17. 
Termes lucifugus. 
Fic. 1.—Larva, with eleven joints, the head undifferentiated. 
Fic. 2.—Large-headed larva, with twelve pilose joints. 
Fic. 3.—Small-headed larva. 
Fie, 4,.—Small-headed larva, with fourteen joints, the fourth not pilose. 


318 B. GRASSI AND A. SANDIAS. 


Fig. 5.—Diagram of a similar but large-headed larva. The hair-lines 
between 2 and 3, and between 4 and 5, apply to each of the two figures. 


Fic. 6.—Larva similar to 4, but larger and with wing-rudiments. 
Fie. 7.—Larva similar to 5, but equal in size to that of Fig. 6. 


Fie. 8.—Anterior portion of a larva similar to that of Fig. 7, but with a 
somewhat larger head ; probably a soldier larva. 


Fic. 9.—Larva, with fifteen joints; the third alone not pilose, and with 
evident wing-rudiments. 


Fie. 10.—Nymph, with long wing-pads. 

Fic. 11.—Fully winged perfect insect. 

Fic. 12.—Perfect insect after shedding the wings. 

Fic. 18.—Adult worker. 

Fig, 14.—Soldier. 

Fic. 15.—Young wingless substitute queen. 

Fie. 16.—Old wingless complementary queen. 

Fic. 17.—Old complementary queen, with slight rudiments of wings. 
Fig. 18.—The same, with more distinct rudiments. 

Fie. 19.—Male nymph of the second form (March Ist). 

Fic. 20. —Outline of a male nymph of the second form (April 11th). 


Fig. 21.—Young complementary queen, derived from a nymph of the second 
form. 


Fie. 22.—Thorax of a nymph of the second form, with the wings more 
developed than those of the example in Fig. 21. 


Fie. 23.—Substitute queen, derived from a nymph of the first form. 


Fic. 24.—Substitute queen, derived from a perfect insect; partly infuscate, 
and with the wings torn (the infuscation is not shown). 


Fic. 25.—Posterior extremity of a very old substitute queen, without sign 
of wings. 

Figs. 1 to 25 are all drawn by the camera lucida to an equal scale, with 
exception of Figs. 16, 17, and 21, which are somewhat more enlarged. 

Fic. 26.—Pronotum of a worker. 3, 4, Kor. 

Fic. 27.—Pronotum of a completely wingless substitute queen. 3, 4, Kor. 

Fic. 28.—Antenna of a nymph of the second form. 3, 4, Kor. 

Fic. 29.—Antenna of a complementary queen. 3, 4, Kor. 

Fic. 30.—Base of a twelve-jointed antenna. 

Fie. 31.—Base of a thirteen-jointed antenna. 

Fig. 32.—Base of a fourteen-jointed antenna. 

Fic. 33.—Base of a fifteen-jointed antenna. 


CONSTITUTION AND DEVELOPMENT OF TERMITES. 319 


Fic. 34.—Base of a sixteen-jointed antenna. 
Fic. 35.—Base of another sixteen-jointed antenna. 
Fic. 36.—Base of a seventeen-jointed antenna (nymph of the first form). 


Fic. 37.—Base of an eighteen-jointed antenna (nymph of the first or 
second form). 


The figures 30 to 37 are all copied by the microscope, 3, 4, Kor., with the 
tube drawn out. 


Fic. 38.—Flattened tube constructed by T., and suspended from a plaster 
cornice. The lower end (to the right of the figure) was open; the upper end 
is broken. 


Fic. 39.—A similar flattened tube—short, enlarged and flask-shaped. A 
small portion of the cornice to which it was suspended is represented. 


Fic. 40.—Portion of a gallery formed by T. in the angle of a wall; seen 
from the inside. 
PLATE 18. 
Termes lucifugus. 
Fie. 1.—Ovary of a perfect insect before loss of the wings. 3, 6, Kor. 
Fic. 2.—Ovaries of a nymph of the second form. 38, 4, Kor. 


Fic. 3.—Ovary of a very old complementary queen. The tubules are evi- 
dently atropied, and the spermatheca was empty. 1,4, Kor. Por. ant. = 
anterior portion. 


Fic. 4.—Left ovary of a nymph of the first form. 3, 4, Kor. 


Fig. 5.—Testis of a perfect insect before loss of the wings. 3, 6, Kor. 
(Spermatozoa are present in the vas deferens.) 


Fie. 6.—Testis of a nymph of the second form. 3, 4, Kor. 

Fic. 7.—Testis of a nymph of the first form. 3, 4, Kor. 

Fic. 8.—Testes, vasa deferentia, and vesicula seminalis (the two latter 
not containing spermatozoa) of a perfect insect before loss of the wings. 
3, 4, Kor. 

In the above eight figures, and in some on the following Plate, the efferent 
ducts are partly shown, in addition to the generative glands. 

Fic. 9.—Developing spermatozoa, in the fresh state. The one indicated 
with the letter a is fully developed. 4, 3 Kor. 

Fic. 10.—Oviduct (¢u.), uterine portion (a. ov.), spermatheca (spé.), seba- 
ceous (colleterial) glands (g/. seb.) of a complementary queen. 3, 4, Kor. 

Fic. 11.—Portion of the intestine at the origin of the four Malpighian 
tubules in a newly-born larva. 3, 4, Kor. 


Fic. 12.—Similar portion, with four large and four small tubules. 3, 4, Kor. 


320 B. GRASSI AND A. SANDIAS. 


Fis. 13.—Superficial portion of a nest of T. constructed in a jar, with exit- 
holes and chimneys for swarming. 

Fic. 14.—Portion of a termitarium found in a bench in the church 
Pesara. 

Fie. 15.—Tubular gallery, similar to that of Fig. 39, Pl. 2, but not flask- 
shaped. 


Fic. 16.—Portion of a partly excavated and partly built nest, found in a 
large cavity in the root of a cactus swarming with T. 


Fie. 17.—Galleries constructed by T. kept in a glass jar; D-shaped, except 
the one projecting over the lip, which is tubular, The shading at the bottom 
indicates the rubbish which harboured the colony. 


PLATE 19. 


The left side of the plate refers to Calotermes flavicollis, except 
Fig. 6, which refers to Termes lucifugus; the right side refers to Embia. 


Lert Sipe (CALOTERMES), 

Fie. 1.—Ovary and colleterial gland of a true queen in the second year of 
maturity. 3, 4, Kor. 

Fie. 2.—Ovaries of a very bulky queen, enlarged about three diameters 
(the tubes do not actually lie in one plane as in the figure). eat. ant. = 
anterior extremity. 

Fie. 3.—Colleterial glands of a true queen. 3, 4, Kor. 

Fic. 4.—Left testis (¢est.), vas deferens (con. def.), and vesicula seminalis 
(ves. sem.) of a true king. 3, 6, Kor. 

Fie. 5.—Testis of a substitute king, several years old. 3, 6, Kor. 

Fie. 6.—Portion of the spermatheca, showing the ducts and orifices of the 
underlying glands (fresh). (Termes lucifueus.) 

Fie. 7.—Salivary glands of a small larva. ves, = reservoir. 38, 8, Kor. 


Fig. 8.—Relations of the hinder part of the chylific ventricle and intes- 
tine. The former contains a blackish, the latter a yellowish detritus. Their 
junction is indicated by the insertion of the Malpighian tubules. 

Fie. 9.—Stomodeum (zt. ant.) and part of the chylific ventricle (xt. med.). 
inv. = invagination of the former into the latter. 

Fic. 10.—Tympanic organ in the fresh state. 3, 75, Kor. 

Fie. 11.—Tuhe-nest. Large portions of the glass are obscured by dis- 
gorged matter. The cork is riddled with burrows, and is represented separately 
on the right. 


Fie. 12.—Spermatozoa from the spermatheca, in the fresh state. 5, 54, 
Kor. 


CONSTITUTION AND DEVELOPMENT OF TERMITES. 321 


Rieut Sipe (Empta). 
Fig. 1.—Ventral ganglionic chain; L. 1 = anterior, L. 2 = middle, L. 3 
= posterior leg. 
Fic. 2.—Dorsal view of the tracheal system (the stigmata are numbered). 


Fig. 3.—Alimentary canal. gl. sal. = salivary glands. int. ant. = oeso- 
phagus and proventriculus. zz¢. med. = chylific ventricle. zt. post. = in- 
testine. 


Fic, 4.—Ovarian tubes and oviducts (ovd.). seg. med. = segment médiaire 
(the successive abdominal segments are numbered). 


Fig. 5.—Male generative organs. 


Fic. 6.—Posterior extremity of the male. proc. dext. = right, proc. sin. 
= left apophysis. cer. dext. = right cercus. 


Fic. 7.—Labium and labial palpus of one side. 

Fic. 8.—Maxillary lobes and palpus. 

Fie. 9.—Anterior tarsus (seen obliquely from the side). 
Fic. 10.—Middle tarsus (idem). 

Fic, 11.—Posterior tarsus (idem). 


PLATE 20. 
Parasitic Protozoa of Termitide. 


The figures are all drawn with a Koritska microscope, ocular 5, objective +, 


Leitz. The dotted lines represent the continuation of the unbroken lines on 
the remote face. 


Fig. 1.—Trichonympha agilis, in optical section, nearly corresponding 
with the middle plane of the body. 

Fic. 2.—The same, at a somewhat different level. 

Fic. 3.—The same, in a superficial plane. 


Fic, 4.—The same, partly schematic, to show the arrangement of the 
flagella. 


Fie. 5.—The same, to give an idea of the complicated arrangement of the 
spicules in some examples. 


Fic. 6.—Joenia annectens. 
Fic. 7.—Idem. 
Fic. 8.—Idem. 


Fie. 9.—Endoskeleton and nucleus of a very large example of Joenia in 
optical section. 


Fic. 10.—Microjoenia hexamitoides. 


322 B. GRASSI AND A. SANDIAS. 


Fie. 11.—Dinenympha gracilis. 

Fie. 12.—Idem. 

Fic. 13.—Idem. 

Fic. 14.—Idem. 

Fie. 15.—Idem. 

Fic. 16.—The same, with spirilla attached to one extremity. 

Fic. 17.—The same, covered with spirilla. 

Fic. 18.—Pyrsonympha flagellata (the majority of the flagella are 
omitted). 

Fic. 19.—The same, in superficial view. 

Fie. 20.—The same, exhibiting the nucleus and rodlets. 

Fic. 21.—Holomastigotes elongatum (the majority of the flagella are 
omitted). 

Fic. 22.—The same, in superficial view. 

Fic. 23.—Plan of the lines of origin of the flagella on both faces of 
Holomastigotes elongatum. 

Fic. 24.—The same, in another individual. 


ON OTENOPLANA. 323 


On Ctenoplana. 


By 


Arthur Willey, D.Sc. 


With Plate 21. 


Tue discoverer of the remarkable genus Ctenoplana, which 
presents affinities both to the Ctenophora and to the Turbel- 
laria, was, as is well known, Professor Alexis Korotneff, 
who obtained only a single specimen off the west coast of 
Sumatra, and described it in the ‘ Zeitschrift fiir wissenschaft- 
liche Zoologie *! for 1886. 

Korotneff found his specimen drifting in a current of the 
sea, in the company of a large number of Porpita. It was 
distinguished by its deep red or crimson colour, and was 
named C. Kowalevskii in honour of the discoverer of 
Celoplana. 

Since 1886 no second record of the occurrence of the genus 
has been made. 

In January of this year (1896), while cruising among the 
islands which form the Hastern Archipelago of British New 
Guinea, in pursuance of zoological work, I was fortunate 
enough to pick up a cuttle-bone which had evidently been 
afloat for a long time, and was being carried along by the 
current off the group of the islands named on the chart the 
Conflict Group.? On the cuttle-bone were numerous minute 

1 Vol. xliii, pp. 242—251, Taf. viii. 
? These islands surround a magnificent lagoon. 


824 ARTHUR WILLEY. 


organisms, including young green-tinted Polychetes, young 
Planaria, Anthozoan larve, young Cirripedes in the Cypris 
stage, and four specimens of Ctenoplana. Three of the last 
were of a pronounced green colour, and the fourth was 
crimson. 

The green Ctenoplana, both on account of its colour and its 
shape, is an entirely new kind, and I shall name it C. Korot- 
neffi, The crimson variety may or may not be specifically 
identical with C. Kowalevskii; but as my sketches of the 
external form differ somewhat from Korotneff’s figures, I think 
it will be well to give it a provisional name with the object of 
engaging the attention of any zoologists who may have future 
opportunities for studying the genus. I propose, therefore, to 
call my crimson specimen C. rosacea. 

As TI had no chance of getting ashore, I had to make the 
best of the limited accommodation supplied by a small cutter, 
in order to observe the appearances presented by the living 
animals and their movements. Although I omitted to make 
accurate measurements of the living expanded animals, their 
average diameter would correspond closely in length with that 
of C. Kowalevskii, which, as stated by Korotneff, mea- 
sured 6 mm. 

Many of the external features of Ctenoplana were correctly 
described and figured by Korotneff, but his specimen appears 
to have been not very active; and although, as he says, he had 
it under observation in the living condition for three to four 
hours, it did not once extrude its tentacles, so that he only 
became aware of their existence in section. The consequence 
of this was that Korotneff was completely wrong in localising 
the plane of the tentacles. He placed them in the plane at 
right angles to that in which they actually occur. 

At the ends of one of the principal diameters of the disc- 
shaped body of Ctenoplana the margin of the body is incurved. 
Korotneff, without the smallest hesitation, describes these 
marginal bays as anterior and posterior in position, while, 
according to him, the tentacles were situated along the diameter 
at right angles to the “‘ antero-posterior ”’ diameter—that is to 


ON OTENOPLANA. oe 


say, in what would correspond to the transverse plane. This, 
however, is quite wrong. 

As will be seen later on, I agree with Korotneff in his iden- 
tification of the antero-posterior axis (in comparison with 
bilateral animals), but it is along this axis that the solid 
tentacles lie. 

Furthermore, Korotneff’s specimen did not give him an 
exhibition of its swimming powers, so that he could not 
observe the movements of the ctenophoral plates, and the 
result was that he formed the opinion that these plates had 
become altered as to their function, and that they hardly 
appeared to serve for the progression of the body. This 
is a curious repetition of the old error with regard to the 
means of locomotion of the Ctenophores, as set forth in 
Chun’s monograph. As a matter of fact, when Cteno- 
plana swims, the ctenophoral plates are its sole means of 
locomotion. 

I now pass on to a systematic account of my own observa- 
tions. 

1. Shape and Movements of the Body.—Like the 
Planarians, the body of Ctenoplana comprises a thickened 
median ridge-like area and two lateral thin skirt-like areas, 
the “Seitenfelder” of Lang. In the attitude assumed when 
crawling, the body is nearly round with the exception of 
the above-mentioned marginal bays, from which I observed 
the muscular pinnate tentacles being constaxtly protruded and 
retracted while the animal was crawling (fig. 1). As also 
observed by Korotneff, Ctenoplana crawls with one of the 
rounded margins of the body directed forwards as a rule. The 
tentacles thus appear remarkably like transversely paired 
structures, and one would naturally at first describe them as 
such. But it must be remembered that the terms anterior, 
posterior, and transverse, as applied to bilateral animals, 
are not applicable to Ctenoplana. 

The tentacles, when extruded, are found to be white struc- 
tures, thus making a marked contrast to the green or red 
colour of the body. They are provided with small secondary 

VOL, 39, PART 3.—NEW SER. Y 


326 ARTHUR WILLEY. 


tentacles or pinne, arranged somewhat irregularly, but in a 
single series. Like the tentacles of Coloplana, described by 
Kowalevsky, they are strictly comparable with the tentacles of 
a Cydippid. Both the tentacles and their pinne are quite solid, 
being completely filled up with a muscular core. Within the 
body each tentacle is enclosed within a hollow sheath which 
opens to the exterior at the end of a small papilla at the 
base of the marginal bay. When retracted, therefore, the 
tentacles form a median axial skeletal support for the body, 
interrupted in the middle region of the body by the aboral 
sense-organ. 

The aboral surface of the body may be at once called the 
dorsal surface, and the oral the ventral surface. 

The possession of a relatively wide Planarian-hke skirt not 
only permits Ctenoplana to crawl about on firm surfaces, but 
enables it also to attach itself, in a highly characteristic 
manner, to the surface-film of water by its ventral surface. 
In this position it greatly resembles Planarians, which are also 
fond of assuming the same position. When lying thus attached 
to the surface of the water the round central oral opening can 
be seen. The mouth can be protruded so as to form a slight 
cone. 

When swimming, Ctenoplana brings the two halves of the 
skirt together so as to form a bell-shaped, or better, a Pilidium- 
shaped structure which progresses very rapidly by means of 
the ctenophoral ‘plates. In swimming, the aboral pole is 
directed forwards as it is in the Ctenophora. 

The ctenophoral apparatus consists of eight small oval 
plates, placed four on each side of the tentacle axis.‘ Across 
each plate run six or seven shallow grooves, from which the 
long cilia arise (figs. 1 and 5). The cilia of each groove appear 
in section to be united usually for some distance from their 
base, and then to separate out into the individual cilia (fig. 5). 
The ctenophoral plates alternate with the lobes of the central 
gastric system (figs. 1—3). I only had a fleeting view of the 


1 The line joining the bases of the tentacles may be called the “ tentacle 
axis.” 


ON OTENOPLANA. 327 


peripheral anastomosing ramifications of the gastric system, 
and have not indicated them in the sketches of the external 
form. 

When Ctenoplana wishes to sink from the surface to the 
bottom it doubles itself up in the usual way, and so sinks 
apparently without employing the combs. This was also ob- 
served by Korotneff. 

2. The Aboral Sense-organ.—As already described by 
Korotneff, the aboral sense-organ consists essentially of an 
otolithic mass, suspended by stiff processes from adjacent cells 
in a cupule, and surrounded by a ring of ciliated tentacles. 
Korotneff figures the latter in the form of a simple circlet. 
This, however, is not the case. The circlet of sensory 
tentacles surrounding the otolith consists of two 
distinct and separate halves, with about nine ten- 
tacles in each half. The one half is placed on one 
side, and the other on the opposite of the tentacle 
axis (fig. 1). 

This is perhaps the most important observation that I was able 
to make on the living animal, and it is a crucial one for deciding 
upon the homologies of the axes of Ctenoplana with those of 
bilateral animals. The division of the circlet of sensory 
tentacles into two portions was remarkably distinct and un- 
equivocal. The sensory or apical tentacles (as distinguished 
from the muscular or terminal pinnate tentacles) are usually 
carried extruded (figs. 2 and 3), but they can be completely 
retracted. 

3. Cilia.—I cannot confirm Korotneff in his statement as 
to the general distribution of cilia over the surface of the body. 
The places where I have observed cilia (apart, of course, from 
the ctenophoral plates) are as follows: (i) on the sensory 
tentacles, (11) on the cells lining the sheaths of the pinnate 
tentacles, and (iii) over a large area of ventral surface (fig. 5). 
I must deny the presence of cilia on the general dorsal 
surface. 


1 As Ctenoplana is semi-opaque, it is difficult to discern much of the 
internal structure in surface view. 


828 ARTHUR WILLEY. 


4. Gastro-vascular System.—The central main portion 
of the gastro-vascular system presents the lobed appearance 
shown in the figures, the lobes being paired about the tentacle 
axis. The middle and largest pair of lobes belong to the 
stomach, and thus serve to mark out the stomachal plane 
(Magenebene of Chun). The stomachal plane, therefore, as in 
the Ctenophores, lies at right angles to the plane of the ten- 
tacles, which corresponds to the funnel plane (Trichterebene) 
of Ctenophores. 

My identification of the stomachal plane in Ctenoplana is 
just the reverse of Korotneff’s, who erroneously placed it in 
the true tentacular plane. 

From the two opposed sides of the stomach a narrow median 
canal leads into the two terminal end-lobes of the central 
gastric cavity (cf. the schematic fig. 11). The two end- 
lobes are in open communication with the peripheral canal- 
system. 

I do not find such definitely circumscribed peripheral canals 
as those figured by Korotneff, but they appear to me in section 
merely as the spaces partitioned off by the dorso-ventral 
trabeculee, which Korotneff describes as dorso-ventral muscles 
(fig. 5). 

The median funnel-vessel was correctly figured by Korotneff. 
It arises from the stomach immediately opposite to the 
mouth, and, proceeding aborally, embraces the sense-organ 
without opening to the exterior. It is very clearly shown in 
section. 

5. Tentacle Sheaths and Musculature.—The tentacle- 
axis is occupied by the sheaths of the tentacles, which are 
hollow tubes lined by ciliated cells lying immediately beneath 
the dorsal surface, and completely separated from one another 
by the aboral sensory complex. The muscles of the tentacles 
form part of the voluminous musculature, which, so far as I 
can make out, effects the retraction of the aboral sense-organ 
and of the ctenophoral plates, which can be completely with- 
drawn into the body (ef. fig. 5). The tentacles were retracted 
in my preserved specimens, and so it was impossible for me to 


ON OCTENOPLANA. 329 


analyse this very complicated longitudinal musculature. The 
musculature! on one side of the stomachal plane is completely 
separated from that on the other side; so that in sections 
parallel with the stomachal plane, passing through the region 
of the sense-organ, no muscles are visible. But at a short 
distance on either side of the median stomachal plane the 
sections in contracted specimens are almost entirely occupied 
by the convoluted bundles of muscles. Again, beyond the 
region of the ctenophoral plates the sections merely show the 
dorsally placed muscular tentacle lying in its sheath (fig. 10). 

6. The Gonads.—By the discovery of the male genital 
organs of Ctenoplana I have brought a welcome additional 
piece of evidence as to the adult character of the organism.” 
The testes are placed at the bases of the two end-lobes of the 
main portion of the gastro-vascular system. Their position is 
indicated by crosses in fig. 1. They thus consist of two pairs 
of organs, paired about the tentacle axis. They may be either 
simple or lobed and subdivided. They may contain practically 
nothing but mature spermatozoa as in fig. 7; or they may 
contain both mature and developing spermatozoa as in fig. 9. 
Finally, they may possess one or several ducts opening to the 
exterior on the dorso-lateral surface of the body below the 
level of the ctenophoral plates. 

The male genital ducts are merely tubular extensions of the 
tunica propria which encloses each testis. 

In the centre of that portion (always the ventral portion) of 
the testis in which the immature sperm-cells (spermatogonia 
and spermatocytes) occur there is usually to be observed a 
cavity surrounded by large clear cells exactly like those which 
line the cavities of the terminal gastric lobes ; and, in fact, I 
have traced this cavity into communication with the gastro- 
vascular system (cf. figs. 5, 6, and 9). 


1 IT do not include the dermal musculature described by Korotueff, about 
which I am at present in the dark. 

2 Unfortunately I can say nothing about the female reproductive organs. 
It seems unlikely that Ctenoplana should be unisexual. More probably it is 
a protandric hermaphrodite. 


330 ARTHUR WILLEY. 


These central cavities in the immature testes may therefore 
be called the genital ceca, and the genital products appear 
to arise as proliferations of the walls of the ceca (figs. 5 
and 6). 

What is very puzzling is the fact that similar proliferations 
occur on the walls of the terminal gastric lobes themselves 
(fig. 5). On the dorsal walls of the terminal gastric lobes the 
cells of these proliferations appear to assume the properties of 
chloragogenous cells, and numerous yellowish refringent con- 
cretions occur in and amongst them. Somewhat similar re- 
fringent particles are to be seen in the cells of the subjacent 
true endoderm. Finally, in connection with this subject I can 
only mention the fact that as the median walls of the neigh- 
bouring terminal lobes fuse together on nearing the median 
canal which connects them with the stomach, the minute 
cellules which compose the greater part of the proliferations 
in question are replaced by long pyramidal cells which com- 
pose a compact gland, having a radiating structure due to the 
peculiar arrangement of the cells. I will call thisa gastric 
gland, and hope that at some future date light may be 
thrown upon its nature. 

What distinguishes the genital proliferations from the 
above-described gastric proliferations, apart from their 
different topographical relations, is the fact that the nuclei of 
the cellules of the former are of different sizes (fig. 9). The 
larger nuclei I interpret as belonging to spermatogonia and 
the smaller to spermatocytes. Unfortunately 1 am unable 
to make out any nuclear structure in my preparations, although 
they are otherwise well enough preserved. 

The spermatozoa form dense clusters with characteristically 
deeply stained heads and unstained tails. The tails are directed 
both outwardly and mesially. When a testis contains only 
mature spermatozoa there is no longer any trace of the genital 
cecum (fig. 7). 

With regard to the ducts in one individual, I counted no 
less than twelve ducts, which were distributed equally between 
the four testes. Of these ducts 1 was able to see the actual 


ON OTENOPLANA. 331 


opening to the exterior in six. In another individual I counted 
seven ducts altogether. 

It should be added that the above description of the male 
gonads applies exclusively to C. Korotneffii, all three speci- 
mens of which possessed them. 

7. Axial Relations—Comparison of Ctenoplana with 
Planaria and Ctenophora.—As already known from the 
work of Korotneff, Ctenoplana agrees with the Ctenophora 
in the possession of a main axis(Hauptachse) which connects 
the aboral pole with the oral pole, the mouth with the sense- 
organ, and that this main axis forms the line of junction of 
the two principal planes—namely, the tentacular plane and 
the stomachal plane. 

Ctenoplana presents remarkable Planarian affinities in re- 
spect of its dorso-ventrally flattened body, in the possession of 
a definite dorsal surface, and a definite ventral or locomoter 
surface, in its habit of creeping, and especially in its habit of 
attaching itself to the surface-film of water. This enumeration, 
to which may be added the partial ciliation of the ectoderm, 
nearly exhausts the list of its strictly Planarian affinities. 

Besides the coincidence of the main axis and principal planes 
of Ctenoplana with those of the Ctenophora, the chief points 
of affinity are the possession of two pinnate tentacles which are 
each retractile within a sheath, the possession of the eight cteno- 
phoral plates, and the presence of the median funnel vessel. 

The two series of sensory tentacles placed on opposite sides 
of the otolith, whose epithelium is directly continuous with 
the epithelium of the cupule of the otolith, are directly com- 
parable with the polar plates (Polplatten) as described by 
Chun in the Ctenophores. In the first place they agree with 
the latter in lying in the stomachal plane, in so far that they are 
paired about the tentacle axis. This is the most important 
point of agreement morphologically, but they also agree in 

1 According to the remarkable observations of Chun, some Ctenophores 
possess this power, effecting it by spreading out the wall of the stomach,— 
sometimes, as in Lampetia Panceri, nearly everting the stomach as far as 
to the origin of the peripheral vessels. 


332 ARTHUR WILLEY. 


some details. The relation to the otolith-bearing portion of 
the sense-organ is identical in both cases. The polar plates of 
Ctenophores are ciliated, as are the sensory tentacles of Cte- 
noplana. Moreover in the Beroide, according to Chun, the 
thickened margin of the polar plates does not form a simple 
ridge, but is raised up into a series of lappets. This is a very 
remarkable correspondence, and after my observation of the 
double, paired character of the sensory tentacles of Cteno- 
plana I think there can be no doubt that the latter are homo- 
logous with the polar plates of Ctenophores. 

The comparison of the gonads of Ctenoplana with those of 
other forms is not such a simple matter. They agree with 
those of the Ctenophora in being developed about the walls of 
diverticula of the gastro-vascular system, and with those of the 
Potyclades in being enveloped in a tunica propria. But they 
differ from both in the possession of ducts opening directly to 
the exterior. In the Ctenophora the genital products fall into 
the meridional vessels, and are discharged through the mouth ; 
while in the Polyclades, according to Lang, the tunice pro- 
prize which envelop the innumerable testes open into a system 
of intra-cellular genital capillaries which eventually convey the 
sexual products to the vas deferens on each side, by which 
they are ultimately led to the ventrally placed external genital 
pore. 

We now come to the critical consideration of the axial rela- 
tions of Ctenoplana. The problem to be solved is the follow- 
ing :—To what do the planes of the tentacles and of the - 
stomach respectively correspond in bilateral animals? Does 
the tentacle plane of Ctenoplana (‘Trichterebene of Ctenophores) 
correspond to the sagittal plane of bilateral animals or to the 
transverse plane ? 

We shall find, if it has not already appeared evident in the 
foregoing pages, that Ctenoplana unequivocally proves, as I 
think, that the tentacle plane or funnel-plane of it and the 
Ctenophores corresponds to the sagittal plane of bilateral 
animals, and not to the transverse plane. 

At present there exist two interpretations of the axial rela- 


ON CTENOPLANA. 3993 


tions of the Ctenophores, namely, that of Chun! and that of 
Lang.? These may be briefly tabulated as follows. 
According to Chun— 


1. Tentacle or funnel-plane of Ctenophores = Sagittal plane of Bilateralia. 
. Stomachal plane of Ctenophores = Transverse plane of Bilateralia. 
3. Main axis of Ctenophores*® = Longitudinalaxis of Bilateralia. 


[a 


According to Lang— 
1. Tentacle or funnel-plane of Ctenophores = Transverse plane of Polyclades. 
. Stomachal plane of Ctenophores = Sagittal plane of Polyclades. 
3. Main axis of Ctenophores becomes bent (geknicht) in Polyclades. 


tw 


From the above it will be seen that, as regards the tentacle and 
stomachal planes, Lang’s interetation is the exact reverse of 
that of Chun; and yet it is singular that there is no mention 
of such a fundamental discrepancy in Lang’s monograph. 

Selenka (quoted by Lang) held the view that the anterior 
end of a Polyclade corresponded to the aboral pole of the 
Ctenophore, and the posterior end of the former to the oral 
pole of the latter. Lang says he himself formerly held this 
view, but afterwards gave it up as being erroneous. It would, 
however, necessarily follow if the main axis of the Ctenophores 
corresponded to the long axis of Polyclades as stated by Chun. 
The latter view, however, is irreconcilable with Chun’s own 
identification of the tentacle plane of Ctenophores with the 
sagittal plane of Bilateralia, and, in fact, it may be dismissed, 
once for all, as erroneous. 

Chun’s other homologies, however, in respect of the tentacle* 
and stomachal planes are fully confirmed by the conditions 


1 Carl Chun, ‘ Die Ctenophoren des Golfes von Neapel,’ 1880. 

2 Arnold Lang, ‘ Die Polycladen des Golfes von Neapel,’ 1884. 

3 Carl Chun, “Die Verwandtschaftsbeziehungen zwischen Wiirmern und 
Ceelenteraten,” ‘ Biol. Centralblatt,’ Bd. ii, 1882-8. In this paper Chun 
intimates that the main axis of Ctenophores becomes the long axis of Poly- 
clades; but I cannot find out how he reconciles some of the views here 
expressed with the previous statements as to the homologies of the planes 
contained in his monograph. 

4 Chun denominated the plane in which the tentacles of Ctenopkora lie 
the Trichterebene, because there are no tentacles in the Beroide. 


334 ARTHUR WILLEY. 


observed in Ctenoplana. Chun was at first in doubt as to the 
criterion by which to homologise the planes of Ctenophores 
with those of Bilateralia, as the axes passing through these 
planes in Ctenophora were equipolar (gleichpolig). 

The way by which he finally arrived at the conclusion that 
the tentacle plane corresponded to the sagittal plane was so 
remarkable that I will give a free translation of his descrip- 
tion. 

“ Naturally,” says Chun, “ we must disregard all accidental 
conditions of asymmetry by which one of the axes (Kreuz- 
achsen) becomes inequipolar. For example, one seldom finds 
a Cestus veneris in which the two band-like halves of the 
body are equal in length. . . . Should, however, one of the 
axes prove to be inequipolar in such a way that constantly an 
essential organ-complex failed to develop on the one half of 
the axis, then we should have a transition to bilateral symmetry 
which would enable us to speak of a dorsal and a veutral sur- 
face. . . . How surprised was I to find a larva which pre- 
sented a remarkable axial disturbance in the funnel-plane [2. e. 
the tentacle plane]! I give it the provisional name of Thoe 
paradoxa, as I have not succeeded in associating it with cer- 
tainty with any adult Ctenophore. It possesses, in fact, only 
a single tentacle apparatus and tentacle [Fangfaden]. Only 
in the course of the later development is a second tentacle 
apparatus differentiated at the other pole of the axis, so that 
the original disturbance becomes gradually levelled out.” 

In Ctenoplana the tentacle axis and the stomachal axis 
are equipolar ; but if we consider about which axis the paired 
structures are situated, we are simply forced to acknowledge 
that the plane of the tentacles corresponds to the sagittal 
plane,—in other words, that the tentacle axis of Ctenoplana 
and Ctenophora corresponds to the longitudinal axis of Bi- 
lateralia. 

Lang’s theory of the origin of Polyclades from Ctenophores 
rests in the first instance on the assumption that the pinnate 
tentacles of Ctenophora and Celoplana are homologous with 
the sensory tentacles of Polyclades; and his above-quoted in- 


ON CTENOPLANA. 335 


terpretation of the axial relations is framed in accordance with 
this assumption. 

In the first place the fact should be emphasised that under 
no circumstances and from no point of view are the tentacles 
of Ctenoplana bilaterally disposed, but they are biradially 
disposed. 

As mentioned above, it cannot be denied that, in the creeping 
attitude, the tentacles of Ctenoplana present to the onlooker 
the appearance of ordinary transversely paired structures, and 
it may seem difficult to imagine an ancestor of bilateral 
animals with an unpaired tentacle in front and an unpaired 
tentacle behind. But the point is that we have not got to 
imagine this, because in the animals with which we are dealing 
there are no such relations as anterior and posterior, right and 
left. 

As regards the particular homology of the pinnate tentacles 
(Greiftentakel) of Ctenophores with the nuchal tentacles of 
Polyclades, so strongly and, it must be added, plausibly upheld 
by Lang, I venture to think that my observations on Cteno- 
plana, especially as to the double character of the aboral 
circlet of sensory tentacles, justifies me in frankly denying its 
accuracy. From their close relation to the central sensory 
apparatus, and the fact that they are paired about the tentacle 
axis, which I regard as equivalent to the longitudinal axis of 
Polyclades, I suggest that it is much more probable, from their 
relations and function, that the paired multiple sensory ten- 
tacles of Ctenoplana and the polar plates of Ctenophora are 
homologous with the sensory nuchal tentacles of Polyclades, 
than that the latter are homologous with the pinnate tentacles 
of Ctenophora, whose chief function is that of seizing objects 
for food. 

Moreover, from their structure and function, their extreme 
retractility within definite sheaths, and their worm.like 
mobility, it would appear that the pinnate tentacles of Cteno- 
plana and Ctenophora belong to a category of structures 
totally different from that of the nuchal tentacles of Poly- 
clades. They belong, namely, to the same category of struc- 


336 ARTHUR WILLEY. 


tures as the proboscis of Nemertines and of certain Rhabdocele 
Planariauns. 

Finally, we may recapitulate the organs in Ctenoplana which 
are paired about the tentacle axis (cf. fig. 1). There are four 
pairs of ctenophoral plates, three pairs of gastric lobes, two 
pairs of gonads with their ducts, and one pair of multiple 
sensory tentacles. 

From what has been said I regard it as proved that— 


1. The tentacle axis of Ctenoplana = the longitudinal axis of Planarians. 

2. The stomachal axis of Ctenoplana = the transverse axis of Planarians. 

3. The main axis of Ctenoplana = according to Lang, the primary main 
axis of Planarians, which becomes bent as the ganglion is shifted 
towards the anterior end of the body. 

4. The main axis of Ctenoplana and Ctenophores = the dorso-ventral axis 
of Bilateralia. 


8. Synopsis of Species of Ctenoplana. 

(i) C. Kowalevskii, Korotneff—Colour crimson, body in 
swimming attitude shaped like a truncated pyramid ; median 
dorsal surface concave ; free margin of skirt frilled. Habitat, 
west coast of Sumatra. 

(ii) C. rosacea, n. sp.—Colour crimson; body in swim- 
ming attitude of a quadrilateral form; median dorsal surface 
convex; free margin of skirt plain. May be merely a variety 
of preceding species. Habitat, Eastern Archipelago of New 
Guinea. 

(iii) C. Korotneffi, n. sp.—Colour green; body in swim- 
ming attitude roof-shaped; median dorsal surface upraised 
with two upright end-knobs; free margin of skirt slightly 
frilled. Habitat, Hastern Archipelago of New Guinea. 

In Korotneff’s fig. 13 there are only eleven sensory ten- 
tacles figured for C. Kowalevskii,! while C. rosacea had 


1 Korotneff describes an aperture to the exterior on each side in the 
neighbourhood of the tentacles in C. Kowalevskii which leads into a 
“system of canals which branch in the body parenchyma.” Korotneff him- 
self says that after he had preserved his specimen in sublimate he was unable 
properly to orientate it. I have already shown that he completely misplaced 
the tentacles. The apertures in question are obviously the openings of the 


ON CTENOPLANA. aol 


about eighteen. But I should hesitate to insist upon this as a 
specific difference. 

9. General Conclusions.—With regard to the systematic 
position of Ctenoplana, which we now know to be an adult 
animal, I am strongly of opinion that it is an ancestral form, 
and not, as some zoologists seem to suppose, a highly modified 
ereeping Ctenophore. By ancestral form I simply mean a 
primitive archaic form belonging to an ancestral type, and of 
course I do not imply that it is the actual ancestor of any- 
thing in the world. That the Planarians and Polyclades in 
particular have close affinities with the Ctenophora there can 
be no doubt, but it is very much open to question whether 
the former are derived from the latter. The view that the 
Polyclades are so derived seems a reversal of the natural order 
of events, which point to the littoral fauna as the origin both 
of the pelagic and of the abyssal fauna. 

Are we to regard the immediate ancestors of the Turbellaria 
as amorphous forms, like Trichoplax, or forms without any 
kind of symmetry, like Planule or the Mesozoa? Or, on the 
contrary, are we not rather to regard their immediate ancestors 
as forms with some kind of radial symmetry ? 

Having regard to the complete bilateral symmetry of the 
flat-worms, and more particularly their well-developed nervous 
system, with cerebral ganglion in even the lowest forms, I 
cannot imagine them to be derived directly from amorphous 
organisms, but rather from animals which possibly, like Cteno- 
plana, possessed a biradial symmetry. 

Ctenoplana approaches more nearly to a condition of bilateral 
symmetry than the Ctenophores do, in that it possesses very 
clearly differentiated dorsal and ventral surfaces. And this is 
exactly what we should expect to find in the littoral or sub- 
littoral ancestor of such purely pelagic forms as the Cteno- 
phora, the pelagic habit, as is well known, often tending to 
produce a more or less radial symmetry. 


tentacle sheaths, and the “system of canals” are the sheaths themselves, 
which do in fact send off occasional diverticula, possibly due to the contraction 
of their walls during preservation (ef. figs, 5 and 10). 


338 ARTHUR WILLEY. 


On the other hand, a biradial form, like Ctenoplana, 
possesses the potentiality of assuming a strictly littoral 
life, in which the ventral surface is the permanent locomotor 
surface, such an existence leading to a condition of bilateral 
symmetry, according to well-understood physiological prin- 
ciples. 

The ctenophoral plates must have put in their appearance 
for the first time in some form or other; and although it is at 
present beyond the limits of our knowledge to explain how 
they arose, yet it is not right to conclude that the ctenophoral 
plates of Ctenoplana are degenerate or reduced structures 
merely because they are smaller than the ctenophoral rows of 
the Ctenophora. 

It is a groundless assumption to say that Cceloplana and 
Ctenoplana are modified creeping Ctenophores. Ctenoplana 
is an expert crawler, it is expert at hanging on to the surface 
film of water, and it is indeed an expert swimmer. Everything 
it attempts it does well in the old primeval fashion, and there 
is nothing degenerate about it. 

If Celoplana and Ctenoplana are neither Ctenophores nor 
Planarians, what are they? I think it is necessary to create 
a new order of Plathelminthes for their reception; and I 
propose to call the new order the Archiplanoidea, and to 
regard it as equivalent to the orders Turbellaria, Trematoda, 
Cestoda, and Nemertina. 

Furthermore, I should look to the Archiplanoidea for the 
ancestors of all the Plathelminthes (including the Nemer- 
tines) on the one hand, and of the Ctenophora on the 
other. 

The resemblance in form and shape, however superficial, 
between Ctenoplana and the Pilidium larva of Nemertines 
should not pass unnoticed; and it is a remarkable fact that 
the main axis of Pilidium passes through mouth and apical 
sense-organ as in Ctenoplana. 

In the Archiplanoidea, therefore, we have organisms pre- 
senting a transition from radial to bilateral symmetry. 

In the Cerianthide, as we know especially from the works 


ON CTENOPLANA. 339 


of Carl Vogt! and Ed. van Beneden,? we have also forms 
belonging to the category of radial animals, which are un- 
doubtedly physiologically radial, and nevertheless present a 
pronounced bilateral symmetry. 

Van Beneden’s views are clearly set forth in the following 
quotation from his memoir above quoted:—‘‘ Je partage 
entiérement l’opinion de Sedgwick et de Caldwell d’aprés 
laquelle le disque qui porte la bouche et les tentacles, chez 
les Actinozoaires, répond morphologiquement a la face neu- 
rale des Annéles, des Arthropodes, et des Chordés. Je pense, 
comme ces auteurs, que la bouche des Cnidaires est homo- 
logue 4 la fente blastoporique des Arthrozoaires. Les diver- 
ticules coelomiques qui sont, ontogéniquement parlant, la 
cause de la segmentation, répondent aux loges mésenteriques 
des Anthozoaires et les cloisons intersegmentaires sont ana- 
tomiquement équivalentes aux sarcoseptes.” 

If we accept these conclusions side by side with those 
derived from the study of Ctenoplana, we are compelled to 
frame the hypothesis, which I believe to be highly probable, 
of the diphyletic origin of Bilateralia. 

The following scheme will make this view clear, and will 
save a long discussion : 

TC tenophora. 


_A rchiplanoidea————__ 
ian __——Plathelminthes. 
Sheet Sn ee OO 
Cer Se ks 


ee —— CL lenin 


I believe this view will be found to be a natural one in every 
respect ; and if it be regarded by morphologists as substantiated, 
it will certainly relieve the science of morphology of several 
burdens. For instance, Hubrecht’s original speculations as to 
a relationship between the Nemertines and the Chordates, as 
well as Bateson’s comparison of the Nemertines with Balano- 

1 Carl Vogt, “ Des Genres Arachnactis et Cerianthus, f. Arch. de.Bioli 
t. ve 1888. 


* Hd. van Beneden, ‘“‘ Recherches sur le développement des Arachnactis,” 
ibid., t. xi, 1891. 


340 ARTHUR WILLEY. 


glossus and many other such like theories, which at the time 
no doubt appeared to be logical necessities, will be quite ruled 
out of the field of possibilities. 

The descendants of the Archiplanoidea have no ccelom and 
no preoral lobe. The descendants of the Cerianthide have a 
celom and also a proral lobe (excluding the Anthozoa). 

It is an interesting parallelism that the criterion for the 
antero-posterior axis, both in the Cerianthidze and in the 
Ctenophora (Thée paradoxa), was provided by what may be 
called a directive tentacle. It may, indeed, be something 
more than a mere parallelism. 

10. Summary of Principal Results.—(1) Discovery 
of one very distinct new species of Ctenoplana, and of another 
somewhat doubtful new species. 

(2) Observation of the movements and of the pinnate tenta- 
cles of the living Ctenoplana. 

(3) Accurate localisation of the pinnate tentacles. 

(4) Discovery of the double character of the circlet of sensory 
tentacles surrounding the otolith. 

(5) Discovery of the male genital organs and ducts of Cteno- 
plana, thus proving that Ctenoplana is an adult animal. 

(6) Description of the tentacle sheaths. 

(7) Account of the genital ceca of the gastro-vascular 
system, about whose walls occur the genital proliferations. 

(8) Chloragogenous tissue and gastric gland. 

(9) The tentacle axis of Ctenoplana corresponds to the longi- 
tudinal axis of Planarians, the stomachal axis of the former to 
the transverse axis of the latter, and the main axis of the 
former to the dorso-ventral axis of the latter. 

(10) The solid pinnate tentacles of Ctenoplana are not dis- 
posed bilaterally, but biradially. 

(11) The ctenophoral plates, gastric lobes, gonads, gonaducts, 
and aboral sensory tentacles are paired about the tentacle 
axis, 

(12) The aboral sensory tentacles of Ctenoplana are homo- 
logous with the polar plates (Polplatten) of Ctenophora, and 
with the nuchal tentacles of Polyclades. 


ON CTENOPLLANA. 341 


(13) The testes of Ctenoplana are enclosed within a tunica 
propria. 

(14) The pinnate seizing tentacles (Greiftentakel) of Cteno- 
plana and Ctenophora, retractile within definite sheaths, belong 
to the same category of structures to which the proboscis of 
Nemertines and certain Rhabdoccele Planarians belong. 

(15) Creation of a new order, the Archiplanoidea, for the 
reception of Cceloplana and Ctenoplana. The order will thus 
contain two families, the Celoplanide and the Cteno- 
planide. 

(16) Hypothesis of the diphyletic origin of Bilateralia. 


In conclusion, as I am about to leave Sydney, I wish to 
repeat my thanks to Professor W. A. Haswell for his kindness 
and hospitality to me. 


SypNEY; June 29th, 1896. 


EXPLANATION OF PLATE 21, 
Illustrating Mr. Arthur Willey’s paper “‘ On Ctenoplana.” 


N.B.—My material was preserved in sublimate and in a sublimate-acetic 
mixture, and stained with alum-cochineal. 

Fic. 1.—Ctenoplana rosacea, n. sp., from dorsal aspect. The two 
pinnate tentacles are extended, the ctenophoral plates alternate with the 
gastric lobes, the two middle large gastric lobes mark the stomachal plane ; 
in the dorsal centrum is seen the otolith, surrounded by an incomplete circlet 
of ciliated sensory tentacles paired about the tentacle axis; the spots round 
the margin of the skirt represent crimson pigment spots. From living speci- 
men. N.B.—The crosses indicate the positions of the male genital organs. 

Fie. 2.—C. rosacea. In swimming attitude. From living specimen. 
Tentacles retracted. ¢.0. Opening of tentacle sheath. 

Fic. 8.—C. Korotneffi, n. sp. In swimming attitude. The space indi. 
cated by dotted line below the aboral sense-organ was dimly seen through 
the body, and probably represents the space into which the sense-organ is 
withdrawn on retraction, although it might represent the funnel vessel. 

VOL. 39, PART 3,—NEW SER. Z 


342 ARTHUR WILLEY. 


From living specimen. Tentacles retracted. ¢.0. Opening of tentacle 
sheath. 


Fic. 4.—Sketch of Ctenoplana as it may be seen when attached to the 
surface-film of water. In the centre is seen the mouth. The two end-lobes 
of the central gastric system show up white. 


Fie. 5.—C. Korotneffi. Section parallel to stomachal plane (i.e. trans- 
verse to tentacle axis) to show the origin of the genital cecum, &c. The 
section is somewhat oblique. cé/. Ciliated epithelium of ventral surface. 
ce. Genital cecum. ch. Chloragogenous cells. c¢.p. Ctenophoral plate, re- 
tracted. d.e. Dorsal spongy vacuolar non-ciliated epithelium, with mucous 
granules at external surface. These cells would seem to be comparable to 
the so-called “ Glanzzellen” of Ctenophora. ext. Colenteron. ep. Diges- 
tive epithelium, nuclei placed near free end of cells. These are clear faintly 
staining cells with indistinct cell outlines, and with a sharply defined non-ciliated 
free margin. gez. Genital proliferation on the wall of the genital cecum. 
g.p. Gastric proliferation. mes. Mesenchymatous tissue. ¢.0. Opening of 
tentacle sheath. ¢.s. Tentacle sheath with its ciliated epithelium; the 
tentacle itself is retracted further back. v.e. Non-ciliated glandular epithe- 
lium of the ventral surface. 


Fic. 6.—The same. Succeeding section to preceding through gonad to 
show the conversion of the genital cecum into a canal. ce. Genital cecum. 
s. Genital proliferation broken up into polygonal groups of sperm mother- 
cells. 


Fie. 7.—C. Korotneffi. Section through the region of the gonad of 
another individual to show the testis full of mature spermatozoa and the 
opening of the genital duct. d. Tangential section of portion of duct. d.e. 
Dorsal epithelium. g.o. Genitalaperture. ¢.p. Tunica propria with flattened 
nuclei in its walls. 

Fig. 8.—The same. Succeeding section to preceding to show junction of 
genital duct with tunica propria. 

Fic. 9.—Section through a gonad of same individual as that from which 
fig. 5 was taken, to show subdivision of the testis. c@. Genital cecum. 
d. Genital duct. s'. Spermatogonia. s*. Spermatocytes. s*. Spermatidia. 
s4. Spermatozoa. ¢.p. Tunica propria. 

Fie. 10.—Section through tentacle and its sheath. c.s. Cavity of sheath. 
s. Wall of sheath composed of ciliated epithelium. ¢. Tentacle with peri- 
pheral nuclei and central muscle-fibres. In the centre of the tentacle runs a 
core of mesenchyme-cells, mes. p. Branch or pinna of the tentacle. 


Fie. 11.—Schematic figure to explain connections of the lobes of the 
central portion of the gastro-vascular system. /. Position of funnel vessel. 


ATTEMPT TO DEDUCE VERTEBRATE EYES FROM THE SKIN. 343 


An Attempt to deduce the Vertebrate Eyes 
from the Skin. 


By 


H. M. Bernard, M.A.Cantab., F.L.S. 


With Plate 22. 


In the February issue of the ‘ Annals and Magazine of 
Natural History’ I published an outline sketch of a new 
theory of vision. It was suggested that eyes arose as local 
modifications of tissue induced by the excessive crowding of 
pigmented granules at spots most frequently and brilliantly 
illuminated, and that this crowding of the pigmented granules 
might be made to explain both the origin of the eye as a 
structure and its functions as an organ. I now propose, as 
time and opportunity allow, to compare a few eyes with the 
tissues out of which they have, or according to the theory they 
should have. arisen, in order to ascertain how far such com- 
parisons support the theory. 

I take the Vertebrata first, not only because the general 
facts of the structure and development of the eyes in this class 
are most widely known, but because of the paramount interest 
attaching to any questions relating to our own highest sense. 

Now, according to the theory alluded to, eyes must have 
developed out of that tissue in which, under the action of 
light, pigment accumulates. That tissue is the skin; hence 
the eyes in the Vertebrata must have arisen as modifications of 
the skin. There are two kinds of eyes in the Vertebrata; the 
more or less vestigial ‘ pineal” eye, and the Vertebrate eye 


344 H. M. BERNARD. 


proper. These eyes are constructed on different plans, and 
must either, according to the theory, be different modifications 
of the same kind of skin, or else modifications of two different 
types of skin. I shall endeavour to show that the latter 
assumption best accords with the facts ; that the change which 
undoubtedly took place in the character of the skin as the 
Vertebrata arose out of their Invetebrate ancestors is sufficient 
to explain the difference between these two types of eye. 

Following the lines of the theory, I make two assumptions :— 
(1) The retina is but a specialised portion of the epithelial 
layer of the skin, between the cells of which the pigment 
granules from the subjacent chromatophoral layer stream 
outwards under the action of light. (2) The retina and the 
chromatophoral layer must have been in intimate and insepar- 
able association through all the stages of the evolution of the 
eye. I have, then, to try to show how far the facts relating 
first to the structure, and secondly to the development of the 
Vertebrate eyes can be harmonised with these assumptions ; how 
far, indeed, it is possible to deduce the Vertebrate eyes directly 
and continuously from the skin. 

A glance at the diagrams (Plate 22) shows that a striking 
parallel between the eyes and the skins, out of which I assume 
them to have arisen, can be easily instituted, but the details 
are by no means easy to work out, and still more difficult to 
demonstrate. The greatest difficulty, of course, lies in the 
fact that, according to their ontogenetic histories, the eyes 
either wholly or partly developed from the brain. It is true 
that no one has yet succeeded in elucidating these embryological 
records ; nevertheless they hold the field, even though amount- 
ing to little more than bald assertions that the eyes developed 
from the brain and not from the skin. Before critically exa- 
mining this adverse testimony I propose to marshal all the 
available facts which seem to connect the eyes directly with 
the skin, and to suggest possible explanations of the various 
modifications and specialisations of the tissues of which our 
eyes are constructed. Only after having shown that it is 
possible to deduce the eyes as continuous structural modifica- 


ATTEMPT TO DEDUCE VERTEBRATE EYES FROM THE SKIN. 349 


‘tions of the skin shall we examine the embryological records, 
and endeavour to show that the facts are, in the main, in har- 
mony with the argument based upon structure. 

I take the pineal eye first, in spite of the disadvantage 
involved in this choice, arising out of the many debatable 
details concerning this organ. Its claim even to be an organ 
of sight at all has been disputed. Assuming it, for the present, 
to have been an eye, we are, it seems to me, justified for many 
reasons, which will be more apparent later on, in regarding it 
as having preceded in course of time the Vertebrate eyes 
proper. Its place in this discussion is, therefore, that which 
is suggested by its history as here interpreted. 

The close resemblance of the pineal organ to an eye of the 
Invertebrate (Molluscan) type was first pointed out by de 
Graaf.| The suggestion has not been accepted, for reasons 
chiefly embryological, nevertheless I am convinced that it is 
correct. If the pineal organ ever was an eye—and all the 
evidence points that way,—it should, according to our theory, 
have developed directly from the skin. With this hypothesis 
it entirely agrees, its structural relationships being precisely 
those which they should be had it arisen simply as an invagina- 
tion of a skin of the Invertebrate type, i.e. of a skin consisting 
of an external palisade layer supported internally by a layer of 
connective tissue (Diagram1). Regarding its structure alone, 
then, and this is the only point that at present concerns us, 
there is no difficulty in believing that the pineal eye deve- 
loped first as an optic pit from the skin of the ancestors of the 
Vertebrata, before that skin had assumed the Vertebrate type, 
i.e. before the palisade layer had become protected externally 
by the mucous and horny layers. 

Dealing with the development of the pineal eye more in de- 
tail,—according to the theory, the retina arose by a secondary 
multiplication of the epithelial sensory cells for the apprecia- 
tion of the variations in intensity of irritation caused by the 
movements of the pigmented granules towards the exterior ; 
while the dioptric apparatus was due to the accumulation 

1 ¢ Zool, Anz.,’ 1886. 


346 H. M. BERNARD. 


at the surface of slime wholly or partially produced by the 
excessive discharge of the pigmented granules. This refrac- 
tive mass, at first probably lying merely at the surface, would 
eventually be enclosed in an invagination, a phenomenon so 
common (e.g. in the formation of glands) that a discussion 
of the mechanics by which it might be brought about need not 
here detain us. 

The formation of a lens, the definite structure of which 
is now perhaps only indicated by the lengthening of the pali- 
sade cells as shown in the diagram, brought the pineal eye as 
an organ of sight to a high functional level. From this it has 
steadily declined, until it persists as a vestige which, in many 
cases, can no longer function in any way for the appreciation 
of variations of light intensity. 

In support of this interpretation of the morphology and 
physiological origin of the pineal eye, I would call attention 
to Spencer’s figures,! a study of which leaves no doubt what- 
ever that the pigmented cells are streaming from the connective- 
tissue capsule round the eye through the retina, just as in the 
rest of the skin they stream out from the cutis through the 
palisade layer of the epidermis. In the series of sections of 
Hatteria, kindly lent me by my friend Mr. Martin Wood- 
ward, in addition to the pigmented cells, there are others 
making their way alike through the retina and the palisade 
layer of the epidermis. These are cells containing enormous 
vacuoles which force the nuclei to one side. On reaching the 
horny layers these cells flatten out, and their vacuoles form 
flat spaces which give the cuticle a sort of false lamination. 

A still closer parallel between the pineal eye and the skin is 
observed in Uromastix, for the loan of sections of which I 
am again indebted to Mr. Woodward. Streaming through the 
cutis are innumerable cells laden with concretions which are 
black by transmitted, white by reflected light. They remind 
one of the guanin granules found in the tissues of Arachnids. 
Now, while the cells containing these white concretions do not 
seem to be able to pass through the palisade layer, but stop 


1 «Quart. Journ. Mier. Sci.,’ xxvii. 


ATTEMPY! TO DEDUCE VERTEBRATE EYES FROM THE SKIN, 347 


short within the cutis beneath it, the cells which contain the 
red pigmented granules pass freely on, and are seen everywhere 
forcing their way into the epidermis, and, I believe, helping to 
build up the horny layer. 

On turning to the pineal eye, what do we find? A close 
network of cells filled with similar white matter enveloping 
the eye externally. The white matter here also fails to climb 
up into the retinal cells, just as in the skin it fails to climb up 
into the epidermis, while on the other hand passing up between 
these cells are others laden with red pigment, which makes its 
way into or between the retinal cells, just as the pigment from 
the cutis passes between and into the palisade cells of the 
epidermis.! 

In addition to the pineal eyes of Hatteria and Uromastix, 
I have examined this organ in Petromyzon planeri, ammo: 
coetes of which I obtained near Jena and fixed in corrosive 
sublimate. As is now well known, there is no black pigment 
in these eyes; they appear to me to have become a receptacle 
for concretions similar to those already noted in the cutis of 
Uromastix.? The condition of the eye itself appears to indi- 
cate extreme degeneracy. Having sunk below the skin it 


1 Comparison of the pineal eye and skin of Uromastix reveals another 
fact of some importance for the theory of light sensation which I have 
ventured to put forward. The pigment granules are in this case much 
larger and more clearly defined in the eye than in the skin. It is a noticeable 
fact that in eyes in general the pigment grains are much more sharply and 
distinctly granular, and also apparently not seldom rather larger than are the 
pigment granules in the rest of the body. ‘This fact is certainly in favour of 
my suggestion that their passage up and down between the rods of the 
retina causes a mechanical stimulation. 

2 [ therefore doubt the wisdom of Gaskell’s description of this matter as 
“white pigment.” It is worth noting also that the white concretions, 
which have no definite shape in the cutis, are pointed and fusiform in the cells 
round the eye in Uromastix. ‘The guanin granules have the same or a 
very similar shape in the argentea of fishes. I have seen similar white 
long-oval plates in the eyes of spiders. These facts suggest the possibility 
of associating these last-named granules with the guanin “crystals” found 
in such quantities in the bodies of these as of other Arachnids. (On their 
origin cf. ‘Journ. Roy. Micr. Soc.,’ 1893, p. 427.) 


348 H. M. BERNARD. 


seems no longer in a position to have much pigment pass out- 
wards through it, and, if any travels through, it probably passes 
right through without being even temporarily arrested, pre- 
sumably because the resistance of the retina to the passage 
of the pigment which in the functional eye is, according to 
the theory, the cause of the irritation and resulting sensation 
has ceased. In the same way the resistance of the retina to 
the entrance of the white matter observed in Uromastix has 
also ceased, so that it streams outwards through the cells in 
single shapeless or angular granules [which are of very differ- 
ent sizes, and have no resemblance to true pigment granules 
either in appearance or distribution], to accumulate at the distal 
ends of the degenerate retinal cells. Beyond this point they 
seem unable to travel. I know of no evidence which would 
lead me to believe that the coagulum in the cavity of the eye 
is due to any transformation of these granules; on the other 
hand, it is quite possible that it may result from the occasional 
passage and conversion into slime of true pigment. 

With regard to Leydig’s doubts! as to whether the pineal 
eye was ever an eye at all, his objections are based upon the 
very facts which, it seems to me, establish the point beyond 
dispute. He appeals to the presence of pigment in the lens. 
But this is exactly what we might expect in an eye going out 
of function. The pigmented cells ball together as they reach 
the free spaces among the retinal cells outside the palisade 
layer; then streaming through the retina they ball again on 
reaching the cavity of the eye, and may accumulate as globular 
masses against and within the lens.2 For some reason or 
other, as the eye ceased to function, the pigmented granules 
ceased to clarify on reaching the cavity of the eye, as they 
appear to do, say, in the eyes of molluscs, of Petromyzon 
(see below), on passing into the vitreous humour through the 


1 © Abh. Senk. Nat. Gesch.,’ xvi, 1890, p. 531. 

2 In some cases the pigment seems (in Spencer’s figures, |. c.) to be forcing 
its way between the lens cells; this we should imagine would be the most 
natural. It is not easy to understand why, in other cases, the cells should 
ball together in the heart of the lens. 


ATTEMPT TO DEDUCE VERTEBRATE KEYES FROM THE SKIN. 349 


cells in the pars ciliaris retine of Vertebrate eyes (see below), 
and as they certainly do in the rete mucosa of the Vertebrate 
skin. That they formerly clarified in the cavity of the pineal 
eye also we have some evidence in another of the facts which 
Leydig adduces as a reason for disbelieving the original ocular | 
function of these structures. The remains of the homogeneous 
clear substance which I assume once filled the cavity of the 
eye still here and there persists, as Leydig himself has shown. 
It appears in some cases as bristle-like streaks or threads of 
clear matter streaming outwards from the surface of the retina, 
or even in a thick layer like a cuticle. That this substance 
ever formed a definite system of rods turned towards the 
cavity of the eye, such as has been suggested by Gaskell,! I 
think somewhat doubtful, although it is quite possible that 
here and there some such differentiation may have taken 
place (cf. the “ rods ” of the Cephalopod eye). 

Summing up this brief sketch of the pineal eye, we note 
that structurally (its embryology will be discussed later 
on) it is quite explicable as a simple invagination of a skin 
of the Invertebrate type, and may well have arisen in the 
manner suggested by our theory, while our ancestors still 
possessed such an undifferentiated epidermis. 

An eye arising in this way would necessarily be what is 
known as a direct vision eye,—that is, the nerves would end in 
the retina without any bending back upon themselves. 

The Vertebrate Eye Proper.—In process of time the 
skin lost the simple character it possessed when, according to 
the foregoing, the pineal eye arose. Cells budded off from 
the palisade layers, while others (e. g. pigment-bearing cells) 
migrated through the palisade layer, and these together, the 

1 *Quart. Journ. Micr. Sci.,’ vol. xxxi. At the same time it is obvious 
that the interpretation of the facts here adopted is hardly reconcilable with 
the deduction of Vertebrates from the Arthropods; a chitinous exoskeleton 
does not lend itself to such simple invagination as we have here assumed. 
I do not think the depression in the centre of the retina of the pineal eye of 
Petromyzon planeri at all justifies Gaskell’s comparison of that eye with 


the eye of the Acilius larva. A comparison with other pineal eyes shows that 
that depression is hardly primitive ; it may even be a result of degeneration. 


390 H. M. BERNARD. 


former being the more important element, built up a layer of 
cells, the outermost of which typically harden into horny 
scales. We thus get the Vertebrate type of skin as shown to 
right and left of Diagram II. The cutis is richly provided with 
blood-vessels (shown diagrammatically as loops), the chroma- 
tophores are numerous, and those which have reached the 
epidermis are seen forcing their way up between the palisade 
cells. Lastly, the ultimate ramifications of the integumentary 
nerves no longer end, as in the primitive skin, in or among 
the palisade cells, but, penetrating that layer, terminate 
among the cells of the rete mucosa. 

In this change in the character of the skin we can perhaps 
find, on the one hand, a partial explanation of the degenera- 
tion of the pineal eye, and, on the other, a clue to the chief 
differences between that eye and the Vertebrate eye proper. 

The pineal eye, sunk below the palisade layer and no longer 
in organic connection with it, would probably suffer by any 
change such as that described in the character of the outer 
skin. The separation of the ocular vesicle from the palisade 
layer would hamper its control over the development of that 
layer, which might or might not run a course favourable to 
the eye as an organ of vision. As far as we can see, the 
secondary thickening of the outermost. layers by stratification 
from the palisade layer was not calculated to benefit the pineal 
eye. It would impede its function, and therefore cause it to 
degenerate. A still more important factor making for degene- 
ration is probably to be found in the rise and development of 
the skull. A review of the available facts relating to these 
eyes involuntarily suggests that they—there were, it seems, 
originally two!'—succumbed, first one and then the other, 
before the advancing edges of the bony plates which developed 
to protect the ever-enlarging brain. This degeneration would 
be hastened if new and perhaps more efficient eyes developed 
to replace and more than compensate for the loss of the old. 

The eyes which appear actually to have replaced the pineal 

1 Owsianikow, ‘Mem. Akad. St. Petersburg,’ 7, xxxvi, 1888; also Locy, 
‘ Anat. Anz.,’ 1894, p. 169. 


ATTEMPT TO DEDUCE VERTEBRATE EYES FROM THE SKIN. 301 


eye (or eyes) would again, according to our theory, have 
developed out of the skin; but, inasmuch as the character of 
the skin had changed, the type of eye would differ from that 
of the pineal eye. The same physiological principle would, 
however, come into play; the cells of the palisade layer 
would be irritated by the variation in the pressure of the 
pigmented matter travelling between them under the action 
of light, and these cells could become sensory cells by asso- 
ciation with the nerve-endings now found among the cells of 
the rete mucosa. The new eye might thus from the first be, as 
indicated in Diagram II, an inverse eye, i.e. the nerve-fibres 
(n), passing outwards towards the surface, would have to bend 
back on themselves in order to become associated with the 
palisade-retinal cells. 

The gradually thickening nerve-strands of this hypothetical 
epidermal eye may be assumed to have ultimately passed in a 
group through the palisade layer at the (? posterior) edge of 
the new specialised sensory area (see diagram). This ever- 
thickening and broadening nerve-strand appears to have carried 
through with it one or more loops of the vascular system of 
the cutis, together with strands of connective tissue. 

The specialisation of a portion of the palisade layer to form 
the retina of a purely epidermal eye such as that figured in 
the diagram would, in course of time, prevent that portion 
from yielding any mucus or horn cells towards the exterior, 
more especially as the palisade cells of the retina must be 
supposed to have early specialised into cuticular rods. The 
deficiency caused by this more or less effective barrier to the 
free passage of pigment granules through the retina towards 
the surface would have to be made up somehow, not only 
to form a dioptric apparatus, but also for the protection of 
the retina. It was along time before I could see how the 
palisade cells surrounding the retina could cover it over with 
a horny layer. The solution to the problem suggested in the 
diagram was arrived at by working backwards from the lens 
of the definitive eye (cf. Diagrams IV, III, II). The palisade 
cells round the periphery of the primitive retina appear to 


ope H. M. BERNARD. 


have lengthened greatly, and bent over the specialised area 
somewhat in the manner indicated. These long cells would 
therefore be merely a modification of the ordinary prickle- 
cells. By the detachment and death of their distal portions 
they would be able to yield a succession of horny cells, so as 
to maintain a continuous hard covering over the retina. This 
covering appears eventually to have formed a primitive surface 
lens, moved by fibrils from the intrusive connective tissue. 

The next stage in the development of this hypothetical 
epidermal eye was probably an invagination of the whole of 
the modified portion of the skin, including the lengthened 
palisade or lens cells. This invagination is shown in Diagram 
III. Two factors may have helped to bring it about. 1. The 
retina becoming bulged inwards by the great accumulation of 
slimy fluid in the cavity of the eye, the tendency of the 
surrounding skin would be to constrict off the vesicle so 
formed. 2. The lens being attached by fibrils to the place of 
entrance of the optic nerve, and probably slightly moveable, 
would tend to be drawn in with the retina. Any way, if the 
eye started at all as we have suggested, we are justified in 
assuming an invagination such as that described, for that alone 
appears able to explain the arrangement of the tissues, among 
which I would call attention to the loops of the vascular system 
of the cutis, which, diagrammatically represented, assume the 
positions. shown by faint dotted lines in Diagram III.! 
Further, the remains of the neck of such an invagination may 
still perhaps be seen in the fibrous ring which persists after 
the formation of the aqueous chamber as the ligamentum 
annulare (fishes) or ligamentum pectinatum (other Vertebrates), 
and which, it is important to notice, connects the iris (and not 
the lens) with the cornea (lig. an., Diagram IV). 

If we picture to ourselves the change of shape which the 
spoon-shaped retina shown in Diagram II would have to 
undergo in passing to the form shown in Diagram III, the 
origin of the choroidal fissure is at once apparent. As the 


1 In this I am following the familiar diagrams of the blood-vessels of the 
Vertebrate eye to be found in text-books. 


ATTEMPT TO DEDUCE VERTEBRATE EYES FROM THE SKIN. 353 


retina bulged further and further inwards, the nerve-strand 
connecting it with the brain shifted down towards the base of 
the cup. It is quite indifferent whether we speak of the nerve- 
strand cutting down the sides of the ever-deepening retinal 
cup, or of the sides of the cup growing up on each side of the 
nerve. The vascular and connective-tissue elements which 
accompanied the nerve (see Diagram II) would now protrude 
into the cavity of the eye. Inthe completely invaginated eye 
(shown in Diagram III) it is obvious that the choroidal fissure 
could originally only have extended from the edge of the long 
palisade or lens cells to the entrance of the optic nerve. 

In Diagram IV I have indicated some of the adaptations 
which would be necessary to complete the Vertebrate eye. The 
lens had to be isolated, and its long cells rolled under it from 
all sides. We need not assume that the earliest isolated lens 
had its fibres arranged in the highly specialised manner 
characteristic of the higher Vertebrates; any folding of the 
long palisade cells under the more hardened central portion 
would be sufficient. The freedom of the lens and folding 
under of its long cells could, it seems to me, be brought about 
if the lens itself, under the action of contractile fibres in- 
truding with the optic nerve into the cavity of the eye, was so 
far moveable that its axis deviated through small angles from 
the optic axis of the eye. 

With regard to this method of accounting for the isolation 
of the lens, the following considerations are worth notice. 
The eye itself would have been immoveable through all the 
earliest stages of its evolution, but the Vertebrate eye is, as 
we know, a highly moveable structure. We are justified in 
concluding that its present powers of movement are but more 
perfect ways of attaining ends which, in its earliest stages, it 
must have striven to attain. It is not unlikely, therefore, that 
the connective-tissue strands which entered the eye in its 
earliest stages became attached to the edges of the primitive 
surface lens, and effected some simple movements. As the 
retina bulged inwards, these simple lens movements might be 
supplemented by slight movements of the whole eye,—these 


354 H. M. BERNARD. 


double movements probably playing some part in bringing 
about the invagination depicted in Diagram III, an invagina- 
tion which, as above insisted upon, must have included the 
lens. Whether the eyeball formed by this invagination ever 
moved with any degree of freedom under the cornea may be 
doubted ; the fibrous connection persisting in the ligamentum 
annulare was almost certainly strong, and may well have been 
rigid from the first. Certain it is that the movements of the 
eye asa whole, i.e. of the ocular invagination together with 
the skin above it (cornea), not only gave rise to the conjunctival 
folds by which the eye is suspended in the skin, but also ren- 
dered the directive adjustment of the lens by contractile fibrils 
unnecessary. As is well known, fibrils for the movement of 
the lens (“ m. retractor lentis ”—Beer) still persist in the eyes 
of the bony fishes, not only for the adjustment of focus, but 
also, according to Beer,! for slight changes in the direction of 
the line of vision. It is in keeping with these facts that the 
eyeball itself in the fishes is much less moveable, and in this 
respect also more primitive, than it is in the higher Verte- 
brates. We shall have presently to note other primitive 
features connected with the eyes in the bony fishes. 

On the isolation of the lens and the consequent abstraction 
of a large portion of the palisade layer from the front of the 
ocular globe, we may assume that this layer was regenerated, 
as shown by dotted lines in Diagram IV. 

The aqueous chamber developed as a space among the fibrous 
tissue forming the neck of the invagination: it would thus 
be morphologically a lymph space in the cutis. I have indi- 
cated the splitting apart of the fibres in this place in Diagram 
III. The aqueous chamber would thus from the first be quite 
distinct from the vitreous chamber, which was primarily a cleft 
in the epidermis (see below). 

The iris could be developed by the thinning away of the 
fibrous substance above the lens, and its radial contraction 
away from the optic axis. This contraction may be supposed 
to have given rise originally to the ciliary processes, as shown 

1 « Pfliig. Arch.,’ Bd. liii, 1894. 


ATTEMPT TO DEDUCE VERTEBRATE EYES FROM THE SKIN. 355 


in the diagram. These are, however, now secondarily spe- 
cialised. 

It was an integral part of my theory, not only that the 
accumulation of pigment granules in the illuminated spots 
gave rise to the sensory visual areas in the epithelium by the 
irritation caused by their passage between the epithelial cells, 
but also that the excessive discharge of such granules supplied 
or helped to supply the refractive matter for a dioptric 
apparatus. This supposition entirely harmonises with this 
suggested deduction of the Vertebrate eye directly from the 
epidermis. 

When a certain area began to form a retina, i.e. in the 
earlier stages of the condition shown in Diagram II, the excess 
of pigment would pass out between the sensory cells, and 
would supply so much matter to the epidermis cells outside 
the retina; here it would be clarified, as the pigment is 
typically clarified in the epidermis, and added to the thickness 
of the refractive layers. As the sensory cells multiplied, and 
were further cuticularised as rods, and the relation. between 
them and the irritating chromatophores became more special- 
ised, the way to the exterior would be barred, and the necessary 
material would have to be brought in from the circumference 
of the retina. 

We have probably to look to the long cells depicted in the 
diagram for the bringing of the matter used in forming the 
harder refractive portions of the primitive lens. But, in 
addition, pigmented cells would almost certainly travel into 
the cavity of the eye along the connective-tissue strands 
which accompanied the nerve and blood-vessels. The rapid 
clarifying of this pigmented matter in the fluid contained in 
the space between the lens and the retina would give rise to a 
slimy mass, the rudiments of the vitreous body. The deriva- 
tives of this intrusive connective tissue would be found in the 
processus falciformis and pecten, which still convey pigment 
into those eyes that retain such structures. As the eye ad- 
vanced in specialisation, and this important channel for con- 
veying pigment into the eye to supply material for the vitreous 


356 H. M. BERNARD. 


humour was almost entirely lost (Diagram IV), compensation 
would be found in the great increase of the inner surface of 
the eye, the posterior half of which alone could function as 
retina; the other, anterior half, consists of the undifferentiated 
palisade epithelium, which allows the pigment to pass 
through as slime into the vitreous humour. It is also pro- 
bable that some material for the same purpose finds its way 
along the optic nerve into the cavity of the eye. 

This suggested origin of the vitreous humour, viz. that it is 
largely due to pigment granules passing into the eye cavity 
and dissolving into slime, is not only in accord with our 
homology of the retina with the epidermal palisade layer which 
gives rise to the mucus layer, but it is also supported by 
certain facts. In sections of the eyes of embryo chickens, 
dendriform exudations of slime, obviously forming the vitreous 
humour, are seen coming from the ciliary portion of the retina. 
The vitreous body is known to be easily detachable from all parts 
of the interior of the eye, except the ciliary region and the 
place of entrance of the opticnerve. The pigment granules are 
known to pass from the pigmented epithelium into the palisade 
layer in the iris, and the process can be seen beginning before 
the lens is reached. The actual microscopic appearances in 
fortunate sections are all in favour of their supplying slime 
for the vitreous humour. In my series of sections of the larva of 
Petromyzon planeri the retina is only in contact with the 
pigmented epithelium in the axis of the eye, only here are any 
rods and cones found ; where the retina is not in contact with 
the pigmented epithelium, not only are no rods and cones 
developed, but the pigment does not accumulate in the choroid 
epithelium. It appears as if, there being no layer of rods and 
cones to hinder its. advance, the pigment granules pass 
freely through, and form masses of semi-clarified granules 
within the space between the retina and epithelium. In the 
ciliary region of the retina, where the two layers—the retina 
(without rods) and the epithelium—come gradually into con- 
tact, all the appearances are as if the pigment granules are 
streaming through the retina into the vitreous humour. The 


ATTEMPT TO DEDUCE VERTEBRATE EYES FROM THE SKIN. 307 


eye is, however, so immature that this latter is not yet a 
clear slimy fluid, but a dense mass of fine granules, quite 
referable, as far as microscopic appearances go, to pigment 
granules only partially dissolved and clarified.! 

This, in brief, is the way in which it is possible to conceive 
that the eye might have been developed directly from the 
skin, as our theory demands. That it actually has developed 
in this way is perhaps difficult to prove in face of the very 
different history suggested by its embryology ; still, when the 
evidence is summed up, it appears to be of considerable 
weight. 

I claim, for instance, that by this method of deriving the 
eye, the arrangement and character of all the more important 
tissues are fully accounted for.” 

As examples I may draw attention to the following points : 

1. The pigmented epithelium is not an ordinary epithelium, 
but rather a close layer of chromatophores, some of which 
are large ten-sided cells with two nuclei,’ which latter feature 
is not infrequent in chromatophores. 

2. The presence of the stellate chromatophores in the 
choroid would be a vestige of the dense chromatophorai layer 
that seems to have been very generally present in the cutis of 
many pre-mammalian Vertebrates. 

3. The fibrous connection between the front surface of the 
iris and the cornea, specially pronounced as the ligamentum 
annulare of fishes, might well be the remains of the neck of 
the assumed invagination. It is attached to the iris and not 
to the lens, which is of significance for this method of deducing 
the eye from the skin. 

4, The difference between the vitreous and aqueous humours 


1 I propose to enlarge on this subject in an illustrated paper dealing 
specially with skins and pigments. 

? I have omitted all mention of the sclerotic because I believe its develop- 
ment was concerned with that of the conjunctival folds, and of the specialised 
muscles for the movement of the eye, which latter did not come within the 
range of the inquiry. 

3 Boden and Sprawson, ‘ Quart. Journ. Mier. Sci.,’ vol. xxxiii. 

VOL. 39, PART 3.—NEW SER. AA 


358 H. M. BERNARD. 


is also accounted for. A comparison of the diagrams shows 
at a glance that, while the latter is simply a lymph space (cf. 
Fuchs’s diagram of the lymphatics of the eye!) arising within 
the connective tissue of the cutis, the former is an epidermal 
space, and is accordingly filled with slimy matter, presumably 
identical with that which gradually turns the cells of the rete 
mucosa into the horn cells. 

5. The shape and curious arrangement of the cells forming 
the lens, the gradual loss of their nuclei, and their progressive 
transformation into hard refractive matter as the centre is 
reached are simply explained by regarding them as the prickle 
and horny cells which formed at one time a primitive surface- 
lens, subsequently engulfed in the invagination. That these 
fibres are closely comparable with epidermal prickle-cells can 
be demonstrated under the microscope ;* they have all the 
known characters of such cells, in addition to their gradual 
decay and conversion into hard refractive matter. The pig- 
mented granules of the iris on the one hand and the vitreous 
humour on the other could obviously supply them with abun- 
dance of slimy matter for this purpose. 

6. I may be allowed again to allude to the arrangements 
of the blood-vascular loops as shown in the diagrams; to the 
ciliary process, which is accounted for as a simple mechanical 
result of the formation of the iris (cf. Diagrams III and IV) ; 
and to such structures as the pecten and processus falciformis, 
and the intra-bulbar blood-vessels still persisting in different 
eyes among the lower Vertebrates. 

In addition to the weighty evidence afforded by these purely 
morphological arguments, it is worth noting that the recent 
discovery of Wolff,? fully confirmed by Miller,‘ as to the re- 
generation of the lens in Triton from the iris, is in the main 
in accord with Diagrams III and IV. The details as given 
by these authors are, however, very curious. The outer of 


' Reproduced in Morris’s ‘ Treatise on Anatomy,’ 1893, p. 889. 
2 Cf. the first note, p. 357. 

3 * Archiv fiir Entwick. Mechan.,’ i, 1895. 

4 Eric Miller, ‘Arch. f. micr. Anat.,’ p. 23, 1896. 


ATTEMPT TO DEDUCE VERTEBRATE EYES FROM THE SKIN. 399 


the two inner pigmented layers of the iris, viz. that con- 
tinuous with the pigmented epithelium, becomes the lens 
epithelium on the distal face of the lens, while the inner 
yields the long lens-cells. This naturally raises the question 
whether the lens epithelium may not be the homologue of 
the pigmented epithelium, the lens fibres alone representing 
the palisade layer. It may be that this regenerative 
process is strictly recapitulatory, and that this is the right 
explanation of the facts. On the other hand, the lens epi- 
thelium passes so gradually into the lens fibres round the 
equator of the lens, that it is difficult to believe that the whole 
structure is not the result of the folding of a single con- 
tinuous epithelium. Again, the lens in the Vertebrate eye can 
hardly be such an entirely new structure, dating, as in the case 
of the lens of the pineal eye, simply from the time of the 
invagination, which this new suggestion would compel us to 
believe. It seems to me far more probable that it was an early 
modification of the external horny layers of the epidermis, 
which, as illustrated in the diagrams, became involved in an 
invagination. A comparative study of lenses, including those 
in degenerate eyes, might throw further light on this subject. 
The embryological development of the lens is, according to 
our view, so purely adaptive that it can hardly help us. 

It is obvious, further, that this method of deducing the 
Vertebrate eye from the skin has the advantage that this organ 
need no longer be any exception to the rule which obtains in 
the animal kingdom, that organs of sense develop directly and 
continuously out of the skin. Lastly, we must not forget the 
significance of the fact that the change of type which can be 
seen to have taken place in the Vertebrate eyes is easily refer- 
able to the change which’we know must have taken place in 
the character of the skin from the Invertebrate to the Vertebrate 
type. 

Having, then, shown that it is possible physiologically and 
morphologically to deduce the Vertebrate eyes from the skin, 
the pineal eye from the skin in its Invertebrate condition, the 
definitive eye from the skin in its Vertebrate condition, we 


360 H. M. BERNARD. 


have to face the embryological histories of these eyes and the 
interpretations usually put upon them. 

Taking the definitive eye first, it is found to arise ontogeneti- 
cally by the union of two distinct structures. The ‘ primary 
optic vesicle,” as it is called, develops as an invagination of 
the brain towards the skin. This collapses into itself so as 
to form a cup or spoon, open towards the skin,and joined to 
the brain by a stalk or handle. The anterior wall, lining the 
hollow of the cup or spoon, becomes the retina and retinal 
palisade layer; the posterior wall, forming the outside wall of 
the cup or spoon, becomes the pigmented epithelium in contact 
with the palisade layer. Into this cup there dips down from 
the outer skin an invagination which becomes constricted off 
and forms the lens. The outer rim of the cup or spoon grows 
over the lens so as to form, with other elements, the iris. 

The question we have to try to answer is, are these pro- 
cesses even approximately historical, or are they purely em- 
bryological adaptations in order to obtain a desired end by a 
series of short cuts? The answer must depend entirely upon 
the weight of the evidence. It must be remembered that we 
have no direct evidence whatever. Our sole guides are infer- 
ences to be drawn from known facts, and these we are only 
roused up to use when discussing some proposed clue to the 
right understanding of the said facts. In the present instance 
a theory of light sensations compels me to assume that the 
Vertebrate eyes have developed from the skin, and in the fore- 
going pages I have endeavoured to show that the known facts 
of structure and function are explicable on this assumption. 
The morphology and physiology of the eye, as the latter is 
interpreted by the theory, go throughout hand in hand. When 
we turn to the embryological development our connected 
story is thrown into confusion. Some of the details, it is true, 
are in most satisfactory agreement, as we shall presently see; 
but the most essential processes, viz. those just described, 
differ entirely from our scheme. 

In the face of this serious difficulty, are we to withdraw or 
boldly to examine the embryological processes in order to 


ATTEMPT TO DEDUCE VERTEBRATE EYES FROM THE SKIN. 361 


ascertain (1) whether any plausible phylogenetic explanation 
can be given them? and (2) whether they cannot be more 
simply explained as adaptations? With regard to the former 
test, I confess at the outset that it seems to me almost waste 
of time to discuss the possibility of these complicated in- 
foldings and fusions of separate parts indicating the history 
of such a singularly compact organ as the eye, especially if we 
are to assume, what we are justified in assuming, an unbroken 
continuity of function from the earliest rudiments to its most 
complicated development. We are, in the first place, pre- 
cluded from believing that the functional eye developed con- 
tinuously from the brain towards the skin which then formed 
a lens, as we should have no explanation of the bending back 
of the nerve-fibrils. In order to get over this difficulty it is 
suggested that the primary optic vesicle, i.e. the invagination 
from the brain, may represent a primitive eye of the type 
known as an “optic pit.” It is true that, according to our 
interpretation of the facts, we have evidence in the pineal eye 
or eyes that such structures did at one time exist on the heads 
of the ancestors of the Vertebrata. Ifa pair of such eyes were 
caught and eventually folded in by the medullary groove, they 
might, it is thought, give rise to these “ primary optic vesicles,” 
with the retine in the right position to become the retine of 
the paired Vertebrate eyes. And, at first sight, there seems 
to be some possibility of deducing the Vertebrate eyes from a 
pair of engulfed, and therefore vestigial, and for the time 
being functionless eyes, which may have started into life once 
more by coming into contact on each side with some thicken- 
ing, or perhaps glandular invagination, of the ectoderm. This 
latter, by condensing the light upon the pigment still present 
in these buried eyes, might once more set it in regular move- 
ments, which would record the variations in the intensity of 
the light. Some such hypothesis, it is claimed, would explain 
what is called the inversion of the eye, for the light would 
now shine through what was formerly the under side of the 
retina. 

The closer, however, this suggestion is examined, the more 


362 H. M. BERNARD. 


unworkable it becomes. Whence, for instance, came the optic 
nerve of such an eye before it was engulfed in the medullary 
groove? If from some primitive ganglionic centre, how came 
it to be transferred to the definitive centre formed by the 
medullary groove? How did this optic nerve manage to 
coincide with the stalk of the invagination? Further, it is 
not easy to deduce in this way the specialised association of 
the chromatophoral layer with the retinal layer. The actual 
evidence against the primary optic vesicles having been me- 
chanically forced in by the lens is very strong, quite apart 
from the difficulty of imagining why a lens-like body should 
crush in the optic vesicle, and, as if by a happy accident, create 
a highly organised eye. Why, again, was the pigment con- 
fined to the (originally) external hemispheres of these buried 
optic pits? Lastly, we do not know to what extent the me- 
dullary groove is historical. Nervous systems usually develop 
from thickenings of the ectoderm; and if a large amount of 
material is required, the thickening may easily become an 
invagination or a groove, and it is by no means necessary to 
suppose that such invagination has any phylogenetic significa- 
tion at all; it may be purely an adaptive process for the supply 
of formative tissue. Indeed, in the Cyclostomata and bony 
fishes, which are very low down in the Vertebrate phylum, 
the central nerve-strand actually arises as a solid thickening. 
But, admitting for the present that the medullary groove may 
have some historical value, eyes are not very likely to let 
themselves be tucked in so as to be functionless, unless there 
is some obstacle to prevent them from shifting. According to 
our theory, the pigment will always be drawn in the direction 
of the strongest illumination, and on changes taking place in 
the form of the body the eyes will shift—through this attrac- 
tive action of the light upon the pigment—into the positions 
in which they can best function as organs of sight.) This 


1 Perhaps one of the most remarkable instances of the power of the eyes to 
follow light is that recorded by Carl Chun (‘ Biologisches Centralblatt,’ xiii, 
1893). Certain Crustaceans have luminous organs which throw light upon 
the ground beneath them, This light has actually drawn down a portion of 


ATTEMPT TO DEDUCE VERTEBRATE EYES FROM THE SKIN. 3638 


renders any theory of the infolding of a pair of functional 
eyes into the medullary groove, to my mind at least, highly 
improbable. 

Indeed, I think it will be freely admitted that it is not easy 
to find any plausible interpretation of the embryological pro- 
cesses as recapitulatory of the actual evolution of the eye. We 
turn, therefore, to ask whether they can be better explained as 
adaptations ? 

At the outset it is obvious that the second assumption with 
which we started, that the pigmented epithelium must have 
been from the first in intimate and inseparable association with 
the retina, compels us to assume that the primary optic vesicle, 
in which the retinal half only secondarily comes in contact 
with the pigmented half, is simply a developmental adaptation. 

Let us see whether the hypothetical history of the eye above 
sketched offers any interpretation of this optic vesicle. The 
great importance of the eye in the economy of the organism, 
leads to its development earlier and earlier in the embryo, 
each shift back admitting of shorter cuts for the attainment of 
the desired end. This may well have been the rule, even when 
the eye was at the simple stage shown in Diagram II, in which 
the retina was spoon- or ladle-shaped, with the nerve represent- 
ing the handle. The most necessary requirement for the eye 
at this stage was undoubtedly the rapid formation of the 
specialised retina involving a great increase of nerve tissue, 
and its intimate association with the chromatophoral layer. 
Secondary infoldings of ectodermal tissue in order to supply 
by a short cut the great increase of sensory cells required in 
the formation of retinz is a well-known phenomenon in the 
development of other eyes, which are more obvious modifica- 
tions of the skinthan arethe Vertebrate eyes. I may mention 
the larger eyes of scorpions and spiders, the specialised retinz 
of which develop embryologically as invaginations in a plane 
the eye on each side, so as to look at the ground illuminated by the animals 
themselves, All stages occur; in some animals the eyes are only elongated 


downwards, in others a piece.of each eye has become detached so as actually 
to form a separate pair of eyes, 


364 H. M. BERNARD. 


more or less at right angles to the optic axis of the definitive 
eye, the two layers of the invagination becoming applied to 
form the more specialised elements of the organ. This is what 
we may assume to have been the first adaptive embryological 
process in the evolution of the Vertebrate eye. Its end was 
not to produce the definitive eye, but the hypothetical eye at 
the stage shown in Diagram II. Hence the primary optic 
vesicle may have originally been a simple ectodermal invagina- 
tion apart from the central nerve system. 

The earlier this retinal invagination is laid down, the more 
chance it would have of being involved in the medullary 
groove. Connection between them could be found, first of all, 
in the rudiments of the optic nerve, which would have been 
laid down as an ectodermal thickening joining the optic invagi- 
nation with the central medullary invagination. A further 
shortening of the process might soon lead to the primary optic 
invagination developing as a lateral offshoot from the medul- 
lary groove.} 

We are not altogether without evidence to support this in- 
terpretation of the origin of the primary optic vesicle, as an 
adaptation for the production of the eye when it was in a 
primitive condition,—such, for instance, as we have depicted in 
Diagram II. I have already called attention to what, accord- 
ing to the history of the eye here described, are certain very 
primitive features in the eyes of fishes. I refer to the accommo- 
dation of the eye by means of the m. retractor lentis (Beer), 
to the persisting power of moving the lens so as to vary the 
direction of sight, to the slight powers of movement of the 
eye itself as a whole, to the small development of the iris, 
and to the persistence of a prominent ring of tissue—the liga- 
mentum annulare—further forward on the iris than is the 
homologous ligamentum pectinatum of higher Vertebrates, 
which represent, according to the foregoing, the neck of the 
ocular invagination (Diagram III). Now it is surely of some 


1 The arrangement sometimes found, in which the optic vesicles only grow 
out after the groove has closed over, can be easily regarded as a secondary 
specialisation. 


ATTEMPT TO DEDUCE VERTEBRATE EYES FROM THE SKIN. 3865 


significance to find that, in these same animals, not only does 
the nervous system arise as a solid strand of ectoderm, but the 
eye and the optic nerve arise “‘ quite high up on the side of 
the brain,” ? and only sink down later into the ventral position 
(with respect to the medullary axis) typical of the higher 
Vertebrates. It seems to me that these facts belong to one 
another, and may be claimed as primitive features retained 
by these lowly Vertebrates.? 

It is, I am aware, usual to regard this development of the 
medullary axis as a solid strand as secondary, but the argu- 
ments are not conclusive. For the reasons above given, I am 
inclined to believe that the solid medullary axis in these cyclo- 
stomes and bony fishes is a direct inheritance from their Inver- 
tebrate ancestors ; and that, conversely, the widely open medul- 
lary groove is a secondary specialisation for the purpose of 
supplying a still larger quantity of material as the nervous 
axis of the Vertebrates became more and more pronounced.® 

According, then, to our interpretation of the facts, the 

1 Hoffmann, ‘ Arch. mik. Anat.,’ xxiii, 1883, p. 45. 

* Evidence of this secondary connection of the optic invagination with the 
brain has, further, been recently deduced from a study of the developing 
brain itself. Waters, from a study of the primitive segmentation of the brain 
(‘ Quart. Journ. Micr. Sci.,’ vol. xxiii, 1892, p. 457), arrives at the conclusion 
that the optic nerve was once serial with the other segmental nerves. The 
fact that the larger proportion of the nerve fibrils develop from the retina 
towards the brain [Assheton, ‘Quart. Journ. Micr. Sci.,’ 34 (1892), p. 85; 
Robinson, ‘ Journ. Anat, Phys.,’ 30, 1896, p. 319] may perhaps represent 
the enormous secondary multiplication of the retinal sensory cells. 

3 The presence of a medullary plate in forms which appear to be still 
lower in organisation than the Cyclostomes is perhaps a difficulty, but the 
very diversity of their specialisations, Urochorda, Cephalochorda, Hemichorda 
(see Lankester’s article “ Vertebrata,” ‘Encyc. Brit.’), suggests that they have 
wandered off in different directions—all, however, towards degeneration— 
from some more typical Vertebrate form. The lowness of their organisation 
may perhaps be paralleled by that of the Copepods, which are apparently 
simpler than the form Apus, which has the best claim to rank as primitive 
among the Crustacea. I would suggest a similar explanation for both cases, 
viz. the fixation and subsequent specialisation of larval forms. I have suggested 
the same also to account for the Acarina, which in many respects are simpler 
than I believe the ancestral Arachnid could well have been. 


366 H. M. BERNARD. 


“primary optic vesicle ” reproduces the retina and specialised 
chromatophoral layer of the eye when it was at the stage shown 
in Diagram II. This would account for the fact that the 
lens, though it appeared, as I think, in the epidermal stage of 
the eye, does not appear ontogenetically with the primary 
optic vesicle, but has to develop separately. 

Some confirmation of this view will be found if we compare 
the shape of the optic cup with the retina at the first hypo- 
thetical stage. The primary optic vesicle never apparently 
forms a true cup; it is far more spoon- or ladle-shaped, the 
sides gradually folding up over the handle or stalk. The 
stalk itself tends to fold, and where the sides of the spoon meet, 
just over the handle, the choroid fissure arises. This is an 
almost exact reproduction of the retina shown in Diagram II. 
As the primitive retina expanded, the optic nerve entering from 
the edge may have been wide and fan-like. As then the retina 
bulged deeper into the body, and its sides folded steeply up, 
the optic nerve would also fold with it near its junction with 
the spoon-shaped retina. This folding would be more marked 
if the point where the nerve joined the retina left the level of 
the skin and travelled inwards towards the base of the in- 
vagination. I claim, therefore, that we have here direct em- 
bryological evidence of the former existence of a ladle-shaped 
retina almost exactly corresponding with that sketched in 
Diagram II, i.e. without a lens in organic connection with it. 

The ontogenetic development of the lens takes place, as is 
well known, by an invagination of the ectoderm ; this obviously 
does not and cannot repeat the evolutionary process as depicted 
in the diagrams. It is, perhaps, probable that the embryo- 
logical invagination which forms the lens may be some faint 
attempt to repeat the invagination shown in Diagram III. 
Any way this method of lens formation would clearly have to 
be adaptive if the account above given of the origin of the 
“primary optic vesicle”’ is approximately correct. As far as 
I can remember, nothing similar to this suggested complica- 
tion has been recorded, although if we knew enough we should 
probably find it to have been very common. We have the 


ATTEMPT TO DEDUCE VERTEBRATE EYES FROM THE SKIN. 367 


slow and uniform development of an organ, viz. the eye, 
through the lapse of time. Its ontogenetic development 
cannot possibly repeat the evolutionary process, because, when 
the eye was in the earliest stages, the embryo early acquired 
the habit of producing the organ in that primitive stage by a 
short cut. This short cut the embryo has never given up, and 
it has thus lost all chance of being able to repeat the subse- 
quent evolutionary processes. The development, therefore, of 
all these later specialisations must necessarily be almost purely 
adaptive. 

With regard to the pineal eye, somewhat the same explana- 
tion of its ontogeny is suggested. The rudimentary optic in- 
vaginations, situated close to the median line, early became 
involved in the developing medullary axis. They were, 
however, not adaptive structures, but the actual optic pits. 
Again, in this case also, the connection between the developing 
optic pit and the medullary axis or groove would be found in 
the optic nerve. The most serious objection to this suggestion 
is probably to be found in the fact that the epiphysis seems to 
appear later than the first rudiments of the Vertebrate eyes, 
whereas it ought, perhaps, to appear earlicr. As a matter of 
fact, very little reliance can be laid upon the order of appear- 
ance. The functional eyes, whether pineal or definitive, seem 
to be hurried on the scene almost before anything else. As 
soon as the pineal eye became functionless or of less importance 
than the definitive eyes, there would be no hurry in its develop- 
ment, and the definitive eyes might appear first, although 
historically much later developments than the pineal eyes. 

It may be objected that there is not the same difficulty in 
supposing the pineal eye to have developed from the brain, it 
being but a single vesicle with a simple nerve-stalk, so that 
its ontogeny might repeat its historical evolution. Spencer, 
who assumes the embryology of the eye to be in the main 
recapitulatory, compares the epiphysis with the tunicate eye, 
but does not believe that it began as an eye. In order 
to get over the difficulties in the way of deducing the pineal eye 
structurally from the tunicate eye, the possibility of which 


368 H. M. BERNARD. 


had been assumed by Lankester, he suggests that the same 
structure which, at an early stage of its development, formed 
the tunicate eye developed into the pineal eye only after it 
had grown out into a long stalked vesicle. The inherent 
objections to this supposition seem to me to be as great as 
those which it was intended to avoid. On the other hand, 
however, the facts appear to me to be all in favour of the 
account given above, the parallel between the retina with the 
retinal pigment and the palisade layer with its subjacent 
accumulation of pigment cells, between the connective-tissue 
capsule of the eye, with its chromatophores and other wan- 
dering cells, and the cutis—the fact that the eye is still em- 
bedded in the thickened cutis in many forms, and that the 
skin above the eye is frequently found transparent,—all seem 
to indicate the skin, and not the brain, as the real mother- 
tissue of the eye. 

Given, then, the possibility of explaining the embryology 
of the pineal eye as adaptive, there seems to me to be no valid 
reason why we should not accept the teachings of comparative 
morphology, and recognise the pineal eye as a relic of the 
simpler type of skin of our Invertebrate ancestors. 

The suggestion that the pineal eye arose out of detached 
portions of the paired primary optic vesicles, I think, defies all 
attempts to work out in detail. Beard’s explanatory diagrams, 
it seems to me, cease to be manageable if, instead of single 
layers of epithelial cells, we try to marshal functional eyes in 
the same way. 

In the foregoing pages I have endeavoured to show how the 
Vertebrate eyes admit of being deduced directly from the skin, 
as my theory of the origin of sight requires. My object has 
been not only to remove a possible objection to the theory 
arising from the embryological development of these eyes, but 
also to show how they might be made even to bear testimony in 
itsfavour. It must be borne in mind that the arguments here 
adopted are not so strong as they might be. They would have 


1 Beard, “‘The Parietal Eye of Cyclostomes,” ‘Quart. Journ. Micros. 
Sci.,’ vol. xxix, pl. vi, figs. A, B, C. ; 


ATTEMPT TO DEDUCE VERTEBRATE EYES FROM THE SKIN. 369 


been greatly strengthened if it had been possible to incorpo- 
rate some of the more direct evidence in favour of the theory 
from which they start. This evidence is, however, cumulative ; 
and in endeavouring, first, to show that the chief types of eye 
are both morphologically and physiologically explicable by this 
theory, I merely begin with that portion of the evidence 
which came readiest to hand. 

In selecting eyes for analysis in subsequent contributions to 
this subject I shall not follow any systematic order; my 
primary object not being any comparative study of eyes, but 
merely the detailed working out of the observations which I 
have made, which seem to me to afford evidence of the truth 
of my theory of the origin of light sensations. 


EXPLANATION OF PLATE 22, 


Illustrating Mr. H. M. Bernard’s paper, “ An Attempt to 
deduce the Vertebrate Eyes from the Skin.” 


Fic. 1.—Diagrammatic representation of a pineal eye regarded as an 
invagination of a simple palisade epithelium. The pigmented contents of the 
chromatophores of the cutis find their way in the skin to the surface, where 
they form a layer of slime (yellow), and in the eye into the cavity of the 
invagination, where they form a slimy vitreous humour. 


Fic. 2.—Hypothetical early stage in the evolution of the Vertebrate eye 
proper. Externally the palisade layer of the skin has budded off layers of 
cells, which as “ prickle-cells”” absorb the pigment and other matter passing 
through the palisade cells, and, dying, become horny scales. The nerves of 
the skin (.) no longer end in the palisade layer, but among the prickle-cells. 
Hence, when the palisade cells become retinal cells stimulated by the passage 
of the pigmented granules, the nerves have to bend back. The retinal pali- 
sade cells early become cuticularised, apparently in order to oppose more 
effectively the outward streaming of the pigmented granules. The nerves 
are already grouped into a solid strand, and form a kind of handle to the 
spoon-shaped retina which is being bulged in by the accumulation of slimy 
maiter in the epidermal cleft which has arisen above the retina. Into the 


370 H. M. BERNARD. 


cavity thus formed, blood-vessels and connective tissue conveying chromato- 
phores intrude, and help to supply the matter for the vitreous humour. The 
horny covering of the retina is supplied by specially long “ prickle-cells,” 
which form a kind of surface lens. 


Fic. 3.—Hypothetical stage following that shown in Fig. 2, the retina 
having bulged still further into the body, and the optic nerve (0. 2.) having 
sunk more towards the base of the invagination, the whole eye, including the 
surface lens, has been invaginated below the skin. In the stalk of the invagi- 
nation, composed of cutis, lymph-spaces commence to appear in the optic axis 
of the eye. ‘The intruding blood-vessels and connective tissue are shown 
entering the cavity of the eye with the optic nerve, and radiating strands 
towards the lens for the movements of the same are diagrammatically 
indicated. 

Fic. 4.—Shows the changes which would be required to complete the 
sensory and dioptric apparatus of the Vertebrate eye. The lymph-spaces are 
specialised into the aqueous chamber; the stalk of invagination persists as the 
ligamentum annulare (or pectinatum). The radial contraction of the tissue 
to form the iris gave rise to a fold—the rudiment of the ciliary processes. 
The lens has become isolated by the rolling round (as indicated by arrows) of 
the palisade layer of the distal portion of the eye. The portion of the 
palisade epithelium thus given up to the lens is represented as having been 
regenerated (dotted lines). 


REPRODUCTION AND METAMORPHOSIS OF COMMON EEL. 37] 


The Reproduction and Metamorphosis of the 
Common Eel (Anguilla vulgaris).' 


By 


G. B. Grassi, 
Professor in Rome. 


Four years of continual researches made by me in collabo- 
ration with my pupil, Dr. Calandruccio, have been crowned at 
last by a success beyond my expectations,—that is to say, have 
enabled me to dispel in the most important points the great 
mystery which has hitherto surrounded the reproduction and 
the development of the Common Eel (Anguilla vulgaris). 
When I reflect that this mystery has occupied the attention of 
naturalists since the days of Aristotle, it seems to me that a 
short extract of my work is perhaps not unworthy to be pre- 
sented to the Royal Society of London, leaving aside, however, 
for the present, the morphological part of my results. 


The most salient fact discovered by me is that a fish, which 
hitherto was known as Leptocephalus brevirostris, is the 


larva of the Anguilla vulgaris. 

Before giving the proofs of this conclusion I must premise 
that the other Murenoids undergo a similar metamorphosis. 
Thus I have been able to prove that the Leptocephalus 
stenops (Bellotti), for the greatest part, and also the Lepto- 
cephalus morrisii and punctatus belong to the cycle of 
evolution of Conger vulgaris; that the Leptocephalus 
haeckeli, yarrelli, bibroni, gegenbauri, kéllikeri, and 
many other imperfectly described by Facciola, and a part of 
the above-named Leptocephalus stenops of Bellotti, 
belong to the cycle of evolution of Congromurena my stax; 
that the Leptocephalus tenia, inornatus, and dia- 

1 From the ‘ Proceedings of the Royal Society,’ November, 1896. 


372 G. B. GRASSI. 


phanus belong to that of Congromurena balearica; that 
under the name of Leptocephalus kefersteini are con- 
founded the larvee of various species of the genus Ophich- 
thys; that the Leptocephalus longirostris and the 
Hyoprorus messanensis are the larve of Nettastoma 
melanurum, and that the Leptocephalus oxyrhynchus 
and other new forms are larve of Saurenchelys cancri- 
vora, and that finally a new little Leptocephalus is the larva 
of Murena helena. 

The form known as Tylurus belongs to Oxystoma, of which 
we unfortunately know nothing more than a figure by Raffi- 
nesque. I have not been able to find the Leptocephalus of 
Myrus vulgaris, of which I have had only a single indi- 
vidual, in which the transformation was already far advanced. 
Neither have I found the Leptocephalus of Chlopsis bi- 
color, a very rare form, which is related to Murena and to 
Murenichthys. As the result of these observations, the 
family of the Leptocephalide has been definitely suppressed 
by me ; the various forms of that family are, in fact, the normal 
larvee of the various Murznoids. 

In regard to the greater part of the above-named species, 
the control has been threefold, namely : 

Firstly, anatomical. I have compared the various stages in 
all their structures, and have made the due allowance for the 
changes brought about by the metamorphosis at the close of 
larval life. 

Secondly, natural. I have found in nature all the required 
transitional stages. 

Thirdly, experimental. I have followed, step by step, the 
metamorphosis in aquariums. 

Therefore the hypothesis of Giinther that the Leptocephali 
are abnormal larvee, incapable of further development, must be 
rejected. All this is related by myself at length, with all 
historical details which concern the question, in a_ large 
memoir which is about to appear in the journal edited by 
Professor Todaro. 

Until now all these facts have been unknown, because nor- 


REPRODUCTION AND METAMORPHOSIS OF COMMON EEL. 373 


mally they can only be observed in the abysses of the sea at a 
depth of at least 500 metres. Fortunately, along a part of the 
coast of Sicily strong currents occur, which must be ascribed 
to the tide, producing very large displacements of the water in 
the narrow Strait of Messina. I shall give further details con- 
cerning these currents in my large memoir. In consequence 
of the strong currents sometimes—I say sometimes because 
there is no regularity, and one may have to wait for.a year with- 
out obtaining any material—not only many deep-sea fishes, but 
also all stages of the development of the Murzenoids are met 
with in the surface-water. To these currents we owe all the 
captures of Murzna helena with ripe eggs, which is in ac- 
cordance with what I had already argued from other facts, 
namely, that the reproduction of the Murznoids takes place at 
great depths of the sea. 

Before I proceed to speak of the common eel, I must pre- 
mise that Dr. Raffaele has described certain pelagic eggs as 
belonging to an undetermined species, putting forward the 
suggestion that these eggs belong to some Murenoid. This 
matter has been investigated by myself, and I have shown that 
the newly hatched larve (called “ pre-larve” by me) derived 
from these eggs have essentially the character of Leptocephali. 

The life history of the Murzenoids, leaving aside for the 
present the common eel, is as follows:—Females can only 
mature in very profound depths of the sea, that is to say, at 
least a depth of 500 metres. This fact I established by find- 
ing well-known deep-sea fishes together with Leptocephali, 
ripe Murzene, and quite ripe eels (see below). The females of 
those species which do not live at this depth must therefore 
migrate to it. The male, however, can mature at a smaller 
depth, and therefore they migrate into the greater depth when 
they are already mature. Fertilisation takes place at great 
depths: the eggs float in the water; nevertheless they remain 
at a great depth in the sea, and only exceptionally, for un- 
‘known reasons, some of them mount to the surface. 

From the egg issues rapidly a pre-larva, which becomes a 


larva (Leptocephalus) with the anus and the urinary opening 
VoL. 39, PART 3.—NEW SER. BB 


374 GB; GRASSI. 


near the tip of the tail. The larva then becomes a hemi-larva, 
the two apertures just named moving their position towards 
the anterior part of the body, which becomes thickened and 
nearly round. By further change the hemi-larva assumes the 
definitive or adult form. The larva, as well as the hemi-larva, 
shows a length of body much greater than that exhibited by 
the young Murzenoid of adult form into which they are trans- 
formed. By keeping specimens in an aquarium I was able to 
establish a diminution of more than 4 cm. during the meta- 
morphosis. With regard to the greatest length which the larva 
can attain in a given species, and the amount of diminution 
which accompanies metamorphosis, there are great individual 
variations. 

The history of the common eel, to which I am now about 
to refer, is very similar to that given above for the other 
Murenoids. The common eel (Anguilla vulgaris) under- 
goes a metamorphosis, and before it assumes the definitive 
adult form it presents itself as a Leptocephalus, which is known 
as Leptocephalus brevirostris. This Leptocephalus was 
discovered in the Straits of Messina many years ago. A speci- 
men was also captured by the ‘‘ Challenger,” and another 
specimen was taken by the Zoological Station of Naples in the 
Straits of Messina. This form is occasionally carried to the 
surface by currents. By exception, in the month of March, in 
the year 1895, we captured several thousands of them in one 
day ; but the best way to secure this Leptocephalus (and a very 
easy one) is to open the intestine of the Orthagoriscus 
mola, a fish which is common in the Straits of Messina, and 
in it one is certain to find a very large number of specimens. 
It must be observed that Orthagoriscus mola is a deep-sea 
fish. The specimens of Leptocephalus brevirostris found 
in the intestine of Orthagoriscus are more or less altered by 
digestion. Those specimens of Leptocephalus brevirostris 
which are taken near the surface in the open sea are in a better 
state of preservation, but, unfortunately, these also frequently 
have the epidermis injured so that they cannot maintain their 
life in an aquarium for more than a few days; they live long 


REPRODUCTION AND METAMORPHOSIS OF COMMON EEL. 375 


enough, however, to allow us to observe that it is their habit 
to conceal themselves in the sand or in the mud as the common 
eel (Anguilla) does. Here it is to be noted that the various 
forms of Leptocephali have habits resembling those of the 
Murenoids to which they belong, i.e. they dig into the sand 
or abstain from doing so according as the adult form has or has 
not this habit. 

I now pass on to the characters of Leptocephalus brevi- 
rostris. I give them here in the same order as I shall use in 
my larger memoir. The length varies from 77 to 60 mm., the 
same extent of variation as observed in other Murenoids. 
The caudal fin tends to assume the form which it has in the 
Elver! or young Anguilla. It is to be noted that in other 
Leptocephali the caudal fin also tends always to exhibit the 
adult form. The lower jaw projects sometimes more than the 
upper jaw, asin Anguilla. The margin of the mouth is wide, 
as in Anguilla. The tongue is free, as in Anguilla. On the 
other hand, the youngest elvers which I have observed have 
smaller eyes than Leptocephalus brevirostris, and this 
need not surprise us, since we know that in other species of 
Murenoids the diminution of the eyes occurs during the 
metamorphosis. The nostrils are separated from one another, 
the anterior tubes are relatively at a considerable distance from 
the tip of the snout and from the rim of the mouth. They 
are in a position in which they are observed in many other 
Leptocephali, which are destined to transform themselves into 
adult forms having the anterior nostrils in nearly the same 
position as in the commoneel. The posterior nostrils, on the 
contrary, are not tube-like, and are in the same position as 
those occupied in the adult Anguilla. It is worth remarking 
that in other Leptocephali also the posterior nostrils have 
already assumed the adult position when the anterior ones are 
still far removed from it. In L. brevirostris I find a larval 
dentition which resembles that of the other Leptoce- 


1 The word “ Elver” is used in this paper in its strict sense, viz. for the 
young form of Anguilla vulgaris as taken when ascending rivers in vast 
numbers. 


376 G. B. GRASSI. 


phali. In’ correspondence with the small size of Lepto- 
cephalus brevirostris the number of larval teeth is 
small. Researches founded, firstly, on the enumeration of 
the myomeres; secondly, upon the enumeration of the dorsal 
and ventral arches of the vertebrae of the caudal extremity 
(hypurals) ; and thirdly, upon the enumeration of the posterior 
spinal ganglia, lead with great certainty to the conclusion that 
the Leptocephalus brevirostris is the larva of a Murenoid, 
the number of whose vertebree must lie between 112 and 117, 
most probably 114 or 115. Such a Murenoid is the Anguilla 
vulgaris. The Murenoid indicated cannot be any other of 
those occurring in the Mediterranean, because they all have a 
number of vertebre higher than 124.! Counting the myomeres 
in Leptocephalus brevirostris, one finds generally only 105 
complete, five others incomplete, and all the others in a state 
of transparency and incomplete formation. These latter are 
fortunately at the posterior extremity, where other criteria 
come to our assistance, namely, the spinal ganglia and the 
vertebral arches. To show how I arrive at the number of 
vertebree which must be possessed by the adult Individual, 
corresponding to a given Leptocephalus brevirostris, I 
quote the following example :—I assume that three vertebree 
develop themselves in correspondence to the first four incom- 
plete myomeres, and that 105 must develop themselves in 
relation to the 105 complete myomeres,—that is to say, between 
the fourth and fifth myomeres, between the fifth and sixth, and 
so on until we reach the 105th vertebra lying between the 
104th and 105th myomeres. I further conclude that seven 
other vertebra are developed at the caudal extremity, as indi- 
cated by the number of vertebral arches and the spinal ganglia 
in that region. We count, therefore, in all 115 vertebra, and 
this is the number which can be easily seen in many specimens 
of Anguilla vulgaris. 

1 Murenesox savanna is said to have 109 vertebre, but it is 
doubtful whether it really occurs in the Mediterranean. The position of 


its nostrils and the number of its branchiostegal rays render its association 
with Leptocephalus brevirostris impossible. 


REPRODUCTION AND METAMORPHOSIS OF COMMON BEL. 377 


Here I must particularly insist that I have ascertained in 
an absolute manner that during the metamorphosis of the 
Murenoids the number neither of the myomeres, nor of the 
vertebral arches, nor of the spinal ganglia is subjected to any 
change. The hypurals of Leptocephalus brevirostris are 
precisely the same as in the elver of Anguilla vulgaris, 
The last hypural which is fused with the urostyle may pre- 
sent itself as a single piece, or may be more or less cleft, 
These are variations which are met with also in the elver. 
Just as in the elver, the last hypural but one is always ex- 
tensively cleft, or, if the expression is preferred, doubled. To 
the last hypural correspond five rays, whilst four correspond to 
the last but one, and one to the last but two, the whole 
structure being identical with that found in the elvers of 
Anguilla vulgaris. Of these ten rays, the eighth, seventh, 
and sixth are bifid, both in Leptocephalus brevirostris and 
in the elvers of Anguilla vulgaris. In the pectoral fin of 
Leptocephalus brevirostris the definitive rays can be ob- 
served, and these are of the same number asin the elvers of An- 
guilla vulgaris. Leptocephalus brevirostris is transpa- 
rent, and has colourless blood. The red corpuscles are wanting, 
but there are present so-called ‘ blood-plates” (“Blutplattchen” 
in German) similar to those of the inferior Vertebrates. The 
bile is also colourless. This fact is observed in all the other 
Leptocephali. Leptocephalus brevirostris is, however, 
the only one which is free from all pigmentation. Corre- 
spondingly, the common eel is the only species of Murznoid 
which at the close of metamorphosis (i.e. in the youngest 
elvers) is devoid of all trace of larval pigmentation. It was 
this observation which first led us to the discovery of the 
relations between Leptocephalus brevirostris and An- 
guilla vulgaris. 

In making transverse sections of Leptocephalus brevi- 
rostris I found other characters which confirm the relation 
between it and the common eel; for instance, the branchio- 
stegal rays are ten to eleven in number, as is also observed in 
the elvers of Anguilla vulgaris. In the common eel the 


378 G. B. GRASSI. 


well-known lateral branch of the fifth pair of the cranial 
nerves exists. It is also found in Leptocephalus brevi- 
rostris. This lateral branch could not be found by Dr. 
Calandruccio in the other common Murenoids of Sicily, and 
is wanting also in the other Leptocephali. 

The mucous-canal-system (sensory canals) in the head are 
already developed, partially in Leptocephalus_ brevi- 
rostris, and are incompletely developed in the elver. As in 
the elver, so in Leptocephalus brevirostris, the pyloric 
ceca are wanting. ‘The blind extremity of the stomach and 
the incompletely developed swim-bladder, which is as yet free 
from contained gas, are present both in Leptocephalus 
brevirostris and in the elver of Anguilla vulgaris. The 
pronephros is in active function, as in the other Leptocephali. 
The Malpighian glomerules of the kidney (mesonephros) are 
lobed as in the eel, and their number corresponds with that 
observed in the Helmichthys stage, of which I will speak 
further on. The genital gland, not yet sexually differentiated, 
is almost identical with that of the same stage. In short, it 
may be said that the whole organisation of Leptocephalus 
brevirostris corresponds with the organisation of the common 
eel, if we make allowance for those changes which are 
observed in the metamorphosis of the other species of Mure- 
noids, such as reduction of the pancreas and of the liver, dis- 
appearance of the protoskeleton, complication of the muscu- 
lature, increase in size of the cerebellum, loss of the larval 
teeth, development of the definitive teeth, &c. 

From the description of these Leptocephali I must pass on 
briefly to speak of the stages nearer to the condition of the 
elver. I am, however, obliged to leave a break in the series, 
which, however little its significance, yet certainly will make 
some impression on the minds of those who do not realise with 
what caution I have formed my conclusions. I must confess 
that since I have learnt how difficult it is to procure an entire 
series of the development of a Murznoid, I am more astonished 
at being able to recognise a single stage in the development of 
a given species than at not finding the whole series. I must 


REPRODUCTION AND METAMORPHOSIS OF COMMON EEL. 379 


point out that the break in my series of the development of 
Anguilla vulgaris would have been much smaller if I could 
have persuaded myself to kill and preserve one of the hemi- 
larve which I happened to meet with at the end of the year 
1892. They were really transitional stages between Lepto- 


WN AIAG << 


\ LE a= 


ae 


SS 
/, 


Fie. 1. 


Gener 


rae MMMM EM A gaa 
IIL 


Fie. 1.—Leptocephalus brevirostris. Natural size. 

Fie. 2.—Leptocephalus brevirostris. Later stage. Natural size. 
Fie. 3.—Anguilla vulgaris. ‘Transition stage. Natural size. 

Fie. 4.—Anguilla vulgaris. Definitive habit (Elver). Natural size. 


cephalus brevirostris and that stage which I shall describe 
further on. I published this fact in a preliminary note in the 
month of May, 1893. They were transparent with almost 
colourless blood, without any trace of pigmentation except at 
the eyes, and had lost all the larval teeth, whilst they possessed 
already very few and very minute teeth of the definitive series. 


380 G. B. GRASSI. 


The body was thickened, and already showed the cylindrical 
form. They measured little less than 8 cm. In short, they 
were Leptocephalus brevirostris on the way to trans- 
formation into Anguilla vulgaris. Asa matter of history 
they actually did transform themselves in my aquarium with 
the usual diminution in their dimensions, and subsequently 
proceeded to increase in bulk.' The metamorphosis took 
place, as usual, without the animal taking in any nourish- 
ment whatever. The resumption of growth was accompanied 
by a resumption of feeding. Unfortunately I had no other 
individuals of this stage. 

The stage which I now pass on to describe (fig. 3) can 
be obtained during the winter in the sea. I have never 
found them at the mouths of rivers. The length varies 
from 54 to 73 mm. Most individuals measured about 65 mm. 
The body is relatively longer than in the elver. It is also 
relatively deeper, as in Leptocephalus. We are reminded 
of Leptocephalus also by the pigment of the eye, the vitreous 
transparency of the body, the swim-bladder being indis- 
tinguishable in the living animal, and the absence of all 
larval pigmentation. The blood is slightly coloured, and the 
bile is already green. Slight pigmentation can be seen along 
the central nervous system, and at the middle part of the 
caudal fin. This commencement of the definitive or adult 
pigmentation in the regions named before it occurs in any 
other part is also seen in other Murenoids. The definitive 
teeth are very minute, and few in number. The intestine 
contains no food. After what I had observed in the other 
Murenoids, the simple observation of the barely indicated 
teeth, and of the absence of aliment in the gut, would have 
been sufficient to convince me that the stage now under notice 
must be preceded by a Leptocephalus phase. Indeed, if we 
did not admit such a preceding history, we could not under- 
stand how this little fish could have attained such a size with- 

1 The fact that I actually have obtained in an aquarium the transforma- 


tion of L. brevirostris into Anguilla vulgaris is of prime importance. 
The time occupied was one month. 


REPRODUCTION AND METAMORPHOSIS OF COMMON EEL. 381 


out acquiring well-developed teeth, and without nourishing 
itself. 

In conclusion, no one would hesitate, even not knowing 
Leptocephalus brevirostris, to refer the stage now under 
discussion to a Murenoid about to complete its Leptocephalus 
metamorphosis, were it not for the fact that there has been so 
much question concerning the reproduction of the common 
eel, and that so many capable observers have failed in dealing 
with it, that every new observation is received with scepticism. 
The stage of which I am now speaking, in the hands of a 
pure systematist, would probably be described as a Helmichthys, 
a genus established for certain forms of Leptocephali far 
advanced in transformation. 

The next forms to which I have to refer are captured in the 
course of migration from the sea into fresh water. When kept 
in an aquarium they assume the characters of the elver, dimin- 
ishing more or less in volume, and without nourishing them- 
selves. The elvers of the common eel can present themselves 
in stages differing little from that last described, as well as in 
a form which has already developed the full pigmentation of 
the adult. Even those which most resemble the preceding 
stage always have a character which distinguishes them easily, 
namely, the presence of a definitive pigment, more or less 
superficially placed on the head, and not to be confounded 
with the pigment round the posterior extremity of the brain, 
which latter is already present in the preceding stage. In 
specimens taken at the mouths of rivers this more or less 
superficial pigment was, so far as I could ascertain, always 
present. 

As the pigmentation develops itself, the little eel gradually 
undergoes a diminution in all its dimensions. It results from 
my measurements, that the fully pigmented elver has an 
average length of 61 mm., while for the more or less colourless 
elver the average length is 67 mm. I found pigmented elvers 
which were reduced in length to 51 mm.,a size which I never 
observed in those elvers in which the development of pig- 
ment had not taken place. 

VoL. 39, PART 3,—NEW SER. cc 


382 G. B. GRASSI. 


The facts which I have stated demonstrate that the eel goes 
through a metamorphosis, and that Leptocephalus brevi- 
rostris is its larva. Some further considerations remain to 
be given, although I believe that zoologists will not consider 
the question still an open one after the record of facts given 
above—facts which any one may verify by examining the 
material which is preserved in my hands. Many to whom I 
have related my discovery of the history of the common eel 
have objected that eels are found almost everywhere, whilst 
Leptocephalus brevirostris is limited to Messina. In 
reply, I must say that, first of all, it is not true that Lepto- 
cephalus brevirostris is limited to Messina; secondly, that 
at Messina there are special currents, which tear up the deep- 
sea bottom which everywhere else is inaccessible; thirdly, 
although it is true that on the coasts of many countries where 
Anguilla vulgaris is found, no one has ever seen a Lepto- 
cephalus brevirostris; it is also true that in no country, 
not even in those where eels are abundant, has anyone ever 
seen an eel of less than 5 cm. in length. Since it has to be 
admitted that no one knows the eel before it arrives at the 
length of 5 cm., there is no greater difficulty in supposing that 
during this unknown period the eel passes through a Lepto- 
cephalus stage than in supposing that it does not do so. The 
critical study of the literature of this subject, and a great many 
continued observations, have occupied me for many years, and 
have been undertaken just in those places where young eels 
are to be found. They enable me, from my own studies, to 
affirm with assurance that young eels with the definitive adult 
form do not exist of less than 5 cm. in length. 

From the study of the memoir of Raffaele on pelagic eggs, I 
have come to the conclusion that the eggs of his undetermined 
species No. 10, having a diameter of 2:7 mm., and differing 
from all the others in the absence of oil-globules,! must belong 
to the Anguilla vulgaris, because from them Dr. Raffaele 


1 Renewed researches have convinced me that this egg is that of An- 
guilla vulgaris. There is, however, another egg belonging to an unde- 
termined Mureenoid which is devoid of oil-drops, and can easily be confused 
with the true eggs of Anguilla. 


REPRODUCTION AND METAMORPHOSIS OF COMMON EEL. 3883 


obtained pre-larvee which had only forty-four abdominal 
myomeres. I endeavoured for two years in vain to study 
these eggs at the Zoological Station of Naples. I found only 
a few of them, and these died prematurely. 

In another point my researches have yielded a very inte- 
resting result. As a result of the observations of Petersen, 
we know now that the common eel develops a bridal colora- 
tion or “ mating habit,” which is chiefly characterised by the 
silver pigment without trace of yellow, and by the more or 
less black colour of the pectoral fin, and finally by the large 
eyes. Petersen inferred that this was the bridal coloration 
from the circumstance that the individuals exhibiting it 
had the genital organs largely developed, had ceased to take 
nourishment, and were migrating to the sea. Here Petersen’s 
observations cease and mine begin. The same currents at 
Messina which bring us the Leptocephali bring us also many 
specimens of the common eel, all of which exhibit the silver 
coloration. Not a few of them present the characters de- 
scribed by Petersen in an exaggerated condition,—that is to 
say, the eyes are larger and nearly round instead of elliptical, 
whilst the pectoral fins are of an intense black. It is worth 
noting that in a certain number of them the anterior margin 
of the gill-slit is intensely black, a character which I have 
never observed in eels which had not yet migrated to the sea, 
and which is wanting in the figures and in the originals sent to 
me by Petersen himself. Undoubtedly the most important of 
these changes is that of the increase of the diameter of the eye, 
because it finds its physiological explanation in the circumstance 
that the eel matures in the depths of the sea. That, as a 
matter of fact, eels dredged from the bottom of the sea have 
larger eyes than one ever finds in fresh-water eels I have 
proved by many comparative measurements, made between eels 
dredged from the sea bottom and others which had not yet 
passed into the deep waters of the sea. Thus, for instance, in 
a male eel taken from the Messina currents, and having a total 
length of 343 cm., the eye had a diameter, both vertical and 
transversal, of 9 mm. and in another eel of 334 cm. the same 


384 G. B. GRASSI. 


measurement was recorded. Ina female eel, derived from the 
same source and purchased in the market, whose length was 
483 cm., the vertical diameter of the eye was 10 mm., and 
the transversal diameter rather more than 10 mm. These 
are not the greatest dimensions which I observed, and I con- 
clude from these facts that the bridal habit described by Petersen 
was not quite completed in his specimens, and that it becomes 
so only in the sea and at a great depth. In relation to these 
observations of mine stands the fact that the genital organs in 
the eel taken in the Messina currents are sometimes more 
developed than in eels which have not yet entered the deep 
water. Thus it has happened that male individuals have 
occurred showing in the testes here and there knots of sper- 
matozoa. These spermatozoa are similar to those of the 
Conger vulgaris, and must be considered as ripe. As is 
well known, so advanced a stage of sexual maturity has never 
before been observed in the common eel. This appears to be 
due to the fact that the males hitherto examined had not yet 
migrated into the deep water of the sea. 

Eels with big eyes taken from the depths of the sea were, 
before the above facts were known, described as a distinct 
species under the name of Anguilla bibroni (Kaup) and of 
Anguilla kieneri (Kaup), not to be confounded with An- 
guilla kieneri (Gunther), which is a synonym of Lycodes 
kieneri. 

In certain cloace of ancient Rome which to-day are disused 
and contain pure water, remarkable eels are found of a length 
of from 20 to 30 em., of a grey colour, without trace of yellow, 
of male and female sex, with enormous eyes, and with more or 
less rudimentary genital organs. They are individuals which, 
confined in a place without light, have acquired prematurely 
one of the characters of the bridal habit without a correspond- 
ing development of the genital organs. These individuals are 
probably incapable of ulterior development, as the condition 
of their genital organs seems to demonstrate. 

Under the name Anguilla kieneri (Kaup) there have 
probably been included some individuals which had acquired 


REPRODUCTION AND METAMORPHOSIS OF COMMON EEL. 385 


big eyes under conditions similar to those described for the 
eels of these Roman cloace. From these and similar observa- 
tions it clearly results that all the European eels must be in- 
cluded under-a single species; and this is an important fact 
from another point of view, namely, that it destroys an 
objection which might be raised against my conclusion with 
regard to the development of Anguilla vulgaris from Lepto- 
cephalus brevirostris, namely, the objection that Lepto- 
cephalus brevirostris belongs not to Anguilla vulgaris, 
but to Anguilla kieneri, or to Anguilla bibroni. 

To sum up, Anguilla vulgaris, the common eel, matures 
in the depths of the sea, where it acquires larger eyes than are 
ever observed in individuals which have not yet migrated to 
deep water, with the exception of the eels of the Roman 
cloace. The abysses of the sea are the spawning-places of the 
common eel: its eggs float in the sea water. In developing 
from the egg it undergoes a metamorphosis, that is to say, 
passes through a larval form denominated Leptocephalus 
brevirostris. What length of time this development re- 
quires is very difficult to establish. So far we have only the 
following data:—First, Anguilla vulgaris migrates to the 
sea from the month of October to the month of January ; 
second, the currents, such as those of Messina, throw up, from 
tle abysses of the sea, specimens which, from the commence- 
ment of November to the end of July, are observed to be more 
advanced in development than at other times, but not yet 
arrived at total maturity; third, eggs, which according to 
every probability belong to the common eel, are found in the 
sea from the month of August to that of January inclusive; 
fourth, the Leptocephalus brevirostris abounds from 
February to September. As to the other months, we are in 
some uncertainty, because during them our only natural fisher- 
man, the Orthagoriscus mola, appears very rarely ; fifth, 
I am inclined to believe that the elvers ascending our rivers 
are already one year of age, and I have observed that in an 
aquarium specimens of L. brevirostris can transform them- 
selves into young elvers in one month’s time. 

VOL. 39, PART 3,—NEW SER. DD 


am 
2 


au 
i. 
=<. 
5 
| 
a 
7 
- 
« 
:, 


a 


- v 
. - 7m : _ 
7 - -* be : om : 
& ae “yt © 
; ay HAE ROUTED 1 A7ROw 
: - hea 
et Py) Tits ee 5 ‘Gnit a an Le f on 
prone aie a a Reena ts 
ei eLwie: yell a ts G2 “unary 
book aimee gee nee ei Pre aiJés us Die} 
ote id me: oft ‘ ets < Tit Fi? ee 7 ius ‘a 
shiver pope Tubs a foes Wl ih Ghee Hf 7 
asta Lye Palsawth. er altt gmc de? é Peas et 
4 eer Toomer Feat aor wy. Th Peat iit) 82 
ae% - ine we " fought £2 o Ob SIN 
-— a Su 7. te oe ee ay f s i 
6 ‘ ’ 4 ' t 7 ' = : 
2 r is fia i? & 
; >. Si 
i ? - 4 
3. | bs 
» - > iP a 
3 bg t 4 7% rs 
: (eeu pe ee 5G 
- : chia Tsyg Viv aaa), |) Sage aay 
aru oe ue , =) , r 502% oid ra 
~~ em 7 ae by es * ages re, Shy i. » ea 
me F es = 
oy: ey BUN, 20 4 yi aie 
= 1 Fr = Vash ir eae Be eee 
® a = * >: 
i 
ee ans 
ee Fa ie a oe, 


FEB 16 1897 


CHANGES IN CELL-ORGANS OF DROSERA ROTUNDIFOLIA. 387 


Changes in the Cell-organs of Drosera rotun- 
difolia, produced by Feeding with Egg- 


albumen. 
By 
Lily Huie, 
Physiological Laboratory, Oxford. 
With Plates 23 and 24. 

ContTENTS. 

PAGE PAGE 
Literature : ‘ : . 888 | Results of the Present Research 400 
Experiments. : ‘ . 393 | Summary and Conclusion . « 423 
Methods . P : 394 | Description of Plates : ~ 424 


General Structure of the Glands 396 


INTRODUCTION. 


Tue subject of this research was suggested to me by Dr. 
Gustav Mann, with the view of throwing some further light 
upon the functions of the nucleus and nucleolus, and to test 
whether his views were correct,—namely, that the chromatin 
segments are organs for the elaboration of food; that the 
nuclear chromatin is not the essential element, inasmuch as 
it only represents partly elaborated material; that the nu- 
cleolus is a storehouse for some elaborated matter, &c.1 By 


1 Gustav Mann, “ The Functions, Staining Reactions, and Structure of 
Nuclei,” ‘ Proc. British Association,’ Section D, 1892, Edinburgh meeting. 


VOL, 39, PART 4,—NEW SER, EE 


888 LILY HUIE. 


the kindness of Professor Gotch I have been permitted to 
work in the Histological Department of the Physiological 
Laboratory at Oxford, under the guidance of Dr. Mann. 


LITERATURE. 


Numerous as have been the papers published on the Dro- 
seracee since Charles Darwin’s classical treatise upon ‘ In- 
sectivorous Plants,’ comparatively few have dealt with 
cytological changes induced by functional activity of the 
glands; and these few have been mainly devoted to the obser- 
vation and elucidation of the “Phenomenon of Aggregation 
of Protoplasm,” a term applied by Darwin! to the following 
phenomena : 

“Tf a tentacle is examined some hours after the gland has 
been excited, . . . . the cells instead of being filled with 
homogeneous purple fluid, now contain variously shaped 
masses of purple matter, suspended in a colourless or almost 
colourless fluid. The little masses of aggregated matter are 
of the most diversified shapes, often spherical or oval, some- 
times much elongated,&c. . . . These little masses inces- 
santly change their forms and positions, being never at rest. 
A single mass will often separate into two, which afterwards 
reunite. Their movements are rather slow, and resemble 
those of amcebe or of the white corpuscles of the blood. We 
may, therefore, conclude that they consist of protoplasm.” 

Cohn, in his review of ‘ Insectivorous Plants, 2 threw 
doubts upon Darwin’s inference that the aggregated masses 
consisted of protoplasm, and called them ‘ Zusammenbal- 
lungen” of the red-cell sap. 

Francis Darwin,® who studied the phenomenon in the cells 
of the stalks of the tentacles, reasserted the view of his father 


1 *Tnsectivorous Plants,’ London, 1875, p. 39. 

2 * Deutsche Rundschau,’ 1876. 

3 Darwin, F., “On the Process of Aggregation in the Tentacles of 
Drosera rotundifolia,” ‘Quart. Journ. Micr, Sci,’ 


CHANGES IN CELL-ORGANS OF DROSERA ROTUNDIFOLIA. 389 


that the aggregated masses were “protoplasm.” He states 
the two opposing theories thus : 

I, “ My father’s view, that the aggregated masses consist of 
protoplasm, and that their movements are simply due to their 
own contractility, excited by various external agencies.”’ 

II. “ Professor Cohn’s view, which appears to be that the 
ageregated masses consist of condensations of cell-sap, and as 
a necessary corollary that the movements are impressed on the 
masses by some kind of protoplasmic action external to the 
masses.” 

That the latter view is the correct one was proved by 
Schimper, Gardiner, and De Vries, all of whom subsequently 
studied the phenomenon, and came to the conclusion that it is 
the contents of the vacuoles, and not the protoplasm, which 
give rise to the phenomenon described by Darwin. 

Schimper! compared the condition of the cells before and 
after stimulation in Sarracenia, Drosera, and Utricularia, and 
found that the protoplasm under the influence of the stimula- 
tion gains an increased power of imbibition. Water is with- 
drawn by it from the cell-sap, and in Sarracenia and Drosera 
the tannin with which the cell-sap is laden becomes concen- 
trated, and gives rise to the appearance Darwin called 
“ Aggregation.” 

De Vries? describes the movements thus: 

1. An increase in the rapidity and vigour of circulation in 
the peripheral layer of protoplasm. 

2. A division of the vacuole into several smaller vacuoles, 
which are each surrounded by part of the original vacuole 
wall. 

3. A very marked diminution of the volume of the vacuole, 
because some of its original constituents are expelled through 
the vacuole wall, and collect between this and the circulating 
protoplasm. The expelled fluid has, at least approximately, 


1 Schimper, A. J. W., “ Notizen iiber insektfressende Pflanzen,” ‘ Bot. 
Zeit.,’ 1882, Nos. 14 and 15. 

2 De Vries, H., “ Ueber die Aggregation im Protoplasma von Drosera 
rotundifolia,” ‘ Botanische Zeitung,’ 1886. 


390 LILY HUIE. 


the same attraction for water as the remaining contents of the 
vacuole, but possesses a different chemical composition. The 
colouring-matter, the tannin, and certain unknown albumins, 
precipitable by ammonium salts, are retained inside the vacuole. 
The grape-sugar and vegetable acids are probably, in part at 
least, expelled. ‘l'urgor is almost the same in stimulated as 
in unstimulated tentacles; it takes, e.g., a solution of between 
2 per cent. and 3 per cent. KNO, to produce plasmolysis in 
unstimulated tentacles. The author finds that aggregation 
takes place first in the cells nearest the glands, and that divi- 
sion of the vacuoles occurs also in the gland-cells themselves, 
although no protoplasmic circulation is to be seen. When 
stimulation ceases the cells return gradually to their original 
condition, while the vacuoles increase in size, and ultimately 
fuse again into a single one. 

This author describes further a similar but distinct pheno- 
menon. Under certain conditions, especially when stimulated 
by salts of ammonia, by free ammonia in weak solution, by 
iodine, by osmic acid, and even sometimes by stimulating a 
fresh leaf with albumin, or by slowly killing a tentacle , by 
drying, the albumins dissolved in the sap of the vacuole are 
precipitated; first in the form of a fine granulation, afterwards 
fusing into larger globules equalling in size the little vacuoles 
formed by aggregation, when it is not easy to distinguish be- 
tween this appearance and that resulting from aggregation. 
But that they are different phenomena may be shown by first 
inducing normal aggregation, and then treating the aggregated 
cells with 1 per cent. ammonium carbonate. 

The ammonia salts did not penetrate the thick cuticle of the 
tentacle stalks, but entered at the cut places and also by the 
gland-cells of the tentacle, and the little mammilla-shaped 
glands situated on the stalk of the tentacle. 

A year previous to the appearance of De Vries’ papers, 
Gardiner! published an account of “The Phenomena accom- 


1 Gardiner, W., “On the Phenomena accompanying Stimulation in the 
Gland-cells of Drosera dichotoma,” ‘Proc. Roy. Soc, London,’ 1885, 
p. 229. 


CHANGES IN CELL-ORGANS OF DROSERA ROTUNDIFOLIA. 391 


panying Stimulation in the Gland-cells of Drosera di- 
chotoma,” in which he described the phenomenon of aggre- 
gation in living stalk cells, much as Darwin had done. He 
also noticed the phenomenon produced by the action of car- 
bonate of ammonia, and called it “ passive aggregation.” A 
spindle-shaped or acicular body in the stalk cells and epidermal 
cells of the leaf called a “rhabdoid” is also described, which, 
as well as the elongated nuclei of the stalk cells, tends to be- 
come spherical after long stimulation. 

Gardiner studied also the changes in the appearance of the 
gland-cells themselves after stimulation. His methods were 
as follows : 

Leaves stimulated from five minutes to seventy-two hours 
by feeding with flies, or small pieces of frog’s muscle, were 
examined fresh, or after fixing in alcohol, picric acid, or osmic 
acid. The best results were got with specimens treated for 
twelve hours with 1 per cent. or 2 per cent. solution of chromic 
acid. He describes a typical resting gland-cell thus :—‘“ The 
protoplasm is arranged in a network or reticulum. The meshes 
of this reticulum are excessively close around the nucleus, 
which is situated at the base of the cell; but towards the free 
surface they are much more open, the close and more open 
arrangement merging the one into the other. The meshwork 
extends through the whole cell-cavity, and the interstices be- 
tween the meshes are occupied by a pink cell-sap, and the 
whole is bounded by a structureless ectoplasm. In neither of 
the three layers of cells covering the tracheidal cells of the 
head could any obvious movement of the protoplasm be de- 
tected.” 

After stimulation for twenty-four hours the following histo- 
logical changes were observed : 

“ A gland mounted in water exhibited a mottled appearance, 
such mottling being caused by a vacuolation of the most peri- 
pheral portions of the protoplasm of the gland-cells. In 
section, such a cell showed that in the course of secretion 
there had been a using up of the cell contents, and instead of 
the meshwork occupying the whole of the peripheral portion 


392 LILY HUIE. 


of the cell, so as to give a fairly homogeneous appearance, 
large spherical cavities had appeared in the reticulum here and 
there, such cavities being occupied by the cell-sap. The sap 
had, moreover, a much darker pink tint. Thus a breaking 
down or destruction of some part of the reticulum has taken 
place. After some seventy-two hours’ stimulation this break- 
ing down of the reticulum had reached to such an extent that 
in the peripheral portion before referred to, all the central 
core of the meshwork had for the most part disappeared, and 
replacing it was a single large vacuole filled with cell-sap. 
The cytoplasm had, moreover, contracted from the upper or 
free surface of the cell-wall. In no case does the destruction 
and consequent vacuolation extend to the base of the cell, 
where the nucleus is situated. The nucleus is always sur- 
rounded by dense protoplasm; and there are grounds for 
believing that after very long stimulation, when all the secre- 
tion has been poured out, and before absorption begins, an 
active growth of the protopiasm takes place around the nucleus, 
and in the more basal portion of the cell.” 

«The view here taken (which is supported by certain of the 
staining reactions) with regard to secretion is that in the 
gland-cells the more peripheral network consists of protoplasm, 
together with a formed substance derived from it, and that the 
outpouring of the secretion is caused by the repeated breaking 
down (owing to stimulation) of the protoplasm into this formed 
substance, which is of a mucous nature, and which rapidly 
attracts water and escapes as the secretion to the external 
surface.” 

Gardiner does not mention what the staining reactions are. 
The paper I quote from is called a ‘‘ Preliminary” one. I be- 
lieve the author has never published a more complete account ; 
and, as far as I know, no one has with modern histological 
methods worked out the minute cytological changes which 
occur in the gland-cells, especially with regard to the nucleus 
and nucleolus and the staining reactions of the protoplasm. 


CHANGES IN CELL-ORGANS OF DROSERA ROTUNDIFOLIA. 3893 


EXPERIMENTS. 


With the view of observing the minute histological changes 
induced by functional activity in the gland-cells of Drosera 
rotundifolia experiments were made as follows. 

I. In Perthshire, Scotland, in the beginning of September, 
1895, leaves were fed with the white of a soft-boiled egg, cut 
into pieces approximately 5 mm. in length by 1 mm. in breadth 
and thickness. The leaves were fixed after the following 
intervals of time :—one hour, twenty-four hours, forty-eight 
hours, seventy-two hours. These intervals were chosen in 
order that the leaves might be fixed at the same time of day 
in each case to avoid all variations due to diurnal changes. 
Young vigorous leaves, which showed no trace of having had 
a previous meal, were likewise fixed at the same times, and the 
whole experiment was conducted while the plants were growing 
in their native habitat, an open moor. 

II. Another series of experiments was made in May, 1896, 
in Berkshire, England, where the leaves were again fed without 
being removed from their natural habitat, and fixed at intervals 
of from one minute up to the time when the leaves had again 
fully expanded and appeared to have regained their normal 
condition, which varied with the vigour of the leaf from four 
to seven days. 

III. Repetition of experiment in August, 1896, in Scotland, 
to determine changes induced by an insoluble tissue, namely, 
boiled pieces of cork. 

IV. A number of plants were removed for convenience to a 
cool greenhouse, and their leaves were fixed at various intervals 
after feeding. 

In each set of experiments 15—20 leaves were used for 
each period of time, e. g. twenty leaves to determine changes 
after one minute, twenty leaves to show alterations after two 
minutes, &c. 

The results of the four sets of experiments coincided on the 


394 LILY HUIE. 


tissues being examined by the methods about to be detailed ; 
but none of my drawings have been taken from material of the 
fourth series, because plants kept for a few days in a green- 
house frequently show, as Darwin pointed out, a loss of the 
bright red colour peculiar to them under natural conditions, 
and they are therefore affected to some extent by their 
artificial surroundings. 


Meruops. 


I used the four following fixing fluids: 

1. One per cent. chromic acid solution, as recommended 
by Gardiner. 

2. Absolute alcohol. 

3. Mann’s picro-corrosive alcohol.? 

4. Mann’s weak watery picro-corrosive fluid. 

The effects of these fluids were very different. 

The chromic acid caused the controls (unfed leaves) to close 
up as if powerfully stimulated ; and on subsequent histological 
examination the glands presented the appearance of having 
been stimulated for a short time,—for example, the nuclei of 
the third layer of gland cells had become ameeboid, or irregular 
in form, The whole tissue was cloudy, and stained with no 
precision. 

The two alcoholic fixing agents preserved the control leaves 
in a beautifully expanded condition, but instantly made the 
glands white and transparent-looking, as if some substance 
had been extracted. Sections of such glands showed more or 
less distortion or slight collapse of the cell walls, and great 
vacuolation of the contents of the apical gland-cells. By the 
picro-corrosive alcohol, however, the large chromatin bodies 


1 Gardiner, W., “On the Phenomena accompanying Stimulation of the 
Gland-cells of Drosera dichotoma,” ‘Proc. Roy. Soc. London,’ 1885. 

2 Mann, G., “On a Method of preparing Vegetable and Animal Tissues 
for Paraffin Embedding, with a few Remarks as to Mounting Sections,’ 
‘Trans. Bot. Soc. Edin.,’ vol. xviii. 


CHANGES IN CELL-ORGANS OF DROSERA ROTUNDIFOLIA. 395 


of the later stages of stimulation were specially well fixed, and 
they stained with great precision. 

Manu’s watery picro-corrosive fluid is in every way satisfac- 
tory, especially in its action upon control leaves. It both keeps 
the leaves expanded and preserves the dark colour of the glands. 
Though it causes some shrinkage of the protoplasmic body 
from the cell wall, it does not destroy any of its characters, 
and the points of attachment that remain in spite of the 
shrinkage serve to bring into evidence the fact that the proto- 
plasmic bodies of neighbouring cells are intimately connected 
with the cell wall at corresponding places, and so are probably 
in connection with each other. Further, material fixed in 
the watery fluid takes the stains with most satisfactory 
precision. 

The formula for this fixing fluid is as follows :—Saturated 
HgCl, in 3 per cent. NaCl, 1 part. Saturated solution of 
picric acid in Aq. Dest., 3 parts. This fluid has a specific 
gravity of 1020. 

The material was left in this for twelve hours, then trans- 
ferred for twelve hours to a saturated solution of corrosive 
sublimate in normal saline, and afterwards dehydrated in 
alcohol of gradually increasing strength, being placed in 50 
per cent. alcohol for four hours; 60 per cent., four hours; 70 
per cent., about ten hours (i.e. all night) ; 80 per cent., five 
hours; 90 per cent., five hours ; absolute alcohol, ten hours ; 
then next day into two further changes of absolute alcohol. 
Chloroform was next introduced into the bottom of the vessel 
with a pipette. An hour was allowed to elapse after the tissue 
had sunk in the chloroform ; the fluid was then all poured off, 
and fresh chloroform substituted, and this was changed once 
more after six hours. Paraffin of 52° melting-point was next 
added in small pieces, until saturation at a temperature of 
30° C. was reached. The tissue was then placed in a warm 
chamber heated to the melting-point of the paraffin, a little 
more melted paraffin added, and then the chloroform was 
allowed to evaporate slowly. Sections were cut 5:08 yu thick 
by a Cambridge rocking microtome, fixed on glass slips in 


396 LILY HUIE. 


ribbons by Mann’s! egg-albumen method, and stained either 
by M. Heidenhain’s iron-alum hematoxylin (with or without 
previously staining in Bordeaux red), which gives very good 
results for the merely morphological aspects; or by Mann’s? 
eosin and toluidin blue method, for the study of the alkaline 
and acid reactions. In every stage of stimulation the staining 
reaction was controlled by having on the same slide a row of 
sections of unfed material. The method of staining with 
eosin and toluidin blue is as follows :—Free the sections from 
paraffin by xylol; remove the xylol by alcohol; place the 
sections in Gram’s iodine solution (double strength) for 
five minutes; wash out the greater part of iodine with 
alcohol. 

Wash in water till the sections are quite white; place in 1 
per cent. watery solution of water-soluble eosin (Gribler) for 
fifteen minutes ; rinse in water; place in 1 per cent. solution 
of toluidin blue for five minutes; rinse in water ; decolourise 
in absolute alcohol till the control sections upon the slide 
appear as on fig. 1, or pale blue to the naked eye. It is very 
essential that the absolute alcohol is pure, or otherwise the 
blue colour will be extracted too rapidly by the lime which is 
used in the distillations of alcohol. Further, do not allow the 
alcohol to drop on the slide, because this causes unequal 
washing out; but always immerse the whole slide in a vessel 
filled with absolute alcohol; clear in xylol by immersing the 
slide, and mount in turpentine or xylol balsam. 


GENERAL STRUCTURE OF THE GLANDS. 


In the heads of the tentacles of Drosera rotundifolia, 
D. dichotoma, &c., we may distinguish between glandular 
and non-glandular elements (see the accompanying figure). 
The latter consist of a group of large tracheids forming the 


1 Mann, G., “ A New Fixing Fluid for Animal Tissues,” ‘ Anat. Anzeiger,’ 


vii Jahrg. (1893), Nos. 12 u. 13, p. 442. 
? Mann, G., ‘ Zeitschrift f. wiss. Mikrose.,’ xi, 1895, p. 489. 


397 


CHANGES IN CELL-ORGANS OF DROSERA ROTUNDIFOLIA. 


* il 7 TU | 


MDT TATRA AN AAR NU VV VY ey /} we ViVi 
WON aA VAAN ANA ARR RRR ROMY 


WITTE) 5 rns ee AKANAAAAN 
A SM se 


<\ 
< 


2 
te) TI 


Tentacle of Drosera rotundifolia constructed from serial sections :— 


cells; C = second layer of 


apical gland-cells; B = lateral gland 


AC 
gland-cells; D 
matogen; G 


der- 


tracheids; F = 


third layer of gland-cells; H 
periblem; H = plerome of stalk of tentacle. 


398 LILY HUIE. 


core of the head, and varying in number with the size and 
vigour of the leaf, and in the centre of which terminates a 
fibro-vascular bundle (consisting of one or two spiral vessels, 
of small lumen, and an accompanying sheath of long narrow 
starch-containing cells) which springs from the vascular 
system of the leaf, and traverses the stalk of the tentacle. 
These non-glandular elements presumably function as a 
channel for the transference of fluids. 

The glandular elements consist of three bell-shaped layers 
of gland-cells covering the tracheids. The external layer is 
composed of cells of two kinds. Those at the apex are thin- 
walled elongated cells, numbering in a longitudinal section 
12—14. The lateral cells of the outer layer are shorter, and 
possess thick walls. Their external walls have tooth-like in- 
ternal projections of cellulose. 

The cells of the second glandular layer resemble closely the 
lateral cells of the superficial layer, but their walls have no 
projections. 

The third layer is composed of comparatively few, but very 
long cells, which are remarkable for the large amount of 
tannin in solution contained in their cell-sap, and for the 
strong cuticularisation of those walls, both longitudinal and 
transverse, which separate them from one another. The 
transverse walls are worthy of special attention, because they 
are composed of two projections, one from each side. The 
projection from the outer side is strongly cuticularised ; that 
from the inner side (next the tracheids) is lignified. Between 
the two there is a minute space through which intercellular 
communication is permitted. This layer of cells is also epi- 
dermal in its origin, and arises from a zone of epidermal cells 
immediately below that forming the initial cells of the lateral 
external layer. Hach cell of this zone elongates inwards and 
then subdivides. 

The entire tentacle head is cuticularised superficially. I do 
not understand how Gardiner’ comes to describe the glands 


1 Gardiner, W., ‘On the Phenomena accompanying Stimulation in the 
Gland-cells of Drosera dichotoma,” ‘Proc. Roy. Soc. London,’ 1885. 


CHANGES IN CELL-ORGANS OF DROSERA ROTUNDIFOLIA. 399 


of Drosera dichotoma as without cuticle, and Goebel! to 
say that the apical gland-cells are not cuticularised. Like 
those of Drosera rotundifolia, the gland is covered with 
cuticle. After prolonged treatment by iodine and sulphuric 
acid all that is left undestroyed of the median longitudinal 
section of a gland of either of these species is the continuous 
outline of cuticle and the cuticularised walls of the third layer 
of gland-cells. 

A method for readily seeing the cuticular envelope in com- 
paratively large pieces is to take an entire leaf that has been 
blanched in alcohol, to soak it in water, then to leave it for 
some minutes in strong iodine solution, to lift it out and wash 
off the superfluous iodine in water. A few of the tentacles 
should then be cut off and laid upon a glass slip; a drop of 
concentrated sulphuric acid be placed upon them, and a cover- 
glass laid over them; and when the reagent is seen through the 
microscope to have taken effect by the usual coloration being 
produced, the cover-glass is to be pressed gently so as to rup- 
ture the apices of the gland and squeeze out the contents. 
The cuticular envelope of the gland usually remains attached 
at its base, and floats out as a semi-transparent membrane 
stained yellow by the iodine. 

Gardiner? describes the gland-cells as “ remarkably pitted 
on their upper or free surface; and Goebel® states that the 
outer walls of the apical gland-cells are dotted. I have exa- 
mined pieces of cuticle obtained in the manner just described, 
with a 1/12 apochromatic oil immersion object-glass by Zeiss 
with oculars Nos. 8 and 12, but have found no well-defined 
pores. I have equally failed on examining sections. The 
internal cellulose protuberances of the thickened outer walls 
of the lateral cells of the external glandular layer often produce 
the appearance of enclosing large deep pits; but it is a decep- 
tive appearance, as a little careful study shows. There is no 


1 Goebel, ‘ Pflanzenbiologische Schilderungen,’ Theil i, p. 198. 

2 Gardiner, W., “On the Phenomena accompanying Stimulation of the 
Gland-cells of Drosera dichotoma,” ‘ Proc. Roy. Soc. London,’ 1885, 

3 Goebel, ‘ Pflanzenbiologische Schilderungen,’ Theil i. 


400 LILY HUIE. 


such appearance in the apical cells. The secretion must, of 
course, leave the apical gland-cells somehow, and the albu- 
men pass somehow into the tentacle, but how it is impossible 
as yet to say quite definitely. Silver nitrate enters the tenta- 
cles most readily through the lateral gland-cells, and clefts 
appear between the apices of the apical gland-cells a few 
minutes after stimulation, and the albumen seems to pass in 
through these clefts, as will be shown later on. 

As Charles Darwin pointed out,! the extreme marginal ten- 
tacles of some leaves of D. rotundifolia differ from the 
other tentacles in exhibiting a unilateral development. Their 
heads are very much elongated, and bear the glandular surface 
on the ventral side only. The plan of structure is, however, 
the same as in ordinary tentacles, all the elements being 
present, and the unilateral development appears to be merely 
the result of pressure during early growth in the bud. 


Results of Histological Examination of Material 
fixed in Watery Picro-corrosive Sublimate, and 
stained with Eosin and Toluidin Blue. 


Before describing these results, it may be well to explain the 
names applied to the nuclear organs. 

Chromosomes are the organs which show the well-known 
affinity for alkaline dyes, and which are situated at the peri- 
phery of the nucleus. 

Nuclear Plasm is the material which in the resting state 
forms the main bulk of the nucleus, and is neutrophile in 
character; it corresponds to what is frequently called the 
nuclear sap. 

Nuclear Sap is the more watery fluid inside the nucleus, in 
which is suspended the nuclear plasm, the latter being precipi- 
table by HgCl, and other reagents. 

Nucleolar Chromosomes or Nucleoli are the spherical 
organs which show a special affinity for acid dyes, and have a 
more or less central position. 

Endonucleoli are spaces inside the nucleolus. 

1 Darwin, C., ‘Insectivorous Plants,’ p. 7. 


CHANGES IN CELL-ORGANS OF DROSERA ROTUNDIFOLIA. 401 


APICAL GLAND-CELLS. 


The Resting State as seen in Unfed Leaves used 
as Controls. Fig. 1, Plate 23. 


The Cell Wall stains pale blue. The free wall of the cell by 
shrinkage has sometimes become slightly concave, but other- 
wise appears turgid. 

The Cytoplasm has retracted from the cell walls except in 
the lower basal third, where it either remains entirely in con- 
tact with the cell wall, or is attached to it by strands, which 
are in line with similar strands in neighbouring cells. The 
general appearance is suggestive of communication in this 
region between the apical gland-cells themselves, and also 
between them and the cells of the second layer. The cyto- 
plasm is apparently homogeneous, and arranged in a foam-like 
manner. It stains pure blue, and contains a coarse granula- 
tion embedded in it which stains a deeper blue, and represents 
some zymogen. The constant position of the nucleus is just 
below the middle of the cell, and the cytoplasm between it and 
the outer wall is generally densest in the middle third, being 
more vacuolated towards the free surface and in the imme- 
diate vicinity of the nucleus. Surrounding the nucleus, and 
partly hiding this looser arrangement, is a more finely granular 
mass of protoplasm which often shows a slight affinity for the 
red stain. This zone gradually merges over into the general 
plasm, but shows a number of distinct radiations passing both 
outwards from the nucleus and towards the basal attachment 
of the cell. The plasm, where retracted from the cell wall, 
sometimes shows minute dark blue granules, which frequently 
also occur on the margins of the vacuoles. Compared with 
the lateral cells of the outer layer and with the cells of the 
second layer, the cytoplasm of the apical gland-cells is much 
deeper in colour, because of its greater abundance in these 
cells. 

The Nucleus is situated just below the middle of the cell. 
It may be spherical or oval. 

The nuclei vary in size, being generally smaller in the cells 


402 LILY HUIE. 


nearest to the apex. They are remarkably plump, exhibiting 
a full and rounded contour. They stain of a purplish tint. 

Whether a definite nuclear membrane exists is doubtful, but 
in mid-focus the nucleus is sharply defined against the cyto- 
plasm, partly because of the difference in tint, and also because 
of the peripherally placed dark blue chromatin granules. 

The Nuclear Chromosomes are very minute granules, appa- 
rently shaped like diplococci, each coccus being slightly flat- 
tened, and attached to its neighbour by its broad side. They 
stain an intense blue without any reddish tint. 

The Nuclear Plasm is granular and always very dense in 
controls. Sometimes the granules appear to radiate from the 
nucleolus, and in other cases their arrangement produces a 
sponge-like appearance; or, again, the granules are too closely 
packed to show any definite arrangement. They stain of a 
purplish tint. 

The Nucleolus is embedded in the nuclear plasm. One, 
sometimes more, very distinct deep red nucleolus is always 
present, surrounded by a narrow Frommann’s clear zone. If 
there is only one nucleolus, it is large; it may be in any part 
of the interior of the nucleus, but is generally about one 
quarter to one-third of the diameter of the nucleus, removed 
from its periphery. 

It always presents a perfectly smooth spherical outline, and 
contains one or more distinct endonucleoli. 


Effects of Stimulation for One Minute on the 
Apical Gland-cells. Type 1. Fig. 2: 


The Cell Wall shows no evidence of diminished turgor, there 
being no collapse. It stains purplish. Between the apical 
thirds of neighbouring cells a very distinct reddish material is 
to be noted, staining, in colour and intensity, exactly like the 
semi-fluid egg-albumen in contact with the head. In tentacles 
stimulated with pieces of boiled cork this red material is 
always absent. We are dealing, therefore, with a substance 
(albumen ?) which is passing in between the cells, and not with 
a substance which is being excreted by the cell. 


CHANGES IN CELL-ORGANS OF DROSERA ROTUNDIFOLIA. 403 


The cytoplasm in this specimen has retracted from all the 
walls equally. 

It is very difficult to distinguish in colour between the 
cytoplasmic granulation and its matrix. The general tone is 
of a deep-blue purple, and the matrix seems to be paler than 
the granulation. Fully half the cell is occupied by very large 
vacuoles filled with a fluid which stains pale blue. The walls 
of the vacuoles are lined, to a great extent, with deep blue 
granules which show no trace of purple or red. The vacuoles 
are largest in the upper third of the cell, but usually one or two 
large ones lie also near the base. The cytoplasmic granulation 
is extremely dense round the nucleus, and stretches between 
the vacuoles in radiating ameba-like processes. 

Characteristic of this stage is the presence in the cyto- 
plasm of one (occasionally two) very apparent blue “ corpus- 
cles,” about twice the size of the nucleolus, and having a dark 
blue granular periphery. The explanation I should like to 
offer for this appearance is as follows :—In the unstimulated 
cell some substance is freely distributed, which helps to give 
rise to the affinity for the blue stain. On stimulation this 
substance is withdrawn from the cytoplasm (which now takes 
a purplish stain), and is collected in vacuoles of various sizes. 
By mordanting the cytoplasm surrounding the vacuoles an 
affinity for the blue dye is established, and the deeply stained 
cytoplasm will give rise to the appearance of a distinct wall to 
the vacuole. The smaller vacuoles, because seen in their 
entirety, appear of a deeper colour than the large ones which 
are seen in section, and could easily be mistaken for definite 
corpuscles. The substance acting as a mordant is in this case 
probably tannic acid. 

The Nucleus maintains its normal position in the cell, its 
normal size, shape, and outline. There is no sign of a collapse, 
it being still plump and well filled out. It stains exactly the 
same depth and tint as the cytoplasm, but is distinctly marked 
off from it by its dark blue periphery. 

The Nuclear Chromosomes are in no way different from 
those of unfed specimens. 

VOL. 39, PART 4.—NEW SER. FE 


404 LILY HUIE. 


The Nuclear Plasm appears very indistinctly granular, as 
if the granules had swelled and fused, but is still extremely 
dense. 

The Nucleolus stains a purple-red. It is still full and 
spherical, and frequently appears even enlarged. It is sur- 
rounded by a broad Frommann’s circle, and does not show 
the endonucleoli as distinctly as in the resting condition. 


Effect of Stimulation for One Minute on the Apical 
Gland-cells. Type 2. Fig. 3 (slightly more ad- 
vanced than Fig. 2). 


The Cell Wall is pinkish purple, as in the last specimen, 
and there is the same appearance of some reddish material 
between the apices of the cells. 

The Cytoplasm is shrunken from the walls, but attached in 
one place to the basal wall, and in two places to one lateral 
wall. Its matrix and granulation are indistinguishable from 
each other. The plasm is very red round the nucleus for a 
distance equal to two thirds of the diameter of the latter. 
Beyond this it is deep blue-purple, as in the last specimen, and 
the apical fourth of the cell is entirely occupied by a network 
of vacuoles, the reticulum between them containing masses of 
the deep blue granules which line the vacuoles as in the last 
specimen. The same granules are distributed, apparently in 
strings, through the thick granulation occupying the middle 
of the cell. A pseudo-corpuscle is present, as in the last 
specimen. 

The Nucleus is normal as to position, size, and shape. It is 
of the same reddish tint as the cell-plasm surrounding it, but 
is clearly defined by its dark blue periphery, and still maintains 
its plumpness. 

The Nuclear Chromosomes are like those of unfed specimens. 

The Nuclear Plasm is of the same tint and general appear- 
ance as the granulation surrounding the nucleus. 

The Nucleolus is dark reddish purple, and is surrounded by 
Frommann’s clear zone. The endonucleoli in all the cells are 
much less distinct than in unfed specimens, 


CHANGES IN CELL-ORGANS OF DROSERA ROTUNDIFOLIA. 405 


Effect of Stimulation for Five Minutes on Apical 
Gland-cells. Fig. 4. 

The Cell Wall is coloured as in specimens fed for one minute. 
The specimen from which fig. 4 was taken had been only 
lightly touched by the egg-albumen, but in specimens in eon- 
tact with large masses of the albumen the same appearance of 
a pink substance between the apices of the apical gland-cells is 
seen as in the former specimen. 

The Cytoplasm has retracted from the walls, but remains in 
contact at one or two points with the basal wall and the lower 
parts of the lateral walls as usual. The coloration generally 
resembles that of fig. 2, though there is less of the scattered 
dark blue granulation. The blue corpuscle-like vacuoles are 
generally present in cells of this stage, though absent in the 
one from which the drawing was taken, which, in fact, shows 
scarcely any vacuolation, perhaps because its contact with the 
albumen used as food was comparatively slight. 

The Nucleus is, with regard to position, size, and shape, 
normal. It is of the same red tint or slightly redder than the 
surrounding cytoplasm, and its dark blue periphery is less 
definite and regular than formerly, as if broken at intervals, 
which may perhaps be due to a change in the nuclear mem- 
brane, the latter no longer staining with the blue. 

The Nuclear Chromosomes.—The individual granules have 
become larger and more conspicuous. The gland from which 
the drawing was taken was, however, the only one in the 
material examined at this stage which showed this development. 

The Nuclear Plasm is as dense as in unfed specimens, but 
the granular structure has fused, as already described in 
material after one minute’s stimulation. 

The Nucleolus is as in specimens fed for one minute. 


Effect of Stimulation for Ten to Twenty Minutes 
on the Apical Gland-cells. Fig. 5. 
The Cell Wall is in some specimens the same as in the last 
description, while in others no penetration of albumen between 
the cells can be made out, 


406 LILY HUIE. 


The Cytoplasm shows, as regards vacuolation, great variety 
in material of this stage. The specimen from which the figure 
(5) was drawn is more vacuolated than most of the cells, which 
latter rather correspond in this respect to earlier stages already 
described. All other details remain as in the earlier stages 
(figs. 4 and 2), except that there are no blue granules, and the 
blue “ corpuscles” are rare. 

The Nucleus becomes less distinctly defined in outline. 
Other details remain as in the last specimen described. 

The Nuclear Chromosomes are distinctly larger and more 
conspicuous than in unfed specimens. 

The Nuclear Plasm shows enlarged pale spaces which, com- 
bined with the less full and rounded contour, give to the 
nucleus a slightly attenuated appearance as compared with 
unfed specimens. In colour it resembles the earlier spe- 
cimens. 

The Nucleolus is unchanged from earlier stimulated spe- 
cimens. 


Effect of Stimulation for Twenty Minutes to One 
Hour on the Apical Gland-cells. Fig. 6. 


Notr.—It is extremely difficult to decide which is the 
sequence of the three types next to be considered. They are 
all characteristic of glands which have been active for periods 
intermediate between twenty minutes and four hours, and it is 
by no means certain that the order here adopted is the correct 
one. The type shown in fig. 6, which was taken from a -ten- 
tacle stimulated for one hour, was seen also in a tentacle fed 
for only twenty minutes. I have, therefore, placed it before 
the other two. 

The Cell Wall is pale blue-purple. In many cells the walls 
appear less turgid than in controls, being often somewhat 
crooked or uneven; this is especially the case in the lateral 
walls. 

The Cytoplasm shows considerable vacuolation, especially 
in the outer third of the cells, but not more so than in some 
previously described earlier stages. Round the nucleus the 


CHANGES IN CELL-ORGANS OF DROSERA ROTUNDIFOLIA. 407 


cytoplasm is dense. The entire cytoplasm stains red, except 
that at the apex of the cell there is a trace of blue. A few 
deep blue granules adhere to the exterior of the apical wall, a 
condition which I leave unexplained. Whether in figs. 6, 7, 
and 8 there is not more protoplasm than in cells stimulated 
1—5 minutes is difficult to say. Probably, as the result of 
stimulation, a temporary increase in the amount of cytoplasm 
occurs. 

The Nucleus is shrunken, leaving a clear white zone be- 
tween it and the surrounding cytoplasm. 

Nuclear Chromosomes appear as large granules, and are 
brought into close proximity to each other by the shrinkage of 
the nucleus. 

The Nuclear Plasm is the same colour as the surrounding 
cytoplasm, but appears deeper because of its greater density. 

The Nucleolus stains the same tint as the surrounding 
nuclear plasm, and therefore is difficult to see. In the speci- 
men figured it is invisible, hidden probably by the enlarged 
nuclear chromosomes. 


Effect of Stimulation for One to Four Hours onthe 
Apical Gland-cells. Fig. 7. 


The Cell Wall is very pale blue. The apical cells appear 
frequently to overlap each other, an appearance due to loss of 
turgor. 

The Cytoplasm is stained of a purple-red, becoming bluer in 
the apical fourth, where it is greatly vacuolated. There are 
very marked red granular aggregations in contact with the 
nucleus on its basal aspect. 

The Nucleus is somewhat swollen. Its position is normal. 

The Nuclear Chromosomes are larger than in unfed leaves. 

The Nuclear Plasm is shrunken to form a dense central 
mass, attached at fairly regular intervals to the nuclear peri- 
phery, so as to present a stellate outline. The vacuoles thus 
formed appear slightly blue. The nuclear plasm, because of 
its density, stains more deeply than the cytoplasm, but of the 
same tone of colour. 


408 LILY HUIE. 


The Nucleolus is undiscernible. It is, however, probably 
present, as in other material, also, showing this retracted con- 
dition of the nuclear-plasm, I have been able to reveal it by 
Heidenhain’s iron-alum hematoxylin. 


Effect of Stimulation for One to Four Hours on the 
Apical Gland-cells. Fig. 8. 

The Cell Wall is very pale blue. 

The Cytoplasm is greatly vacuolated. The main body of 
the plasm is red, shading to blue round the vacuoles. 

The Nucleus is spherical and swollen. Its outline is very 
indistinct, and its periphery undefined. 

The Nuclear Chromosomes are deep blue. ‘They show a 
tendency to be displaced or arranged irregularly. 

The Nuclear Plasm is purple. It appears thin because spread 
out over a larger space than before. White spaces occur in it 
here and there. 

The Nucleolus is small and pale red. 


Effect of Stimulation for Twenty to Thirty Hours 
on the Apical Gland-cells. Figs. 9a and 10. 


Note.—The state about to be described is characteristic of 
glands that have been active for twenty to thirty hours, but is 
sometimes reached in twelve hours. 

The Cell Wall is pale blue. It presents the same appearance 
of loss of turgor as noticed in specimens fed for one to four 
hours. 

The Cytoplasm is an extremely scanty network staining red, 
or with a trace of blue here and there in the parts furthest 
from the nucleus. 

The Nucleus is normal as to size and outline, or is shrunken 
so as to form an oblong body with its long axis at right angles 
to that of the cell. 

Nuclear Chromosomes now consist of eight large dark blue 
segments. They are remarkably distinct and conspicuous in 
uushrunken nuclei, and appear to be V- or U-shaped, reminding 
one of the well-known stages in karyokinesis. In shrunken 


CHANGES IN CELL-ORGANS OF DROSERA ROTUNDIFOLIA. 409 


nuclei they are brought very close together, and help to give 
the nucleus a dark purple appearance. 

The Nuclear Plasm is red. It is vacuolated in the un- 
shrunken nuclei. It is dense purple-red in the shrunken 
nuclei, which are therefore much darker than the cytoplasm. 

The Nucleolus is pale red, inconspicuous, sometimes very 
small in the unshrunken nuclei. It is undiscernible in the 
nuclei that have shrunken. Endonucleoli are indistinguish- 
able, owing to the paleness and transparent appearance of the 
nucleoli. = - 

Fig. 98 is a nucleus from an apical gland-cell of a leaf that 
had been fed with yolk of boiled egg, and which after thirty 
hours was fixed in Mann’s picro-corrosive alcohol. It is re- 
markable for the large size of the chromosomes, which seem 
to be approximately V-shaped. It is also remarkable for the 
absence of nuclear plasm. The cytoplasm is in the same con- 
dition as is shown in figs. 9 and 10. 

In those experiments where pieces of cork were laid on the 
leaves to produce stimulation, complete exhaustion of the cell, 
similar to fig. 10, resulted as regards the cytoplasm; while the 
nucleus, although shrivelled and shrunken, stained a pure red, 
there being not a trace of blue colour in the chromosomes. 
This peculiar behaviour of the chromatin segments I have also 
seen, though not frequently, in leaves which had been fed 
twenty to thirty hours previously with egg-albumen. I men- 
tion this fact because there may be some analogy between this 
affinity for red dyes as seen here, and also met with during the 
middle period of mitosis—a view which seems to be supported 
by figs. 9a and B, where one of the features of mitosis is re- 
produced. 


The Apical Gland-cells Two to Three Days after 
Feeding (the Leaf beginning to reopen). Fig. 11. 
The Cell Wall is pale blue. 

The Cytoplasm shows the usual general retraction and basal 
attachments. It is extremely dense and granular round the 
nucleus, forming a well-defined zone which occupies the middle 


410 LILY HOIE. 


third of the cell, so that cells of this stage can be easily re- 
cognised with low powers. The thick granulation stains a 
deep purple. The matrix appears to stain the same colour, 
but fainter. Below the nucleus are one or two very large 
vacuoles, Above it the space is either occupied by one or two 
large vacuoles, or, as in the cell figured, a protoplasmic net- 
work traverses the upper third of the cell, enclosing numerous 
smaller vacuoles in its meshes. 

The Nucleus is in position, size, and shape as in unfed 
specimens. It stains of the same tint as the cytoplasm, and 
therefore lacks the clear definition which is characteristic of 
the resting state. 

The Nuclear Chromosomes are in the form of numerous 
granules of slightly larger size than in the resting state. They 
have a somewhat undefined appearance, and want of clearness 
of outline. 

The Nuclear Plasm is granular. It resembles very closely 
the cytoplasmic granulation in every way ; so that the nucleus 
appears to be only a portion of the cell-plasm enclosed by a 
broken ring of chromatin granules. 

The Nucleolus is either very pale and transparent, and 
usually very small, or is quite undiscernible. 


Apical Gland-cells, Seven Days after feeding (the 
Leaf thoroughly opened up, and glistening with 
exuded drops). Fig. 12. 


The Cell Wall is pale blue. 

The Cytoplasm has retracted as a whole from the cell walls, 
but is attached in the basal third at various points, as is the 
general rule, both in controls and fed leaves. It is densest 
near the nucleus and below it. In the rest of the cell it is 
largely vacuolated. In structure and colour it is exactly like 
that of unstimulated glands. The vacuoles are large and 
numerous, and their contents are perfectly homogeneous, and 
are unstained ; though usually the vacuoles appear to have a 
bluish tinge in mid-focus, caused by underlying or overlying 
plasm. 


CHANGES IN CELL-ORGANS OF DROSERA ROTUNDIFOLIA. 411 


The Nucleus is in position, size, and shape as in unfed speci- 
mens. They also agree with the resting nuclei in their general 
appearance of plumpness and turgor. 

The Nuclear Chromosomes are peripheral granules which 
are either equally small or very slightly larger than those of 
specimens that have never been fed. They stain the same 
deep blue, and are similar in shape and arrangement. 

The Nuclear Plasm is granular. Its general arrangement 
produces a sponge-like appearance, and is therefore similar to 
the least dense specimens among controls. It shows affinity 
for both dyes, as does the nuclear plasm of controls. 

The Nucleolus is exactly as in controls,—that is, very evident, 
with distinct endonucleoli. 


LATERAL SUPERFICIAL GLAND-CELLS. 


The Resting State, seen in Unfed Leaves used as 
Controls. Fig. 13, Plate 24. 


The Cell Wall is turgid, and stains pale blue. 

The Cytoplasm is either not retracted at all from the cell 
wall, or is retracted as in the apical cells, except at certain 
points where it is applied to the basal wall and to the lower 
parts of the lateral walls. It is always attached by strands to 
the projections from the external wall. 

The Cytoplasm is exactly similar to that of the apical cells 
in its structure and staining properties, but frequently 
appears paler because it is less dense. It is very dense round 
the nucleus, but the more peripheral parts are largely vacuo- 
lated. A reddish tint such as sometimes occurs in apical cells 
round the nucleus is hardly ever seen in lateral ceils. 

The Nucleus is placed in the centre of the cell or slightly 
below it. It is similar to the nuclei of the apical gland-cells, 
though sometimes slightly paler owing to the nuclear plasm 
being somewhat less dense. 

Nuclear Chromosomes are like those of the apical cells, or 
sometimes the granules are slightly larger. 


412 LILY HUIE. 


The Nuclear Plasm resembles that in the apical cells, or 
occasionally is less dense. 
The Nucleolus is as in the nuclei of the apical cells. 


The Effect of Stimulation for One Minute on the 
Lateral Superficial Gland-cells. Type 1. Fig. 14. 


The Cell Wall is pale purplish pink. Its turgor is un- 
changed. 

The Cytoplasm is attached to the basal and lateral walls, but 
shrunken towards the apex. The distribution is like that of 
unfed specimens. Its coloration is pinkish purple, and fairly 
uniform throughout the cell. A blue ‘ corpuscle” is usually 
present in the cell-plasm. 

The Nucleus is normal as to position, size, and shape. It is 
slightly deeper in colour than the cytoplasm surrounding it. 

The Nuclear Chromosomes are as in unfed specimens. 

The Nuclear Plasm is unchanged in structure, but shows 
slightly more affinity for red than in unfed specimens. 

The Nucleolus is deep purple-red, and is surrounded by 
Frommann’s zone. The endonucleoli are less distinct than in 
unfed specimens. 

Nore.—The lateral superficial gland-cells at the base of this 
tentacle-head show a much redder coloration than the other 
cells of the layer, and therefore resemble fig. 15. 


Effect of Stimulation for One Minute on the Lateral 
Superficial Gland-cells. Type 2. Fig. 15. 


The Cell Wall is extremely pale purple. There is no red 
coloration between the cells. 

The Cytoplasm is retracted from the basal and lateral walls, 
but not from the apical to which it is closely applied. The 
general distribution of the cytoplasm is like that of controls. 
In colour it is like that of the apical cells,—that is, red 
round the nucleus, becoming bluer towards the periphery. 
The blue “ corpuscles” are present as a rule. 

The Nucleus is normal as to position, size, and shape. It 
is of the same shade of colour as the cytoplasm surrounding 


CHANGES IN CELL-ORGANS OF DROSERA ROTUNDIFOLIA. 413 


it, or sometimes rather bluer, but is sharply defined by its 
dark blue periphery. 

The Nuclear Chromosomes are as in unfed specimens. 

The Nuclear Plasm is less distinctly granular, though as 
dense as in unfed specimens. 

The Nucleolus is deep purple-red, and is surrounded by 
Frommann’s zone. The endonucleoli are not distinct. 


Effect of Stimulation for Five Minutes on the 
Lateral Superficial Gland-cells. Fig. 16. 


The Cell Wall is pale pinkish purple in the specimen figured. 
In other specimens more densely surrounded by the albumen 
used as food, the external walls, and the upper part of the 
lateral walls are very red. 

The Cytoplasm has apparently not diminished in mass, or 
only slightly so. Its coloration is as in apical cells of this 
stage. A blue ‘‘ corpuscle ” is frequently present. 

The Nucleus is normal as to position, size, and shape; but 
its blue periphery is more or less irregular. 

The Nuclear Chromosomes are like those of the apical cells, 
appearing as large, conspicuous, generally distinctly double 
granules. 

The Nuclear Plasm in general arrangement is as in unfed 
specimens. In some cases it is slightly less dense. Its colour 
is like that of the surrounding cytoplasm. 

The Nucleolus is unchanged from the one-minute stage. 


Effect of Stimulations for Ten to Twenty Minutes on 
the Lateral Superficial Gland-cells. Fig. 17. 


The Cell Wall, as in the last case, stains either pale purple 
or red. In the specimen chosen the walls are pale purple. 

The Cytoplasm is in general aspect, bulk, distribution, and 
coloration similar to that seen in specimens fed for shorter 
periods, 

The Nucleus and Nuclear Chromosomes show no change 
since the stage last described. 


AL 4 LILY HUi#. 


The Nuclear Plasm is less dense than formerly. Its colour 
is like that of the surrounding cytoplasm. 
The Nucleolus is unaltered from the last stage described. 


Effect of Stimulation for Twenty Minutes to One 
Hour on the Lateral Superficial Gland-cells. 
Fig. 18. 


Cell Wall is pale purplish blue. The walls of these cells do 
not show the same loss of turgor that is seen in the apical cells. 
Owing to their greater thickness they are more resistant. 

The Cytoplasm is reddish purple. It seems to cover as 
great an area as in control leaves, but is everywhere much 
less dense. There is no shrinkage from the cell wall. The 
peripheral layer is very blue. A blue corpuscle-like vacuole 
is sometimes present. 

The Nucleus is swollen, and its outline is very indistinct. 
The periphery is apparently broken, and there is no trace of a 
nuclear membrane. 

The Nuclear Chromosomes are in the form of a great number 
of irregular granules larger than in unfed specimens, but 
somewhat indistinct in outline, and staining a less intense 
blue. 

The Nuclear Plasm is less dense than formerly, because of 
the increased size of the nucleus. In it there are large white 
spaces. It stains a bluish purple, and so gives the nucleus a 
bluer tint than the surrounding cytoplasm. 

The Nucleolus is pale red. Frommann’s clear zone is not 
seen. The endonucleoli are not well defined, or are absent. 


Effect of Stimulation for One to Four Hours on the 
Lateral Superficial Gland-cell. Fig. 19. 

The Cell Wali is very pale blue. 

The Cytoplasm is red or reddish purple. 

The Nucleus may be said to be perfectly wanting in outline. 
It is chiefly defined from the cytoplasm by its distinct blue 
chromosomes upon their light background of sparse nuclear 
plasm, and large white spaces. 


CHANGES IN CELL-ORGANS OF DROSERA ROTUNDIFOLIA. 415 


The Nuclear Chromosomes appear as a few large segments. 
They are fewer and larger than those of the apical gland-cells. 
They stain a pure deep blue. 

The Nuclear Plasm is very scanty, and of the same tint as 
the cytoplasm, but large white spaces occupy most of the 
interior of the nucleus, which has therefore a washed-out, 
empty appearance. 

The Nucleolus is small. 

Notr.—It seems as if the lateral superficial cells and the 
cells of the second layer in this case were in advance of the 
apical cells; for the collection of the chromatin into a few 
large segments, and the vacuolation of the nucleus, are cha- 
racteristic of more advanced stages of glandular activity. 


Effect of Stimulation for One to Four Hours on the 
Lateral Superficial Gland-cells. Fig. 20. 


The Cell Wall is pale blue. 

The Cytoplasm is red, shading to a bluer tint towards the 
periphery of the cell. The red colour is deepest close to the 
nucleus, but there are no such well-defined masses of red 
material as are seen in the apical cells of the same gland 
(fig. 7). A blue “corpuscle” is sometimes present. 

The Nucleus, Nuclear Chromatin, and Nuclear Plasm are 
the same as in the apical cells. 

The Nucleolus is srnall, and also indistinct because it re- 
sembles in colour the surrounding nuclear-plasm. Fig. 20 
may be an earlier stage than fig. 19, because the nuclear 
chromosomes have not as yet aggregated into eight big seg- 
ments. 


Effect of Stimulation for Twenty to Thirty Hours 
on the Lateral Superficial Gland-cells. Figs. 21 
and 22. 


The Cell Wall is pale blue. 
The Cytoplasm is as in the apical Cells. 
The Nucleus is represented by the two types found in the 


416 LILY HUIRE. 


apical cells. The unshrunken nuclei are in all respects similar 
to those of the apical cells of a corresponding period. The 
shrunken nuclei differ somewhat in the two layers. In the 
lateral cells they seldom, if ever, become elongated, but remain 
approximately spherical. 

The Nuclear Chromosomes, the Nuclear Plasm, and the 
Nucleolus resemble those of the apical cells of same period, 
the formation of eight chromosomes being again very evident. 


The Lateral Superficial Gland-cells Two to Three 
Days after Feeding. Fig. 28. 


The Cell Wall is pale blue. 

The Cytoplasm exhibits a granulation resembling that of the 
apical cells (fig. 11), but is scarcer, and therefore paler; it 
surrounds the nucleus, and stretches away from it, especially 
towards the base of the cells. Cytoplasmic strands also run 
out from the nucleus towards the apical peripheral layer, in- 
tersecting the cell and dividing the space into large vacuoles. 

The Nucleus resembles those of the apical cells, but is here 
somewhat darker, because relatively denser than the surround- 
ing cytoplasm. 

The Nuclear Chromosomes resemble those of the apical 
cells, or are slightly larger; they are apparently spreading 
out to form less compact bodies than in figs. 21 and 22. 

The Nuclear Plasm and Nucleolus resemble those of the 
apical cells of the same period. 


The Lateral Superficial Gland-cells Seven Days 
after Feeding. Fig. 24. 

The Cell Wall is pale blue. 

The Cytoplasm is exactly like that of controls, or is some- 
what less abundant, with larger vacuolated spaces. 

The Nucleus, Nuclear Chromosomes, Nuclear Plasm, and 
Nucleolus resemble those of the apical cells of same gland and 
those of controls. 


OHANGES IN CELL-ORGANS OF DROSERA ROTUNDIFOLIA. 417 


Tue Sreconp LAYER or GLAND-CELLS. 


The Resting State, seen in Unfed Leaves used as 
Controls. 

The Cell Wall is pale blue. 

The Cytoplasm is very scanty compared to cells of the 
outer layer. 

A thin peripheral plasmic layer always remains applied 
to the wall of the cell. Cytoplasmic strands suspend the 
nucleus in the middle of the cell, and radiate from it to the 
peripheral layer lining the wall. Large vacuoles occupy the 
spaces between these radiating strands. The structure and 
staining properties of the cytoplasm are exactly the same as in 
the apical and lateral gland-cells of the superficial layer. A 
reddish tint round the nucleus is sometimes (though rarely) 
seen. 

The Nucleus, Nuclear Chromosomes, Nuclear Plasm, and 
Nucleolus are like those of the lateral superficial gland-cells. 


The Effects of Stimulation for One Minute to 
Twenty Minutes upon the Gland-cells of the 
Second Layer. 

All the cytological changes correspond to those of the 
lateral superficial cells, except that the cell walls are always 
pale purple, never red. 


Effect of Stimulation for Twenty Minutes to One 
Hour on the Gland-cells of the Second Layer. 


The Cell Wall and Cytoplasm are like those of the superficial 
cells stimulated for a slightly longer period. 

The Nuclei of the cells below the apical gland-cells have 
shrunken in the same way as the apical nuclei, while the lateral 
cells of the second layer have their nuclei in the same con- 
dition as the lateral superficial cells, or only very slightly 
shrunken. 

The subsequent changes in the cells of the second glandular 
layer agree in all respects with those of the lateral superficial 
gland-cells. 


418 LILY HUIE. 


GLAND-CELLS OF THE THIRD LAYER. 


The Resting State, seen in Unfed Leaves used as 


Controls. ‘Fig: 254, Nueleus seen in Pomer 
tudinal section; 258, the same in transverse 
section. 


I have only figured the nuclei of this layer, because the cells 
are very large, and the cytoplasm is too scanty to show any 
characteristic changes, except those of the staining reaction. 
In tentacles fixed in absolute alcohol, however, a dense plasm 
is seen filling the cell, and resembling mucus. 

The Cell Wall.—The outer wall of a longitudinal section 
stains pale blue. The inner wall next the tracheids stains 
deep blue. The cuticularised parts do not stain at all. 

The Cytoplasm is merely a thin peripheral layer, slightly 
granular, staining pure blue. The layer lining the wall next 
the tracheids generally shrinks away from it to the centre of 
the cell, which it traverses like a cord, attached at its ends to 
the point in the transverse wall between the cuticularised por- 
tion and the lignified projection. As this is the case in all the 
cells of the layer, a cord, or rather a lamella, appears to stretch 
continuously from cell to cell. 

The Nucleus is spindle-shaped. Its position is peripheral. 
It lies either in the layer of cytoplasm lining the external wall, 
or in that which lines the internal wall. Should the cyto- 
plasm shrink to the centre of the cell, the nucleus is taken 
with it. Its long axis corresponds to the long axis of the cell. 

The Nuclear Chromosomes are minute granules which gene- 
rally appear to be double. They stain deep blue. They are 
peripherally placed, and their arrangement frequently suggests 
that they are disposed like beads upon invisible threads, which 
describe a wide-meshed reticulum. 

The Nuclear Plasm is granular; and the granules appear 
more or less distinctly to be arranged in chains, or to form a 
kind of open network with the spaces occupied by a homo- 
geneous substance. Both granules and spaces stain blue, 
without any reddish or purple tinge, thus differing from the 


CHANGES IN CELL-ORGANS OF DROSERA ROTUNDIFOLIA. 419 


nuclei of the other gland-cells. This is the general rule; but 
in leaves taken in autumn the nuclei of this layer occasionally 
correspond in tint with those of the other gland-cells. 

The Nucleolus.:—There is one large nucleolus, or two smaller 
ones. They stain deep red, and contain one or more distinct 
endonucleoli. 


Effect of Stimulation for One Minute on the Cells of 
the Third Glandular Layer. Type 1. Fig. 26a, 
Nucleus seen in longitudinal section; Fig. 26B, 
in transverse section. 


The Cell Wall is pale purple; deeper in tint and bluer on the 
side next the tracheids. 

The Cytoplasm is as in unfed specimens, except that it stains 
pinkish purple. The occasional presence of a blue ‘‘corpuscle” 
is doubtful. 

The Nucleus is normal as to position and size. It shows no 
alteration in shape, except that it is slightly less pointed at the 
ends. Its general tint is pinkish purple; and it shows more 
affinity for the red stain than the majority of the nuclei of the 
other gland-cells. 

Nuclear Chromosomes are as in unfed specimens. 

The Nuclear Plasm is of a reddish-purple tint. It is dis- 
tinctly granular, and is arranged in an open network in the 
meshes of which is a pale purple homogeneous substance. 

The Nucleolus is dark purple-red. The endonucleoli are 
indistinct or absent. 


Effect of Stimulation for One Minute on the Cells 
of the Third Glandular Layer. Type2. Fig. 27. 


The Cell Wall stains pale blue-purple, showing a much bluer 
tint than the walls of the other gland-cells. 

The Cytoplasm is pale pink, but otherwise is like that of 
unfed specimens. There is no blue “corpuscle” present. 

The Nucleus is normal as to position and size. It is less 
finely pointed at the extremities than in controls. It exhibits 

VOL. 389, PART 4.—NEW SER. aa 


420 LILY HUIE. 


a strong affinity for the red stain. This affinity is markedly 
greater than in the nuclei of the other gland-cells. 

The Nuclear Chromosomes are apparently as in controls. 

The Nuclear Plasm is distributed as in unfed specimens, but 
the granules stain red. The substance which occupies the 
meshes of the reticulum is also red. 

The Nucleolus stains dark crimson. The endonucleoli are 
badly defined. 


Effect of Stimulation for Five Minutes on the Cells 
of the Third Glandular Layer. Fig. 28. 


The Cell Wall and the Cytoplasm resemble those last 
described. 

The Nucleus has become much broader and shorter, and 
shows an intense affinity for the red stain. 

The Nuclear Chromosomes are very distinctly double. They 
stain very dark blue. 

The Nuclear Plasm stains red. It is a very open reticulum, 
composed distinctly of granules strung together in a moniliform 
manner. 

The Nucleolus resembles those of specimens stimulated for 
one minute. 


Effect of Stimulation for Ten to Twenty Minutes 
on the Cells of the Third Glandular Layer. 
Fig, 29. 

The Cell Wall and Cytoplasm resemble those of specimens 
fed for one minute. 

The Nucleus has become still shorter and broader. It 
shows the same affinity for the red stain as the last type. 

The Nuclear Chromosomes have slightly increased in size, 
but in other respects they are like the Nuclear Plasm and 

Nucleolus as in the last type. 


Effect of Stimulation for Twenty Minutes to One 
Hour on the Cells of the Third Glandular 
Layer. Fig. 30. 


The Cell Wall.—The external longitudinal wall stains pale 
blue; the wall next the tracheids dark blue. 


CHANGES IN CELL-ORGANS OF DROSERA ROTUNDIFOLIA. 421 


The Cytoplasm stains very pale bluish purple. 

The Nucleus is almost spherical, and stains very pale blue- 
purple. 

The Nuclear Chromosomes are conspicuous deep blue 
granules of various sizes, and considerably larger than in the 
control. 

The Nuclear Plasm stains reddish purple, and is gathered 
together in the centre of the nucleus, instead of being distri- 
buted in chains. It is still apparently granular. 

The Nucleolus is pale red, and much smaller than in the 


control (fig. 25). 


Effect on the Cells of the Third Glandular Layer 
of Stimulation for One to Four Hours. Fig. 31. 
The Cell Wall and Cytoplasm are as in the last type. 

The Nucleus is oval, spherical, or irregular in shape. It 
stains a more or less intense blue, sometimes very deeply. 

The Nuclear Chromosomes are conspicuous dark blue 
granules. 

The Nuclear Plasm is scanty and granular, staining reddish 
purple. It is collected to form a little cloud round the nu- 
cleolus. 

The Nucleolus stains purplish red. It has a transparent 
appearance, and contains one or more endonucleoli. 


Effect of Stimulation for One to Four Hours on the 
Cells of the Third Glandular Layer. Fig. 82. 

The Cell Wall, Cytoplasm, Nucleus, Nuclear Chromosomes, 
aud Nuclear Plasm as in the last description. 

The Nucleolus is purple-red, and is indistinct because it 
agrees in colour with the granules which closely surround it. 
One or more endonucleoli are present. In the figured speci- 
men the endonucleolus is large. 


422 LILY HUIE. 


Effect of Stimulation for Twenty to Thirty Hours 
on the Cells of the Third Glandular Layer. Fig. 33. 


The Cell Wall.—The external wall is pale blue; the internal 
wall is deep blue. 

The Cytoplasm is very scanty. It stains a very pale blue, 
thus differing in a very marked manner from that of the other 
gland-cells. Sometimes there is a trace of red, or an appear- 
ance of a few red granules here and there. 

The Nucleus is irregular in shape, and stains pale pinkish 
purple, and apparently quite homogeneous. 

The Nuclear Chromosomes consist of a number of dark blue 
granules of varying size which are peripherally placed. 

The Nuclear Plasm is altogether absent, at least no granu- 
lation can be made out, only a diffuse stain. 

The Nucleolus is pale red and transparent-looking. It 
possesses one or more endonucleoli. 


Cells of the Third Glandular Layer Two to Three 
Days after Feeding. Fig. 34. 


The Cell Wall.—The external wall is pale blue, the internal 
wall deep blue. 

The Cytoplasm stains pale blue. 

The Nucleus is irregular in shape, slightly oval, and stains 
blue. 

The Nuclear Chromosomes occur as a number of scattered 
granules, each apparently double. 

The Nuclear Plasm consists of blue granules, arranged in 
rows. It is rather scanty, and stains much less deeply than 
the nuclear chromosomes. 

The Nucleolus is large, stains purplish red, and contains one 
or more endonucleoli. 


CHANGES IN CELL-ORGANS OF DROSERA ROTUNDIFOLIA. 423 


Cells of the Third Glandular Layer Seven Days 
after Feeding. Fig. 35. 


The Cell Wall is pale blue. 

The Cytoplasm is pale blue and somewhat granular, as it is 
in leaves that have never been fed. 

The Nucleus is spindle-shaped in some tentacles, but not 
drawn out at the ends into such fine acicular points as in con- 
trols. The spindle-shaped nuclei are coloured blue like those 
of controls. Nuclei like fig. 34 are also common. 

The Nuclear Chromosomes are exactly as in specimens that 
have never been fed. 

The Nuclear Plasm consists of blue granules like those of 
control specimens, arranged in moniliform rows. In some 
nuclei the appearance is most suggestive of a spiral arrange- 
ment of these chains of granules round the nucleolus. 

The Nucleolus is large and deep red as in controls, with 
clear endonucleoli, and surrounded by Frommann’s clear 
zone. 

In general appearance the cells, as a whole, are very lke 
those in the control specimens. 


SUMMARY AND ConcLUDING REMARKS. 


As the result of an appropriate stimulation of the gland- 
cells of Drosera by food, it is seen that the basophile cytoplasm 
becomes used up, and is represented ultimately by a very 
scanty eosinophilous plasm. The restoration of the cytoplasm 
is brought about by the nucleus absorbing food material, 
metabolising it, and then excreting it into the cytoplasm, as is 
proved by— 

(1) The great increase in the bulk of the basophile nuclear 
chromosomes during a period preceding the restoration of the 
cytoplasm. 

(2) The scanty amount of eosinophilous nucleolar chro- 
matin, and small size of the nucleolar organ during the same 
period. 

(3) By the neutrophile cytoplasm first appearing immedi- 


424 LILY HUIE. 


ately round the nucleus; and this extra-nuclear plasm cor- 
responding in colour reactions (chemically) and in size of 
granules (morphologically) with the intra-nuclear plasm. 

(4) By the diminution of the nuclear chromatin, and in- 
crease in the amount of nucleolar chromatin, whenever the 
cytoplasm has been restored. 

That the extra-nuclear plasm undergoes a further change on 
its own account seems to be proved by the purple colour 
gradually being changed into blue. 

Whether the nuclear or the nucleolar chromatin is the 
primary product of metabolism; whether they are formed 
simultaneously by the nuclear or nucleolar organs respectively ; 
whether the one or the other plays the part of a ferment- 
secreting organ in relation to nuclear metabolism, are questions 
only to be answered after further investigation. 

The aggregation of the nuclear chromatin into a definite 
number of V-shaped segments—eight in Drosera—proves that 
such a change is not a feature characteristic of mitosis, but 
simply a sign of great activity in the nuclear organs. 

I intend continuing my research on Drosera with the view 
of determining whether food materials which differ chemically 
will produce characteristic changes in the gland-cells. 


DESCRIPTION OF PLATES 23 and 24, 


Illustrating Lily Huie’s paper, “ Changes in the Cell-organs 
of Drosera rotundifolia produced by feeding with 
Egg-albumen.”’ 


All the figures were drawn with Zeiss’s camera lucida, Zeiss’s 3, apochro- 
matic oil immersion objective, and No. 8 compensating ocular. The tube 
of the microscope was drawn out to its full extent, and the drawing-paper was 
laid on the table. The figures of the lateral superficial cells, and those of 
the cells of the third glandular layer, were almost invariably taken from the 
same gland, as the corresponding figures of the apical gland.-cells, 


CHANGES IN CELL-ORGANS OF DROSERA ROTUNDIFOLIA. 425 


Fries. 1—12 show the cytological changes in the apical gland-cells, pro- 
duced by feeding with egg-albumen : 

Fig. 1, The resting condition. 

Fig. 2. One minute after feeding. 

Fig. 3. One minute after feeding. Slightly more advanced. 

Fig. 3a. Appearance of semi-boiled white of egg, fixed in picro-corrosive 
and stained exactly like the sections. No blue colour visible. Granu- 
lation much finer than in the cells. 

Fig. 9B. Nucleus from cell, fed for thirty hours on yolk of egg. 

Fics. 13—24 show the cytological changes in the lateral superficial gland- 
cells. 

Fig. 13. The resting condition. 

Fig. 14, One minute after feeding. 

Fig. 15. One minute after feeding. Somewhat more advanced. 

Fics. 25—35 show the changes in the nucleus of the cells of the third 
glandular layer. 

Fig. 25a. Nucleus cut longitudinally. 

Fig. 258. Nucleus cut transversely. 

Fig. 26a. Nucleus cut longitudinally. 

Fig. 268. Nucleus cut transversely. 


eae 
\ ae ane 


i a ayy re PA itt; ae 
pees a4 ithe os np ae fa. Col Seo Jae 
¥ APs ~~ ay aa) », a oe a at, ah 4% an “a _ Vaary 

/ ge Sh Sas Set Bega 


< ' 
A ; a ’ a > vi ry 4 Thy ih ar eye? ru ‘aie 


Potted! at rulad® 


(7%) ie jae. ep Tf) 1? 7 oul ay 


DEVELOPMENT AND SUCCESSION OF TEETH IN PERAMELES. 427 


Observations upon the Development and Suc- 
cession of the Teeth in Perameles; together 
with a Contribution to the Discussion of 
the Homologies of the Teeth in Marsupial 
Animals. 

By 
J. T. Wilson, M.B., 
Professor of Anatomy, 
and 


4 


J. P. Hill, F.LS., 
Demonstrator of Biology in the University of Sydney, New South Wales. 


With Plates 25—32. 


Part I.—IntrRopvuction. 


REcENT years have witnessed a noteworthy revivai of interest 
in the morphological interpretation of the facts of mammalian 
dentition. This renewed interest is traceable to the more 
systematic and thorough-going application of the methods of 
embryological study to the distinctively morphological, and no 
longer merely to the histogenetic aspects of the phenomena of 
tooth development. These methods have yielded a valuable 
supplement to the work of the systematist and paleontologist, 
and have provided a very necessary basis for a critical estimate 
of views arrived at by investigators working along such dif- 
ferent lines of research. 

As a result of the fresh activity to which we have alluded 
there has sprung up a pretty copious literature, which, added 
to that previously existing, renders the study of the problems 
of dentition at the present day by no means a light one, 


42.8 J. ZT: WILSON. AND J, P. HIth. 


It is not, however, the object of the writers to attempt to 
give any comprehensive sketch of the evolution of the problems 
met with in such a study. Nor do they profess to supply any- 
thing lke a complete and adequate discussion, even of the 
present position of these problems. Their aim has from the 
outset been a much more limited one, and they feel absolved 
from making any such attempt all the more that general 
literary surveys of the kind referred to, along with sufficient 
bibliographical data, have been published comparatively re- 
cently by more than one distinguished authority in this 
department of science. Special reference may here be made 
to the address by Professor H. F. Osborn at the American 
Association in August, 1893, specially important from the 
palzontological side, to the masterly résume given by 
Professor Gustav Schwalbe in an address to the “ Ana- 
tomische Gesellschaft” at Strassburg in May, 1894; to 
the brief but lucid and interesting account by Mr. M. F. 
Woodward in ‘Science Progress’ for July, 18945; and, lastly, 
to that given by Professor W. Leche in his recent important 
monograph on “ mammalian tooth-development in the ‘ Bib- 
liotheca Zoologica,’ 1894-5. 

But although in the following pages no effort is made after 
historical completeness, abundant reference must necessarily 
be made to the various phases of scientific opinion upon the 
most important dentitional questions, and we trust that our 
statement of these questions will be found to be not only accu- 
rate, but sufficiently detailed for the purposes of discussion. 

In spite of the substantial enlargement of our knowledge of 
the mammalian dentition attained of late years, it must be 
admitted that no final settlement of some of the more im- 
portant issues has yet been arrived at. On the contrary, 
certain of the more novel phenomena recently brought to light 
have so far served rather to complicate than to simplify the 
problems involved. 

No mammalian order has called forth more discussion in 
respect of its tooth equipment than the Marsupialia. Pri- 
marily, no doubt, this is owing to the inherent peculiarities of 


DEVELOPMENT AND SUCCESSION OF TEETH IN PERAMELES. 429 


the dentition in this order. But in part also it is due to the 
fact that the more or less peculiar characteristics of the denti- 
tion in the Metatheria (and, a fortiori, in the Proto- 
theria) have, on account of the usually accepted zoological 
position of the group, been approached with a quite special 
interest, as likely to afford important clues to the phylogenesis 
of the teeth in Mammalia generally. A consideration of the 
literature of this branch of the subject might indeed suggest 
that something like unanimity has been attained upon several 
important matters, e.g. the serial homology of the deciduous 
premolar of Marsupials with the more anterior premolars. As 
will appear in the sequel, however, it is the opinion of the 
present writers that, even on the points referred to, the dis- 
cussion is very far from being finally or satisfactorily con- 
cluded. As regards several other weighty questions, no 
pretence of agreement is made by various investigators; and 
in not a few instances the accounts of what are obviously 
identical phenomena reveal actual discrepancies of observation, 
as well as very diverse interpretations of those facts in regard 
to which their observations do coincide. It has accordingly 
appeared to the authors highly desirable that the body of fact 
and observation available for the construction of adequate 
hypotheses concerning the marsupial dentition should be 
further supplemented from the developmental side ; and, being 
comparatively favorably placed for obtaining suitable ma- 
terial, they have been induced to undertake the task of, at 
least, providing further reliable records of observations upon 
tooth development in Marsupials. With this object the poly- 
protodont genus Perameles was chosen in the first instance, 
and P. nasuta has throughout furnished the major part of 
our material, and formed the basis of most of the descriptive 
part of this paper. Other forms, however, have been examined 
with reference to particular points, e.g. Dasyurus viver- 
rinus, Phascologale cristicauda, Trichosurus vulpe- 
cula, and Macropus ruficollis. 

Perameles was specially chosen as our type for study and 
description as representing a fairly generalised marsupial type 


430 J. T. WILSON AND J. P. HILL. 


which has not hitherto been made the subject of any complete 
or extended investigation, though several authors (Rése [1], 
Woodward [2], and Leche [3] ) have made reference to obser- 
vations made by them on isolated individual stages of its 
development. 

The series of specimens at our disposal has been a fairly 
complete and satisfactory one, and in its collection we are 
indebted to the following HERLESET for material generously 
placed at our disposal : 

The Trustees of the Australian Museum, Sydney; Pro- 
fessor W. A. Haswell, University of Sydney; Professor W. 
Baldwin Spencer, University of Melbourne; Messrs. C. W. 
De Vis, Brisbane Museum; A. G. Hamilton, George Masters, 
A. M. lea, and Dr. R. Broom. 

Review of Past and Present Opinion.—Whilst we 
have disclaimed any attempt after historical completeness, we 
nevertheless deem it expedient, or even imperative, before 
attempting to set forth our own observations, to make such 
reference to past and present views as may be necessary to pro- 
vide a basis for the future discussion of our own observations. 

In 1867 the law of the succession of the teeth in the Mar- 
supialia was definitely formulated by Flower (4). 

This author claimed to show by means of his own researches 
and those of others that, wherever a tooth-change can’ be 
shown to occur in marsupial animals, such tooth-change in- 
volves only one tooth—the last of the premolar series. 

This view being accepted, the question naturally arose— 
“to which set of teeth in non-marsupial Mammals do the 
non-changing antemolar teeth of Marsupials correspond ? ” 
Flower concluded that they answered to the permanent or 
successional teeth of higher Mammalia, and that thus only 
one true milk-tooth was present in any Marsupial. This in- 
terpretation, taken along with the more primitive character of 
the marsupial organisation in other respects, led him to suggest 
that “ the milk or deciduous teeth may rather be a set super- 
added to supply the temporary needs of Mammals of more 
complex dental organisation ” (4, p. 639). 


DEVELOPMENT AND SUCCESSION OF TEETH IN PERAMELES. 431 


The views of Flower were subsequently adopted and ex- 
tended by Oldfield Thomas (5), who adduced in their support 
certain important paleontological facts. 

Upon the relation of the Eutherian to the Metatherian 
stocks, Thomas is notably explicit. He considers that “the 
change from Metatherian to Eutherian” occurred, “in all 
probability,’ within the limits of the family Dasyuride. 
His conclusions, therefore, that “the rudimentary tooth- 
change of Marsupials represents an early stage in the first 
formation ” of a complete change, and that the Marsupials 
are “still in a backward condition out of which the Kutheria 
have long since passed,” are of quite fundamental significance 
in his interpretation of the phenomena of the dentition of 
Mammals generally. 

The general argument in favour of Flower’s view of the 
primitive character of the single tooth-change in Marsupials, 
derived from the zoological position of the order, is expressed 
thus by Thomas :—“ When we consider that in every character 
of their organisation the Marsupials are infinitely behind and 
at a lower stage of evolution than the placental Mammals, it 
would appear to be a total subversion of all the ordinary rules 
to suppose that in this one character of their dentition they 
should have passed on in advance of all the other Mammals, 
and, having gone through the condition in which the latter 
now are, should have again nearly evolved away that process 
of tooth-change which is to its placental possessors so evidently 
advantageous ”’ (5, p. 450).! 

A more special argument is also derived from the fact “ that 
five out of the six families of Marsupials, natives both of 
Australia and America, have, with the comparatively unim- 
portant exceptions already noted as occurring among the 
Dasyuride, arrived at precisely the same stage of tooth- 
change ”—a circumstance which would be unlikely to occur if 


1 It may be remarked that this is precisely what has actually occurred in 
Dasyurus and Sarcophilus in relation to other Dasyuride, according 
to Thomas himself, i.e. after first evolving a tooth-change (single, of course), 
they have next proceeded to get rid of it. 


432 J. I. WILSON AND J. P. tbh, 


the modern tooth-change were a remnant of a fuller one, for 
then “ we should naturally expect that, under the very various 
conditions of the struggle for existence, equally various 
degrees of reduction would have been attained to.” It may 
be pointed out that the cogency of this argument would 
entirely disappear if it could be shown to be probable that the 
(hypothetical) reduction was dependent upon conditions of 
life common to, and peculiar to, the entire group. 

In professing his firm adherence to Flower’s view that the 
teeth of Marsupials in front of the last premolar represent the 
permanent teeth of other Mammals, Thomas states that he 
was led to that opinion by “ finding the impossibility of work- 
ing out the general homologies of the teeth on the basis of 
the opposite view,” and by very extended observations of 
specimens. 

It is perhaps unnecessary to follow Mr. Thomas through 
his somewhat elaborate and rather far-fetched doctrine of a 
retardation which ought to have occurred in Marsupials 
if their teeth had formerly possessed milk predecessors which 
were subsequently lost,—a retardation parallel to that which 
occurs in the case of the last permanent premolar; and which 
he alleges as also occurring in the case of the first incisor in a 
number of Marsupials, probably by way of preparation for the 
acquirement of a milk-tooth. Thomas himself remarks upon 
‘the difficulties in the way of understanding how the ordinary 
processes of evolution” could first have “ brought about such 
a preliminary retardation,’—a remark with which his readers 
will readily agree. 

The above views have now an interest for us which is largely 
historical, since Thomas has explicitly surrendered the main 
point of his position in view of later discoveries. 

The views of Flower, prior at least to Thomas’s advocacy of 
them, had not as a whole received general approval. His 
identification of the non-changing antemolar teeth of Marsu- 
pials with those of the permanent series of Placentals was, 
indeed, practically universally received. The view, however, 
that the seemingly almost complete monophyodontism of 


DEVELOPMENT AND SUCCESSION OF TEETH IN PERAMELES. 433 


Marsupials was a primitive condition, and that the original 
mammalian stock was actually monophyodont, had not met 
with general acceptance previous to the publication of Thomas’s 
paper in 1887. Notably Winge (6)! maintained, as against 
Flower, that in all probability the condition in Marsupials was 
secondary, and due to an almost complete loss of the milk 
dentition in that group; and this opinion of Winge possessed 
many adherents. 

The emphatic reiteration of the views of Flower by so emi- 
nent an investigator as Oldfield Thomas, upon the grounds of 
his own extensive researches, seems to have served as a stimulus 
to renewed consideration of the subject of marsupial dentition, 
especially from the developmental point of view. The first- 
fruits of the latter research appeared in December, 1891, in 
the form of a paper by Kiukenthal (7) on tooth development 
in Didelphys. 

In his introductory remarks Kikenthal indicates the Flower- 
Thomas theory of the primitive character of the single tooth- 
change in Marsupials as the prime consideration with refer- 
ence to which his investigations were directed. Thomas had 
admitted that a discovery of successional germs accessory to the 
non-changing marsupial (antemolar) teeth would be fatal to 
the theory in question, and the result of Kiikenthal’s exami- 
nation of Didelphys was to reveal the presence, at the lingual 
sides of nearly all the developing teeth in that animal, of 
epithelial ingrowths of the primitive dental lamina. These 
ingrowths he naturally interpreted as rudimentary enamel- 
germs of successional teeth, in series with the undoubted 
enamel-germ of the actual successional premolar (p 3, i.e. 
pm. 4 of Thomas). 

Kukenthal thus believed that he had established on a per- 
fectly secure basis, the view that the existing teeth of Marsu- 
pials in front of the last premolar are in reality milk-teeth in 
series, not with the persisting last premolar, but with its milk 

1 We have been unable to consult Winge’s original paper, and have had to 


rely for our knowledge of his views upon the accounts of Kiikenthal and 
Leche. 


434 J- T. WILSON AND J. P. HUGE. 


predecessor. Kiikenthal’s claim was promptly admitted by 
Thomas himself (8) as apparently incontrovertible, and he 
therefore no longer refuses to subscribe to the view of a primi- 
tive diphyodontism of mammals, ‘‘ probably in direct suc- 
cession to the irregular polyphyodontism of their reptilian 
ancestors,”—a diphyodontism which “ may even have existed 
in what were in other respects members of the latter class.” 
Nevertheless Thomas is far from admitting that the whole 
problem of the phylogeny of the dentition in Mammalia 
is thereby cleared up. He regards, e.g., the fact that the 
Mesozoic Triconodon changed a single tooth only, as ren- 
dered inexplicable by the newer theory. Still he definitely 
adopts the latter, merely passing on to a criticism of Kiken- 
thal’s views upon other issues, i.e. the homologies of the 
individual premolars, and the questions of the nature and mode 
of origin of the molars. 

An important and elaborate memoir on the development of 
the human teeth by Carl Rose was published in 1891 (9), and 
apparently independently of Kiukenthal, and contempora- 
neously with him, Rése had also worked out the tooth develop- 
ment of Didelphys as well as of several other marsupial 
forms. The results of this work were published in September, 
18921) 2 

As regards the question of the presence of rudimentary 
enamel-germs of successional teeth, Rose largely confirms the 
observations of Kiikenthal upon Didelphys,” and he showed 
that lingual ingrowths of the dental lamina, similar to those 
found in Didelphys, were present in all other Marsupials he 
examined. 

During 1892-3 Rose continued to publish the results of an 
active investigation into the phenomena of tooth development 
in a number of other forms, both mammalian and reptilian. 


1 Rése’s views upon the marsupial dentition were, however, outlined in his 
paper on tooth development in Edentata, published in July, 1892 (34). 

2 An unfortunate error in Rose’s paper, subsequently acknowledged by 
him, was his mistaking the enamel-organ of the deciduous premolar for that 
of the first molar. 


DEVELOPMENT AND SUCCESSION OF TEETH IN PERAMELES. 435 


Rose has also pushed boldly into the realm of theory, and 
has presented hypotheses to account not only for the more 
divergent types of mammalian dentition, but for the derivation 
of the latter from that of polyphyodont and homodont 
ancestors. We cannot avoid the feeling that in his views ~ 
upon these matters the quantity of theorising is somewhat 
disproportionate to the extent of the basis of fact and ob- 
servation. His schematic representation (29), for example, 
of the mode of derivation of the human dentition from a lowly 
reptilian-like type appears to us to be a rather striking 
instance of a highly ingenious but a somewhat uncritical 
exercise of the scientific imagination. 

Rose’s contributions to our knowledge of tooth development 
are, however, both extensive and of very great interest and 
importance, and his interpretations are, as a rule, entitled to 
the utmost respect. 

As regards the mammalian dentition, Rése is a powerful 
advocate (a) of the primitive diphyodontism of the mam- 
malian stock, (0) of the doctrine that the multituberculate 
teeth of the heterodont dentition originated by a process of 
fusion of several individually distinct and simple conical 
members of a homodont dentition. 

In these views Rose has, of course, had many predecessors, 
Kikenthal among others, but Rose’s attempts to further estab- 
lish the validity of these propositions deserve special recognition. 

As regards the marsupial dentition in particular, Rése’s 
general conclusions may be thus summed up :—That dentition, 
like that of mammals generally, was originally diphyodont: 
** Allein die Thatsache, dass hinter allen Zahnen zeitweilig eine 
zusammenhingende, bandférmige Zahnleiste verlauft, beweist 
uns, dass die Beuteltiere von diphyodonten Saéugern und 
weiterhin von polyphyodonten reptilienihnlichen Vertebraten 
abstammen” (34, p. 508). And again: ‘Das Milchgebiss 
der Siugetiere ist nicht eine Neuerwerbung, sondern eine 
phyletische Vererbung und ist entstanden durch zusammen- 
draugen mehreren reptilienahnlichen Zahnserien in eine 
Einzige” (ib., p.509). Thus it is held that during phylogeny 

VOL. 39, PART 4,—NEW SER. HH 


436 Ju T. WILSON AND J.P. Hib, 


the Marsupials have suffered loss of all the teeth of the “ per- 
mauent” series in front of the last premolar, with, indeed, the 
doubtful exception, according to Rose, of the last upper incisor 
in some forms. 

In this connection Rése remarks upon the “ noteworthy con- 
stancy of the group with reference to the single tooth-change 
from Mesozoic up to modern times—a constancy which had 
seemed to Thomas so inexplicable,—and he offers the following 
explanation :—“ Die Beuteltiere sind aber bei der Reduktion des 
vielfachen Zahnwechsels der reptilienahnlichen Vorfahren der 
heutigen Sauger gleichsam iiber das Ziel hinausgeschossen und 
haben sich in eine Sackgasse verrannt, aus der kein Ruickweg 
moglich ist” (1, p. 705). He also suggests that such a reduc- 
tion could only be serviceable if at the same time the milk- 
teeth came to grow from persistent pulps,—a condition which, 
he remarks, is attained only by the wombat. 

Rose agrees with Kikenthal in referring the molars to the 
same series as the other teeth, i.e. to the milk or first denti- 
tion, if they are to be referred specifically to one or other of 
the two series represented by the antemolar teeth. Into his 
later view respecting the possibility of the molars representing 
*end-members of separate dentitions’’ we need not now enter. 

In 1893 a highly interesting and important paper was pub- 
lished by Rose (11) on the subject of tooth development in 
Phascolomys. We are of opinion that the full significance 
of the observations therein recorded has been apprehended 
neither by subsequent writers nor by the author himself, though 
the latter undoubtedly attached considerable importance to 
them. 

The results of Roése’s research (for which unfortunately only 
one pouch specimen, of 19 mm. body-length, was available) 
would seem to tend towards a confirmation of a long discredited 
statement of Owen’s (15), to the effect that the incisor teeth 
and the first molar are changed in the young animal. It 
appears from Rose’s investigation that at a comparatively early 


1 This view regarding the nature of the last polyprotodont incisor has 
received no confirmation whatever from any later observer. 


DEVELOPMENT AND SUCCESSION OF TEETH IN PERAMELES. 437 


stage in its tooth development the wombat possesses rudiments 
of teeth entirely unrepresented in the adult dentition; and 
that, further, certain of these teeth are precociously calcified, 
and, according to the author, obviously belong to a generation 
antecedent to that to which the permanent teeth of the animal 
belong. In addition to this he finds, in a position in the jaw 
corresponding to that occupied later by the single adult pre- 
molar, a single pointed, precociously calcified tooth, provided 
with a large and “ bulbous” lingual germ (“Ersatzleiste”), 
which latter he regards as the true germ of the permanent 
premolar; the rudimentary but well-formed calcified tooth 
beside it being the deciduous predecessor to which Owen 
apparently referred. 

In summing up his results Rose first emphasises the fact of 
the possession by the wombat of two typically distinct den- 
titions “ ganz iihnlich wie die placental en Sauger.”” Of these 
two incomplete series he points out that all the members of 
the first, and several of those of the second, must be resorbed 
in early life. 

He then proceeds to put the important question, ‘ which of 
the two tooth-series of wombat corresponds to the milk series 
of other Marsupials?”? And just here he finds himself in- 
volved in some little difficulty. For, while he finds it easy to 
homologise the posterior elements of the series with those 
found in other Marsupials by reckoning the “ Ersatzleiste” 
of the small calcified “‘ premolar” as the germ of the persist- 
ing one, he feels almost compelled to identify the calcified and 
precocious germs of the first generation in the anterior seg- 
ment of the jaw with the “milk” (persisting) dentition of 
other Marsupials. He is thus led to postulate a fundamental 
difference in homology between the teeth of the anterior and 
posterior segments of the jaw respectively, and thus also be- 
tween the anterior persisting teeth of wombat and those of 
other Marsupials. “ Wahrend die Schneidezahne der polypro- 
todont Beutler zur ersten oder Milchzahnserie gehoren, rechnen 
diejenigen vom Wombat zur zweiten oder bleibenden.” He 
adds, however, that it is not improbable that the condition 


438 J. T. WILSON AND J. P. HILL. 


found in wombat may ultimately be found characteristic also 
of some other diprotodont Marsupials. 

It will be noted that if the view for which Rose indicates 
his preference be correct, the acceptance of it implies at least 
a partial surrender of what we may call the Kiikenthal-Rose 
position in relation to the marsupial dentition generally. 

In his recent work (8, pp. 100, 101), Leche has definitely 
adopted that interpretation of Rése’s observations which Rose 
himself somewhat reluctantly put aside, and has thus attempted 
to conserve in its integrity the ruling modern theory of the 
marsupial dentition as a true ‘ milk ”’ series. 

The solution which Leche confidently advocates involves the 
interpretation of the prematurely calcified teeth of wombat, 
not as homologues of milk-teeth, but as vestigial remains 
of a “prelacteal” series inherited from the polyphyodont 
ancestors of the Mammalia. 

It will appear in the sequel that the present writers are 
very strongly inclined to adhere to that interpretation of these 
rudimentary teeth which Rose decides to accept in the paper 
under consideration. But in so doing it will appear that they 
also advocate a much wider extension of the hypothesis in- 
volved in that interpretation than was contemplated by Rose. 

In the ‘ Proc. Zool. Soc.,’ May 2nd, 1893, there appeared a 
paper by Mr. M. F. Woodward (2) on ‘“ The Development of 
the Teeth of the Macropodide,” in which the author gave 
an account of his discovery in certain members of that family 
of a number of small calcified teeth supplementary to the 
proper rudiments of the adult dentition. 

Thus in the upper jaw of Petrogale he found, in addition 
to the germs of the three adult incisors, three “ minute 
calcified rudimentary (or rather vestigial)” teeth. So, again, 
in the case of the lower jaw he found two vestigial teeth 
in addition to the enamel-germ of the permanent lower 
incisor. From his facts Mr. Woodward draws certain 
conclusions regarding the relations of the dentition of 
Macropods to that of Polyprotodonts and primitive Mar- 
supials, with which we are not here concerned. It is to be 


DEVELOPMENT AND SUCCESSION OF TEETH IN PERAMELES. 439 


noted that the author refers the adult incisors of Macropods 
to the first dentition, accepting, as he does, the criterion 
which had seemed sufficient to Kiikenthal and Rose. He 
says, ‘‘If these various and often minute cord-like down- 
growths of the dental lamina (lingually to the enamel-germs 
of the adult incisors) are to be interpreted as representing 
rudiments of teeth, as seems probable from comparison with 
the known rudiments of the first or second dentition in other 
mammals, then we find that in the kangaroos the incisor 
teeth all belong to the first dentition, that the relations of the 
canines are uncertain,” &c. &c. Again, on account of the 
presence of a “distinct but small” lingual downgrowth by 
the side of the enamel-germ of the foremost upper vestigial 
tooth, Woodward referred that tooth to the first dentition 
equally with the germs of the adult teeth; and he extends the 
same determination to the other two, partly from their 
general analogy to the first, and partly on account of certain 
observations as to their relative position in reference to the 
dental lamina, and to the enamel-germs of the neighbouring 
adult teeth. 

Similar reasonings prevailed with him in regard to the 
lower incisors, and accordingly he is led to interpret the 
vestigial incisors simply as decadent members of the same 
series to which the adult teeth belong. 

Mr. Woodward’s paper constitutes the earliest publication 
in which were recorded observations upon the presence of 
undoubted embryonic vestigial teeth in Marsupials. Rdése’s 
paper, dealing with such structures in wombat, was not pub- 
lished till 4th August, 1893. 

On the same day there appeared in the ‘ Morphologisches 
Jahrbuch,’ Bd. xx, a paper (12) by Professor W. Leche, which 
bore the date January, 1893. This contribution was supple- 
mentary to his paper (13) in the preceding volume on the 
subject of mammalian tooth-development, and it contains an 
account of his researches into the development of the teeth of 
Myrmecobius, to which in his previous paper he has 
referred as still in progress, 


440 J. T. WILSON AND J. P. HILL. 


In Myrmecobius Leche had discovered the presence of 
calcified structures connected with the dental lamina, and 
placed labially with reference to it and to the enamel-germs 
of the adult teeth. These calcified tooth-remains, though less 
perfectly formed, or perhaps we should say more completely 
degenerated, than those described by Mr. Woodward, bear a 
striking general resemblance to his, both in respect of their 
topographical relationships and of their structural features. 
Leche’s interpretation of them was, however, widely different 
from that given by Woodward in the case of the Macropod 
vestigial teeth, for he viewed the calcified structures as the 
sole remains of an entire “ prelacteal” dentition which had 
otherwise become suppressed. To this view Leche still 
adheres in his latest work (3), where he claims that both 
Woodward’s and Rose’s observations just referred to are to be 
explained along similar lines. 

Woodward himself (14) has accepted Leche’s notion 
of a “prelacteal” dentition, so far at least as Myrmeco- 
bius is concerned, and he believes that he has obtained 
confirmation of the view through a similar discovery in 
Phascologale. 

By far the most important recent contribution to the lite- 
rature of this subject is the comprehensive monograph by 
Professor W. Leche already referred to (3). 

In addition to a rich collection of observations copiously 
illustrated, this book contains a systematic discussion of all 
the more important issues raised by a study of the ontogeny of 
the mammalian dentition. In the latter portion of this paper 
Leche’s views are dealt with in detail in so far as they have a 
bearing upon the subject-matter of this work. Meanwhile we 
may remark that one of the most important features of Leche’s 
contribution is his criticism of the commonly prevalent but 
loose and unreliable notion as to what really constitutes a 
morphological tooth-germ. 

Such a critical determination is urgently needed if any 
further progress is to be made in investigation of the dentition 
along the lines of embryological research. In the development 


DEVELOPMENT AND SUCCESSION OF TEETH IN PERAMELES. 4A1 


of our own views we have been largely stimulated and aided by 
Leche’s masterly essay. 


GENERAL ACCOUNT OF THE LEADING CONCLUSIONS FROM OUR 
OWN INVESTIGATION. 


In approaching the subject of our own researches one or 
two matters call for special remark. 

In the first place it is necessary or desirable that we should 
indicate the general grounds upon which we have felt compelled 
to depart from the now usually accepted views of the marsupial 
dentition. The attitude we have felt bound to adopt practi- 
cally implies a return to Winge’s views on the main question 
involved; but with the notable corroboration of that theory 
derived from our interpretation of the so-called ‘ prelacteal”’ 
vestigial teeth as remains of the true milk series, which, 
according to Winge, had been entirely lost from the marsupial 
dentition, except in the case of the deciduous last premolar. 

At the time our research was begun, and for some time 
afterwards, we were fully persuaded of the truth of the 
Kiikenthal-Rése theory, and found no reason to question 
Leche’s identification of his vestigial teeth in Myrmecobius 
as ‘“prelacteal,’ to which category we felt, with Leche, 
strongly disposed to refer the vestigial calcified teeth dis- 
covered by Woodward in the Macropodide. 

We were accordingly highly delighted to (apparently) con- 
firm the accepted views by the further discovery in Pera- 
meles of a number of quite similar “ prelacteal”? rudimentary 
teeth. 

The examination of the earlier stages of tooth development 
in Perameles, however, forced upon us irresistibly the con- 
viction that the deciduous premolar in that animal 
must belong to the same series as the so-called 
* prelacteals.” 

For in Perameles the deciduous premolar is differentiated 
from the dental lamina contemporaneously with the so-called 
 prelacteal”’ teeth, and at a period prior to the differentiation 


442 J. T. WILSON AND J. P. HILL. 


of any other tooth of the adult dentition. This early forma- 
tion of the deciduous premolar is not peculiar to Perameles. 
It is important to note that it occurs also in Didelphys. 

But its precocity cannot be explained in Perameles, as it 
usually has been in Didelphys, by a reference to the future 
size of the tooth, since, as is well known, it isin Perameles 
by far the smallest of the whole cheek-tooth series, and, in- 
deed, is absolutely a small-sized tooth. Neither can the early 
development be explained by its position in the jaw. Indeed, 
from its relatively posterior position in the young and develop- 
ing jaw, quite the contrary, i.e. a retardation, might have 
been expected, such as affects the hinder (molar) elements of 
the dental series. The discussion of this important matter 
will be resumed further on in this paper. So much it was 
essential to state at the outset to render our position in- 
telligible. 

It was only with great reluctance that we arrived at this 
apparently inevitable conclusion that the deciduous premolar 
was a member of the “ prelacteal” series, and although it 
seemed impossible to interpret the facts in the case of Pera- 
meles in any other way, it seemed doubtful whether we 
should be able to reconcile this view satisfactorily with the 
facts observed in the case of other Marsupials, more particu- 
larly in Didelphys. At least, it seemed difficult to do so in 
a way that would carry conviction to those whose views on the 
marsupial dentition were moulded upon the conditions repre- 
sented in Didelphys. In that animal it has indeed been 
shown (cf. Kikenthal [7] ) that the deciduous premolar is, at 
the first, the most advanced of all the teeth; but the subse- 
quent course of its development, resulting in the production 
of a relatively large molariform tooth, seemed anything but 
favorable to a theory according to which it must be regarded 
as belonging to an otherwise degenerated “ prelacteal ” order 
of teeth. 

It may now appear somewhat curious that the very obvious 
possible explanation of the whole difficulty did not dawn upon 
us at the very outset. But so completely were we under the 


DEVELOPMENT AND SUCOERSSION OF TEETH IN PERAMELES. 448 


influence of the ruling theory of the homology of the perma- 
nent marsupial dentition to the milk dentition of higher 
Mammals, that it was some little time ere it occurred to us 
even to call in question the grounds of this identification, and 
at the same time the “ prelacteal”’ character of the vestigial 
teeth so frequently met with in Marsupials. But we very soon 
saw that, could the lacteal theory of the permanent mar- 
supial dentition be successfully impeached, all difficulties 
would be at once removed from the theory we felt bound to 
advocate from a consideration of the condition in Perameles, 
where the deciduous premolar and the so-called “ prelacteals ” 
evidently belong to one and the same dentitional series. This 
series would of course no longer be regarded as “ prelacteal,” 
but as the true homologue of the milk series of the higher 
Mammalia. 

It was while in this mental attitude to the question that we 
were first appealed to by Leche’s sceptical critique (3, pp. 132, 
136, &c.) of the criteria usually relied upon as sufficient to deter- 
mine whether or not any given lingual downgrowth of the 
dental lamina is morphologically the enamel-germ of a 
replacing tooth. For the fuller elaboration of our final 
attitude to the questions thus suggested, and for the evidence 
upon which our conclusions are based, we must refer to the 
subsequent sections of this paper. Here it is only necessary 
to say that we have been led to reject the presently prevailing 
Opinion regarding the more or less swollen downgrowths 
formed at the lingual sides of the developing marsupial teeth. 
It is our emphatic opinion that these are not the repre- 
sentatives of the replacing teeth of higher mammals, nor, 
indeed, are they actual individual tooth-germs at all, but mere 
residual appendages of the lamina, though they may, in a 
sense well recognised by Leche, contain the “ promise and the 
potency” of a possible tooth-generation. But the adult teeth 
in front of the last premolar are the genuine homologues of 
the replacing teeth of higher mammals, and the milk-teeth of 
the latter are represented, in front of the deciduous premolar, 
by the so-called “ prelacteal” vestiges, of which, as we shall 


444 J. T. WILSON AND J. P. HILL. 


have occasion to notice, there are abundant remains in 
Perameles. 

We are convinced that this view is amply justified by the 
considerations set forth in this paper, and that, when 
deliberately weighed, it will at once be found agreeable to all 
the facts of development of the marsupial dentition, and to 
afford a far more natural and unstrained explanation of these 
facts than the hypotheses heretofore dominant. We cannot 
but believe that the enormous simplification which would be 
effected by the adoption (if satisfactorily demonstrated) of the 
view we now advocate must be appreciated on every hand. In 
its light the peculiarities of the marsupial dentition may be 
rendered easily intelligible, and the difficulties of interpreta- 
tion hitherto associated with these peculiarities forthwith 
disappear. 

In particular the new view would seem to afford the only 
adequate explanation of Rése’s highly interesting discoveries 
in the wombat. We have already seen how this observer felt 
bound to interpret the rudimentary teeth he found in his 
foetal specimen as belonging to a milk series; and although 
Schwalbe and Leche have both subsequently expressed the 
opinion that these rudiments are to be interpreted as ‘‘ pre- 
lacteal,’’? we are of opinion that this view presents many 
difficulties which are obviated by the simpler and more natural 
view we feel bound to advocate on the grounds of our own 
observations. We at least find it now impossible to adopt any 
different conclusions on the points in question, respecting the 
morphology of the marsupial dentition, than the following : 

1. The permanent teeth of Marsupials are the homologues 
of the permanent or replacing teeth of higher mammals. 

2. The deciduous premolar is a true milk-tooth, and it is 
not the sole representative of the series to which it belongs, 
since the so-called ‘‘ prelacteal ” teeth are in reality milk-teeth 
which have undergone reduction, and have well-nigh wholly 
disappeared, under the operation of influences unfavorable to 
their development. 

3. The lingually situated downgrowths of the dental lamina 


DEVELOPMENT AND SUCOESSION OF TEETH IN PERAMELES. 445 


by the sides of the developing teeth are no rudimentary 
enamel-germs. They are merely portions of a quite indifferent 
“yesidual dental lamina” becoming liberated on differ- 
entiation of the permanent teeth from the parent lamina. The 
swelling of the distal portion of the downgrowth is a mere 
thickening of the free border of the residual lamina, and exhibits 
no differentiation which is really characteristic of the pro- 
duction of actual enamel-organs. 

It is possible that its presence is to be explained in terms of 
a continued formative activity on the part of an unexhausted 
dental lamina, an activity which might conceivably issue in 
the production of enamel-germs of a third series, such as 
Leche has shown exceptionally to occur amongst other 
mammals. We are, however, unaware of any single instance 
in which a differentiation genuinely characteristic of the for- 
mation of an enamel-organ of a third dentition has been 
observed among the Marsupialia. 

We may here touch upon one other strong point in the 
prima facie case we have been attempting to set forth in 
favour of the serial homology of the deciduous premolar with 
the so-called “ prelacteal” teeth. This is brought to light in 
the investigation of those marsupial forms in which the de- 
ciduous premolar is inconstant in its presence or altogether 
absent from the normal dentition. Thus in Dasyurus viver- 
rinus, in spite of the absence in the adult of any representative 
of the last premolar, either deciduous or successional, we find 
the deciduous premolar to be constantly present in the young 
mammary foetus as a small and precociously calcified vestigial 
tooth (figs. 80 and 81). It is indeed considerably larger than 
the vestigial ‘‘ prelacteal”? teeth met with elsewhere, as in 
Perameles, or evenin Dasyurus itself (see di 3, fig. 23). 
But in all other essentials it agrees with the “ prelacteals,” 
e.g. in time of differentiation, relative situation, and in the 
fact of its absorption during foetal life. 

An essentially similar condition we find in Phascologale 
cristicauda, in which also no trace of any milk premolar 
appears in the adult, This form, however, differs from 


446 J. T. WILSON AND J. P. HILL. 


Dasyurus in the presence in the adult upper jaw (though 
only occasionally, according to Baldwin Spencer, 23, pp. 23 
and 26) of the successional last premolar as a “ minute and 
tubercular ” tooth (i.e. ‘‘p. 4,” cf. 22, p. 276). This tooth is, 
however, absent from the adult lower jaw. 

Thylacinus offers a still more advanced phase of develop- 
ment, for in it the successional premolar is now a well- 
developed tooth, larger than the second premolar (“p. 3” of 
Thomas), and is preceded by a milk premolar, though that is 
still only a minute and rudimentary one, and is shed during 
infancy (22, p. 255). 

In these respects Perameles may be regarded as standing 
just above Thylacinus, for while the last permanent pre- 
molar is well developed in both forms, its deciduous predecessor 
in Perameles is a well-formed though still relatively a small 
tooth, and is not extruded by the eruption of p. 3 (‘Thomas’s 
“py, 4”) until the animal has attained about three fourths 
adult size. The condition in Perameles may thus be re- 
garded as itself intermediate between that in Thylacinus 
and that in Didelphys, where the deciduous premolar is a 
very large and multicuspidate molariform tooth. 

It is beyond question that these several conditions repre- 
sent stages in a process of reduction affecting first the 
deciduous, and then the successional premolar, until the 
Dasyurinz condition is reached. 

Is it not in the highest degree probable that in this series 
we have to recognise, as it were going on under our eyes, the 
same process of reduction which, at an earlier epoch, has 
brought the other antemolar milk-teeth down to the condition 
of mere calcified vestigial structures—the so-called “ prelac- 
teal”? rudiments? In other words, the deciduous premolar is 
simply the hindmost member (so far as we know) of a dental 
series exhibiting various stages of a retrogressive process, some 
of the more anterior members of the series being represented 
by various ‘ prelacteal ” rudiments, 


DEVELOPMENT AND SUCCESSION OF TEETH IN PERAMELES, 447 


Tue Question or NoMENCLATURE OF THE PREMOLARS. 


After careful consideration we have not thought it well to 
adhere to Thomas’s (5) numerical determination of the teeth in 
the premolar region. That is doubtless an attractive theory, 
which aims at enabling us to homologise the premolars of 
modern Marsupials with those of the higher mammals; but 
we are as yet unconvinced that the case has been sufficiently 
made out. 

Admitting the probability that one of the premolar series 
has been lost in the ancestors of modern Marsupials, we 
cannot regard Thomas’s contention that it is p. 2 which has 
disappeared as placed beyond all reasonable doubt. Thomas’s 
case for the homology of the last premolar of modern Mar- 
supials to p. 4 of other mammals is made up of the following 
factors : 

(a) The existence of variations in the way of an occurrence 
of the hypothetically missing p. 2; to which may be added— 

(6) The alleged occurrence in ontogenetic development of 
a possible rudiment of an enamel-germ in the position of 
Thomas’s “ p. 2.” 

(c) The probable phylogenetic relation of Triconodon to 
modern Marsupials. 

The present writers are, however, of the opinion that there 
are too many uncertainties connected with each of these 
factors to warrant our basing any system of nomenclature 
upon a theory so conditioned. 

With regard to the first of these, Bateson (21) has well 
shown how indecisive is the evidence derived from a study of 
tooth-variations in determining individual homologies of teeth. 
In Bateson’s judgment ‘the system elaborated by Thomas 
breaks down ; not because there is any other system which can 
claim to supersede it, but because the phenomena of variation 
are not capable of this kind of treatment,” because “it is not 
possible to apply any scheme based on the conception that 
each tooth has an individual homology which is consistently 


44.8 J. T. WILSON AND J. P. HILL. 


respected in variation.”” It will be noted that these conten- 
tions do not necessarily affect our belief that the normal teeth 
are individually homologous in different forms; but only 
lead to the belief that the facts of variation do not yield reli- 
able evidence in favour of such an homology in any given case, 
If this be so, then the first and a very important support for 
Thomas’s theory gives way. 

Even less weight can be attached to the second considera- 
tion adduced in support of Thomas’stheory. Rose does indeed 
mention a possible rudiment of Thomas’s “ pm. 2” in Didel- 
phys. But for this Kikenthal sought in vain, and, as a 
result of a later and more extended research on Didelphys 
and other marsupial forms, Leche formulates his judgment 
upon the matter in the following passage: 

“‘TIn diesem Zusammenhange mochte ich ausdricklich her- 
vorheben, dass die ontogenetischen Untersuchungen bisher 
keinen Aufschluss tiber die Homologien der einzelnen Zahne 
der Beuteltiere und derjenigen der Placentalier gegeben habe. 
Auch die von Thomas versuchte Homologisirung der Pramo- 
laren der Beuteltiere gewinnt durch die ontogenetischen 
Befunde keine Stiitze ” (8, p. 107). 

Woodward, however, has recorded his discovery in Ma- 
cropus giganteus (2, p. 463) of “numerous small enlarge- 
ments and irregularities ” in the diastema between the canine 
and the “third” premolar of the upper jaw, “some of which 
may possibly represent the missing premolars.” This is vague 
enough. But he goes on to say, ‘In the lower jaw, however, 
there is a very distinct vestige of a tooth in the form of an 
irregular enamel-organ with enamel epithelium and pulp (fig. 
14). This, from its proximity to the third premolar, must re- 
present pm. 2.” Since, however, there is another typical 
marsupial premolar to be accounted for somewhere in this 
situation, it seems to us not a little rash to pronounce the 
enamel-organ in question—merely on account of its proximity 
to p. 83—to be “pm. 2,” a tooth admitted to be otherwise 
entirely absent from the marsupial dentition. 

Like Kiikenthal and Leche, we have sought in vain for any 


DEVELOPMENT AND SUCCESSION OF TEETH IN PERAMELES. 449 


confirmation (in Perameles) of the view that Thomas’s 
“pm. 2” is present in the developing marsupial jaw. 

We must therefore hold with Leche that embryological in- 
vestigation yields no support to the theory of the presence of 
this hypothetical element of the marsupial dentition; and 
though its absence may not be conclusive against that theory, 
it forms an additional difficulty in the way of entertaining a 
system of homologies based upon the ideal existence of such 
an element. : 

Only the third consideration quoted above in support of 
Thomas’s nomenclature remains for consideration, viz. that 
derived from the dental formula of Triconodon. This would 
certainly seem to render it highly probable that the ancestors 
of Marsupials were possessed of four premolars. Thomas’s 
argument rests upon the usually accepted close phyletic or 
even phylogenetic relationship between Triconodon and 
modern Marsupials, but it also requires the identification of 
the last premolar in the latter with “pm. 4” in the former. 
Now both of these propositions may doubtless be justified as 
highly probable, but neither, surely, can so far be regarded as 
scientifically certain. And, in particular, we hold that the 
evidence so far brought forward is insufficient to determine 
which premolar, if any, has disappeared in the course of 
evolution. 

On the whole we believe it to be safer in the meantime to 
designate the premolars simply numerically in the order of 
their occurrence in the modern marsupial type, instead of 
founding a system of nomenclature upon the condition of 
Mesozoic forms whose precise zoological relationships can only 
be inadequately determined. 

And, in adopting in this paper the older method of numera- 
tion, we do so the more readily that Thomas’s system has not 
as yet gained general acceptance among Continental writers. 


450 J. T. WILSON AND J. P. HILL. 


MatveritaAL AnD Meruops. 


For the purposes of our research we have had at our com- 
mand fourteen different stages of Perameles young. Among 
these stages P. obesula was represented by three specimens, 
while the remainder of the material was furnished by P. 
nasuta. 

The youngest was an intra-uterine specimen of P. obesula, 

rand from this early condition onwards we possess a pretty 
complete series of stages. 

The material representative of the more important stages 
has been sufficiently abundant to enable us to obtain complete 
series of sections in different planes. 

Decalcification was effected chiefly by means of nitric 
alcohol ; sections were cut in paraffin except in a few cases 
where celloidin was preferred. The stains used were Gre- 
nacher’s borax-carmine, hematoxylin, hematoxylin and eosin, 
hematoxylin and picric acid; the hematoxylin stains were 
chiefly used, and are greatly to be preferred, especially the 
double stains named. We find a very dilute solution (very 
faintly claret-coloured) of Renaut’s hzmatoxylic glycerine to 
give the best results. The stain should be light if to be fol- 
lowed by eosin, and deep (i.e. of considerably longer duration) 
if to be followed by picric acid. The latter (as also the eosin) 
is used in a weak solution in 90 per cent. alcohol. Sections 
were generally stained on the slide, being fixed by Mayer’s, 
or latterly by Maun’s albumen method. 


1 Owing to the enormous number of serial slides we had to work with, each 
slide containing very numerous sections, it was rather important to have a 
short and convenient method of designating any particular section in our 
notes and descriptions. We have found it exceedingly useful to employ a 

18 
~ 6-IN=E' 
number of the section in any given row, counting from the left-hand end. 
The Arabic numeral below the line refers to the number of the particular row 
of sections on the slide, counting from the top. The Roman numeral indi- 
cates the number of the slide in the series ; and the capital letter specifies 
the series itself. For ordinary work the last may often be omitted. The 


formula thus: Here the figure above the line indicates the 


DEVELOPMENT AND SUCCESSION OF TEETH IN PERAMELES, 401 


Part II.—Description oF Stages, 
Stage I.—Perameles obesula: intra-uterine embryo. 


Length from anterior convexity of head to curved hinder 
extremity of body ' : ; . 875 mm. 
Coronal and sagittal series examined. 


This stage is that in which an allantoic placenta has been 
described by one of us (H., 30). 

The margins of the gap are not yet soldered together to 
form the “ Saugmund.” 

As regards its tooth development, this stage appears to 
correspond pretty closely to that described and figured by 
Rose as his earliest stage in Didelphys (1, p. 641, fig. 1). 

The structure referred to by Rose as “ einer leicht ins Kie- 
fermesoderm eingesunkenen Epithelwucherung,”’ and labelled 
“7.1.” in his fig. 1, is present in our specimens in both upper 
and lower jaws. In the upper jaw a very shallow groove like 
that shown in Rose’s figure appears abruptly a little behind 
the anterior end of the jaw. ‘This faintly indents the surface 
of a slight thickened ingrowth of the oral epithelium into the 
mesoderm. Both groove and solid cellular ingrowth when 
traced backwards gradually fade away, the superficial groove 
disappearing sooner than thesolid proliferating band of epithelial 
cells. 

In coronal sections of the lower jaw a lens-shaped thicken- 
ing of the oral epithelium is first met with anteriorly. Soon 
a shallow groove appears over this, and the thickening then 
dips more deeply into the mesoderm, becoming almost tri- 
angular in cross-section. Further behind, the groove becomes 
description of a slide is thus easily condensed into a series of notes, each pre- 
faced by a formula expressive of the individual section to which the note 
refers. Frequently one wishes to refer to a number or groups of sections at 
one time. ‘This is obviously to be accomplished by extending the upper line 
18-26 
6—-lI-E 
sixth row of the third slide of series E is conveniently indicated. This 


method is specially useful for reference from drawings to the sections 
figured. 


VOL. 39, PART 4.—NEW SER. 11 


of the formula thus :— Here a group of eight sections in the 


4.52 J. T. WILSON AND J. P. HILL. 


slightly deeper and opener, the ingrowth at the same time 
becoming thicker. Finally the groove opens out and dis- 
appears, leaving only the thickened ingrowth, again lenticular 
on cross-section, which in turn disappears a few sections further 
back. 

Rose’s fig. 1 (1) sufficiently illustrates the condition found 
by us in this early stage. 


Stage IJ.—P. nasuta: mammary fetus. 


Length from vertex to root of tail . : . 17 mn. 
Head length . : ; ‘ Tor as 
Transverse (coronal) and sagittal series studied. 

The dental lamina is already pretty fully developed, though 
the differentiation from it of enamel-organs has not progressed 
very far, so that only a very few enamel-organs can be recog- 
nised as such, and there are relatively long stretches of quite 
undifferentiated lamina. The latter is relatively thick (fig. 
1, dl.), especially at its so-called “neck.” In other words, 
the groove in the mesoderm occupied by the elongated mass 
of ectodermal cells forming the lamina is a widely open one, 
its lips not yet being approximated so as to thin out the 
lamina towards its connection with the deep surface of the 
oral epithelium. 

Upper Jaw.—In the upper jaw the dental lamina extends 
anteriorly almost, but not quite, to the mesial plane: mesially 
it is separated by an interval from the lamina of the opposite 
side. In this anterior incisor region the structure is best 
studied in sagittal sections, for in the sagittal plane we obtain 
nearly true transverse sections of the lamina for some distance 
outwards. Fig. 6 shows a sagittal section parallel with the 
mesial plane which just shaves the margin of the orifice of 
Stenson’s duct (S.d.), and cuts the dental lamina (di/.) trans- 
versely to its axis some distance in front of this. The lamina 
is seen to be surrounded by young connective tissue, and the 
whole is capped by an imperfect dome-shaped roof of bone— 
the commencing alveolar portion of the premaxilla (pmz.). 
This dome-shaped roof in turn causes an upward inflection of 


DEVELOPMENT AND SUCCESSION OF TEETH IN PERAMELES. 453 


the cartilage of the nasal floor (a/.) in this its most anterior 
portion. The upward inflection of the cartilage of the nasal 
floor is visible also in coronal sections some distance in front 
of the naso-palatine foramen. In serial sagittal sections, as 
one traces them outwards from the mesial plane, the dental 
lamina soon after its commencement is seen to become dis- 
tinctly thickened and enlarged. This enlargement extends 
outwards for some distance, and occupies a considerable por- 
tion of the anterior transversely lying portion of the lamina. 
This we take to be the representative of the first upper incisor, 
although there is as yet no differentiation of an enamel-organ 
distinct from the dental lamina, and therefore the tooth, as 
a distinct organ, cannot yet be said to have come into ex- 
istence. 

(It may not be out of place here at the outset to remark 
that, where we have given interpretations of the facts and 
phenomena observed [as distinguished from simple records of 
observations], these interpretations are expressions of judgments 
arrived at after careful collation of the facts, not only in the 
stage under notice, but in those both preceding and succeeding. 
An explanation may thus occur which would not be warranted 
by the facts noted in the immediate context alone.) 

Laterally to the swelling mentioned the lamina is again 
shallower and less bulky for a short distance. Soon, however, 
it again swells out to form an enlargement of considerably less 
transverse extension than the first. This we believe to be the 
representative of the second incisor. It is developed about 
the point where the lamina begins to bend backwards in a 
sagittal direction, as a comparison of coronal and sagittal sec- 
tions indicates. Immediately external to it (or behind it) the 
lamina bends much more acutely, and assumes definitely the 
general sagittal direction which it maintains during the re- 
mainder of its course backwards. And just at this acute 
portion of its curve the lamina expands rather suddenly into a 
large swelling, oval on cross-section, which is evidently repre- 
sentative of the third incisor tooth. This swelling is of con- 
siderable interest, because in connection with it we have a 


454 J. T. WILSON AND J. P. HILL. 


most important and striking feature present. This consists in 
the differentiation, towards its anterior end, of a shallow but 
quite definite mesodermal papilla, which indents its labial side, 
forming the rudiment of the true milk predecessor of the per- 
manent third incisor, whose Anlage forms the bulk of the 
swelling. 

The series of drawings (figs. 1—5) completely illustrates the 
characters of this milk rudiment (di) and its relations to 
the swollen portion of the dental lamina representing the 
Anlage of the permanent tooth (¢3). 

Thus fig. 1 shows the lamina (dl.) in advance of the 73 
region destitute of any differentiation. In the next succeeding 
section, fig. 2, the labial outgrowth (d23) has appeared very 
abruptly growing out from the “ neck” of the lamina. In the 
following section (fig. 3) the milk enamel-organ (di3) is 
broader, and its relation to the lamina is slightly modified. 
The connective tissue on its labial aspect promises to form a 
papilla (mp.), and there is a slight tendency to cupping of the 
enamel-organ itself. As yet the main portion of the dental 
lamina exhibits very little modification, but from this point 
backwards it swells considerably, so that in fig. 4, three sec- 
tions further back, it is decidedly more massive, whilst at the 
same time its differentiation from the milk enamel-organ is 
hardly visible, being indicated only by the arrangement of the 
nuclei of the constituent cells. Still the milk papilla (mp. di 2) 
is distinct though very shallow. Three sections still further 
back (fig. 5) the milk enamel-organ has wholly disappeared, 
leaving, however, a much-thickened dental lamina, representing 
the hinder and main portion of the Anlage of the permanent 
third incisor (72). 

The precise position of the milk-tooth rudiment on the 
antero-lateral aspect of the entire third incisor Anlage is defi- 
nitely established, not merely by an examination of the serial 
sections figured, but by comparative study of sagittal series, 
which affords complete corroboration of the statements just 
made. 

It will be shown in connection with the description of the 


DEVELOPMENT AND SUCCESSION OF TEETH IN PERAMELES. 455 


next stage that the milk enamel-organ just described under- 
goes a rapid but most imperfect differentiation, and that in 
later stages (from Stage rv onwards) it has entirely disap- 
peared. At no time is there any formation of hard dental 
tissues in connection with it. Comparison of the present 
stage (11) with the next (111) proves that the labially placed 
papillary indentation now described has nothing to do with the 
formation of the papilla of the permanent third incisor. 

Behind the site of 22 the dental lamina is again much 
reduced in size for a few sections, but presently it enlarges 
slightly in the succeeding ones and again decreases. This 
almost certainly indicates the site of the future 74+. Behind 
this the lamina again enlarges, its cross-section becoming 
swollen and rounded, and after continuing thus through a 
number of sections it changes its shape, but without decreas- 
ing in sectional area, so as to form a figure elongated in cross- 
section. This gradually undergoes still further enlargement 
and elongation by a markedly deeper ingrowth of the lamina: 
this enlargement we have definitely ascertained to be the 
canine rudiment. Whether the shallower thickening of the 
lamina continuous with it in front has any significance as the 
Anlage of 72 we cannot be certain, though probably it may be 
thus interpreted. 

The shape of the canine rudiment is quite distinctive in 
early stages, for already the part of the dental lamina answer- 
ing to it manifests that tendency to disproportionately deep 
ingrowth into the surrounding connective tissue which is indi- 
cative of the future deep-seated position of the differentiated 
tooth. Its development may be followed with ease through 
all the stages from this onwards. 

But, as now differentiated, the markedly enlarged canine 
moiety of the dental lamina is not to be regarded as merely 
the rudiment of the permanent canine. Subsequent stages, 
especially that immediately succeeding, prove that from the 
present Anlage a rudimentary deciduous (milk) tooth (d*) is 
also formed. Indeed, the formation of a d* is already fore- 
shadowed by the existence of a labially directed outgrowth 


456 Jw. T. “WIESON- AND rd. “PL HERE. 


(vide fig. 7) from near the basal portion (i. e. the “ neck”) of 
the elongated laminar ingrowth. A comparison with fig. 13, 
which shows the exactly.corresponding structure in the case of 
the lower canine rudiment, will sufficiently demonstrate the 
first steps in that evolution which results in the interesting 
condition so beautifully shown in the next stage, where we 
have the well-formed miniature enamel-organ of a deciduous 
canine tooth above and below (figs. 20 and 37). 

On account of the deep extension of the canine Anlage it 
forms a prominent feature in sagittal sections which pass 
through its plane, all the more that a long extent of the 
dental lamina behind it, though quite continuous throughout, 
is relatively very shallow and insignificant. This latter por- 
tion of the dental lamina, representing the whole of the pre- 
molar region between the canine and the deciduous premolar, 
is at the present stage absolutely void of any special structural 
differentiatiou. It is, as already said, comparatively shallow 
(though a “neck”’ is slightly indicated), and extends back- 
wards through very many sections to the region of dp3. 

Here it again deepens considerably, its deeper part or fundus 
becomes swollen, and ere long it is seen to be deeply indented 
on its labial aspect by a large papilla, giving it the typical form 
of a cup-shaped enamel-organ. This is represented in fig. 8, 
in which it is to be noted that the process of formation of a 
cup-shaped enamel-organ has gone on without any trace, so 
far, of the emancipation of the developing organ from the 
parent lamina. An early stage of the latter process may, 
however, be recognised in fig. 9 (rdl.), which represents the 
corresponding organ in another embryo from the same pouch. 
(For the more advanced stage of the same process compare the 
next stage as illustrated in fig. 24.) 

Special attention may once more be called to the fact that 
although dp= in Perameles is never other than a small and 
insignificant tooth compared with the other functional members 
of the dental series, yet nevertheless at this early stage it is 
already large and well advanced in its development as an 
enamel-organ; and it is the only antemolar tooth-rudiment, 


DEVELOPMENT AND SUCCESSION OF TEETH IN PERAMELES. 457 


amongst those which subsequently attain maturity, which as 
yet exhibits the least trace of papillation. The precocious 
character of its development is well illustrated in fig. 10 
(dp*), showing in sagittal section its relation to the oral 
epithelium (0.e.) and to the molar portion of the dental lamina 
(dl.) behind it, which is cut vertically in the sagittal plane of 
the section. 

The molar portion of the dental lamina is massive and well 
developed for quite a considerable distance behind dp3, as 
may be seen in fig. 10. At first it is merely flask-shaped on 
transverse section, but soon its cross-section alters in shape 
by its distal margin bulging out obliquely in a labial direction. 

This constitutes the Anlage of m+, and is seen in fig. 1] to 
exhibit a very slight depression at its fundus. . . . Opposite 
this depression the connective-tissue cells are aggregated to 
form the very earliest Anlage of a dermal papilla (mp.m+). 
Both in front and behind, the swollen Anlage gradually 
diminishes in thickness. It is worthy of special remark that 
here we have the very youngest stage in the differentiation of 
a separate molar enamel-organ, and that in this there is not 
the slightest indication of a composite character of the dermal 
papilla whose earliest stage of evolution is here exemplified. 
It is, from the first, single and individual. In the next 
succeeding stage we shall find that this incipient enamel-organ 
has rapidly attained a typical cup-shaped or even “ bell- 
shaped” (‘‘glockenformig”’) character, and has “caught up 
with” dp2 in its progress towards full differentiation. 

Behind the Anlage of m+ the molar lamina gradually 
diminishes in size, and finally comes to a rather abrupt termi- 
nation in a somewhat truncated hinder extremity. 

It will be evident from fig. 10 that at least the major portion 
of the dental lamina in the molar region is directly continuous 
with the deep surface of the oral epithelium. If the third suc- 
ceeding section of this series were also figured, it would be seen 
that this continuity of the lamina with the oral epithelium 
extends right up to the very hindmost end of the molar lamina. 
It is thus evident that the dental lamina in the molar region 


458 J. T. WILSON AND J. P. HIM. 


at the present stage has arisen by direct proliferation and in- 
growth of the cells of the Malpighian layer of the oral epithe- 
lium, as is the case in front of the molar region. 

An examination of later stages proves that the line of conti- 
nuity between the oral epithelium and the molar lamina grows 
backwards, for some time, pari passu with that rearward 
progress of the lamina itself which accompanies the gradual 
elongation of the jaw. 

At the present stage the entire dental lamina has retained 
its connection with the oral epithelium. 

In a very young mammary foetus of Dasyurus the molar 
dental lamina does not end abruptly, but, after somewhat 
suddenly becoming shallower, undergoes a further very gradual 
diminution in depth, until it fades away into a thickening 
of epithelium, lens-shaped in cross-section, like that which 
has been described by Rose as constituting the earliest trace 
of the dental lamina in the anterior region of the jaw. 

Lower Jaw.—The armature of epithelium on the oral 
surface of the lower jaw near its tip is only moderately thick, 
and at first it thins slightly as it is traced backwards; then 
suddenly on each side a localised thickening of epithelial cells 
appears, which invades the underlying connective tissue, and 
suggests the character of an extremely thick, broad, and some- 
what shallow dental lamina. In reality, however, the definitive 
dental lamina appears abruptly a short distance behind as a 
further downgrowth from the thick cell-mass in question. 
(The precise significance of the latter is doubtful. As it is 
traced further back it becomes largely continuous with the 
corresponding mass of the opposite side. Further back still, 
the fused structure is seen to be absolutely continuous with 
the epithelial cell-mass forming the tip of the tongue, which 
is thus firmly glued to the upper surface of the lower jaw. As 
the tongue is traced backwards it gradually frees itself from 
the latter, and the thickening of the jaw epithelium is then 
seen to have largely disappeared.) The dental lamina itself, 
springing abruptly as aforesaid from the deep surface of the 
epithelial armature of the jaw, becomes almost at once some- 


DEVELOPMENT AND SUCCESSION OF TEETH IN PERAMELES. 459 


what enlarged, and forms a swelling whose outer aspect 1s 
somewhat flattened, while its fundus is more rounded and is 
somewhat mesially inflected. The outer, more flattened por- 
tion now comes to project labially, and is seen to consist 
mainly of cells, the chromatin of whose nuclei is evidently in 
the active phase. This feature causes this portion of the mass 
to appear pretty definitely differentiated with a haematoxylin 
stain. Some sections further back the labial projection of the 
lamina is deeply indented by a dermal papilla (fig. 12), so as 
to form a labially directed enamel-organ (d7;) in the cup- 
shaped stage. This is as yet incompletely differentiated from 
the common mass of the swollen dental lamina. The entire 
enlargement of the lamina in this situation corresponds to the 
region of the first incisor, and the labially cupped enamel- 
organ, in process of differentiation, is the Anlage of the first 
milk incisor (diz). This rudimentary tooth is shown in 
subsequent stages to undergo a precocious though imperfect 
development into a minute calcified tooth. 

Opposite the papilla of di; the main mass of the dental 
lamina has become considerably enlarged, and the increase in 
bulk continues behind the rudiment of the milk-tooth. Part 
of this posterior thickening doubtless represents the hinder part 
of 7, but the more posterior portion must provide the material 
for the development of tz. This is rendered certain by com- 
parison with the succeeding stage (111), for in both stages the 
dental lamina makes an abrupt turn outwards and runs trans- 
versely for a very short distance, instead of antero-posteriorly. 
Now in Stage 11 it becomes certain that 7, is developed at 
the postero-external limit of this transversal part of the lamina, 
and 7; at its antero-internal extremity. Further, in the 
present stage the short transversely running part of the 
lamina commences immediately posterior to the thick part of 
the lamina which lies behind di; ; whilst at the outer end of the 
same short transverse part of the lamina appears a new 
thickening. This, as we shall see, is the only other swelling 
in front of the canine, and it corresponds exactly in position 
to ig in Stage 111. 


460 J. T. WILSON AND J. P. HILL. 


The posterior limit of this rudiment of 7; may be recognised 
by a marked diminution in the sectional area of the lamina, 
~ which continues back quite small until it again expands into 
the typically elongated canine Anlage. The lamina in this 
region is relatively much elongated on transverse section. It 
does not, however, penetrate the deeper tissues in a vertical 
direction, but is diverted rather mesially, parallel with, and 
pretty closely adjacent to, the deep surface of the oral 
epithelium. In it a very interesting step in differentiation 
may be observed in the shape of avery slight but unmistakable 
attempt at the formation of the enamel-organ of the milk 
canine (fig. 18, d.;). This is just on the point of becoming 
cup-shaped, and the dermal papilla (mp.d;) is beginning to 
form. In the succeeding stages this Anlage -will be followed 
through the course of its evolution. Here, as in the other 
examples of milk Anlagen, the structure is first developed not 
only on the labial aspect of the common canine enlargement 
of the dental lamina, but opposite its anterior part, the 
swelling attaining its maximum development just behind the 
milk rudiment. 

As is the case in the upper jaw, the premolar portion of the 
lamina behind the canine Anlage is small and presents no 
local differentiation until we reach the region of the last 
premolar (dp;). Here the dental lamina becomes much 
enlarged in all its dimensions, and the bulbous mass becomes 
indented labially by the large dermal papilla of d.p.z (the 
“milk molar”). Its characters are so closely similar to 
those of the corresponding organ in the upper jaw that the 
figures of the latter may suffice for both (figs. 8 and 9). 
Fig. 14, however, illustrates the appearance seen in sagittal 
section. 

The molar region of the dental lamina is in a precisely 
similar condition to that exhibited in the upper jaw. There is 
only a slight indication of the differentiation of the first molar 
enamel-organ. For the most part the molar lamina forms 
simply a slab-like lamina of considerable dimensions, extend- 
ing backwards, and continuous in front with the milk premolar. 


DEVELOPMENT AND SUCCESSION OF TEETH IN PERAMELES. 461 


It is throughout in direct connection with the deep surface of 
the oral epithelium, as serial sections, sagittal and transverse, 
show most definitely. But as in the upper jaw, so also here, 
there is found at a short interval behind dp.z an outward pro- 
jection or bulging of the labial aspect of the lamina entirely 
comparable to that which in the upper jaw serves to broaden 
the lamina transversely. Here, however, it is placed much 
nearer to the “ neck ” of the lamina, and throws a very slightly 
cupped surface, with the rudimentary papilla in relation to 
it (mp.m;z), more distinctly on to the labial aspect of the 
Anlage. Fig. 15 illustrates the condition described, and may 
be advantageously compared with the figure of the lower 
canine Anlage (fig. 13). 

Behind the Anlage of m; the labial projection disappears, 
and the lamina is left as a simple but fairly thick band lying 
parallel with the oral epithelium, to which it is attached by a 
“neck” curved almost at right angles with the dorso-ventral 
axis of the lamina. As in the upper jaw, this disappears alto- 
gether with considerable abruptness. 

Our readers will thus observe that in following the develop- 
ment of the dental lamina in Stage 11 we have been able to 
trace what to all appearances are the earliest steps in the 
direction of individual tooth-differentiation. During this 
period the dental lamina still includes the Anlagen of all the 
future teeth; none of these have as yet been emancipated 
from it, and the great majority of the Anlagen have as yet 
little or no claim to be considered as individualised tooth- 
germs. 

Regarding the morphological interpretation of our observa- 
tions, especially in the way of the determination of tooth 
Anlagen, we believe that in the case of a number of these 
there can be difference of opinion. It has already been indi- 
cated that the justification of our interpretation in certain 
other cases depends to some extent upon our observations in 
subsequent stages of development. A fuller discussion of 
these more disputable points must therefore be reserved till 
later. 


462 J. T. WILSON AND J. -P.- HILL. 


Stage 11].—P. obesula: mammary feetus. 


Length from vertex to root of tail . . . 22 mm. 
Head length : : ‘ . ~ oe 
Studied in coronal and horizontal series. 


The examination of this stage reveals features of uncommon 
interest and importance. 

Although our specimen representing this stage is not speci- 
fically identical with those of the stages immediately pre- 
ceding and succeeding, we are confident that this difference is 
of no consequence, and does not in the least prejudicially 
affect the conclusions arrived at. So far as it is possible to 
ascertain and to judge, the condition presented by the stage 
answers exactly to what one would expect to find, intermediate 
between Stages 11 and tv. The correspondences, indeed, are 
throughout unmistakable. 

Here for the first time the great majority of the permanent 
teeth are sufficiently differentiated to be recognisable as dis- 
tinct and individual enamel-organs. Most of these, however, 
as we shall see, are in a very early stage of their individual 
differentiation, and are not yet emancipated from the dental 
lamina. The latter structure is relatively massive and abso- 
lutely continuous throughout. Its line of attachment to the 
oral epithelium (proximal margin or “ neck ”) is, however, in 
parts considerably attenuated; and in a few places, for a 
section or so, it is just possible to recognise a discontinuity 
between these structures. This is doubtless due to the com- 
mencement of the process of resorption which Rose has shown 
(9) to produce fenestration of the lamina. 

With these extremely rare and slight interruptions the at- 
tachment to the Malpighian layer extends up to the posterior 
end of the molar part of the lamina. 

Upper Jaw.—In the incisor region the extreme anterior 
end of the dental lamina (fig. 16, d/.) is directed somewhat 
mesially, but it no longer lies so distinctly in a transverse 
direction as in the preceding stage. * 

The Anlage of the first permanent incisor (fig. 16) is indi- 


DEVELOPMENT AND SUCCESSION OF TEETH IN PERAMELES, 463 


cated close to the anterior extremity of the lamina by the 
presence of a subglobular outgrowth (i+), bulging in an an- 
tero-lateral direction, i.e. at right angles with the axis of the 
lamina (d/’.)in this situation. The lamina itself is here plump 
and swollen, so that on section the two together form a bilobed 
mass. The labial lobe, however (“7+ ”’), is merely a localised 
swelling, while the lingual (dl.), becoming more attenuated 
behind, is continued for some distance onwards to connect 
with the second incisor region. 

It will be remembered that, in the preceding Stage 11, i+ 
was represented by a mere uniform enlargement of the dental 
lamina. The outgrowth of the labial mass we have now de- 
scribed constitutes therefore the next step in the differentia- 
tion of the enamel-organ of the first incisor. And the growth 
of this labially projecting mass involves the development, 
between it and the main body of the lamina, of a depression. 
This is seen both in front, behind, and especially above, where 
it is somewhat more pronounced. The surrounding capsule of 
connective tissue, where that is in relation with the depression 
in question, exhibits distinct though slight evidences of cel- 
lular differentiation. Fig. 16 shows a horizontal section of 
the first incisor region with the sulcus extending on to the 
anterior face of the Anlage. Here the connective tissue is 
less fibrous, and its cells are more closely aggregated to- 
gether to form what is in all probability the Anlage of the 
dermal papilla. The latter indeed projects slightly into the 
sulcus. 

The formation of the labial lobe above referred to, and the 
associated production of a depression between it and the main 
portion of the dental lamina, are of great importance. We 
find that when the earliest stage of the process of cupping of 
an enamel-organ can be traced, that process is commonly, if 
not invariably, initiated by an overgrowth of the cells of the 
dental lamina so as to form a projection on the labial aspect ; 
and that the future cup-like depression into which the dermal 
papilla subsequently fits is foreshadowed by such a sulcus as 
we have above described. And when the cupped enamel- 


4.64 J. LT. WILSON AND <3. OP. ATE 


organ is at length evolved by this process, the outer part 
of the rim of the cup is constituted by the labially 
outgrowing lobe, and the inner part of the rim of the 
cup is formed by the main body of the dental lamina 
itself. It will be observed that this explanation is essen- 
tially in agreement with the idea expressed in Rése’s simile 
regarding the relation of the cupped enamel-organs to the 
dental lamina, when he says they are arranged “like swallow- 
nests on a board.” Our present stage illustrates the condition 
found just prior to the actual ‘“swallow-nest” stage of 
differentiation of the enamel-organs. At this stage, indeed, 
“sulcus” is the term which more adequately expresses the 
character of the depression corresponding to the future cup of 
the enamel-organ. It is open both fore and aft, because it 
owes its first formation to a mere labial outgrowth of some- 
what inchoate character. It may be noted further that this 
sulcus may be very shallow, or even absent, if the outgrowth 
of cells takes place near to the free margin of the lamina and 
extends simply outwards. Thus in the figure given (fig. 11) 
of the first upper molar Anlage in Stage 11 the labial promi- 
nence is so wide, and extends so near to the free margin of the 
lamina, that the intervening sulcus is broad and extremely 
shallow. The cellular condensation indicative of the future 
dermal papilla is, however, unmistakable. So also, with some 
difference, in the Anlage of m4, Stage 11 (fig. 15). 

The definitive cupping of such a bilobed enamel-organ of 
course involves the subsequent appearance of a margin or rim 
to the cup both in front and behind, in consequence of the 
more pronounced local limitation of the processes of enamel- 
organ formation,—processes which at first operate along a 
stretch of dental lamina indefinitely limited fore and aft. 

As may be gathered from Rése’s ‘ swallow-nest ” compa- 
rison, the lingual wall of the cupped enamel-organ is formed at 
first simply by the dental lamina,—the “ board ” against which 
the “‘swallow-nests ” are placed. But by-and-by, as we shall see 
hereafter, this dental lamina undergoes further differentiation, 
whereby a proper and distinct inner wall is contributed to 


DEVELOPMENT AND SUCCESSION OF TEETH IN PERAMELES. 465 


complete the enamel cup, leaving a more or less liberated dental 
lamina independently of the enamel-organ, and situated at its 
inner or lingual side. This we term the “residual dental 
lamina.” 

After this somewhat necessary digression—partly antici- 
pative of what we shall have to point out in detail later on— 
we may return to the description of the lamina in Stage 111. 

Behind the posterior limit of the Anlage of i+ the dental 
lamina becomes itself progressively reduced in sectional area 
until it approaches the region of the second incisor, its axis 
being directed somewhat outwards as well as backwards to 
reach the latter. 

Here it quickly increases both in height and in thickness, 
and is seen to form the lingual portion (d/2) of another 
bilobed mass (fig. 17), comparable to that in the region of 7+. 
To this bilobed swelling a similar interpretation must be given 
as that employed in the case of the first; i.e. its lingual por- 
tion is simply the enlarged dental lamina (d/?), while the 
labial portion of the main mass projecting outwards (“72”) 
is the outer or labial portion of the enamel-organ of 12. Here 

_again the lobes are separated by a groove (fig. 17) passing 
over the summit of the Anlage, and reaching its anterior and 
posterior aspects. 

In fig. 18 “II” represents a horizontal section of the whole 
epithelial mass in the region of the second incisor. Here dl? 
indicates the lingual lobe of the mass corresponding to the 
swollen continuation of the dental lamina itself. This hori- 
zontal section is taken above the level which would show the 
continuity anteriorly of dl2 with dil+ in fig. 16, aud about 
the level indicated by the line ad. in fig. 17, which shows a 
coronal section through the same region. (It will be remem- 
bered that between the developing teeth the dental lamina is 
considerably lower than opposite them.) In this figure of a 
horizontal section there is seen to project from the labial 
aspect of i2 a small epithelial mass (di2), which we believe to 
be the representative of the enamel-organ of the second milk 
incisor. The same projection is visible in fig. 17, showing a 


4.66 J. T. WILSON AND J. P. HILL. 


transverse (coronal) section of the same region.! The back- 
ward recurvature of this small labially projecting cell-mass, 
seen in fig. 18, must not be mistaken for an indication of com- 
mencing papillation from behind forwards. The series of sec- 
tions proves that the projection in question is simply being 
gradually pinched off from the main body (22). Higher up, 
the horizontal sections show only the two lobes of the Anlage 
of i2 (d/2 and ‘‘22”’), It will presently appear that all these 
features are still better marked in connection with the region 
of the third incisor, which is partly illustrated in the same 
drawing (fig. 18, “ III’). 

Horizontal sections taken at a lower level than fig. 18 show 
the dental lamina (d/2) in the region of the second incisor to 
be prolonged backwards into continuity with d/3 in the third 
incisor region. 

Horizontal sections passing through the upper part of the 
third incisor-mass closely resemble those through the upper 
part of the second—viz. in each case we have lingually the 
lobe formed by swollen dental lamina, and labially the lobe 
formed by the differentiating labial moiety of the incisor 
Anlage. As in the cases of the two more anterior tooth- 
germs, the two lobes are separated by a slight groove or 
sulcus, and, corresponding to this, over the summit of the 
Anlage the connective tissue shows a slight cellular aggrega- 
tion indicating the commencement of the dermal papilla. 

Horizonal sections taken a little lower down begin to show 
a narrow labial projection in a position corresponding to that 
of di2 in the second incisor region. Still lower there suddenly 
appears, closely external to the last-named outgrowth, an 
isolated mass of epithelial cells. In sections taken still more 
inferiorly, as shown in fig. 18, “III,” this cell-mass (d73) is 
found to be really connected with the labial aspect of the main 
mass of 72 by a narrow neck (cb.). This labial cellular out- 


1 The single coronal section here figured might seem to suggest that the 
di% projection simply forms the lower limit of a papillated depression on the 
labial aspect. This is a merely accidental resemblance, as a study of the 
neighbouring sections in the series amply proves. 


DEVELOPMENT AND SUCCESSION OF TEETH IN PERAMELES. 467 


growth is further seen to be the seat of retrogressive processes, 
for in neighbouring sections it shows invasion by connective 
tissue, partially breaking it up into smaller cell-groups. 

This cell-mass is incontrovertibly the serial homologue to 
the projection figured under “ di2” in the region of the second 
incisor in figs. 17 and 18, though plainly it is in a more ad- 
vanced phase of evolution, as its greater isolation from the 
main mass and condition of degeneracy testify. We are 
satisfied that this cell-mass can be regarded as none other 
than the degenerating remains of that very early, labially 
placed enamel-organ, which we have described and figured 
above (p. 454, and figs. 2, 3, and 4) in connection with a pro- 
minent swelling of the incisor region of the dental lamina in 
the preceding Stage u. It will be remembered that the 
localised thickening of the dental Jamina was set down as the 
Anlage of 78, and the labially developing enamel-organ, im- 
perfectly approximating to the cupped stage, was regarded as 
that of di3. The judgment there expressed on the strength 
of our observations upon the differentiation of the dental 
lamina in that stage, is now seen to be borne out by the degree 
of development subsequently attained by the presumably ho- 
mologous structures. 

It is interesting to note that in the present stage, as in the 
preceding, the third incisor Anlage as a whole is still the most 
bulky of the incisor series, although the disproportion between 
it and the other incisor Anlagen is now greatly diminished. 

It must be noted in regard to fig. 18 that, in a longitudinal 
section passing so near to the base of the dental lamina in the 
73 region as that figured, it is only the basal portion of the 
Anlage of 73, and the vestigial remains of d2, which are 
recognisable. Itis only at higher planes that the differentiation 
of the Anlage of i2 into labial and lingual lobes is as recog- 
nisable as it is, e.g. in the case of 72 in the same section. 
Lower down in the case of the latter region, indeed, the labial 
portion of the Anlage of 72 similarly disappears from the 
section as a distinct lobe, leaving only the common lamina 
plus the labial projection of di2. 

VoL. 89, PART 4, 


NEW S5ER. K K 


468 Jj. La WALSON AND J. P: HILL; 


The structure of the differentiating dental lamina in the 
third incisor region will be rendered further intelligible by a 
reference to fig. 19, which illustrates a coronal section through 
this region. Here again the bilobed character is only ob- 
scurely manifested, but in the next succeeding section it is as 
plainly seen as in 72 in fig. 17, and the cellular aggregation 
at the site of the future papilla is also clearly seen. The line 
ab. shows the approximate level of the horizontal section in 
fig. 18. Here the almost isolated remains of the enamel-organ 
of di3 are well shown. 

Tn this figure (19) it will be observed that the attacliment of 
the whole mass to the oral epithelium is already narrowed. 
The deep notch between oral epithelium and dental lamina on 
the lingual side represents the result of the process of pinching 
off from the oral epithelium which is now in progress. Origi- 
nally the attachment no doubt corresponded to the whole 
breadth of the base of the mass lingually. As a result of the 
pinching off, the primitive continuity of the dental lamina on 
its lingual side with the oral epithelium is now obscured. 

The plane of the horizontal section through the region of 73 
(fig. 18) is such that although it does not shave the upper free 
border of the low dental lamina between the low-lying 72 and 
72, it does cut the lamina between 72 and i4, obliquely for a 
certain distance (fig. 18 d/4). The fourth incisor Anlage 
itself, however, lies wholly at a higher level. 

It is unnecessary to enter upon a detailed description of the 
Anlagen of the fourth and fifth incisors. The following state- 
meuts will suffice. The dental lamina is continuous throughout, 
both in its own extent and with the oral epithelium. It has 
undergone enlargement at two places corresponding respec- 
tively to the fourth and fifth incisors. These localised 
enlargements are for the most part of a general character, 
though bulging mainly to the labial side, except that in each 
case there is already to be found, projecting labially from near 
the base of each rounded swelling, a somewhat attenuated 
epithelial process, precisely comparable in general appearance, 
structure, and relations to that marked ‘‘ di? ” in figs. 17 and 


DEVELOPMENT AND SUCCESSION OF TEETH IN PERAMELES. 469 


18. And as in the latter case, so here, we take these rudi- 
mentary labial outgrowths of epithelial cells to represent the 
abortive enamel-germs of true milk incisors, di4 and di. 

It is, however, when we come to deal with the canine rudi- 
ment in this stage that we really obtain the most emphatic 
testimony to the validity of the interpretation which we have 
given of these interesting vestiges. 

From the fifth incisor the dental lamina, though small, is 
continued back without interruption, and very soon it exhibits 
that great vertical enlargement which we find to be character- 
istic of even the earliest canine rudiments in both upper and 
lower jaws. 

In fig. 21 we show a horizontal section which in front 
passes through the upper part of the fifth incisor mass (28). 
This is seen to bulge out labially (but this is quite above the 
level of rudimentary d=). From the inner or lingual part of 
the Anlage of 22 the dental lamina is continued backwards for 
a short distance, to lose itself, in this sectional plane, in a 
broad area of epithelium (0.e.). This, as may be learnt from 
the drawing, is simply the deep layer of buccal epithelium 
sliced parallel with its free surface. A section or two higher 
this epithelial patch disappears, giving place to the prolonga- 
tion backwards of the narrow dental lamina. Still higher in 
the series the lamina is seen to undergo enlargement so as to 
form the most inferior portion of the canine swelling of the 
lamina (ef. fig. 20). This enlargement markedly increases in 
thickness as it is followed upwards through the horizontal 
series. But although the lamina is so definitely enlarged, the 
differentiation of the permanent tooth has hardly yet set in; 
certainly no distinct outgrowing labial lobe is apparent, like 
that observable in the anterior incisor germs. Itseems highly 
probable, however, that the upper and labially deflected portion 
(marked “c” in fig. 20) of the whole Anlage is really 
equivalent to the labial lobe of the bilobed incisor masses, and 
is thus about to form the labial portion of the cupped enamel- 
organ of the permanent canine. If this be so, then the 
proper morphologically distal margin of the dental lamina is 


4.70 J. T. WILSON AND J. P. HILL. 


formed by the elbow-lke projection marked “c.dl.”? in the 
figure, and the flattened surface between c.d/. and the apical 
point of the whole enlargement would represent the site of the 
future enamel cup. However this may be, the one specially 
striking and important feature of this general canine Anlage 
is the presence, near its “ neck,” of a very perfect miniature 
enamel-organ in the cup-shaped stage. ‘This structure 
projects directly outwards in a labial direction from the 
proximal portion of the dental lamina, and in it we 
recognise, in accordance with the views we have set forth 
in the Introduction, the true rudimentary “milk” canine 
(fig. 20, d®). 

In fig. 22 we have shown another horizontal section through 
the upper jaw of this stage, through the same region as, but at 
a higher level than, the section illustrated in fig. 21. In this 
section we remark the canine thickening of the dental lamina 
¢, and its more attenuated prolongation backwards into the 
premolar region (p+di.). Beside the thickened canine portion, 
the enamel organ of d£is seen. Further important reference 
will be made to this figure later on. In the meantime we 
would draw attention to the fact that the minute, and up 
to this stage perfect, enamel-organ described above 
is entirely homologous in position and relations 
with those small lobular outgrowths of the epi- 
thelium of the dental lamina situated labially from 
the Anlage of the four hindmost incisors, which we 
have interpreted as representatives of enamel- 
organs of deciduous incisors. And it is not mate- 
rially diverse from these in regard to its chrono- 
logical position. 

Now if the serial homology with the incisor vestiges be 
granted—and we think it is impossible to avoid such a con- 
cession—this at once establishes the title of these vestiges to 
be regarded as true representatives of the enamel-organs of a 
tooth-generation preceding the permanent teeth. And in our 
opinion this generation is homologous to the milk series of 
Eutheria. 


DEVELOPMENT AND SUCCESSION OF 1 EETH IN PERAMELES. 471 


So far as we can ascertain, there is not to be discovered 
either in the present or in the preceding stages any such dif- 
ferentiation in the first incisor region as might be taken to 
represent there the rudiment of a true milk-tooth, i.e. one 
which belongs to a tooth generation preceding that to which 
the first permanent incisor belongs. Nevertheless there is 
some reason for the belief that traces of an attempt at the pro- 
duction of such a tooth do appear at some time during the 
development of the teeth in Perameles. And if any such 
rudiment does occur, then the cell material of which it consists 
must be derived by outgrowth from the labial side of the base 
of the mass we have described as constituting the Anlage of 
the first incisor. It has been stated above that in its size and 
degree of differentiation (including that of a “milk” rudi- 
ment) the third incisor Anlage is the most developed member 
of the upper incisor series at the present stage. The second 
is less developed, but yet exhibits features similar to those in 
the third. So also in the case of the fourth and fifth incisor 
Anlagen. Although, therefore, no such labial remains occur 
in connection with the first incisor Anlage in Stage 1ir as we 
have been able to show associated with the incisors behind it, 
and more especially in the canine region, we consider ourselves 
justified in the belief that certain cellular processes situated 
on the labial side of the root of the dental lamina in the first 
incisor region of the next stages (1v and v) may possibly 
represent the remains of the missing rudiment of a dit. In 
the present stage we may suppose that the latter is not struc- 
turally defined from the basal portion of the common first in- 
cisor Anlage. 

In further corroboration of our views upon the incisor 
development in Perameles it may not be inappropriate to 
refer to the condition we have found in another polyprotodont 
Marsupial, Dasyurus viverrinus. Among our series of 
stages of development of this form we possess one which, in 
respect of the third incisor, exhibits a general stage of de- 
velopment intermediate between the present stage (111) and 
the following stage (1v) of Perameles. 


472 J. 2. WILSON’ AND? 7, °P. Bibi: 


As in Perameles, but in much more marked degree, the 
third is the largest and most advanced in its development 
among the upper incisors. Here (fig. 23) the definitive cup- 
ping of the enamel-organ of 2%, has at last been effected, and 
a well-marked dermal papilla (mp. i*) formed, while at the 
same time the main body of the dental lamina has undergone 
such differentiation that the mass of cells forming the lingual 
portion of the cupped enamel-organ is now becoming distinct 
from the rest of the dental lamina. The latter is, in fact, in 
process of liberation as a “residual”? dental lamina (rdi.). 
(In an antecedent stage the Anlage of 73 is found to be differ- 
entiated to exactly the same extent as the corresponding 
Anlage in Stage 111 of Perameles.) A remarkable feature 
in this section is the rudimentary di3. Whatever may be 
said in detraction of the claims of the structure we have 
described in connection with 72 in Perameles, to figure as a 
vestigial enamel-organ, it is impossible to deny that here in 
Dasyurus we have an indubitable vestigial predecessor to 
28, It seems to us impossible to gainsay the homology 
testified to by this comparison between Perameles and 
Dasyurus. And it is superfluous to point out that an ad- 
mission as regards the identity of the structure labelled di2 
in Perameles with the vestigial tooth in Dasyurus involves 
the admission of a similar homology for the structures in 
Perameles in series with “di3.” Here in Dasyurus 
(fig. 23) di= is not simply a degenerate cell-mass, but has 
assumed specific dental characters by the precocious evolution 
of a small but thick dentine cap, which again is covered bya 
hood of regularly arranged enamel-cells. Instead of forming 
a projection from the rest of the incisor mass, the vestigial 
tooth rather lies embedded in the labial aspect of the latter 
near its base. Perhaps this may be explained by a reference 
to the great labial outgrowth of the Anlage of the permanent 
incisor which must have occurred to bring about the condition 
figured, together with the later stationary and unprogressive 
character of the calcified vestigial di%. 

The examination of an earlier stage of Dasyurus has 


DEVELOPMENT AND SUCCESSION OF TEETH IN PERAMELES. 4738 


shown us the third milk incisor at a stage prior to its calcifica- 
tion as a deeply cupped enamel-organ structurally comparable 
to the enamel-organ of d‘, figured by us in Perameles 
(fig. 20). Thus we have shown that, in Dasyurus, the rudi- 
mentary enamel-organ of di2, does not undergo the rapid 
degeneration which is exhibited by di3, in Stage 11 of 
Perameles, but passes into the regular cup-shaped stage which 
we have seen in the upper jaw of Perameles to be attained 
only by d3. 

A thoughtful comparison of the following series of figures 
will render it easy to understand the derivation of the condi- 
tions found in the third incisor region of Dasyurus from the 
earlier condition illustrated in Stage 11 of Perameles: 


Cf. Perameles, Stage II, 73 and di in figs. 3 and 4, 


" epee LUN) Lares in figs. 17 and 18. 
9 oo sae eo in figs. 18 and 19. 
” ” » +(C and ds in fig. 20. 


In fig. 21 it is seen that the dental lamina is continued 
backwards, diminished but unbroken, from the region of the 
canine (marked “0. e.”) to a point where it somewhat suddenly 
expands to form the Anlage of the first premolar (p+). It 
will be evident from the figure that, though this enlargement 
appears as a general expansion of the lamina, the addition or 
increase of cell material is towards the labial side, so that the 
bulging is in that direction. On coronal section, indeed, one 
obtains an appearance not far removed from that exhibited in 
this same stage by 74, although it can hardly be said that a 
definite furrow has appeared between the labial bulging and 
the main body of the dental lamina. Both in figs. 21 and 22 
the stretch of dental lamina extending from the canine to the 
first premolar is recognisable. 

In the first-named figure, however, the section passes along 
the more constricted basal portion or ‘ neck” of the lamina. 
In the second figure the section passes through the lamina 
near its free margin or “ fundus,” where that reaches a high 
vertical extension in the neighbourhood of the canine. Further 
back the lamina disappears from the section by becoming ver- 


A474 J. 1. WILSON: AND. J. P, HELL. 


tically shortened, the horizontal section thus cutting it obliquely. 
The hinder part of this obliquely cut portion of the lamina, as 
shown in fig. 22, does to some extent coincide with the anterior 
part of the first premolar swelling in fig. 21. It constitutes, 
indeed, as coronal sections prove, the lingual portion of the 
latter mass. The labial bulging of the Anlage is lower down, 
and is not cut through at this level (cf. fig. 21). 

The interval between the tapering ends of the Anlagen of 
the first and second premolars is a comparatively short one. 
Here the dental lamina, although quite continuous, appears 
low and relatively insignificant.1_ This is especially interesting, 
as it is just here that one might expect some trace of Thomas’s 
hypothetical ‘ pm2.” 

A glance at fig. 21 is sufficient to show how far forward the 
anterior tapering end of the second premolar enlargement of 
the lamina extends. This enlargement is, indeed, of very 
considerable antero-posterior extent. Anteriorly its cross- 
section is rounded, but posteriorly it is much more vertically 
extended, gradually narrowing at the same time, to be con- 
tinued on as the vertically elongated but otherwise unmodified 
lamina, which is found in the region between the second and 
third premolars (cf. text Fig. 1, a—e). 

Fig. 22, representing a horizontal section taken at a higher 
level than fig. 21, shows the higher hindmost portion of p2 
cut somewhat obliquely, and gradually passing back into con- 
tinuity with the freed dental lamina on the lingual side of the 
well-developed milk premolar, dp. 

On the labial side of the basal margin or “neck” of the 


1 The lamina looks lower than it really is because it is somewhat folded in 
its length, so that in cross-section the long axis (being attached to free 
margin) is not straight, but crooked. Such bending of the lamina near the 
neck is a common enough phenomenon, and here it is strongly marked. An 
outwardly projecting elbow in cross-section thus produced may easily enough 
be mistaken for a true outgrowth of cells in some cases, and such an error 
must be carefully guarded against. The separation of the lamina from the 
oral epithelium frequently leaves such an elbow projecting freely, and this 
may form another source of error in interpretation, 


DEVELOPMENT AND SUCCESSION OF TEETH IN PERAMELES. 475 


(G) 


Fie. 1.—a—c. Series of coronal sections (each three sections apart) from 
premolar region of upper jaw of Stage 111, and passing through dental 
lamina (d./.) between p2 and dp3. (x 90 diameters.) 


476 J. tT; WILSON AND JS. P.-Hiih. 


dental lamina, opposite the anterior half of p2, there is to be 
seen through several sections a small mass of cells projecting. 
It is possible that this is representative of dp2. If so, it is 
very degenerate, and there is no similar structure beside the 
Anlage of p+, though a similar small cell-mass is present in 
the lower p2 (see below). 

The epithelial projection in question is at all events not due 
to a bending of the lamina, whose “neck” is short, straight, 
and perfectly preserved in the region referred to. 

It will be remembered that in the preceding Stage 11 no 
trace of localised thickening or other differentiation was ob- 
servable in the premolar region of the dental lamina in front 
of the site of the third premolar, but that here a tooth Anlage 
in the cup-shaped stage was already present. And whilst in 
the present stage the rudiments of the two anterior premolars 
(p+ and p2) are sufficiently defined as such by localised en- 
largement of the dental lamina, the Anlage of that premolar 
tooth which was already cup-shaped in the preceding stage is 
now still further advanced, showing differentiation of the 
stellate reticular tissue of the enamel-organ (figs. 22 and 24). 

Furthermore it is to be remarked that the process of libera- 
tion of dp® from the lingually-placed dental lamina, which was 
stated to have just set in the preceding stage, has progressed 
considerably, so that the freed dental lamina (fig. 24, rdJ.) 
now forms a partly independent epithelial ingrowth by the 
lingual side of the tooth rudiment. The latter is still seen 
(fig. 24) to be somewhat broadly attached to the labial aspect 
of the liberated lamina at some distance from its apex, or, 
strictly speaking, from its free distal margin, and extending 
nearly to its attached proximal margin or “ neck.” 

A careful investigation of the earliest stages of 
premolar tooth-differentiation has thus served to 
show that this third in order among the premolars is 
not differentiated at all contemporaneously with the 

1 The question of the significance of certain outgrowths of the dental 
lamina on its labial side is discussed below in connection with the molar 
problem. 


DEVELOPMENT AND SUCCESSION OF TEETH IN PERAMELES. 477 


other premolar germs. On the contrary, it is com- 
paratively far advanced in structural differentiation 
by the time there appears any trace at all of the 
anterior premolars, or more than very slight traces 
of the formation of the enamel-organs of any perma- 
nent tooth whatever. 

The vertical elongation of the cross-section of the dental 
lamina just anterior to dp? has been referred to. Just in 
front of the plane at which the enamel-organ of the latter 
tooth comes into view the dental lamina is higher than at 
any more anterior part of the premolar region, and as high as 
it is at the canine thickening. This height is retained by the 
residual lamina opposite to dp®. Towards the hinder part of 
the tooth the lamina becomes separated off from the oral epi- 
thelium. 

A couple of sections or so behind the disappearance of the 
last trace of the enamel-organ of dp3, the free margin of the 
dental lamina undergoes an abrupt further elongation, i.e. it 
suddenly penetrates considerably deeper into the connective 
tissue to be prolonged into the rudiment of m+, which forms 
immediately behind as a large cup-shaped enamel-organ. 
This is not visible in fig. 22, since the plane of that section 
passes altogether beneath m+. Fig. 25, however, represents a 
section taken at a higher plane, where the extreme upper or 
basal portion of dp2 is cut through, along with its associated 
residual lamina (rdl.). The latter is seen to be continued 
backwards into the first molar, and the lower face of the 
enamel-organ of this tooth (m+) is seen to be shaved through. 

Fig. 26 shows a still higher plane, where therefore the con- 
tinuity with the lamina in the premolar region is not visible, 
but where the lamina (m2 dl.), ascending posteriorly, and 
becoming freed from the hinder part of the developing tooth, 
is continued backwards into the region of the future second 
molar. Here the first molar is seen to be a deeply indented 
(“glockenformig”) enamel-organ (cf. fig. 27), which in its 
degree of maturity now rivals dp. It will be remembered 
that in the preceding stage the enamel-organ of m+ had only 


478 J. LT. WILSON AND J. P. HILL. 


just become recognisable as a distinct Anlage, though dp2 was 
already a cup-like enamel-organ. 

Figs. 28 and 29 represent coronal sections through m4, 
rather behind the middle of the tooth, and through the 
posterior part of the tooth, respectively. In these the process 
of differentiation of the tooth Anlage from the parent lamina 
is seen to be well advanced, so that the residual dental lamina 
(rdl). which remains, after separation from it of a tooth 
rudiment, appears with its free distal margin or fundus some- 
what swollen, and freely projecting by the side of the first 
molar Anlage. 

As the residual lamina is traced backwards from the region 
of m+ (ef. figs. 25 and 26) it retains for a time on its labial 
face some irregular projections over the area corresponding to 
that occupied further forwards by the differentiated enamel- 
organ of m1. A comparison of figs. 29 and 30 (the latter 
being the tenth section behind the former) will serve to 
explain this statement. It will also appear that the residual 
dental lamina of the first molar region has very slightly 
increased in thickness. Here, of course, it simply constitutes 
the primitive and undifferentiated lamina intermediate between 
the region of the first and second molars. 

Fig. 31, taken some little distance further back, shows con- 
siderable broadening of the lamina at its free margin, after 
somewhat the same fashion as we have seen to characterise the 
Anlage of the first molar in the preceding Stage 11. 

Here, again, the thickening of the fundus appears to be 
chiefly due to a proliferation of the cells at the labial lip of the 
free margin of the lamina. 

Fig. 32 shows a section immediately in front of the plane at 
which the molar lamina abruptly terminates. 

In the last three sections (cf. especially figs. 80 and 31) the 
prominent labially directed process (/.0.) near the base or 
attached margin of the dental lamina is the cross-section of a 
continuous secondary lamina in this region (m+ to m2), 
whose significance will be discussed below in connection with 
the questions of molar homology (vide Part II). 


DEVELOPMENT AND SUCCESSION OF TEETH IN PERAMELES. 479 


Lower Jawof Stage III.—The phase of tooth develop- 
ment exemplified in the lower jaw is on the whole a little in 
advance of that found in the upper. This is more notable in 
some regions than in others. During the period of development 
which intervenes between Stages 11 and 111, rapid advancement 
has been made in the differentiation of the first lower incisor 
Anlage. An attentive comparison of our figures of sections 
through this region in the two stages (figs. 12 and 388) will 
chiefly show that the early papillated Anlage of di, in Stage 1 
has rapidly passed through the succeeding stages of its 
development up to and including the development of a perfect, 
though thin, dentine cap for the papilla. No enamel has been 
formed, but the enamel-organ—apparently arrested in its 
evolution—has become separated from the main mass of the 
dental lamina by a considerable interval, remaining connected 
with it only by a delicate and disintegrating strand of 
epithelial cells forming a “ Verbindungsbriicke” (fig. 
33, cb.). 

It may be noted that the position of diz relative to the 
Anlage of 7+ appears to have altered during the transition from 
Stage 11 to Stage 111. In the former the rudiment of di> 
occupied the labial aspect of the more anterior portion of the 
swelling corresponding to 77. Here, however, the differentia- 
tion and segregation of diz have allowed of a growth forward 
of the bulky Anlage of i;, unhindered by the presence of any 
more anterior dental element. Hence, in the present stage, 
the vestigial dit has come to lie opposite the more posterior 
moiety of its large morphological successor. The relative 
forward extension of the latter appears to continue at least up 
to a period represented by our Stage 1v (see fig. 56). 

We have in the next place to point out that, in the present 
stage (111) the dental lamina, which was almost wholly 
undifferentiated from the enamel-organ of diz in Stage 11, has 
gone on to form the large papillated dental germ of 7; (fig. 33). 
But as yet there is no trace of differentiation of a residual 
dental lamina from the Anlage of 71. In the succeeding Stage 
tv, however, a residual lamina, with its free marginal portion 


4.80 J. T. WILSON AND J. P. HILL. 


distinctly swollen, is exceedingly well developed (fig. 55) 
along the whole extent of the tooth (¢;) so that the latter 
has come to appear as a huge appendage of the liberated 
lamina. 

In the present stage no such differentiation has occurred, 
and the whole of the large dental lamina remaining over after 
the differentiation of diz appears for the present totally con- 
verted into an enamel-organ for 77. Thus the sequence of 
phenomena observable in this series of three stages (11, III, 
and iv) in the case of this particular region goes far to illus- 
trate and support our contentions as to the nature and con- 
ditions of occurrence of a residual dental lamina, the so-called 
“ Krsatzleiste ” of Rose. 

As the papillated Anlage of 7; is traced backwards it 
gradually diminishes in sectional area, becoming at the same 
time more consolidated by loss or marked reduction of the 
looser epithelial cells in its interior, which become the steilate 
tissue. When the last trace of papillation has disappeared 
posteriorly, the structure is seen to be again a mere undif- 
ferentiated dental lamina transitional from the first to the 
second incisor region. In the latter situation it swells out 
again, and at the same time becomes slightly indented below. 
There is also a somewhat denser aggregation of connective- 
tissue cells in relation to this indentation than elsewhere in 
the fibrous capsule, and the cellular accumulation forms a 
faint elevation, corresponding to the indentation of the enamel- 
germ. 

This condition of affairs corresponds to what we have 
already noted in connection with the upper incisors. Beyond 
the Anlage of 73 the dental lamina runs for a short distance 
almost transversely outwards, and then swells into the Anlage 
ofiz. This relationship is very striking. At first we were 
puzzled by the appearance of the coronal sections in this region, 
for both 7; and zz appear in the same section (fig. 34), and 
the relation of the intervening part of the dental lamina seems 
obscure. <A casual examination would suggest the idea that 
the Anlage of i; is morphologically in the same transverse 


DEVELOPMENT AND SUCCESSION OF TEETH IN PERAMELES. 481 


segment of the dental lamina as éz, and is therefore related to 
it as belonging to a labial or elder generation. Its superficial 
position would almost seem to lend support to this view. But 
it is absolutely negatived by the facts elicited by more careful 
study. The true explanation was arrived at by following the 
dental lamina forwards and backwards in the coronal series, 
by careful collation with the appearances seen in the corre- 
sponding locality in coronal series in Stage 11, and above all 
with those found in the examination of horizontal sections of 
Stage 111 itself. By means of such observations we have 
been able to determine with certainty that the longitudinal 
continuity of the dental lamina is maintained nearly in the 
antero-posterior direction from 7; to zs, then almost trans- 
versely outwards for a short distance between 75 and 73, and 
then again backwards and rather inwards into the canine 
region. 

In other words, zz is situated upon the dental lamina pos- 
teriorly, in the morphological sense, to ¢z, and not morpho- 
logically on its labial side, as a coronal section by itself might 
suggest. 

In illustration of these facts, fig. 34 of a coronal section 
through 7; and 7; may be compared with figs. 35 and 36 of 
horizontal sections. These latter are taken at slightly dif- 
ferent levels, six sections intervening. In fig. 85 the swollen 
Anlagen of 7; and 7; are seen connected by the outwardly 
directed dental lamina. The peaked anterior extremity of the 
tz swelling indicates the continuation forwards of the dental 
lamina (cut obliquely) into the first incisor region. The 
rounded hinder end of the Anlage of iz is seen to end 
abruptly at this level, but this is only because the plane of 
section prevents its backward continuation from appearing. 
In fig. 86, which is closer to the oral epithelium, the lamina is 
seen to be prolonged back as a well-marked band, from the 
hinder end of the upper part of the i; swelling, through the 
canine region, to reach the first premolar Anlage (p;). 

The swollen Anlage of iz exhibits the same kind of indenta- 
tion of its fundus as that of 7;, and there is similar evidence 


482 J. T. WILSON AND J. P. HILL. 


of preparation for the formation of a dermal papilla. It forms 
a shallower Anlage than 7, besides being placed at a higher 
level. 

In one or two sections we find close under the oral epithe- 
lium, and placed in the angle between the latter and the labial 
side of the enamel-organ of iz, a somewhat flattened mass of 
epithelial cells surrounded by condensed connective tissue. 
This appears to represent a rudimentary diz. In the follow- 
ing stages its form and relations become still more manifest, 
though it never undergoes calcification. It does not appear in 
the section shown in fig. 34. 

No such rudiment is observable in connection with 73. 

Behind 7,, the dental lamina, which is there considerably 
displaced outwards from the mesial plane, now bends slightly 
inwards again (cf. fig. 36), and becomes markedly reduced in 
height. After continuing for a short distance as an extremely 
shallow and insignificant structure, its cross-section rather 
suddenly undergoes marked elongation, penetrating further 
and further into the surrounding connective tissue, not verti- 
cally but rather mesially, in a direction almost parallel with 
the deep surface of the oral epithelium. At the same time it 
swells out markedly at its free distal margin or fundus, and at 
its point of maximum development exhibits some flattening, or 
even slight depression, towards its labial side (fig. 37). There 
is present also just the faintest trace of the differentiation of 
the connective tissue foreshadowing the growth of the dermal 
papilla, though as yet there is no papillary projection what- 
ever in this situation. The proximal or attached portion of 
the lamina (or “ neck ”’) is elongated and mueh constricted, 
and is pretty closely appressed to the deep surface of the oral 
epithelium. Close to its point of continuity with the latter is 
attached the small but deeply cupped enamel-organ of de. (fig. 
37), which has been figured in its earlier stage of differentia- 
tion in Stage 11 (fig. 13). 

The superficial resemblance between fig. 34, showing rela- 
tion of i; to zz, and fig. 87 showing the relation of d; to; is 
somewhat striking; but, as has been pointed out, the funda- 


DEVELOPMENT AND SUCCESSION OF TEETH IN PERAMELES. 483 


mental relationships are totally distinct in the two cases. The 
appearance is explained by the kink in the lamina which has 
been described. 

In fig. 86 the dental lamina is seen continued back as a 
well-marked band from the hinder end of 7; through the 
canine region to the first premolar Anlage. The intermediate 
part of this elongated portion of the dental lamina really con- 
sists of the thinned-out ‘neck ” portion of the dental lamina, 
where that constitutes the canine Anlage. Its topographical 
relation to the vestigial milk canine d; is well shown. 

Behind the canine region the dental lamina is again reduced 
in size, until it is continued, as may be seen in the horizontal 
section figured (fig. 86), into the Anlage of py. Here it 
attains considerable bulk, and, unlike the corresponding 
Anlage in the upper jaw, it shows slight but distinct indenta- 
tion at its base, and the usual indication of the earliest com- 
mencement of a dermal papilla, in the form of a deeply stained 
cellular aggregation opposite the depression in the enamel- 
germ, and slightly projecting into it. 

Beyond pz the dental lamina becomes attenuated before it 
once more enlarges to constitute the Anlage of pz. This 
closely resembles pz, but is even more distinctly though still 
only slightly cupped (fig. 88). As ps is traced backwards the 
labial lobe, which forms the outer boundary of the shallower 
depression of the enamel-organ, disappears, but the main body 
of the dental lamina continues backwards, uniformly plump 
and well developed, for a very considerable distance; nor, 
indeed, does it lose its character, but merely undergoes further 
vertical elongation at its continuation into the residual dental 
lamina, which now exists in a well-developed condition by the 
side of the highly developed enamel-organ of dpz. Here it is 
longer, and, though a little less thick, it is still markedly 
swollen in its marginal portion, so that upon cross-sec- 
tion it appears like a large bud-like lobe (fig. 39, rdl.), 
and this is surrounded by a capsule of condensed connective 
tissue. 

The development of dpz so closely resembles that of the 

VOL. 39, PART 4+,—NEW SER. Lt 


484, J. T. WILSON AND J. P. HILU. 


corresponding tooth in the upper jaw that little special descrip- 
tion need be given. It is indeed just a little in advance of its 
opponent. In particular we find that in the hinder part of 
its enamel-organ an important differentiation has set in, of 
which no trace is as yet visible in the upper tooth. Fig. 40 
shows that the cup-like cavity of the enamel-organ is now in 
process of subdivision into two, and that, corresponding to this 
the dermal papilla exhibits two very slight pointed projections. 
Here plainly we have the earliest stage in the production of 
more than one cusp in the case of an enamel-organ which, at 
the previous stage of its development, was a perfectly simple 
cup, with a correspondingly simple papilla. The figure repre- 
sents, in fact, an appearance which has been already claimed 
by Rose as illustrating the earliest and most primitive con- 
dition of a multicuspidate tooth. In the case of the deciduous 
premolar in Perameles, it can be proved that such is not the 
case; and that here, at least, cusp-formation is a secondary 
complication, introduced at a period subsequent to the forma- 
tion of the primitive dermal papilla. 

As in the upper jaw, the dental lamina behind dpz is the 
prolongation of the residual lamina in the region of the latter, 
and some distance behind it is continued into the Anlage of 
the first molar. There is not the same sudden elongation here 
as in the upper jaw, and it can therefore be plainly seen that 
the inner lip of the most anterior part of the cupped enamel- 
organ of mz is the direct continuation of the dental lamina, 
the rest of the enamel-organ having plainly arisen by out- 
growth from the labial surface of the lamina, as we have seen 
to be the case in other tooth rudiments. On the other hand, 
the residual dental lamina has not liberated itself from the first 
molar Anlage in the lower jaw as in the upper, except in so far 
that the enamel-organ gradually diminishes posteriorly, and 
finally disappears as a mere labial excrescence of a thickened 
lamina. The latter is continued backwards, and very shortly 
behind it develops another laminal excrescence, which when 
followed is seen to form the labial boundary of the cup of the 
papillated enamel-organ of mz (fig. 41). This tooth now ex- 


DEVELOPMENT AND SUCCESSION OF TEETH IN PERAMELES. 485 


hibits a much more advanced condition than does the corre- 
sponding Anlage in the upper jaw. 

The labial portion of mz in turn disappears rather abruptly 
(see fig. 42), leaving a simple elongated dental lamina (dl.). 
The latter has now become separated from the oral epithelium, 
and soon disappears entirely from the section, without having 
given any indication of the future formation of the third molar. 


Stage 1V.—P. nasuta: mammary fetus. 


Length from vertex to root of tail. : . 34mm. 


Head length : : : : oe On sss 
Coronal, horizontal, and sagittal sections examined. 


The transition from Stage 111 to Stage rv represents a greater 
advance in development than that from Stage 11 to Stage 111. 

Upper Jaw.—The dental lamina (fig. 43, di.) begins some 
distance in front of the anterior end of the first incisor, as a 
downgrowth from the lingual angle of a broader epithelial 
invasion of the connective tissue (ca.). The possible signifi- 
cance of this will be adverted to in the sequel. In fig. 44, 74, 
is seen to be connected with the dental lamina by a bridge of 
cells (cd.), or“ Verbindungsbriicke” of Rose. This struc- 
ture alone bears witness to the earlier condition manifested in 
Stage 111, where the enamel-organ of 7+ was still an integral 
portion of the dental lamina. Beyond the attachment of the 
connecting bridge (c.d.) the “ residual” dental lamina (rdi.) 
extends as a free laminar ingrowth. 

Springing from the labial angle of the broad epithelial 
ingrowth above referred to may be seen (fig. 44) a hooked 
cellular process (ep.). And, between this and the proper 
dental lamina itself, the surface of the broad epithelial ridge is 
beset with irregularities. In many sections, indeed (cf. fig. 
46, cp.), completely isolated epithelial cell groups may be seen 
in the connective tissue to the labial side of the “ Verbin- 
dungsbriicke” of z+, i.e. between the latter and the irre- 
gular edge of the broad epithelial mass which forms the basis 
from which the proper dental lamina springs. In other sec- 


486 J. T. WILSON AND J. P. HILL. 


tions somewhat similar epithelial cell-masses are seen still 
connected with the Malpighian layer of the epithelium. 

We cannot give a positive opinion respecting the signifi- 
cance of these more or less isolated cell-groups. Either they 
are partially strangled epithelial ingrowths, such as we shall 
later on have occasion to refer to as dipping more or less 
deeply into the connective tissue from the deep surface of that 
lamina which forms the Anlage of the lip-furrow (labio- 
alveolar furrow) ; or else they may have to be interpreted as 
degenerated remains of the connections of a first milk incisor. 
If the latter, then the irregular area with which they are related 
must be interpreted as forming, morphologically, a part of the 
labial surface of the original dental lamina, whose basal part 
has, as it were, opened out towards the mouth cavity. We 
think that consideration of the appended figs. 43—46 will 
support this as at all events a plausible explanation. 

It will be remembered that in Stage 111 no rudiment of 
di+ was found. But the connection of the dental lamina (at 
its first incisor enlargement) with the oral epithelium was a 
very broad one, and, in the reduction to its present dimen- 
sions, it might well have left cellular vestiges of a rudimen- 
tary di+, stranded, as it were, on the labial side of the now re- 
latively constricted base of the proper dental lamina. 

In fig. 45 the residual dental lamina (rdl.) in the region of 
74 is well seen. Its fundus is distinctly enlarged (‘‘kolbig”), 
but it is merely the thickened free margin of the continuous 
lamina seen in front in fig. 43 and the succeeding figures. 
Behind this point, indeed (i.e. fig. 45), it loses its independ- 
ence for a few sections, when it is more intimately fused with 
the side of the enamel-organ of é+ (fig. 46), but almost imme- 
diately it again extricates itself as a free but less prominent 
residual lamina. Shortly behind this again, opposite the 
posterior part of 24+, the lamina is practically suppressed 
altogether. Traces of its former presence remain in the shape 
of slight and inconstant cellular projections from the oral epi- 
thelium, and from the lingual aspect of the enamel-organ of i+, 
respectively. 


DEVELOPMENT AND SUCCESSION OF TEETH IN PERAMELES, 487 


Soon after the appearance of 72 in the sections the lamina 
completely reappears and relates itself to that tooth in a 
manner similar to that observed in the case of i+, a residual 
dental lamina, “ bud-like ” on cross-section, being present here 
also. 

The enamel-organ of 7+ is relatively shallow, having only 
attained cup-like form, while 72 is deep and _ bell-shaped, 
with pointed apex. In neither case are hard dental tissues 
present. 

We have been rather surprised to find that the enamel- 
organ of 23 in the present stage is less advanced in deve- 
lopment than that of 72. It is even a trifle less mature 
than 78, 

In the two previous stages it was the largest of the incisors, 
and, at least in Stage 11, the most advanced. Now it is found 
to lag behind as a much shallower cup-shaped enamel-organ. 
It is, however, somewhat deeply placed, and is attached near 
to the distal end of a dental lamina which is, as a whole, more 
massive here than in front. 

Another evidence of relative immaturity is found in tlie 
fact that the residual dental lamina, though distinctly indicated, 
is only as yet imperfectly liberated by differentiation from the 
enamel-organ. 

Opposite the hinder end of 7% the deep and bulky dental 
lamina abruptly diminishes in depth, and is prolonged 
backwards into the less deeply situated 74, with the lingual 
part of whose cupped enamel-organ it is directly continuous, 
much as was seen to be the case with the second lower 
molar of Stage 111. 

Behind 74 the dental lamina relates itself in like manner 
to the closely similar cupped enamel-organ of i8. Neither 
of these Anlagen as yet exhibits any traces of a residual 
dental lamina, i.e. they are as yet destitute of “ Ersatz- 
leisten.” They are decidedly less advanced in development 
than the anterior incisors. 

No trace of a milk predecessor is now found in connection 
with any of the incisors, with the doubtful exception of 


488 J. T. WILSON AND J. P. HILL. 


the first. We have seen above that the vestiges of such 
present in Stage 111 were already retrogressive, that of di3 
being in process of invasion and dispersion by connective- 
tissue ingrowth. 

The canine calls for little remark. It is a large and deeply 
cupped enamel-organ (fig. 47), and its future residual dental 
lamina is only very partially indicated as a somewhat bulky 
projection from the side of the enamel-organ. The drawing, 
however, contains a highly interesting feature in the shape 
of an attenuated cord of cells (d‘) with a slightly clubbed 
termination, springing from the labial side of the dental 
lamina near where it is continuous with the oral epithelium. 
This appears in only a very few sections, abruptly disappear- 
ing fore and aft, and is without doubt the retrogressive 
vestigial representative of the perfect miniature cupped 
enamel-organ of the milk canine figured in Stage m1 (fig. 
20 d?). Since this is all that is left of such a perfect rudiment 
as that in Stage 111, itis not surprising that the much more 
rudimentary milk incisor structures seen in Stage 111 should 
have wholly disappeared. 

Behind the canine the dental lamina, losing much of its 
depth, but still remaining pear-shaped on cross-section, is 
continued bodily into the first premolar Anlage. . 

Both the first and second premolars are now deeply cupped 
enamel-organs of au equally advanced stage of development to 
that of the canine (fig. 47), which they now much resemble. 
In the case of each a residual dental lamina is just being differ- 
entiated. In this respect, but not otherwise, p+ is just a little 
in advance of p2. 

At its hinder end p2 tapers away into the undifferentiated 
dental lamina, which here remains tolerably thick and club- 
shaped on cross-section. On its being traced backwards, there 
soon appears by the side of the dental lamina, but at a consi- 
derable distance from it, the enamel-organ of the deciduous 
premolar. This is not only placed labially, but is situated low 
down, i.e. close to the oral epithelium. It has now attained 
some degree of complexity. One can easily distinguish a 


DEVELOPMENT AND SUCOESSION OF TEETH IN PERAMELES. 489 


protocone, deep and pointed, from the shallower and less 
pointed paracone and metacone, in front and behind. 

The inner epithelium is elongated and cylindrical, especially 
at the apex of the protocone, where the cells are about as long 
again as elsewhere, and their inner halves are pale and un- 
stained, the nuclei forming an outer marginal band. Covering 
the tip of the papilla of the protocone there is an exceedingly 
thin dentine cap, the only one hitherto encountered in the 
upper jaw (fig. 48). Outside the inner enamel epithelium the 
stratum intermedium forms a very distinct layer of cubical 
or flattened cells, and outside this again the stellate enamel 
tissue is reduced to a comparatively narrow layer. 

As it is traced backwards into the region of dp®, the dental 
lamina increases in thickness, and opposite the paracone it 
begins to exhibit remarkable modifications of its swollen free 
marginal portion in the shape of alternate constriction and 
enlargement, the constrictions being due to ingrowth of con- 
nective tissue on the labial aspect of the lamina. The thickest 
portion of the lamina is opposite the protocone (fig. 48, 7di.), 
where also the connection between dp* and the swollen dental 
lamina becomes apparent. 

Fig. 49 represents a horizontal section through the region 
between p2 and m+; the level is too high to show the very 
lowly situated dp® at all well, it is indeed only shaved through 
near its higher or basal limit. 

The tapering of p2 into the dental lamina, and the prolon- 
gation of the latter by the sides of dp? and m4, is strikingly 
shown; also the peculiar modification of the lamina which has 
been referred to as taking place opposite dp. 

It can hardly be gainsaid that here we have the earliest 
promise of the specific differentiation of p®. But we are not 
at all clear as to the significance to be attached to the con- 
stricting ingrowths of connective tissue described and figured. 
They are also present in similar shape in Stage v (fig. 50). 
At first sight they suggest papilla, but not only are they histo- 
logically entirely unlike the Anlagen of papille (for instead of 
being cellular growths they are fibrous bands), but an exami- 


490 J. T. WILSON AND J. P. HILL. 


nation of much later stages shows that they disappear as such, 
and that the enamel-organ of p3, when finally elaborated, is 
a simple and single rounded cup-like structure. 

A comparison of sagittal series with the horizontal and 
coronal of this stage proves that the ingrowths of connective 
tissue shown in the horizontal section are neither mere pits 
nor simple cross-furrows, but that they are circular furrows 
surrounding, and partially isolating, two rounded projections 
of the labial face of the lamina marked a and 6 in fig. 49. 
How these two knobs—for such they are definitely proved to 
be—are related to the process of formation of p2, we are not 
in a position to say. That they are no mere accidental out- 
growths is shown by the fact that they are present, under only 
slight modification in form, in the next stage (v, fig. 50). 
Nevertheless, in still later stages the regularity of the arrange- 
ment is lost, and the lamina simply appears irregularly thick- 
ened, until finally out of this thickening a quite simple and 
normal cup-like enamel-organ arises. 

A superficial criticism, having regard only to our fig. 49 of 
the present stage and fig. 50 of the next, Stage v, might be 
disposed to claim the indentations there represented as exam- 
ples of supposed primitive papillze whose fusion would consti- 
tute p32. It is just remotely possible that here we do have 
‘phenomena that ought to be interpreted by some fusion 
hypothesis, though we are strongly inclined to think otherwise. 
In any case if we do here meet with a case of fusion the 
structures that are fused are not dermal papille, for the 
structures which in some sections are suggestive of papille are 
merely cross-sections of the fibrous connective tissue occupy- 
ing certain circular grooves which circumscribe knob-like pro- 
minences of the dental lamina near its fundus. 

We may further point out that so far, and after careful study, 
especially of these earlier and plastic stages, we can (with 
Leche) find no support for Woodward’s statement that p3 
(i.e. the successional tooth) takes its origin from the dental 
lamina morphologically in front of dp2 (i.e. the milk molar), 
and is therefore probably in series with it. 


DEVELOPMENT AND SUCCESSION OF TEETH IN PERAMELES. 491 


In his paper already referred to Mr. Woodward makes the 
following statement:—“It is interesting to note that in 
Perameles the large supposed successional tooth is quite 
distinct in origin from the small fourth premolar which is shed 
[dp.*]|; it is, in fact, formed from the dental lamina situated 
immediately behind pm 3 [p2] and morphologically in front 
of the fourth premolar [ p2]” (2, p. 467). It is unfortunate 
that Mr. Woodward makes this statement without amplifying 
it by a more detailed description or illustrating it by drawings 
(for Perameles). 

From our own minute and detailed study in Perameles of 
the region of the dental lamina concerned we are convinced 
that Mr. Woodward has misinterpreted the appearances there 
met with. We do not deny that the latter may sometimes 
seem to suggest the view that Woodward adopts, but upon 
further examination it becomes untenable. In the absence of 
any figures the readers of Woodward’s description would 
necessarily suppose that in Perameles one meets with just 
such an enlargement of the dental lamina between p2 and 
dp2 (his “pm. 3” and “dpm 4”’) as is represented in the wax 
model of the corresponding region of Petrogale figured by 
him. But this is by no means the case. In connection with 
the description of Stage 111 of Perameles we have drawn 
attention to the fact that the dental lamina behind p2 gradu- 
ally tapers backwards from the thickening which represents 
the enamel-organ of the latter tooth, and that, as it approaches 
the region of dp3, it becomes somewhat more slender, but at 
the same time distinctly more elongated vertically. (See text, 
Fig. 1, a—c.) There is no localised enlargement whatever of 
the stretch of lamina between p2 and dp3. It is true that it 
is plump and well developed, but this does not specially distin- 
guish it at the stage in question from stretches of the lamina 
between other tooth Anlagen. 

We have already pointed out that in fig. 22 the apparent 
swelling of the dental lamina in front of the region of dp2 is 
due merely to its plane of section. To understand it properly 
the gradual vertical elongation of the lamina behind p2 must 


492 J. T. WILSON AND J. P. HILL. 


be borne in mind. The plane of the section lies too high to 
intersect the main mass of p2, which is relatively low and 
transversely elongated. Still the obliquely cut anterior por- 
tion of the segment of the lamina figured really corresponds 
to the tapering posterior end of p2, which is higher (deeper) 
than the rest of the Anlage. In like manner the narrowing of 
the dental lamina, seen in the figure as occurring towards dp, 
is to be explained not as an actual diminution of the sectional 
area of the lamina, but only of its width at this horizontal 
level. And if sections through exactly the same region but 
in a higher horizontal plane be examined, it will be found 
that, when the fundus of the free dental lamina close beside 
dp® is cut through, its sectional outline is quite comparable to 
that seen further forwards. There is, in short, just such a 
gradual heightening of the lamina, when it is traced backwards 
into the dp= region, as there is in front of the premolar region 
when the lamina is traced forwards from p+ towards the 
canine. 

It has already been shown that a residual dental] lamina is 
liberated by the side of dp? at a very early period, and that 
this soon becomes thickened at its margin and “ bud-like” on 
cross-section. It is from this liberated lamina that p? takes 
its origin in Perameles. 

Even in Petrogale, upon which Woodward chiefly founds 
his case, we cannot allow that the conclusion that p> (his p+) 
is morphologically in front of dp> (his dp*) is at all so in- 
evitable as he regards it. The proofs drawn from the mere 
appearance of the dental lamina in transverse sections are, in 
our opinion, of very little value. A quite similar development 
of the deeper part of the dental lamina, and an identical 
development of loose and stellate-like epithelial tissue in its 
interior, may frequently be met with where there is no 
question of the presence of a tooth at all. 

It is only the definite localised elevation, sharply limited 
fore and aft, figured in the drawing of a wax model of this 
region of the lamina in Petrogale which seems to us atall a 
formidable evidence of the view put forward. And even this 


DEVELOPMENT AND SUOOESSION OF TEETH IN PERAMELES, 493 


we cannot regard as conclusive. For, according to Mr. Wood- 
ward’s statement and figures, the enamel-organ of dp2 (his dp*) 
at the stage referred to has not yet differentiated itself as an 
organ distinct from the dental lamina. The whole lamina in 
the dp* region appears as if converted into the enamel-organ, 
as is commonly the case in the earlier stages of tooth develop- 
ment. This phase of development is, however, only a passing 
one. By-and-by the separation of the enamel-organ from the 
lamina which gives it birth is indicated, and the resulting 
appearance of a residual lamina or “ Hrsatzleiste” we 
regard as the first condition of the appearance of a succes- 
sional tooth. 

This first condition is as yet absent in the early stage of 
Petrogale under notice. Mr. Woodward in effect maintains 
that in its absence we must forthwith accept the presence of 
an enlargement of the lamina in front as the true Anlage of 
p* (his p*). We hold, on the contrary, that until we are in 
a position to trace the fate of the residual lamina, which will 
assuredly manifest itself by the lingual side of dp, it is pre- 
mature to decide upon what constitutes the earliest germ of 
p® (“pt”). If the evolution of the swelling figured by 
Woodward can be clearly traced up to the formation of the 
enamel-organ of p= (“p+”), without suspicion of a confusion 
with the subsequent evolution of a true “Ersatzleiste” of 
dp2, then Mr. Woodward’s case for Petrogale will have been 
made out. 

We therefore insist that it is necessary to have recourse to 
somewhat later stages for corroboration, or otherwise, of Wood- 
ward’s view. And what is the result of such a reference? We 
shall quote Mr. Woodward’s own words. “In a slightly older 
embryo the same condition was observed, but owing to the de- 
velopment of the third and fourth premolars, especially of the 
latter, they somewhat overlap the lamina connecting the two, 
and it in consequence becomes displaced to the inner side of 
these structures. So much does the fourth premolar grow 
forward with age that it appears as if this lamina was 
a downgrowth from the inner side of the enamel- 


494, J. T. WILSON AND J. P. HILL. 


organ of pm [the spaced type is ours]; this, however, 1s 
really not the case; the lamina is morphologically in front of 
that tooth, and only attains a secondary connection with it” 
(2, p. 458). Now, if we disentangle the statement of facts 
from the accompanying explanatory hypothesis, we recognise 
that in the later stage, i.e. when dp? has become thoroughly 
differentiated, a free dental lamina is found by its lingual side 
appearing “as if this lamina was a downgrowth from the inner 
side’ of its enamel-organ. 

But this is precisely the condition we find in Perameles, 
where the appearance of such a laminar “down- 
growth” is most assuredly not due to any process 
of overlapping or disproportionate growth of any 
kind. Nor is there in that form any antecedent enlargement 
of the lamina in front of the dp3 region. We infer from his 
language that Mr. Woodward found some slight difficulty in 
explaining the fact of an actual connection between dp3 
and the lamina by its lingual side, on his theory of a forward 
growth producing overlapping, since he denominates such a 
connection as “secondary.” Now we are firmly convinced 
that no truly secondary connection is ever de- 
veloped between a dental lamina and an enamel- 
organ which has previously been differentiated off 
from the lamina. Itis,a priori, most unlikely that any 
such secondary connection should be established. And we 
believe that, wherever a connecting bridge is present, it will 
be found necessary, and not very difficult, to interpret it as a 
remnant of the original primary connection between enamel- 
organ and parent lamina. We must believe that Mr. Wood- 
ward’s method of explaining away the importance of the con- 
nection he found in his later stage of Petrogale rests upon 
an erroneous interpretation of the facts. We hold that in all 
probability he is wrong in supposing that the part figured in 
the earlier stage in front of dp> is the same as the part found 
in the later stage by the side of that tooth. The latter part is 
simply the free dental lamina of the region of dp® itself, now 
liberated by a more complete differentiation of the enamel- 


DEVELOPMENT AND SUCCESSION OF THETH IN PERAMELES. 495 


organ of the latter; and it is from it, and not from the lamina 
morphologically in front of dp, that p? is subsequently de- 
veloped. We have put our re-interpretation of Mr. Wood- 
ward’s facts in frankly dogmatic form for the sake of clearness. 
There is too much unavoidable uncertainty about any question 
of this kind to encourage a genuinely dogmatic attitude in 
regard to it. And we may, indeed, with Leche, admit after 
all the possibility of dp® arising in front of p32 without 
concluding, with Woodward, that the teeth so arising belong 
to the same series. At the same time the facts in Perameles 
do not permit us to entertain any doubt with regard to the 
origin of dp*, however it may be in the case of Petrogale. 
Nor do our investigations in Marsupials lead us to side with 
Leche in his view that the germs of the permanent teeth do 
arise in front of those of the milk-teeth. Judging from the 
observations already detailed with regard to the position of 
the early rudiments of the vestigial milk incisors, we must be- 
lieve that the Anlagen of the permanent teeth arise side by 
side with those of the milk-teeth, though the latter are placed 
rather opposite the anterior portions of the former. This we 
have already seen to be originally the case in the region of the 
first lower incisor of Perameles. Nevertheless at a com- 
paratively early period the vestigial milk enamel-organ comes 
to lie opposite the hinder end of the permanent tooth, so that 
here the primitive relation comes to appear reversed. 

Woodward calls attention to the anomalous eruption of p3 
in Perameles rather in front of its “ supposed” pre- 
decessor dp. 

But, after all, the point of eruption of a tooth is of 
secondary importance, and we find, on the other hand, that 
not only at its first appearance, but during the very long 
latent period which precedes its elaboration as a cupped 
enamel-organ, the germ of p3 is placed, not in front of, but 
opposite to dp3, 

it would be difficult to draw any other conclusion from our 
fig. 48, and here we are in a position to state that there is 
absolutely not the slightest scrap of evidence tending to prove 


496 Je Lf. WILSON AND J; PP: “HU: 


that any such “ secondary attachment” has been entered upon 
as Woodward has imagined to take place in his parallel case. 
We are perfectly confident that our interpretation is the 
correct one for Perameles. 

From the description of dp2 given above, it will be observed 
that it is still in a more advanced phase of evolution than 
the two anterior premolars. But the difference is rapidly 
becoming less marked. The comparatively rudimentary 
character of the deciduous premolar of Perameles is now 
expressing itself in a much less rapid increase in size, so that 
it is now being rapidly overhauled in its progress by the 
enamel-organs of p+ and p2. In consequence of this fact, the 
very striking discrepancy in the periods of appearance of the 
enamel-organs of dp> and of the anterior premolars respectively, 
is, from now on, less and less apparent. It is thus quite 
easy to understand how observations upon a series of stages 
which did not include any younger than the present (Stage rv) 
might lead to the conclusion that these teeth were serially 
homologous. 

The appearance of the upper first molar in Stage tv is 
represented in fig. 51. Here, as in dp3, a thin dentine cap is 
present over the summit of the protocone. A well-marked 
residual dental lamina (rdl.) is present, whose significance will 
be discussed in connection with the general question of molar 
homology. ‘This figure alone is sufficient to disprove Wood- 
ward’s view (2) that “lingual downgrowths of the dental 
lamina” do not occur in connection with the molar teeth. 

Two drawings (figs. 52 and 53) are given of the enamel- 
organ of m2 from distinct series of sections (and not quite at 
corresponding points). In each case a well-marked swollen 
(“knospenférmig ”) “Ersatzleiste” or residual dental 
lamina (rdJ.) is visible. 

There is also a small labial “ sprout ” or “ outgrowth ” (/.0.) 
opposite a knee-like bend or “ Knickung” of the lamina, the 
significance of which is elsewhere discussed. 


1 Woodward has apparently given up this view, so far at least as certain 
rodents are concerned (cf. 18). 


DEVELOPMENT AND SUCCESSION OF TEETH IN PERAMELES. 497 


After the enamel-organ of m2 disappears from the serial 
sections, the dental lamina continues on, still for the most 
part in connection with the oral epithelium, and by-and-by it 
becomes swollen to form the undifferentiated Anlage of m3 
(fig. 54). In this its hinder portion the lamina preserves the 
kink or flexure above referred to, as well as the (laminar) labial 
outgrowth (/.0.) from this. The latter has, we are convinced, 
nothing to do with tooth production (cf. p. 564 et seq.). 

Lower Jaw.—In the lower jaw the dental lamina arises 
somewhat abruptly a little distance in front of the first incisor 
region. Here it is fairly thick and plump. This port on 
must be regarded as having been produced by a slight exten- 
sion forwards of the lamina since the period represented by 
Stage 111, occurring pari passu with the continued elonga- 
tion of the jaw. When the region of 7; is reached the lamina 
is seen to be continued on into a thickened residual dental 
lamina by the lingual side of 7; (fig. 55, rd/.). The enamel- 
organ of the latter is large and well developed, and is con- 
nected with the lamina by a thin epithelial connecting bridge. 
Further back the latter is lost, while the lamina itself has 
become much elongated and thinned, and here (fig. 56) is 
found attached to its neck a delicate strand of cells (cd.) 
forming a “ Verbindungsbriicke” for the small vestigial 
ealcified dt;. Only a faint trace of a dentine cap is found at 
the apex of 7, in spite of the large size of the tooth. 

Behind the first incisor region the very deep and thin lamina 
is continued on, and soon the enamel-organs of iz and i; 
appear almost simultaneously in the coronal series as a result 
of the peculiarity in the topographical relations of these teeth 
referred to in connection with the earlier stages. 7; is placed 
more superficially and labially, and it would have been impos- 
sible from an examination of this stage alone to have rightly 
interpreted the appearance presented (fig. 57). In this figure 
there will be observed a distinct labial outgrowth (diz) from 
the root of the dental lamina. Adjacent sections show it 
better as a thickened cell-mass attached to the root of the 
lamina by a thinner connecting bridge, and they also prove 


498 J. T. WILSON AND J. BP. HILL. 


that it is related more intimately to 7, than to 7,, as might be 
suggested by fig. 57. We can only interpret this as a vestige 
of a diz, of which a trace was found in Stage 111 in the form 
of a flattened mass of epithelial cells labially to the Anlage 
of 73. 

As iz and is disappear posteriorly, the dental lamina is 
greatly reduced in depth for some distance, and then gradually 
deepens again, becoming more and more elongated and swollen 
as the canine region is approached. Here (fig. 58), but still 
some distance in front of the actual canine enamel-organ, we 
find the vestigial d; attached to the lamina by a delicate epi- 
thelial strand (cd.). 

It is of interest to note that in one of our series of sections 
of this stage this is a precociously calcified tooth lke di; 
(fig. 58), while in another series it exists in the form only of 
a loose and degenerating epithelial mass. 

The identification of this vestigial tooth as d; is rendered 
easy by a comparison of its relations in the earlier stages 
where no possible doubt could be felt, and where its present 
relative dislocation is evidently in progress. 

The lower permanent canine has not progressed very rapidly 
during the period intervening between Stages 111 and Iv. 
Although it has attained considerable dimensions, its papilla is 
still very rudimentary, and its enamel-organ is only slightly 
cupped. Furthermore it is not yet differentiated off from the 
lamina, so that the residual lamina has not yet come into 
existence. Only a short stretch of lamina intervenes between 
the last trace of the canine and the commencement of the 
first premolar (p;). The latter tooth is more advanced in 
development than the canine, though not greatly so. It is 
deeply cupped by a prominent papilla, and the differentiation 
of the residual dental lamina has begun. 

The second premolar is in a precisely similar condition to 
that of the first, tapering away at its posterior end into a fairly 
deep and thick dental lamina, which is continued back into the 
third premolar region. In this situation dp; comes into view, 
placed away to the labial side of the lamina. Since the 


DEVELOPMENT AND SUCCESSION OF TEETH IN PERAMELES. 499 


appearance of this tooth (except that it as yet shows no calci- 
fication) and its relations to the liberated lamina are exactly 
as described and figured in the upper jaw, no further par- 
ticulars need be given. 

Behind the third premolar region the lamina (still con- 
tinuous throughout with the oral epithelium) undergoes 
further deepening and becomes continuous with the large 
enamel-organ of the first molar, which is larger and more 
complicated than the corresponding upper tooth. Here the 
dental lamina is for some distance separated off from the oral 
epithelium. A residual dental lamina is well developed, as in 
the upper jaw, and its freely projecting margin is moderately 
swollen. Towards the posterior end of m; the dental lamina, 
now unconnected with the tooth, is also for the most part 
separated off from the oral epithelium, and is continued into 
the region of the second molar. Here again it shows a labial 
connecting strand (“ Verbindungsbriicke”) joining it to 
the enamel-organ of mz; whilst its free marginal portion 
forms a large residual dental lamina, just as in the upper jaw. 
Towards the hinder end of mz the dental lamina again becomes 
connected with the oral epithelium, exhibiting near its point 
of connection a somewhat elongated labial laminar outgrowth. 
Further back the lamina entirely loses its connection with the 
oral epithelium, becoming more deeply placed among the con- 
nective tissue of the jaw. Its distal portion is converted into 
the papillated enamel-organ of mz, whose posterior end forms 
the present termination of the lamina. The liberation of a 
residual dental lamina by the side of mz has already been par- 
tially effected. 


Stage V.—P. nasuta: mammary foetus. 


Length from vertex to root of tail . é . 37mm. 
Head length ‘ : : ass 
Coronal and horizontal series examined. 


This stage represents only a slight advance on the preced- 
ing. 

Upper Jaw.—All three first incisors are now well developed, 

voL. 389, PARY 4.—NEW SER. MM 


500 J. T. WILSON AND J. P. HILL. 


the third especially showing a distinct advance from the pre- 
ceding stage. It is now triangular and pointed in coronal 
section, and possesses a distinct dentine cap, as also do the first 
and second. 

The dental lamina is definitely interrupted between i+ and 
12, but otherwise it is continuous, though often greatly re- 
duced in size, throughout the rest of the incisor region, and it 
retains its connection with the oral epithelium. 

Towards the anterior part of the first incisor region irre- 
gular epithelial cell-processes spring from the oral epithelium 
close to the root of the dental lamina, similar to those men- 
tioned in connection with the corresponding region in Stage 
Iv (figs. 43—46). 

The residual dental lamina is still clearly marked by the 
sides of 74 and 72, but it is now considerably reduced in size, 
absolutely as well as relatively to the tooth-germs. The resi- 
dual dental lamina in the region of i has become definitely 
established, and indeed it now closely resembles that of 7+ 
and i2, In each case, though forming a distinct and freely 
projecting laminar downgrowth beyond the connecting stalk 
of the tooth-germ, it is already a somewhat attenuated and 
insignificant structure. We should certainly not be war- 
ranted in following Leche by reading any special importance 
into its persistent presence at this stage. It is plainly steadily 
disappearing. 

The fourth and fifth incisor enamel-organs are in a much 
less advanced stage than those in front. They are still rela- 
tively shallower and cup-like, with rounded papillz, much as 
they were in Stage tv. They still appear as modifications of 
the dental lamina which is directly continuous with the lin- 
gual lip of their cup-like body, and no residual lamina has as 
yet differentiated itself from them. As one might expect, 
the portion of the lamina with which they are continuous is 
bulkier, i.e. less reduced, than it is in regions where tooth- 
germs have attained a more advanced stage of development. 

The lamina behind 72 passes directly and with a somewhat 
abrupt deepening into the canine lamina. This, with the 


DEVELOPMENT AND SUCCESSION OF TEETH IN PERAMELES, 501 


associated tooth-germ, is shown in fig. 59. Behind the canine 
the lamina is abruptly shortened, and is continued after a 
short interval into the bell-shaped enamel-organ of p+, whose 
papilla is destitute of any dentine cap. 

The differentiation of the residual dental lamina has set in 
by the lingual side of the enamel-organ. 

The second premolar exhibits a condition precisely similar 
to that of the first, except that the faintest trace of a dentinal 
cap to the pointed papilla may be observed. The residual 
dental lamina is fairly developed. 

Behind p2 the dental lamina becomes gradually larger on 
cross-section, the germ of dp= soon appearing in the coronal 
series. By the side of the latter tooth the dental lamina 
exhibits precisely the same kind of modification as was seen in 
the preceding stage in the corresponding situation (cf. fig. 49). 

Fig. 50 shows a horizontal and fig. 60 a coronal section of 
the structures concerned. It will be noted that in this stage 
also the relative development of dp* is considerably retarded 
as compared with that exhibited in earlier stages. Its dental 
cap is only a trifle in advance of that of p2. From its present 
condition alone it would be quite impossible to arrive at that 
conclusion with regard to its serial homology which we find 
it necessary to adopt in view of the earlier characters, both of 
the tooth itself and of the rest of the dentition. 

In the molar region the connection of the dental lamina 
with the oral epithelium is frequently interrupted. 

The first two molars are large and well developed. Lingually 
from each there is a large and well-developed residual lamina 
comparable to that figured for the last stage. The Anlage of 
the third molar is slightly more bulky than in Stage rv, but 
is still represented by a mere distal swelling of the lamina, as 
represented in fig. 61. In this figure the projection marked 
‘cb. m2” is seen by comparison with the anterior sections to 
correspond to the root of the laminar connection (“ Ver- 
bindungsbriicke”’?) with m2. 

Lower Jaw.—The dental lamina in front of the region of 
the first incisor shows some further antero-posterior extension 


502 J. I. WILSON AND J. P: HILL. 


since the period represented by Stage 1V. This anterior seg- 
ment of the lamina now exhibits an attenuation of its “ neck,” 
while its fundus shows a very definite oval enlargement, which 
disappears as the anterior plane of 7; is reached. Here, as in 
the preceding stage, the lamina is continuous with the “ re- 
sidual” lamina by the lingual side of that tooth. 

The first incisor is a large pointed germ, with a fairly thick 
dentine sheath to the papilla. The milk incisor, di; (fig. 66 a) 
is present as a strongly calcified miniature tooth, placed at 
the labial side of the apex of 7. It has a small but distinct 
dermal papilla, and the dentine shell is surrounded by a sheath 
of small epithelial cells, which is attached by a laminar con- 
nection to the root or attached border of the dental lamina 
beside 23. 

As in the preceding stages, 73 occupies a markedly labial 
position, and owing to the kink of the lamina, whose occur- 
rence has been already described, its anterior portion has come 
to lie opposite the posterior end of the large iz. Posteriorly 
it occupies a similar position with reference to nearly the 
whole of tz. Indeed, in a few sections all three germs are 
visible in the same coronal plane. Both iz and 7z possess 
dentine caps, and end behind in nearly the same coronal plane, 
though iz extends a trifle more posteriorly than 73. The con- 
tinued dental lamina is, however, placed beside zz, the mor- 
phologically posterior member of the series. 

A long though attenuated residual lamina appears by the 
lingual side of 7;, projecting far beyond the connecting bridge 
between the latter and the dental lamina (fig. 66, rd/.). The 
residual lamina is also present beside zg and iz, but, owing 
to the dislocation of iz, the lamina has undergone corresponding 
dislocation and is partly interrupted. As already indicated, 
it is resumed beside the posterior end of 73, from which it 
extends backwards. 

A delicate epithelial lamina is to be seen springing from 
the root of the dental lamina (i.e. from its attachment to the 
oral epithelium) opposite 7g and 73, and extending labially to 
i, This ends in a minute mass of cells, and it persists through 


DEVELOPMENT AND SUCCESSION OF TEETH IN PERAMELES. 503 


a considerable number of sections, and evidently answers to 
the labial outgrowth figured in connection with Stage iv 
(fig. 57) as a vestigial diz. It is now in process of absorption, 
and apparently has never shown any sign of calcification. 
From its first appearance in Stage 11 it has never attained 
the dignity of tooth-like differentiation of structure. 

Behind the third incisor region the dental lamina again 
deepens, though its connection with the oral epithelium is 
interrupted for a time. Its free margin thickens, and is con- 
tinued directly into the canine Anlage, but very soon is 
partially freed from it, being prolonged as a residual dental 
lamina. The canine enamel-organ is now bell-shaped, and as 
yet is uncalcified. 

In the anterior premolar region the oral connection of the 
dental lamina is frequently interrupted ; p; is slightly calcified, 
while pz has as yet developed no dentine. 

The deciduous premolar and its relations to the Anlage of 
pz resemble those in the upper jaw, except that the pz thick- 
ening does not exhibit the peculiar indentations present in the 
upper jaw, and that dp is distinctly less advanced and is still 
uncalcified ; in fact, it appears slightly less mature than in the 
preceding stage. 

The molar region shows an advance upon the condition in 
the lower jaw of Stage rv, in that mz is now a fairly large 
and deeply papillated tooth showing traces of differentiation 
of cusps. A well-developed residual dental lamina is present 
on the lingual side of each molar, and extends far beyond the 
epithelial connection with the enamel-organ. 


Stage VI.—P. nasuta: mammary foetus. 


Length from vertex to root of tail . : . 44mm. 
Head length : ‘ ; . 24 


= ” 
Lips now separated to form the ‘‘definitive Mundspalte.” 


This stage exhibits a marked advance upon the last in 
regard to tooth development. In it we may be said to en- 
counter for the first time throughout the adult form of the 
various teeth, except in the most posterior molar region all 


504 J. T. WILSON AND J. P. HILL. 


the teeth are well calcified, and have now attained such size 
relatively to the jaws that, especially in the incisor region, a 
good deal of overlapping occurs in both jaws. 

Further, in Stage v the dental lamina, though frequently 
interrupted and generally exhibiting signs of reduction and 
commencing resorption, was, on the whole, fairly complete. 
In the present stage (v1), on the other hand, the lamina is 
exceedingly imperfect and scrappy ; and even where it appears 
it usually displays much attenuation and irregularity of form 
from the operation of the degenerative process. 

Upper Jaw.—Beside the first incisor of the upper jaw 
only comparatively slight traces of the dental lamina are 
visible, and these are now quite unconnected with either the 
tooth or the oral epithelium. It is impossible to distinguish 
any specific remains of the residual dental lamina as such. 
Beside i2 it is just possible to recognise portions of lamina 
corresponding to ‘ Ersatzleiste” and “Verbindungs- 
briicke” in a few places ; while in the region of 23 no resi- 
dual lamina can be identified. 

The fourth and fifth incisors are now well calcified, though 
smaller than the anterior ones. Their more juvenile character 
is perhaps evidenced by the persistence, in more regular form, 
of the dental lamina with its residual prolongation. But these 
latter also persist to a large extent in connection with the 
canine, as shown in fig. 62 (rdl.), which represents the anterior 
portion of that tooth (and therefore does not show the pointed 
and strongly calcified character of its apex). 

In the region of the anterior premolars the dental lamina is 
scrappy, but vestiges of the free residual lamina remain here 
and there, as well as portions of the rest of the lamina and of 
the epithelial connecting bridge. 

The lamina becomes more regular and complete behind p2, 
and in the region of dp* its marginal portion is swollen, form- 
ing the Anlage of p2 as shown in fig. 63. In adjacent sections 
here and there it exhibits some indentations, but of irregular 
character. In this its swollen portion it is surrounded by 
markedly condensed connective tissue. 


DEVELOPMENT AND SUCCESSION OF TEETH IN PERAMELES. 505 


The thickening of the residual lamina becomes even more 
sharply accentuated than in fig. 63 in sections further back. 
The deciduous premolar is a well-calcified tooth having a 
sharply pointed chief cusp, while the anterior portion through 
which the section passes in fig. 63 shows a low and flattened 
crown. 

Behind dp® the continued dental lamina becomes more uni- 
form in thickness. In the region of m+ traces of the connec- 
tion with the latter tooth are discernible, as also in region of 
m2, <A residual lamina is present in both regions. 

Towards the end of the region of the second molar the well- 
marked dental lamina becomes gradually more highly de- 
veloped, still retaining a well-marked laminar projection 
towards m2—the remains of its original connection with the 
dental lamina (fig. 68, cb, m2). Then, somewhat abruptly, the 
residual lamina beyond (distal to) the ‘ Verbindungs- 
briicke’ develops a marked labial bulging, so that it appears 
as a knobbed swelling, flattened or even slightly concave dis- 
tally, forming the Anlage of m3 (fig. 684). The connective 
tissue exhibits a very evident condensation and cellular ac- 
cumulation where it is in relation with the flattened or slightly 
concave surface—the rudiment of the papilla of m3. 

Lower Jaw.—The undifferentiated segment of dental 
lamina in front of the first incisor is again in evidence, and its 
swelling is now more marked than in the preceding stage, 
though, as in the latter, it tapers away and becomes more 
flattened and attenuated before being continued into the 
residual dental lamina beside 77. This residual lamina is now 
greatly reduced in thickness, and no longer exhibits traces of 
its former connection with 7. 

Beyond the progressive enlargement and perfecting of the 
three incisors there is little further to note. The vestigial 
diz is still present as a strongly calcified toothlet on the labial 
side of 7; near the oral surface. Its outer dentine layer appa- 
rently forms now a complete shell surrounded by a layer of 
deeply staining cells. The central nuclei appear altered and 
degenerated. 


506 J. T. WILSON AND J. P. HILL. 


The canine is not so far advanced as the incisors. It is 
smaller, and its dentinal covering much thinner. 

Beside p; a trace of the free residual portion of the dental 
laminais still to be seen. This tooth is further advanced than 
the canine, for both its dentinal and enamel layers are thicker. 

Beside pz there is a small but distinct residual lamina pro- 
jecting beyond the point of attachment of the connecting 
bridge. The tooth is somewhat less advanced than pz. 

The third premolar region is similar to that in the upper 
jaw. In one place there is a very distinct indentation of 
the thickened lamina on the lingual side of dpz. This 
indentation persists through a considerable number of sections, 
and is filled by loose connective-tissue cells, the whole being 
surrounded by a condensed connective-tissue capsule (fig. 64). 
This represents the earliest trace of the definitive papilla of pz. 

In the posterior molar region the anterior cusp of mz is 
now calcified. The tooth is still attached to the persistent 


lamina, which is prolonged beyond the point of attachment as 
a residual lamina. 


Stage VII.—P. nasuta: mammary feetus. 


Length from vertex to root of tail . ‘ . 56 mm. 
Head length . : : ; . 25 


The incisor region of this stage derives some special interest 
from comparison with the previous stage. On the whole the 
teeth are somewhat larger, indicating greater maturity. But 
75 forms an interesting exception. Here it is not yet calcified, 
although it was quite strongly calcified in Stage vi. It is, in 
fact, only slightly ahead of the condition in Stage v. Further- 
more its topographical relations differ from those seen, not 
only in Stage v1, but also in Stages v and viii, all of which 
agree with each other. 

Prior to our Stage v there is no overlapping of the posterior 
incisors, which are small; and 24 is simply placed in front of 
i5, But in later stages the enlargement of the upper 
incisor teeth relative to the size of the jaw causes very 
considerable overlapping, and in Stages v, vi, and vir the 


DEVELOPMENT AND SUCCESSION OF TEETH IN PERAMELES. 507 


arrangement is a determinate and identical one. It is briefly 
as follows : 

i2 appears above the posterior portion of 74, but ona slightly 
more lateral (labial) plane. 

i2 appears directly above the posterior portion of 2. 

74 appears below the posterior end of 73, but in a slightly 
more lateral plane. 

15 appears mesially to, and on a higher plane than, 74. 

Now in Stage vir the above arrangement is departed from 
in that 74 appears not below but above 7, as well as laterally 
from it, and 7° appears directly below 7+ instead of being 
over it. 

We are unable to explain these differences (i.e. the weak 
development of 25 and the aberrant position of i4 and 7©) and 
can only set them down as instances of individual variability. 

Fig. 59 a represents the canine of the upper jaw, and may 
be compared with fig. 59, showing the same tooth in Stage v. 
These figures are especially intended to illustrate the gradual 
reduction of the dental lamina, including its “ residual’? pro- 
longation. Along with them may be taken fig. 598 from 
Stage vu, showing a still more advanced condition of 
resorption and degeneration. The figures are introduced 
apropos of Leche’s reiterated and emphatic insistence upon 
the long persistence and large size of portions of the residual 
dental lamina. We do not find these to be at all remarkable, 
though we have chosen for illustration the case of the canine, 
which provides one of the best marked and most persistent 
residual laminze amongst the anterior teeth. The three 
figures given simply show a progressive, if not very rapid, 
degeneration of the superfluous epithelial residua of the dental 
lamina. 

A residual dental lamina beside p1 and p2 is now very 
distinct. Beside dp? the Anlage of p= is labially knobbed, 
and at its fundus is present a distinct dermal papilla (cf. 
succeeding stages). Well-marked residual laminz are present 
by the lingual sides of the first two molars; and m® closely 
resembles the corresponding tooth-germ in Stage v1 (fig. 68 a). 


508 J. ‘I. WILSON AND J. P. HILL. 


Lower Jaw.—The dental lamina begins only a very short 
distance in front of the anterior plane of 7;, so that the ante- 
rior segment of the dental lamina referred to in the preceding 
stages is greatly foreshortened. 

Fig. 65 represents a coronal section passing through the 
dental lamina which just shaves the forwardly directed apical 
part of the first incisor. The dental lamina in this situation 
now shows a fairly well differentiated epithelial enamel-germ 
(7. e.0.), which we are inclined to interpret as that of a lost 
incisor in front of 7;. The facts relating to this important 
structure are deserving of very careful attention, and are as 
follows: 

During the period which elapses between Stage 111 on the 
one hand and Stages v and vi on the other, the dental lamina 
in front of <; appears to undergo an elongation from behind 
forwards. This anterior segment of the lamina exhibits, in 
Stages v and vi, a very well-marked thickening, which de- 
creases posteriorly where it is continuous with the residual 
lamina of 7;. 

In Stage vir the anterior growth of the large 77 has 
encroached upon the region in question, and accordingly the 
latter portion of the dental lamina is both relatively and ab- 
solutely shorter, antero-posteriorly, than it was in the imme- 
diately preceding stages. At the same time, however, it has 
become much more definitely thickened, and now seems to form 
the rudiment of a distinct enamel-organ with a flattened or 
slightly concave base (fig. 65). 

This enamel-organ now appears as if formed by the most 
anterior extremity of the residual dental lamina of iz. At this 
stage posterior sections of the rudiment in question would 
certainly suggest the idea that it is related successionally to 7;. 
But a consideration of the antecedent conditions of this struc- 
ture, arising as it does entirely independently of and quite in 
front of 7;, seems to us sufficient to dispose of such an inter- 
pretation in view of the very ample grounds, both general and 
special, for maintaining the truly successional nature of i; 


DEVELOPMENT AND SUCCESSION OF TEETH IN PERAMELES, 509 


itself. No one viewing our fig. 33 can doubt that 7, isin some 
sense successional to our vestigial diz, as there illustrated. 

It must be borne in mind that the first lower incisor of the 
adult is a large and forwardly growing tooth. We have already 
seen that its Anlage is at first placed behind that of its small 
and rudimentary milk predecessor, but that subsequently the 
latter appears as if displaced backwards, owing to the anterior 
extension of the large permanent tooth. In lke manner, we 
believe that this forward growth of 7; accounts for the note- 
worthy change in the later topographical relations of 2; to the 
rudimentary enamel-germ produced from a segment of the 
dental lamina, which in earlier stages is entirely in front of it. 

We think it most probable that the rudiment in question 
really represents an incisor which has been lost, in front of 
2,, and which appears in imperfect form, and for a short time 
only, during ontogeny, before the very great forward extension 
of 7; wholly prevents its further development. Its continuity 
posteriorly with the residual lamina of 7; according to our 
views proves nothing one way or another with re- 
gard to its successional character, for everywhere we 
find that it is the residual lamina in the region of a tooth 
which is the true morphological continuation of the undif- 
ferentiated dental lamina in front or behind it. 

It is interesting to compare the condition in Perameles 
with that described by Mr. Woodward for Petrogale (2). 
In that form he found two small calcified vestigial incisors. 
The anterior of these was entirely in front of the large per- 
manent lower incisor, which latter he therefore regarded as 
really an ?z. 

In connection with his vestigial first incisor (regarded by 
him as belonging to the first dentition) he describes and figures 
a lingual downgrowth of the dental lamina showing clearly 
the papillated enamel-germ of a successional tooth (see his 
fig. 10). Now we have shown (see our fig. 33) that our vestigial 
di, is related thus, not to a rudimentary successional tooth- 
germ, but to the adult 7;. It is evident, therefore, either 
that 7; of Perameles is not homologous to the adult per- 


510 J. T. WILSON AND J. P. HILL. 


manent lower incisor of Petrogale, or that our vestigial 
di is not homologous to the first vestigial incisor of Petro- 
gale. May it not be that an anterior milk incisor in Pera- 
meles has been entirely suppressed, and that its successful 
germ, serially homologous with 7;, attains a development in 
front of the first lower incisor, comparable to that of the suc- 
cessional germ connected with the first vestigial incisor of 
Petrogale? 

And may it not be that Woodward’s second vestigial incisor 
(his 7;), to which he was unable to recognise any successor, is 
really the homologue of our di; whose legitimate successor is 
Woodward’s iz, the almost certain homologue of our ‘‘7,” of 
the adult Perameles? 

An interesting modification in the relations of the vestigial 
first incisor (d?;) is seen in this stage. In the early stages it 
was seen to be attached intimately and directly to the neck of 
the main dental lamina opposite its connection with 2; (see 
fig. 83). In Stage v (fig. 66.) its attachment to the neck of 
the dental lamina is being, as it were, stretched away labially, 
thus tending towards the acquisition of an apparently inde- 
pendent attachment to the oral epithelium. This may be 
explained as effected, either figuratively or literally, by an 
opening out upon the deep surface of the oral epithelium of 
the most superficial part of the common lamina. In the 
present stage this process has been completed, and in fig. 66 
the vestigial di; is seen to possess an independent lamina of 
attachment to the deep surface of the oral epithelium labially 
from the line of attachment of the main lamina. 

It is such a process of opening out and labial dislocation as 
this, which we have already suggested as explaining the 
peculiar downgrowths on the labial side of the root of the 
dental lamina present in the first upper incisor region in 
Stages 1v and v (figs. 43—46 and p. 485). The present 
instance is an absolute proof of the possible occurrence of just 
such a dislocation of the primitive connection of a labial and 
more superficial member of the dental series. 

In this stage may be noted for the first time the differentia- 


DEVELOPMENT AND SUCCESSION OF TEETH IN PERAMELES. 511 


tion of the posterior secondary cusp of iz. Both in 7; and ty 
the anterior part of the tooth is most strongly calcified. The 
former tooth has the thickest dentine cap, and it extends, 
though somewhat thinned out, practically to the posterior 
limit of the papilla of the tooth. But in 7; the dentine cap 
does not yet cover the posterior part of the papilla. In both 
teeth (¢; and i) the very large and pointed papilla is seen to 
become less pointed as it is traced backwards, until its summit 
becomes flattened, and this flattening persists in iz till the 
hinder boundary of the papilla is reached. In the case of é5, 
however, the papilla, after remaining flattened for some 
sections, quite abruptly shoots up a second pointed apex, 
which, however, at this stage is destitute of any dentinal 
covering. This second point to the papilla is merely super- 
imposed upon the massive basal part of the papilla common 
to the entire tooth, and is a mere outgrowth of the primary 
papilla. 

There are, in the incisor region of our series of Stage vil, 
considerably richer remains of the dental lamina than in 
Stage vi. Indeed, in the former there is still a very well- 
developed residual lamina beside 2;, which retains its connection 
nearly all along with the oral epithelium, and extends into the 
region of 73 and 75. 

This relative immaturity also characterises the more posterior 
teeth in the lower jaw, including dp,, the dentinal covering 
of whose papilla is much thinner than that of the same tooth 
in Stage vi. We have already seen that in respect of the 
upper jaw the present stage exhibits some anomalous features. 

The Anlage of pz is large—even bigger than in Stage vi— 
and, as in the latter, exhibits a distinct indentation with 
rudimentary papilla. 

In the molar region mz is a large enamel-organ. As in 
Stage vi the anterior cusp alone as yet is provided with a 
dentinal cap. The dentinal lamina on its lingual side is 
deeply placed and quite unconnected with the oral epithelium. 
Before it terminates posteriorly it becomes markedly thickened. 
Opposite the anterior part of the tooth it forms a residual 


HQ J. T. WILSON AND J. P. HILL. 


lamina for mz, but posteriorly it becomes free from its con- 
nection with the tooth. 


Stage VIII.—P. nasuta: pouch specimen. 


Length from tip of snout to root of tail: ; . 82 mm. 
Head length . - 5 : « OO es 


Upper Jaw.—tThe incisors are all strongly calcified. Their 
relations to one another are much the same as in Stages v and 
v1, only 74 is practically in the same horizontal plane as 23 and 
placed labially. The fifth incisor, however, as in the stages 
mentioned, appears distinctly on a higher plane as well as 
lingually from i+. There is a hiatus between 72 and £ in 
which no tooth appears in the coronal sections. Here the 
dental lamina is present in a disintegrating condition. Its 
deepened character becomes apparent as it approaches the 
canine, and its distal portion is somewhat swollen but quite 
irregular in outline, due to invasion by connective tissue 
assisting in its disintegration. Occasionally traces of its ori- 
ginal connection with the Malpighian layer may yet be seen. 

The second premolar and the first molar have become so 
approximated that they partly overlap. The deciduous pre- 
molar occupies a position labially and superficially to the 
hinder part of the former and directly superficial to the an- 
terior end of the latter, all three teeth appearing together in 
a number of the coronal sections. The Anlage of p3 is strongly 
developed by the lingual side of dp#, especially opposite its 
posterior part. The residual dental lamina, of which it forms 
the deepest part, has lost its superficial connection with the 
oral epithelium, though traces of such a connection here and 
there appear. The greater part of the superficial moiety of 
the lamina shows the extreme irregularity of outline due to 
progressive disintegration, and a number of rounded cell- 
nests or ‘epithelial pearls” are developed in connection with 
it. The deeper (distal) portion, forming the Anlage of p4, is 
very thick, and is found to develop quite abruptly a labial 
lobe. A few sections behind, this is seen to form the labial 


DEVELOPMENT AND SUCCESSION OF TEETH IN PERAMELES, 518 


boundary of a cupped depression filled with connective-tissue 
cells. The whole Anlage is surrounded by an envelope of con- 
densed connective tissue, and is separated by a bony lamella 
from the overlying hinder part of p2. No drawing is given, 
but the condition may be understood by reference to fig. 67 
(Stage 1x) and fig. 73 of Stage x. The deciduous premolar 
persists in the sections for some distance behind the point of 
disappearance of p? and beneath the rapidly enlarging m+. 

In the region of the first two molars traces of the dental 
lamina persist, sometimes as small irregular groups of epithelial 
cells, and sometimes as large and rounded “ epithelial pearls.” 
The lamina, however, assumes more definite form posteriorly 
opposite the hinder end of m2. 

The third molar is still very much in the condition it ex- 
hibited in the immediately preceding stages (fig. 69), i.e. it 
shows commencing papillation, somewhat more marked than 
formerly ; but the more superficial (more proximal) part of 
the lamina now exhibits a number of irregularities in place of 
the simple labial projection corresponding to the root of the 
laminar connection for m2. These irregularities are secondary 
developments, probably degenerative in character, but they 
might easily be taken for evidences of plurality of papille 
were their history not accurately traced. The connective 
tissue is condensed around the proper Aulage of m®, and 
shows a true rudimentary dermal papilla filling the basal 
depression of the Anlage. No such condensation of papillary 
differentiation is present in relation to the pseudo-papilliform 
modifications of the proximal part of the lamina referred to 
above (fig. 69). 

Lower Jaw.—In the series at our disposal the region in 
front of i; was not available. The anterior region of 7; itself 
was also injured. Abundant traces of a fairly bulky residual 
lamina undergoing disintegration were, however, discernible, 
and these are continued into the region of 7,. 

The deciduous first incisor di; is still in evidence as a small 
and strongly calcified toothlet. ‘The second incisor now ex- 
tends forwards some distance in front of tz. The base of i; 


514, J. T. WILSON AND J. P. HILL. 


extends backwards for a considerable distance below both the 
succeeding teeth, and the third now extends considerably be- 
hind the second incisor. 

The deciduous premolar is now calcified right down to its 
root. The Anlage of pz exhibits a distinct but small and 
simple dermal papilla, while the proximal (“neck”) part of 
the lamina is irregular and contorted from the operation of 
degenerative processes. The three anterior molars are well- 
formed calcified teeth, and the dental lamina towards the 
posterior end of mz becomes well developed as a residual 
lamina for that tooth, and then exhibits the differentiating 
Anlage of mz. 


Stage [IX.—P. nasuta: pouch specimen. 


Length from tip of snout to root of tail 5 . 91mm. 
Head length. : 5 : . 40 


Upper Jaw.—The incisors are now more elongated and 
relatively narrower than in the preceding stages. Hach 
appears, in coronal sections, above the hinder part of the tooth 
in front, overlapping the latter considerably. The relation of 
15 to i4 is similar to that of 74 to 73, &c. 

The tips of all the incisors le near the oral epithelium. 
The canine appears above the hinder part of 23, overlapping it ; 
p+ occupies a similar relation to *, and p? to p+. 

Up to the premolar region the dental lamina is represented 
only by fragmentary and degenerated remains in the shape of 
broken-up and isolated groups of epithelial cells, which are 
best marked beside 72 and‘. In the anterior premolar region 
these degenerated remains continue to appear, though very 
sparsely in the region of the anterior part of the second pre- 
molar. After dp? appears in the section the remains of the 
residual dental lamina become more definite, and soon are 
arranged in definite laminar form, but are deeply placed and 
quite unconnected with the oral epithelium ; the relations of 
dp to the hinder part of p2 and to the residual dental lamina 
with the Anlage of p® are illustrated in fig. 67. The most 


DEVELOPMENT AND SUCCESSION OF TEETH IN PERAMELES. 515 


anterior sections of dp*, which cut the apical part of the 
tooth somewhat obliquely, show that it has already cut the 
gum. As it is traced backwards the crown recedes from the 
surface, and its crown assumes the appearance shown in the 
figure (fig. 67). 

More posteriorly it is narrower than in the drawing, but is 
still much elongated. In this stage dp® is seen to be a very 
much smaller tooth than the anterior premolars. It does not 
appear in the sections until behind the plane of the maximum 
development of p2, but nevertheless the posterior part of p2 
extends backwards slightly behind it, being still visible when 
dp> has totally disappeared. 

The character of m2 is shown in fig. 70. It exhibits only a 
very slight advance on the preceding stages of the same tooth. 

Lower Jaw.—In sections passing through the anterior 
tip of 7; the same rudimentary enamel-organ is encountered 
as was noted in Stages 1v—vit. It is, however, more irre- 
gular than in Stage vir, and is evidently undergoing resorp- 
tion, without having shown the least trace of calcification. 

The tip of 7; has now grown far forward by the labial side 
of z;, and has insinuated itself between the latter tooth and 
its vestigial milk predecessor diz. This is a clear case of 
relative displacement of ¢; similar to that which we have 
shown to occur in the case of 2; with reference to the rudi- 
mentary enamel-organ morphologically in front of it. 7, at 
this stage shows the now calcified posterior secondary cusp 
very distinctly. di; is in statu quo as regards its struc- 
tural characters, only its relative topographical position is 
slightly altered by forward growth of iz. 7; appears further 
back by the labial side of 7,. It bears the same topographical 
relation to 7; as the latter does further forward to 7;, so that 
all trace of the early peculiarities in the relative situation of 
these teeth is now lost. 

The basal part of 7; may be traced far back, lingually to, 
and in a deeper plane than, 7; and iz, all three appearing to- 
gether in the sections throughout a considerable extent of the 
anterior part of the jaw. 

VOL. 89, PART 4,—NEW SER. NN 


516 J. T. WILSON AND J. P. HILL. 


dp is now a narrow tooth much elongated vertically, which 
has cut the gum. Its residual lamina first becomes discernible 
opposite its anterior end. 

The Anlage of pz; is now more bulky than in preceding 
stages. It is placed opposite the hinder part of dpz, and 
though of larger size exhibits only a very slight depression 
answering to the cupping of the enamel-organ ; and here its 
papilla can hardly be said to exist as such, though the con- 
nective tissue is evidently tending towards its formation by 
local cellular accumulation opposite the depression. 

In the posterior molar region the dental lamina ends as a 
free laminar mass by the lingual side of mz, and in front of 
the middle of that tooth. The Anlage of mz does not seem to 
have made progress in the interval which separates this from 
the preceding stage. 


Stage X.—P. nasuta. 

Length from tip of snout to root of tail ; . 118 mm. 
Head length . : Se SON: 

This stage is distinctly eamtior naan on the whole than 
the preceding one, but the deciduous premolars are only just 
cutting the gum, while in Stage 1x they had already protruded 
somewhat. 

In the upper jaw no teeth have undergone eruption save dp3, 
but in the lower jaw the lower incisors have broken through in 
addition to the deciduous premolar. 

In the region of the canine there are still to be found abun- 
dant remains of the residual dental lamina in the shape of 
detached cell-masses, embedded in the fibrous connective tissue 
lingually from the well-developed canine. These disappear 
and reappear frequently in the serial sections, and, in various 
sections widely separated from one another, one of the detached 
cell-masses assumes a very definite pear-shaped form, and is 
surrounded by a distinct connective-tissue capsule (fig. 72). 
This pear-shaped mass is always the deepest mass present, and 
obviously corresponds to the fundus-portion of the residual 
lamina. Its appearance is highly suggestive of the “ knospen- 


DEVELOPMENT AND SUCCESSION OF TEETH IN PERAMELES. 517 


formig” tooth-germs of other writers. But the fact that 
such appearances occur intermittently throughout nearly 
the whole extent of the lengthy canine region seems sufficient 
to negative this idea. We can point to quite a number, no 
less than seven, absolutely separate and distinct “bud- 
hike enamel-germs” of this sort, existing in antero-pos- 
terior series by the side of the canine of this stage. Each of 
these is fairly bulky, pear-shaped, and concentrically encap- 
sulated. It is beyond doubt that they are nothing more than 
portions of the swollen free distal margin of the residual lamina 
which persists for some time after the antero-posterior con- 
tinuity of that margin has been interrupted by regressive pro- 
cesses. It is equally difficult to imagine a production of several 
rudimentary “ bud-like enamel-germs of replacing teeth” by 
the side of the single canine, or to imagine a primitively single 
“bud-like”’ germ subdivided at intervals into a lengthy longi- 
tudinal series of bud-like masses. 

Let it be specially noted that in the case before us we have 
presented the phenomenon of “knospenférmig” cell-masses, 
long persistent, and distinctly encapsulated by the surrounding 
connective tissue, i.e. offering the only credentials to which 
Leche appeals for the determination of such structures as true 
tooth-germs. The demonstration of their multiplicity in such 
a case as the present seems to us to be the reductio ad ab- 
surdum of the replacing-enamel-germ theory. 

It seems to us that the presence of these structures at a 
comparatively advanced stage of the canine tooth development 
is largely explicable by reason of the considerable bulk of the 
entire canine Anlage, to which attention has frequently been 
called in these pages. The notable size of the canine lamina 
is doubtless associated with a higher degree of formative 
activity, which is longer in being exhausted here than else- 
where. In any case we can attach no special morphological 
significance to the continued presence at this period of even 
such definite epithelial structures as those described. 

The two anterior premolars are large pointed teeth, and cause 
marked projection of the gum. Vestiges of the residual lamina 


518 J. T. WILSON AND J. ‘P. HILL. 


are here and there discernible by the lingual side of the teeth 
as in the canine region, but on a smaller and less impressive 
scale. Here and there the form of the residual lamina (pear- 
shaped, with long stalk) is preserved, but for the most part the 
vestiges—generally absent altogether—are in the form of scat- 
tered minute epithelial cell-groups. 

The deciduous premolar is in the act of cutting the gum with 
the most anterior part of its crown. 

The process of eruption is very clearly seen to be attended 
by the flattening out of the enamel epithelium covering the 
tooth crown and its conversion into squamous epithelium with 
concomitant formation in it of epithelial “ nests’ or “pearls” 
close to the tooth cusps. This nest-formation also proceeds in 
the oral epithelium over the tooth. Eventually the epithelial 
pearls become placed quite near the surface, and they then 
undergo rupture and disintegration, their cores of concentric 
epithelial cells being lost and the tooth crown exposed. These 
different stages can be plainly followed in this one case by 
tracing dp? backwards from its point of actual eruption to- 
wards the posterior more deeply placed part of the tooth, where 
the preparations for eruption are proceeding. The cell-nests 
here noted as forming in the enamel epithelium are in all pro- 
bability similar to those epithelial nodules noted by Poulton 
(25) as appearing in the “ most superficial part of the middle 
membrane of the enamel-organ immediately over the apex of 
each chief cusp of the large broad posterior tooth” of Orni- 
thorhynchus (see his pl. ii, figs. 4,5, 11, and 12, and pl. 
ili, fig. 6). It may be noted, however, that in this stage of 
Perameles the tooth is very much more advanced than those 
figured by Poulton in Ornithorhynchus, and the middle 
stratum of the enamel epithelium can no longer be distin- 
guished as such. All we can say is that in the midst of the 
epithelial investment of the tooth-crown, formed by the now 
flattened enamel epithelium (as well as in the neighbouring 
oral epithelium), these cell-nests are formed, and that the 
process of eruption of the tooth under notice is accompanied 
and so far conditioned by their denudation and disintegration. 


DEVELOPMENT AND SUCCESSION OF TEETH IN PERAMELES. 519 


It is noteworthy that, in respect of its enamel formation, 
dp in Perameles displays a striking degeneracy. At no 
time do the cells of its inner enamel layer attain any great 
degree of elongation as compared with those of other teeth, 
and the layer of enamel actually formed is a very thin lamella 
indeed. 

The Anlage of p2 with its distinct though small papilla is 
well represented in fig. 73. 

In the molar region the residual dental lamina has prac- 
tically disappeared from the first molar, but it is present by 
the sides of m2 and of m2, in which latter situation it is shown 
in fig. 71. 

This section passes through the plane of the hinder end of 
m3, which lies considerably to the labial side of the part 
figured. It will be observed that the connective tissue shows 
a definite concentric arrangement around the residual dental 
lamina in this figure. Such an arrangement persists through- 
out the whole extent of the residual lamina, though it is only 
considerably behind the plane figured that it constitutes the 
Anlage of m4. 

Lower Jaw.—The first two incisors have already completely 
broken through the epithelium at their apices. The anterior 
primary cusp of 72 is just in process of eruption, and here there 
may be observed at its tip a follicular or nest-like arrangement 
of epithelial cells which is in the act of opening out on the 
surface by disintegration, as was noted in the case of dp> in 
the upper jaw. 

di; is still observable, but it is now more degenerate in 
character, merely consisting of asolid dentinal nodule, in which 
no trace of a papilla or cellular core can be recognised. 

dp; is in process of eruption, and this is associated with cell- 
nest formation between the surface and the erupting tooth 
point. The Anlage of p; resembles that of the upper jaw 
pretty closely. 

The fourth molar is a simply but deeply papillated enamel- 
organ with the middle layer of stellate tissue well developed. 
A residual dental lamina has already appeared lingually as a 


520 J. T. WILSON AND J. P. HILL. 


result of the differentiation of the anterior part of mz; the 
hinder part of the tooth is not yet constricted off from the 
parent lamina. 


Stage XI.—P. nasuta: pouch specimen. 


Length from tip of snout to root of tail : . 127 mm. 
Length of skull . : ‘ ‘ ~ AS Ss 


This stage is only slightly, if at all, older than Stage x. 
Indeed, in certain respects development is less advanced, for in 
the upper jaw no teeth have yet broken through the gum, 
though the canine and premolars are causing very marked 
projection of the latter. 

Both the deciduous premolar and the Anlage of p3 closely 
resemble those of the preceding stage. The proximal part 
(or “‘neck”’) of the residual lamina, where that forms the 
Anlage of p3, offers a very striking example of cell-nest for- 
mation as a comcomitant of the process of disintegration of 
the lamina. Here we have the papillated Anlage of p2 borne 
upon an attenuated lamina, whose proximal border towards 
the oral epithelium exhibits a number of large globular epithe- 
lial swellings showing the concentric arrangement of the con- 
stituent cells characteristic of epithelial “ pearl ” formation. 

The molars resemble those in Stage x. 

In the lower jaw diz is in the same degenerate condition as 
in Stage x, but the dentinal nodule representing it is even 
smaller. Otherwise the lower jaw presents no noteworthy 
features. 


Stage XII.—P. obesula: pouch specimen. 


Length from tip of snout to root of tail ‘ . 120 mm. 
Length of skull . : : 5 s Agen. 

The lower jaw only of this specimen was studied. Though 
its dimensions do not exceed those of Stage x1, it represents 
a decidedly older stage. This is explainable by difference in 
specific characters. 

The Anlage of pz still shows a rudimentary papilla project- 
ing into the slightly cupped enamel-germ. The whole Anlage 


DEVELOPMENT AND SUCCESSION OF TEETH IN PERAMELES. 521 


has an irregular appearance in many sections, owing to the 
presence of irregular appendages derived from the “ neck” 
portion of the dental lamina. 


Stage XIII—P. nasuta: pouch specimen. 


Length from tip of snout to root of tail ‘ - 180 mm. 
Length of skull . A ; : Gnas 

It will be seen from fig. 75, representing the upper jaw of 
this stage, that all the antemolar teeth (except p®) have cut 
the gum, and most of the general characters of the adult den- 
tition are recognisable. Of the molars two in the upper jaw 
and three in the lower have broken through. The very small 
dp is seen occupying the slight interval between p2 and m1. 

The Anlage of pz in the lower jaw is shown in section in 
fig. 74. It is now a well-developed cupped enamel-organ, 
and, in one of the two series cut, the dermal papilla is much 
larger than in the section figured, filling up the whole of the 
somewhat deeper cup of the enamel-germ. 

In the latter the stellate tissue of the middle layer is now 
well developed. In fig. 74 it will be observed that the enamel- 
organ is now in process of constriction off from the lamina, 
whose free end is already beginning to project as a new resi- 
dual lamina (rdl.). That this interpretation is the correct one 
is proved by reference to Leche’s figs. 140 and 142 (here- 
with reproduced for comparison as figs. 76 and 77), showing 
the corresponding enamel-organs of p? in Phascolarctus. 
In connection with the latter a very evident and swollen 
residual dental lamina has anew developed itself by the side 
of p3. This, according to Leche, provides for the possibility 
of the production of a tooth of the third (his “ fourth ”’) denti- 
tion. We believe that this possibility must not be lost sight 
of, but in the meantime we prefer to point out that the regular 
development of such an Anlage as has been discovered and 
figured, both by Leche and by ourselves, by the lingual side 
of p&, serves in the most striking way to confirm our view of 
the homology of p? to the other adult antemolar teeth, whose 
residual dental lamine (‘‘Hrsatzleisten ”’—auct.) are there- 


522 J. TIT. WILSON AND J. P. HILL. 


fore merely equivalent to that which later develops 
beside the third permanent premolar. 

It may be noted that in Phascolarctus p? appears to 
differentiate at a much earlier period than does the corre- 
sponding tooth in Perameles; for though the stage figured 
by Leche is considerably less advanced in its general develop- 
ment, and as regards the rest of its dentinal characters, than 
the present stage of Perameles, yet the Anlage of pz has 
attained a somewhat higher degree of organisation. 


Stage XIV.—P. nasuta: dried skull of a young adult 

about three quarters grown. (See figs. 78 and 79.) 
Skull length : 5 

Here (fig. 78) in the upper jaw the fourth molar is in course 
of eruption, and the premolar tooth-change is now almost 
completed. The third permanent premolar, p2, has broken 
through, and its crown has nearly attained the level of the 
anterior premolars. By the eruption of its posterior portion 
it is displacing the small dp®. On the right side the latter 
tooth is missing, but may have been accidentally lost. On the 
left side dp2 is still present, lodged in the slight hollow over 
the posterior end of p3. In the lower jaw (fig. 79) the erup- 
tion of pz is not so far advanced, but the relations are other- 
wise entirely similar to those on the upper jaw. The crown 
of dp? is worn and flattish or slightly hollowed. In this con- 
nection we may note that Flower (4, p. 635, and Pl. XXX, 
fig. 1) describes and figures the tooth-change in Perameles 
in an animal not quite full grown. His description of the 
condition is as follows :—‘‘ The permanent incisors, canines, 
and two anterior premolars are in place. Behind these in each 
jaw is a very minute, rather compressed tuberculated tooth, 
succeeded posteriorly by the true molars of the permanent 
series. In the alveolus above this minute tooth, which is the 
temporary or deciduous molar, is lodged the germ of the 
posterior permanent premolar, a tooth having a large com- 
pressed, pointed, triangular crown, with small anterior and 
posterior basal tubercles.” (PI. XXX, fig. 1.) 


70 mm. 


DEVELOPMENT AND SUCCESSION OF TEETH IN PERAMELES. 523 


PART III. 


Discussion oF GENERAL PROBLEMS OF THE MARSUPIAL 
DENTITION IN THE LIGHT OF THE FOREGOING OBSER- 
VATIONS. 


Our own conclusions may, perhaps, be most conveniently 
introduced in the course of a discussion of the views of pre- 
vious writers—already partly outlined in the introductory 
portion of this paper. And in order to focus the discussion, 
and to set forth its issues more clearly and definitely, we have 
thought it useful to formulate certain leading questions, and 
to define the attitude which has been assumed towards these 
by other observers whose views are related to our own either 
by way of agreement or contrast. In the formulation of these 
questions we have pretty closely followed the lines, and even 
in part the language, of writers who, like Thomas, Schwalbe, 
and Leche, have so greatly advanced the intelligent discussion 
of the subject of mammalian dentition. 

I. “Do the existing teeth of Marsupials in front of 
the last premolar correspond to those of the milk or 
of the successional dentition of higher mammals?” 

It will presently appear that the discussion of this question 
involves a consideration of those criteria which may enable us 
to determine a given embryonic structure as the germ of a 
successional tooth. 

II. How are we to explain the single tooth-change 
or succession (i.e. in the case of the last premolar), 
hitherto almost universally recognised as the sole 
instance of tooth-change among Marsupials? 

III. To which dentition belong the molars of Mar- 
supials and of mammals generally? and— 

IV. What is the nature of that generally admitted 
process of evolution by which multicuspidate teeth 
have been derived from a primitively simple uni- 
cuspidate type? 


024 J. T. WILSON AND J. P. HILL. 


I. Homology of Antemolar Teeth of Marsupials. 


In reference to the first question, it has already been pointed 
out that Flower (and formerly Thomas) held that the non- 
changing antemolar teeth of Marsupials answered to the 
permanent or successional teeth of the Eutherian orders. This 
determination was adopted with great unanimity up till the 
period of publication of Kiikenthal’s discoveries, even by those 
who dissented on other fundamental points from the Flower- 
Thomas hypothesis regarding the dentition of Metatheria. 

But from the time that Kiikenthal first recognised the per- 
sistence of somewhat swollen (‘‘kolbig ”) downgrowths of the 
dental lamina at the lingual sides of the developing teeth of 
Marsupials, these downgrowths have on every hand been re- 
garded as the true equivalents of the Eutherian successional 
series, and the entire theory of the marsupial dentition has 
accordingly been remodelled to suit the newer interpretation. 
Up to the present, so far as we know, not a single investigator 
has ventured to proclaim his dissent from the latter-day creed 
in this respect.1 

If, then, we proceed to inquire upon what foundation this 
belief is based, we find that it rests solely on the occurrence, 
in Marsupials, of the epithelial downgrowths referred to, taken 
along with the similarity of the latter (@) to the earliest known 
rudiments of successional enamel-germs in other mammals, 
and (4) to the earliest stage of the one successional tooth 
admittedly existing in the marsupial jaw. But the question 
cannot, we believe, be thus summarily disposed of. It must 
be subjected to a much more thoroughgoing criticism, such 
as is embodied in the very weighty and thoughtful discus- 
sion of the entire subject by Professor Leche in his recent 
monograph (8). 

In that work Leche has pointed out that, in order to decide 
in any given case whether a tooth belongs to one or other 
dentition, certain criteria have been employed whose worth 
and validity are of very different degrees. He attempts to 

1 But see addendum, p. 581. 


DEVELOPMENT AND SUCCESSION OF TEETH IN PERAMELES. 525 


determine which of these may legitimately be utilised in 
arriving at such a decision. 


(1) The Criteria of Serial Tooth Homology as set 
forth by Leche. 


We propose first of all to summarise Leche’s treatment of 
this fundamental problem. After reviewing his position the 
extent of our agreement and difference with him will become 
apparent in the course of discussion. 

In the first place Leche discusses (a) the significance of 
the occurrence of epithelial downgrowths of the 
dental lamina lingually to developing enamel-organs. 

That the presence of swollen epithelial downgrowths in- 
ternal to the enamel-organs of mammalian teeth was sufficient 
to stamp the latter as belonging to the milk dentition, no one 
apparently, until Leche, had thought of questioning. 

Woodward (2), indeed, is somewhat guarded in his language 
when he writes, “If these various and often minute down- 
growths of the dental lamina are to be interpreted as repre- 
senting rudiments of teeth, as seems probable from comparison 
with known rudiments of the first or second dentition in other 
mammals,” &c.; but, nevertheless, he appears to accept in 
their entirety those dentitional theories which depend wholly 
upon the determination in question. 

Leche, however, enters fully into the question of the validity 
of the assumption that Kiikenthal’s swollen downgrowths 
actually represent successional enamel-germs; and his con- 
clusions are by no means decisive in favour of that assumption. 
The following passages will serve to illustrate his attitude :— 
‘“Da, wie bereits erwahnt, die Differenzirung des Schmelz- 
keimes ausschliesslich oder doch vorzugsweise an der labialen 
Flache der Schmelzleiste erfolgt, so ist auch von vorneherein 
zu erwarten, dass das tiefe Ende der Schmelzleiste lingualwarts 
vom Schmelzkeim auftritt. Ist also jene ‘ Knospe,’ welche 
von demselben verdichteten Mesodermgewebe welches das 
Zahnsackchen bildet, umgeben ist (figs. 9, 10), nichts anderes 


526 J. T. WILSON AND J. P. HILL. 


als das zuerst sichtbare Product des Abschniirungsprocesses 
des Schmelzkeimes von der Schmelzleiste, so legt schon diese 
Thatsache den Schluss nahe, dass die ‘Knospe’ nicht, wie 
noch mehrfach auch von denneuesten Autoren angegeben 
wird, an und fiir sich identisch mit einem Schmelzkeim, re- 
spective einer Zahnanlage sein kann. Dieses giebt auch 
daraus hervor, dass, wie die vorstehenden Untersuchungen 
lehren, die Entstehung einer ‘ Knospe’ nicht an eine be- 
stimmte Dentitionsreihe gebunden ist: sie tritt nicht nur neben 
den typischen Milchzaihnen sondern auch neben solchen Zahnen 
auf, die in der Regel ohne Nachfolger sind wie die Ersatzzahne 
(figs. 55, 78, 79, u.a.) und die Molaren (text—fig. 2).” 

The substance of this criticism is frequently repeated by 
Leche, thus (8, p. 183):—**Steht es somit fest, dass das 
Auftreten einer ‘ Knospe’ zunachst nur den beginnenden 
abschniirungsprocess des Schmelzkeimes von der Leiste kenn- 
zeichnet, ohne dass dadurch unbedingt ein neuer Zahn zu 
Stande kommt,” &c.; and again (3, p. 1386), “ Das Vor- 
kommen einer ‘ Knospe’d. h. das mehr oder weniger frei 
hervortretende Schmelzleistenende neben einen Schmelz- 
keime keineswegs beweist, dass der letztere zur ersten Den- 
tition gehort. Ja, wir konnen weiter gehennicht einmal die 
weiterbildung dieser ‘ Knospe’ zu einem wirklichen Schmelz- 
keime berechtigt zu dem Schlusse, dass der mit einer solchen 
Zahnanlage ausgestattete altere Schmelzkeim unbedingt der 
ersten Dentition angehort, da, wie wir aus den obigen Unter- 
suchungen wissen, auch lingualwarts von typischen und 
unbestrittenen Reprasentanten der zweiten Dentition solche 
knospenformige Schmelzkeime vorkommen ké6nnen, welche 
sich in einigen Fallen zu vollstandigen Zahnen ausbilden.” 

Having thus provisionally disposed of the assumption that 
the mere presence even of a ‘ bud-like” lingual downgrowth 
of the dental lamina is sufficient to establish the “milk” 
homology of the tooth beside which it is found, Leche goes on 
to pass in review other criteria which might be taken as afford- 
ing the necessary test of a supposed serial homology. He 
dismisses (b) “simultaneousness of function” as afford- 


DEVELOPMENT AND SUCCESSION OF TEETH IN PERAMELBS. 527 


ing quite incompetent evidence of serial homology between 
individual teeth, on the ground that in some mammals (notably 
in Erinaceus) certain of the milk-teeth function along with 
successional teeth. 

Again, (c) the different depth at which one enamel- 
germ springs from the dental lamina compared with 
another can, in Leche’s view, afford no reliable criterion of 
serial homology. Thus in Erinaceus the dental lamina in 
the situation of several of the antemolar teeth is markedly 
shallowed, and is wholly transformed into an enamel-organ 
very superficially placed. But it is found that the dental 
lamina, at the places referred to, has to produce only one 
tooth, and the anomalous position of the enamel-germ is held 
to be explained merely by the absence of tooth-change in that 
particular segment of the jaw. 

While agreeing in general with the judgment expressed 
above, we are bound to add that the reasoning appended to 
it is not conclusive to us, for the absence of a tooth-change 
will not account for discrepancies of situation in many cases, 
e.g. the striking depth of the upper canine Anlage frequently 
referred to in this paper may be contrasted with the relative 
shallowness of the enamel-germs of the two anterior premolars 
which are without doubt serially homologous with the canine. 
On the other hand, we have found that enamel-organs which 
at first are close under the oral epithelium, and which have 
(as in Leche’s description) no dental lamina or only a very 
short one, superficial to them, do afterwards exhibit quite a 
well-marked dental lamina, and from this by-and-by a residual 
lamina is differentiated. This is what happens in the case of 
the anterior premolars of Perameles, though they never attain 
the great depth of the canine Anlage. 

We are of opinion that the relative depth of the various 
enamel-organs depends upon several factors, partly, perhaps, 
intrinsic to the lamina itself, and dependent upon the degree 
of its formative activity as an epithelial mass; and partly 
extrinsic, and dependent upon its topographical relations to 
various other structures. 


528 J. T. WILSON AND J. P. HILL. 


(d) The time of development of a tooth, in so far as 
this takes account merely of the attainment of maturity and of 
the period of eruption of a tooth, fails, in Leche’s judgment, 
to provide the necessary evidence, and constitutes a quite un- 
reliable test of the serial identity of different teeth. 

This test is, however, constantly relied upon by Rose, appa- 
rently without any misgivings as to its validity. Thus, con- 
cerning the last premolar and third incisor of Phalangista 
Cookii, he says (1, p. 702), ‘‘ Diese Zihne sassen noch tief in 
ihren Alveolen, die Farbe und Dichtigkeit ihrer Schmelzkappe 
wies darauf hin, dass sie von bedeutend jiingerer Bildung 
sind als ihre Nachbarn. Von einem Zahnwechsel, d.h. der 
Resorption eines Zahnes der ersten Reihe konnte auch hier 
keine Andeutung angefunden werden. Es scheint mir sehr 
wahrscheinlich, dass auch bei Phalangista Cookii nicht 
allein der letzte Pramolar, sondern auch der dritte Incisivus 
des Oberkiefers aus der zweiten Zahnreihe entsteht,” &c. 

(e) But, on the other hand, the time of development of 
a tooth takenin a different sense furnishes a test which 
Leche recognises as of the highest importance, if not of para- 
mount authority. Its value depends upon the validity of the 
following generalisation :— Die Anlagen der zu derselben Den- 
tition (Zahngeneration) gehohrigen Zahne differenziren sich 
gleichzeitig oder nahezu gleichzeitig an der Schmelzleiste” 
(3, p. 137).: 

To this law of contemporaneousness of origin of the mem- 
bers of the same dentition Leche does indeed attach certain 
qualifications, but it nevertheless remains for him the most 
valid criterion in the determination of the serial homologies 
of the different teeth. 


(2) The Serial Homology of the Marsupial Antemolar 
Teeth as tested by Appropriate Criteria. 


We may now turn to the results of Leche’s application of 
his critical method just expounded to the general question 
under discussion, viz.: Are the existing teeth of Marsupials 
in front of the last premolar homologous to the first or to the 


DEVELOPMENT AND SUCCESSION OF TEETH IN PERAMELES, 529 


second dentition of higher mammals? And it is noteworthy 
that Leche’s conclusions are to a very large extent negative as 
regards the value of just those developmental features upon 
which alone Kiikenthal and Rose relied to establish the “milk” 
homology of the non-changing antemolars of Marsupials. 

It had simply been assumed by Kiikenthal and Rése that 
the presence of a residual lamina (‘‘ Ersatzleiste” of Rise), 
more or less swollen, by the lingual side of an enamel-organ 
gave sufficient assurance of the ‘ lacteal” character of the 
latter. Leche’s weighty and, as we believe, entirely valid con- 
tentions have, at any rate, opened the door to scepticism with 
regard to this doctrine. It remains to be seen upon what 
grounds Leche after all gives in his adhesion to a view which 
has been rendered assailable with the weapons he has himself 
provided. 


The Nature and Morpologieal Value of a 
“Residual Lamina.” 


In one of the numerous passages in which he discusses the 
morphological value attachable to the liberated residual dental 
lamina, Leche utters his opinion as follows :—‘* Das durch die 
Emancipation des Schmelzkeimes freigewordene Schmelz- 
leistenende bildet an und fir sich nur die Voraussetzung fiir die 
Entstehung einer Zahnanlage : je bedeutender dieser Schmelz- 
leistentheil ist, desto grésser ist die Pradisposition fiir die 
Bildung eines neuen Zahnes. Und wir k6onnen hinzufigen: 
ist dieses Schmelzleistenende wirklich kolbenoder knospen- 
formig angeschwollen und von einem Zahnsackchen umgeben 
erst dann ist diese Moglichkeit als realisirt zu betrachten, erst 
dann konnen wir von einer (knospenformigen) Schmelzkei- 
manlage reden, einerlei ob diese Anlage sich spater weiter 
entwickelt oder nicht ” (8, p. 134). 

It will be remarked that this passage is by itself quite inde- 
cisive. It merely expresses the author’s view that any free 
end of the dental lamina swollen (i.e. “kolbig” on section), 
and surrounded by an investment of condensed connective 
tissue, has a right to be regarded as an enamel-germ, though 


530 J. T. WILSON AND J. P. HILL. 


it may be an abortive one. And, since the author holds that 
such abortive germs are at least occasionally met with beside 
the true successional teeth of higher mammals, it is evident 
that we are not enabled by their mere presence to 
determine which tooth generation, if any, they re- 
present. 

Elsewhere, however, in a very important passage Leche 
explicitly sums up the argument in favour of the true “suc- 
cessional” character of the lingual downgrowths beside the 
various marsupial teeth in the following words: 

*“‘ Charakteristisch fiir die Beutelthiere sind also sowohl das 
constante Vorkommen und die scharfe Auspragung dieser 
Schmelzkeime und ihre Uebereinstimmung mit dem Schmelz- 
keim des p. 3, als auch ihre lange Permanenz, welche Higen- 
schaften diese Gebilde nicht unwesentlich von den lediglich 
durch die Emancipation der Zahnanlagen von der Schmelzleiste 
entstandenen ‘ Knospen’ unterscheiden, ein Punkt, den ich 
hier ganz besonders betonen méchte. An einigen dieser Schmelz- 
keime sind deutliche Zahnsackchen vorhanden” (38, p. 103). 

Now we cannot but feel that if Leche’s criticisms detailed 
above be well grounded, and if the sceptical lesson he repeatedly 
inculcates have taken hold upon the reader, the latter will find 
it difficult to discover in the summary quoted anything at all 
sufficient to reassure him of the certainty of that identification 
which the author advocates. 

It is difficult to understand how “ constancy of occurrence ” 
—‘‘a sharply stamped character ’’—and “long persistence ” 
can possibly constitute a ‘not unessential” difference. To 
us there appears nothing at all “essential”? about such dis- 
tinctions, and thus far we must hold that Leche’s own verdict 
ought to have been, at most, “ not proven.” 

In one of the passages just quoted from Leche (on p. 108) 
we find him asserting that whenever the free end of the 
(residual) dental lamina is actually swollen in a “kolben” 
or “ knospenformig” manner, and when in addition we can 
recognise the presence of an investment of condensed connective 
tissue to form a “‘ tooth-sac,” then we have a right to speak of 


DEVELOPMENT AND SUCCESSION OF TEETH IN PERAMELES, 531 


the structure as an “‘ enamel-germ,” whether it develops further 
or not. 

With regard to the first of these two alleged guarantees 
of identity as genuine enamel-germs, it is held by the present 
writers that undue importance has been attached to the dis- 
tinction between the free end of the dental (residual) lamina 
while in the condition of a mere prolongation beyond the de- 
veloping enamel-organ, and the same prolongation when it has 
undergone some slight enlargement. With most writers, it is 
true, that distinction is not a very prominent one. Indeed, it 
would appear that, to the majority of them, practically any 
lingual prolongation of the dental lamina beyond its point of 
connection with a developing tooth is to be regarded as an 
“Krsatzleiste,” and therefore as the homologue of a suc- 
cessional tooth. In Kikenthal’s earliest reference to the 
marsupial condition (24) we find him basing his view of the 
“milk ” homology of the marsupial incisors upon the possession 
by them of an “Ersatzleiste” with a swollen (“kolbig’’) 
free end. On the other hand, neither Rose nor Woodward lay 
any particular stress upon the enlargement of the free end of the 
residual laminal downgrowth. They do not ignore its frequent 
occurrence, but neither do they seem to regard that feature as 
asine qua non of its title to consideration as an enamel- 
organ. The passage already quoted from Woodward (p. 525) 
will sufficiently bear out this statement as far as that writer is 
concerned. 

But according to Leche a mere lingual prolongation of the 
dental lamina beyond the point of attachment of the enamel- 
organ of a tooth gives of itself no certain promise of future 
tooth formation, or perhaps one should say, no reliable evi- 
dence of its morphological value as an enamel-organ of a 
succeeding tooth generation. Leche is thus forced to base a 
fundamental morphological distinction in large part upon that 
swollen (“kolben” or “knospenférmig”) character 
which is held to distinguish the residual dental lamina beside 
the teeth of Marsupials. It may be quite true that such 
thickening of the free end of the residual lamina is an invari- 

VoL. 39, PART 4,—NEW SER. 00 


532 J. T, WILSON AND J. P. HILL. 


able antecedent to the first formation of a true successional 
tooth. But it is not equally permissible to assert, conversely, 
that wherever such a feature is perceptible there we must 
recognise the presence of a rudiment of an enamel-organ 
belonging to a definite tooth generation. The fact is that 
under nearly all the discussions touching the ‘‘ swollen ” resi- 
dual lamina or “Ersatzleiste”’ of Marsupials—not even 
excluding Leche’s more critical remarks—there appears to us 
to lurk an assumption which is traceable to the use of some- 
what ambiguous descriptive terms. Thus the term “ bud” 
(“ Knospe’’) seems to us wholly misleading as descriptive of 
the somewhat thickened downgrowth (our residual dental 
lamina) so frequently met with by the sides of the developing 
enamel-organs in the marsupial jaw. The word ‘ bud-like ” 
(““knospenférmig”) may indeed describe the outline ap- 
pearance of the lingually prolonged lamina as seen in a 
cross-section, but it is thoroughly inapplicable to the actual 
solid form which that downgrowth really possesses, 

We must therefore raise a most emphatic protest against 
the use of a phraseology such as that employed by Leche in 
the following sentence :—‘‘ Zuniachst ist zu betonen, dass bei 
den Beutelthieren die Schmelzkeime der zweiten Dentition 
sich meist langer als die Schmelzleiste erhalten, also ganz wie 
bei einer Anlage, aus der ein Zahn sich wirklich entwickelt : 
hatte die Zahnanlage jede Bedeutung eingebusst, so ist 
schwer einzusehen, wesshalb sie als knospenformiger Schmelz- 
keim sich langer als ihr Mutterboden, die Schmelzleiste erhalten 
sollte” (8, p. 105). Here the distinction assumed between 
‘“enamel-germs of the second dentition” and the ‘ dental 
lamina” is a most unjustifiable one. The idea that, during 
disintegration of the latter, we can distinguish “ bud-like 
enamel-germs ”’ which are preserved longer than “the parent 
structure—the dental lamina” is a wholly fanciful one, de- 
pendent upon the confusion of the outlines of cross-sections 
with those of solid extended structures. 

In his recent criticism of Leche’s general attitude towards 
the question of marsupial teeth (27) Kiikenthal employs the 


DEVELOPMENT AND SUCCESSION OF TEETH IN PERAMELES. 533 


sane objectionable form of expression— bud-like enamel- 
germs”: “Es sind nun, wohlgemerkt, nicht etwa die Enden 
der frei endigenden, einfachen Zahnleiste, die nach innen von 
den durchbrechenden Zahananlagen der Beuteltiere liegen ” 
[but they are really nothing more], “ sondern knospenformige, 
teilweise von verdichtetem Bindegewebe umgebene Schmelz- 
keimstadien,” &c. (27, p. 655). 

Localised serial swellings of the continuous 
dental lamina in its liberated or residual phase, 
such as we can justifiably speak of as separate 
enamel-germs, do not exist, either distinguished 
from, or as integral parts of, the residual lamina 
aforesaid. 

An exception to this statement must of course be made in 
favour of p. 3, which arises in the first place as a localised 
thickening of the residual lamina by the side of dp. 3. But 
p. 8, as we shall attempt to prove more conclusively below, is 
serially homologous with the enamel-organs of the other ante- 
molar teeth of the adult, and not with the residual lamina 
developed by their sides. 

We cannot insist too strongly on the laminar character of 
the residual lamina. As we have again and again shown in 
the course of our description of stages, it arises as a result of 
the constriction process whereby the permanent enamel-organs 
are differentiated from the main dental lamina. And, when 
thoroughly established, it is in perfect morphological con- 
tinuity with the undifferentiated dental lamina fore and aft, 
intermediate between the region of one tooth and those of the 
next in front and behind. This is well brought out in our 
fig. 82, which is a combination drawing from several suc- 
cessive horizontal sections of the upper jaw of our Stage v. 
The complete antero-posterior continuity of the dental lamina 
in the anterior premolar region could not be demonstrated in a 
single section, but in the combination figure no violence has been 
done to any essential relationship. And the figure will be suffi- 
cient to illustrate the manner in which the free residual lamina by 
the lingual side, say, of p+, simply forms the direct continuation 


5384 J. T. WILSON AND J. P. HILL. 


of the undifferentiated dental lamina before and behind, which, 
in turn, is continued into the residual lamin of £ and of p?. 
(We have carefully verified the planes of section, and have 
definitely determined that they correspond to the free portion 
of the epithelial [residual laminar] downgrowth. But, indeed, 
the view here expressed is supported by the whole body of our 
observations.) 

The merest glance at our figure will serve to show that, at 
least in this region, the term ‘‘ bud-like ” applied to the resi- 
dual laminar downgrowth is ludicrously inappropriate. The 
swollen “ bud-like ” outline seen on cross-section is due to a 
general and continuous thickening of the free marginal portion 
of the lamina. 

It may with justice be contended that a perfectly similar 
marginal thickening of the primary dental lamina may be 
noticed prior to the differentiation from it of the earliest 
enamel-organs (though then the whole structure is much fuller 
and plumper than is the residual lamina of later stages), and 
that the later recurrence of such a condition may well be taken 
as heralding the advent of a new dentitional series. To this 
we reply that that general and continuous marginal thickening 
of the lamina, primary or residual, of itself signifies 
nothing at all in the way of tooth differentiation. 

It is possible, indeed, to affirm—but this is equally true of a 
non-swollen residual dental lamina destitute of any tooth-sac— 
that the epithelial residuum is the potential equivalent of 
further successional teeth. But in our view there is no more 
reason to identify the residual dental lamina with any one tooth 
generation than there is to identify the primary and undifferen- 
tiated dental lamina of a higher mammal specifically with the 
“milk,” to the exclusion of the “‘ permanent ”’ series of mam- 
malian teeth. 

We regard the lingually placed ‘epithelial downgrowths ” 
beside the normal adult enamel-organs merely as portions of a 
dental lamina not yet wholly exhausted of its formative acti- 
vity, and therefore possibly still capable, exceptionally, of 
providing material for the production of teeth homolo- 


DEVELOPMENT AND SUCCESSION OF TEETH IN PERAMELES. 535 


gous with a third Eutherian dentition (fourth dentition of 
Leche). 

The hypothetical product of the residual dental lamina may 
perhaps be represented as tooth-series to the mth power, 
rather than as any definite series. Kiikenthal (24) aptly 
remarks—and in this he is following the lead of Baume (26)— 
that the two mammalian dentitions are sisters whose mother 
is the undifferentiated dental lamina. This is true on the 
whole, but according to our theory it requires the further 
comment that the “Ersatzleiste” in Marsupials is analo- 
gous to the mother after the second parturition (the 
product of the first pregnancy having been early 
aborted), and ought not to be mistaken, as heretofore, for 
the younger daughter. The theoretical possibility of the 
occurrence of a third pregnancy need not be excluded. Leche 
has shown that occasionally in mammals such an event does 
actually occur; but so far as our experience goes the formal 
possibility or promise of a third dentition has been absolutely 
unfulfilled in Marsupials. In this connection we may ask the 
reader to compare Leche’s figures (e.g. his figs. 106, 113, 115, 
116, 118, 121, 124, and 137) of residual laminar downgrowths 
in various marsupials, which he labels as “ knospenformig,” 
or even actually as ““Schmelzkeime,” with his figures (e. g. 
figs. 28—380, 74, 84, 94, and 95) of residual downgrowths of 
the lamina by the lingual sides of the true permanent or resi- 
dual teeth of certain other mammals, and also with the two 
figures (our figs. 76 and 77) we have reproduced from his figs. 
140 and 142, showing the similar appendage to p3 in Phasco- 
larctus. It will be seen that no structural differences serve 
to distinguish the downgrowths in these different cases. 

In view of the foregoing discussion the term “ knospen- 
formig ” will be seen to be thoroughly out of place as applied 
to the slightly thickened laminar downgrowths figured in 
cross-section in Leche’s figs. 187 and 140. 

It has been shown in connection with our Stages 11 and 111 
that the first rudiments of the enamel-organs of the individual 
teeth arise as localised enlargements of the primary lamina 


536 J. T. WILSON AND J. P. HILL. 


especially in the form of labial outgrowths therefrom. And 
if it could be shown that the continuous residual lamina of 
Marsupials does actually develop regular localised enlarge- 
ments, bud-like or other, by the lingual sides of the enamel- 
organs of the adult teeth,—of similar character to that which 
we have seen to constitute the first rudiment of p3, then the 
modified views of Leche would receive some support. But 
such is not the case. Of definite and regular “ bud-like ” 
swellings, other than expressions of that general and conti- 
nuous marginal enlargement to which we have already alluded, 
we find no trace, unless indeed the production of isolated 
groups of cells and the formation of epithelial “ pearls’? and 
*‘cell-nests”’? during the final disintegration of the residual 
lamina were to be reckoned as such. And in the latter event 
we have shown that no less than seven rudimentary succes- 
sional germs must be held to be present in the region of the 
upper canine of Perameles. 

It is true that the residual lamella is not absolutely even 
thoughout its entire extent, but presents various slight irregu- 
larities and inequalities, both in contour and in the thickness 
to which its free marginal portion attains here and there. 

These, however, never amount to localised swellings, which 
could possibly be spoken of or regarded as in any sense 
differentiated from the lamina. (The mode of origin of p.3 
will receive special attention.) Where local variations may be 
recognised they are extremely indefinite, and their only sem- 
blance of regularity of arrangement results from the fact that 
the level of the free distal edge of the residual lamina is to 
some extent affected by the depths of the enamel-organ, beside 
which it lies ; hence, e. g., it is much deeper opposite the canine 
than beside the last incisor. These differences have an obvious 
and easy mechanical explanation. Nothing further issues 
from the structural condition thus brought about, except 
perhaps a slightly longer persistence of degenerating remains 
of the deeper lamina. 

The view here set forth is not only borne out by our own 
pretty extensive observations, directed with special reference to 


DEVELOPMENT AND SUCCESSION OF TEETH IN PERAMELES, 537 


this point, but also by examination of Rése’s figures of his models 
of the developing dentition in Didelphys (11; cf. especially 
his fig. 12). In none of the stages there represented do we 
get any indication of “ bud-like” or partially isolated struc- 
tures arising from differentiation of the residual lamina, which 
is there quite clearly shown (cf. his figs. 12 and 15) as a con- 
tinuous band. It is true that the contour of the latter, whilst 
plain and fairly even in parts, is in other places crooked and 
irregular; but the latter condition is obviously attributable, 
chiefly if not entirely, to the overcrowding of the developing 
teeth (cf. especially the incisor region of his fig. 12). 

In Rése’s models of the early stages of the human dentition 
(9) there may be followed the process of formation of a residual 
dental lamina, which up to a certain point exhibits characters 
very similar to those we find in the case of the residual dental 
lamina of the marsupial jaw. But the structural similarity 
between the indifferent stage of the human residual lamina 
and that of Marsupials really yields no evidence either for or 
against their actual and strict homology. For Leche has 
clearly proved! that residual laminze may be developed by the 
side of the enamel-organs of the second (permanent) dentition 
of higher mammals. And indeed, as already pointed out, he has 
figured a residual laminar downgrowth by the side of the one 
undoubted successional marsupial tooth in Phascolarctus 
(figs. 76 and 77), an observation which we have to a large 
extent been able to confirm for Perameles (cf. p. 521 and 
fig. 74). 

It is, to say the least of it, very suggestive that just in 
Phascolarctus, where the deciduous premolar is very rudi- 
mentary, and where, therefore, we may suppose the formative 
activity of the dental lamina to be less exhausted than in other 


1 See especially his notes on the third incisor and canine of Hrinaceus (6) 
on pp. 25, 26 of his monograph (8), and his illustrative figures, 28—30. Cf. 
also fig. 74 showing residual lamina (“ Schmelzkeimahnliche”) beside 
p. 1 in Phoca; fig. 94 showing same in Desmodus; and fig. 97 showing a 
similar condition of residual lamina beside i2 of the same animal. Cf. also 
Marett Tims’ researches on the dentition of Canis (87). 


538 J. IT. WILSON AND J. P. HILL. 


forms, we find, by the lingual side of the early maturing 
p. 38, a well-formed residual lamina as figured by Leche. So 
also the preservation of the dental lamina in Desmodus and 
Phoca, referred to and figured by the same author, is doubtless 
due to the weaker development of both tooth series, i.e. there 
is more formative material left over than in other cases. 


Connective-tissue Condensation around a Supposed 
Enamel-organ. 


Leche, as we have seen, attached considerable importance to 
the occurrence of a “rudimentary tooth-sac,” in the shape of 
a condensation of the connective tissue around a supposed 
enamel-germ. But while he utilises it for diagnostic purposes, 
he is yet by no means blind to its equivocal character. ‘ Dass 
jene Verdichtung des Mesoderms an sich durchaus nicht immer 
die Anlage einer Zahnpapille oder eines Zahnsackchens zo 
sein braucht, dass sie vielmehr das rein mechanische Produkt 
des Eindringens des Ektoderms ist, geht ausser aus den 
obigen Thatsachen auch aus dem Umstande hervor, dass, wie 
schon Baume (p. 66) beobachtet hat nicht nur der Schmelz- 
keim sondern auch die Schmelzleiste, falls sie geniigend tief 
in das Mesoderm eindringt, von verdichteten Mesoderm-gewebe 
umgeben ist’ (3, p.15). And again :—~ Sowohl aus Baume’s 
als meinen Untersuchungen geht hervor, dass die Schmelz- 
leiste tiberall da, wo sie geniigend tief in das Me- 
soderm eindringt, eine Verdichtung in diesem her- 
vorruft. Wie ich oben (p. 15, figs. 4 and 5) naher ausge- 
fihrt habe, ist also diese Verdichtung und Abplattung der 
Mesodermzellen durchaus nicht immer die Anlage eines Zahn- 
sickchens oder einer Zahnpapille, sondern vielmehr als das rein 
mechanische Product des Eindringens der Ectodermleiste 
aufzufassen. An den Stellen, wo die Schmelzkeime entstehen, 
schreitet durch den verstarkten Druck, welchen diese auf die 
umgebenden Mesodermzellen ausiiben, die Verdichtung und 
Abplattung der letztern weiter zur Bildung von Zahnsickchen 
und Zahnpapille, wahrend durch die Riuckbildung der 
Schmelzleiste in den Zwischenréumen zwischen den Schmelz- 


DEVELOPMENT AND SUCCESSION OF TEETH IN PERAMELES. 539 


keimen die von jener hervorgerufene Differenzirung im Meso- 
derm wieder ausgeglichen wird ” (8, p. 131). 

These quotations from Leche are sufficient to dispose of the 
claim that a mere dermal condensation around an epithelial 
downgrowth can give any reliable guarantee of its nature. 
Our own observations have fully convinced us of the worthless- 
ness of this feature as a test character. Connective-tissue 
condensation, in which the cells become flattened and con- 
centrically arranged around an epithelial downgrowth or 
enamel-organ, is beyond doubt simply the response to the 
stimulus supplied by the continued proliferative activity of 
the epithelium. Naturally such condensation will precede 
tooth formation, but it will also accompany every other con- 
tinued growth of the epithelial lamina to whatever cause it is 
due; and, as a matter of fact, it is most markedly seen around 
those proliferating epithelial masses or cell-nests which so 
frequently appear during the absorptive and retrogressive 
metamorphosis of the dental lamina. 

Even on the ground of his own admission, then, we may put 
aside as incompetent that test of the genuineness of an enamel- 
germ which Leche bases upon the presence or absence of a 
rudimentary tooth-sac. And there remains only the swollen 
(“kolbig” or “knospenférmig”’) character of the epi- 
thelial downgrowths of the dental lamina, already criticised 
at some length. 

That Leche should find it necessary to appeal to the “ con- 
stancy of occurrence,’ “sharply stamped character,” and 
* long persistence ” of the supposed successional germs seems 
to us to indicate some mistrust of the judgment founded upon 
their structural features. But if the latter are insufficient to 
suggest that theory of their homology which he espouses, we 
are at a loss to see how the theory can be strengthened by the 
appeal to the vague characteristics referred to. Nor have the 
characteristics themselves greatly impressed us. The struc- 
tures in question do, doubtless, “‘ constantly occur,” but they 
are no more “sharply stamped” in character than the fact of 
their mere existence would demand ; nor do they “ persist ” 


540 J. T. WILSON AND J. P. HILL. 


beyond a period consistent with the general process of dis- 
integration of the dental lamina of which they form parts. 


Worthlessness of Proof from Agreement of Develop- 
ment of p. 83 with that of Supposed Enamel-germs. 


In the passage quoted above from Leche (p. 198) it will be 
found that in discussing the residual laminar downgrowths 
he mentions their agreement with the enamel-germ of p. 3 as 
tending to establish their claim to be regarded as themselves 
rudimentary enamel-organs. 

It is doubtless true that in its evolution p. 3 arises out of 
the residual dental lamina of dp. 3, and that that lamina at an 
early period exhibits the same developmental phase exhibited 
by the residual lamina subsequently developed by the lingual 
sides of the other enamel-organs. But in view of our previous 
contentions the fact of this general agreement can carry no 
weight as a proof of true serial homology. 


Application of the Criterion of “ Contemporaneous- 
ness of Origin.” 


Leche in the next place passes on to apply in this connec- 
tion his favourite criterion of serial homology, viz. that of 
Contemporaneousness of origin of the different An- 
lagen from the dental lamina (cf. p. 528). 

Remarking that the centre of gravity of the whole question 
of the homology of the persisting antemolar teeth of Mar- 
supials lies in that of the relations of dp. 3 and p. 3 to the 
other antemolar teeth, Leche proceeds to apply to this question 
the above-mentioned test. And instead of obtaining a clear 
and unquestioned verdict in his favour on this head he finds 
himself from the first involved in a difficulty, which has to 
be obviated by an important qualification of his doctrine of 
*contemporaneousness.” He finds, in fact, that the Anlage 
of dp. 8 is, after all, not contemporaneous in its first differ- 
entiation with the Anlagen of the other antemolar teeth with 
which he is attempting to demonstrate its serial homology, 
but is really in advance of the latter. The further hypothesis 


DEVELOPMENT AND SUCCESSION OF TEETH IN PERAMELES. 541 


invoked in explanation of this is that the stage of differentia- 
tion of the Anlagen is conditioned by the future size and grade 
of development of the teeth concerned. And just in so far as 
this contention is valid, the value of Leche’s paramount test 
is of course discounted. The author’s own words in this im- 
portant connection may be quoted: “ Nun ergiebt sich sowohl 
aus dem von Rose (vi, fig. 4) abgebildeten Modelle eines 
153 mm. langen Didelphys-Jungen als auch aus der 
Untersuchung meines Stadium B (17 mm. lang), dass alle 
Zahnanlagen dieselben Beziehungen zur Schmelzleiste zeigen, 
dass aber die Entwicklungsstufe der verschiedenen Anlagen— 
und dies geht auch aus Kiikenthal’s Mittheilungen (I, pag. 
662) hervor—schon auf diesen zeitigen Stadien der Kiinftigen 
Grosse und Ausbildung des betreffenden Zahnes entspricht. 
So ist der grosste von ihnen, namlich Pd. 3,—ich sehe natur- 
lich von den Molaren ab,—auch Zugleich der am weitesten 
entwickelte, dann kommt der nachstgrdsste (Cd.), welcher 
weiter entwickelt ist als die schwacheren Pd. 1, Pd. 2, und 
die Schneidezahne. Man hat daher ebensowenig Recht, dem 
Pd. 8 seiner hoheren Entwicklungsstufe halber, einer alteren 
Dentition Zuzuzahlen, wie wenn man aus demselben Grunde 
in Frage Stellen wollte, dass der Kckzahn zu derselben Denti- 
tion gehort wie die tibrigen Zahne (Pd. 1, 2, und Schneide- 
zahne). Bei dem etwas alteren Stadium C von Didelphys 
(siehe oben pag. 88) finden wir sogar, dass Pd. 3 weniger 
entwickelt ist als z. B, Pd. 2. Es bildet also dieser Umstand 
eine Sttitze fiir die Zurechnung des Pd. 3 zu derselben Denti- 
tion wie die persistirenden Ante-Molaren ” (3, p. 102). 

Now it may be quite true that the future size of a tooth 
does, to some extent, influence the period of the differentiation 
of its Anlage from the dental lamina, i.e. the mere bulk of 
the lamina in a given position may determine the more hasty 
evolution of a tooth-Anlage in that situation. We may 
further point out that yet another factor does, at least in the 
case of the molars, determine the order of appearance of the 
Anlagen of teeth of the same series, viz. the relative 
position in the jaw. And it is at least possible that this 


542 J. T. WILSON AND J. P. HILL. 


factor may have some influence in the case of the antemolar 
teeth, seeing that the first formation of the dental lamina 
itself progresses from before backwards. But the significantly 
early differentiation of the enamel-organ of dp. 3, which we 
have shown to occur in the case of Perameles, obviously 
cannot be explained either by position in the jaw or by future 
size of the tooth. It will be remembered that dp. 3 in 
Parameles is the smallest of all the (functional) 
teeth in that animal. And yet it is differentiated from 
the dental lamina before any of the enamel-germs of the adult 
dentition have taken origin. Leche has been misled in the 
case of dp. 3 through the accident of its large size in Didel- 
phys. Had he been conversant with the facts of develop- 
ment as we have ascertained them in Perameles, he could 
not have rested satisfied with his suggested explanation. 

The relatively early differentiation of the adult canine 
Anlage, to which Leche refers in Didelphys in order to 
illustrate the law that future size influences period of differen- 
tiation, is not to be remarked in Perameles in our, for this 
purpose, most critical Stage 111.1 On the other hand, it is a 
fortunate circumstance that just in this stage we possess a 
perfect example of the differentiation, in connection with the 
canine Anlage of an enamel-organ which is truly synchronous 
with the enamel-organ of dp. 38, and which cannot, in our 
judgment, be regarded as other than that of a genuine canine 
“milk” tooth. This fact will, however, be more fully dis- 
cussed in the sequel. 

The fact, to which Leche several times refers, that in some- 
what later stages of Didelphys (e. g. his Stage C) dp. 3 is 
less advanced than, e. g., the second premolar (p. 2), is of no 
avail in diminishing the importance of the fact of its early 
appearance. 

So also in Perameles we have found that dp. 8 in its 

1 In our Stage 11 the common canine Anlage is certainly very large, but it 
is yet in an indifferent phase. In the lower jaw it may be seen that an 


enamel-organ is in process of differentiation from its labial surface, but it is 
not that of <, but of d:. 


DEVELOPMENT AND SUCCESSION OF TEETH IN PERAMELES. 543 


subsequent evolution and maturation lags behind the neigh- 
bouring and more juvenile teeth, and is finally outstripped by 
them in its developmental progress. But Leche has himself, 
as we have seen, shown the worthlessness of the criterion of 
period of maturity, and we cannot follow him in his virtual 
application of that criterion in the passage above quoted. 
The subsequent behaviour of a developing tooth does not 
cancel, or otherwise alter, the significance of its first appear- 
ance, and we have shown that the latter cannot be disposed of 
in the way Leche suggests. We are far from affirming the all- 
sufficiency in every case of Leche’s test of synchronicity as 
applied to the earliest differentiation of enamel-organs. No 
doubt other considerations, such as those above referred to, 
must in many cases be taken account of. Here, however, we 
believe that we have a case in which the value of the criterion 
of contemporaneousness is clearly illustrated; and it teaches, 
in our view unmistakably, the doctrine that dp. 3 is not 
serially homologous with the other antemolar members of the 
adult dentition. 

This decision is of course the key of the whole position in- 
volved in the question under consideration, viz.:—‘‘ Do the 
existing teeth of Marsupials in front of the last premolar cor- 
respond to those of the milk or of the successional dentition of 
higher mammals ?” 

It can hardly be denied by anyone that the deciduous pre- 
molar of Marsupials is homologous to the milk-teeth of other 
Mammalia.’ A recognition that dp.3 is not homologous with 
the other antemolar teeth, but belongs to an “ earlier” cate- 
gory, practically implies the serial homology of these teeth 
with the successional p.3, and therefore their true homology 
with the successional teeth of higher mammals. 

The view that p.3 is in series with the teeth in front of it 
may at first appear, when viewed in isolation, an improbable 
one. But we think it may be taken as established by the 
facts before us, and at the very least estimate it agrees better 
with the facts in Perameles than the alternative theory. 

1 But see above, p. 441. 


54.4, Je TL. WILSON AND J. P. HILL. 


Kiikenthal, who laid special emphasis upon the precocious 
differentiation of dp.8 in Didelphys, held that the early 
liberation of the residual dental lamina entailed by the con- 
striction off of the enamel-organ of dp.3 could only be con- 
ceived as the differentiation of the first Anlage of the succes- 
sional tooth (7). It will be plain from the views we have 
enunciated throughout this paper that we cannot assent to 
this interpretation. The residual lamina beside dp. 31s not to 
be conceived as actually at its first appearance a rudimentary 
enamel-organ. It is at first indifferent. But it is not long 
before a definite and progressive localised thickening and en- 
largement sets in (cf. Stage 111) similar to that which constitutes 
the earliest rudiments, say, of the other premolars. With the 
latter, indeed, the rudiment of p. 3 is in direct serial continuity. 
We may therefore very well consider the rudiment of p.3 as 
appearing tolerably early in development, though its further 
structural evolution is greatly delayed by the development 
beside it of its predecessor dp. 3. 


II. Interpretation of the Tooth-change in 
Marsupials. 


In reference to the second question formulated, it will be 
borne in mind that all recent authors are agreed that in those 
Marsupials whose dentition has been satisfactorily investigated, 
one tooth only is replaced by a successor, and that this tooth 
is invariably the last premolar. 

The statement of Owen (15) that in Macropus, and also in 
wombat, the milk incisors are shed during the mammary 
foetal stage, has never been confirmed by any subsequent 
observer, and has been universally discredited, the more readily 
that the statement is made without any definite indication of 
the details of such a process having been actually observed. 
It is, nevertheless, hard to believe, as Woodward (14) has 
pointed out, that such an accurate observer should hazard a 
statement of this kind absolutely without evidence. And in 
view of the discoveries of Woodward and Rise regarding the 
presence of vestigial teeth in the forms named, it may yet be 


DEVELOPMENT AND SUCCESSION OF TEETH IN PERAMELES, 545 


found that phenomena in the way of shedding or absorption of 
rudimentary teeth actually do occur, such as might naturally 
be interpreted by Owen as cases of ordinary, if precocious, 
tooth-succession. 

As yet, however, we can only deal with the one undoubted 
manifestation of tooth-change admittedly occurring in Marsu- 
pials. Is such tooth-change to be regarded as the survival of 
an originally more complete succession? With the single ex- 
ception of Professor Leche all the recent authorities would 
seem to adopt the affirmative answer to this question, thus 
taking their stand upon the primitive diphyodontism of Marsu- 
pials, and inferentially of mammals generally. We have 
already seen how Mr. Oldfield Thomas (8) definitely, if rather 
reluctantly, gave in his adherence to this theory as a conse- 
quence of Kukenthal’s researches. Any hesitation he exhibits 
is confessedly due to the difficulty he finds in explaining upon 
this hypothesis the striking fact that Triconodon already in 
Mesozoic times exhibited just the same fulness (or meagre- 
ness) of tooth-change which to-day characterises the marsupial 
group. But after all it does not appear that the peculiar fact 
mentioned (even if Triconodon is admitted as really of 
marsupial organisation, as seems probable) is easier of explana- 
tion on any alternative hypothesis. Surely itis just as difficult 
to account for the fact that a new acquisition of milk-teeth 
(as supposed by the older Flower-Thomas hypothesis) should 
have stopped short with Triconodon in the marsupial group, 
as to believe that the process of degeneration (supposed by the 
newer hypothesis) should have been suspended as far back as 
the Mesozoic age. For any secondary condition capable of 
accounting for the former suspension, such as the peculiar 
nutritional conditions of Marsupials, may without much diffi- 
culty be applied so as to account for the latter. 

Leche has more recently come forward as an advocate of the 
alternative hypothesis that the tooth-change in the case of 
the last premolar is not the remnant of a more complete change, 
but the first appearance of a new dentition foreign to the 
primitive marsupial organisation. His theory, however, is 


546 J. T. WILSON AND J. P. HILL. 


markedly distinct in other respects from the original view of 
Thomas and Flower, for it is framed in full view of the pre- 
sence of those ingrowths of the dental lamina which, in 
common with Kikenthal, Rose, and others, he regards as rudi- 
mentary enamel-germs. These, however, he interprets not as 
the retrogressing vestiges of an older, but as the products of a 
progressive evolution heralding the advent of an entire new 
dentition, which among the Eutheria has arrived at a con- 
dition of greater or less completeness. We have seen, too, that 
Leche’s theory is widely different from that of Flower and 
Thomas in that he holds fast the homology of the marsupial 
teeth to the milk dentition of Eutheria; in his opinion it is 
the “second” dentition, not the “ first,” which is in process 
of acquisition. 

In supporting his theory Leche refers to the difficulty which, 
as we have already seen, was experienced by Thomas in re- 
conciling the condition observed in Triconodon with Kiken- 
thal’s theory of the vestigial character of the second dentition 
in modern Marsupials, and he claims that the difficulty vanishes 
when his theory is adopted, since he is able to point to the 
peculiar adaptation of the marsupial mouth to the sucking 
function as a cause adequate to account for the prolonged failure 
to develop a second dentition in the anterior region of the jaw. 

Leche would thus figure as a supporter of the original 
monophyodontism of Marsupials, were it not for the considera- 
tions introduced by his highly significant discovery of calcified 
vestigial teeth. These he interprets as belonging to a “ pre- 
lacteal ” dentition, since they are evidently antecedent in de- 
velopment to the existing adult teeth in these animals. 


Leche’s Denial of the Vestigial Character of the 
Supposed Second Dentition of Marsupials. 


If the theory of the persisting teeth of Marsupials, which is 
held by Leche in common with every recent investigator, be 
conceded, it is not easy to follow Leche in his opposition to 
the Kiikenthal-Roése doctrine of the vestigial character of 
the supposed successional enamel-germs. It is true he simply 


DEVELOPMENT AND SUCCESSION OF TEETH IN PERAMELES. 547 


refuses his assent until further proof. But the reasons he 
urges in support of the alternative theory do not appear to us 
to be of a very convincing character. Those sceptical argu- 
ments which alone appear to us to be specially cogent as applied 
to the theory of Kiikenthal and Rése, and which we owe largely 
to the logic of Leche himself, appear to us to be almost equally 
destructive of the presuppositions of his own position. 

Leche first calls attention to the long persistence of the 
so-called “ enamel-germs of the second dentition.” We have 
already adversely criticised the idea, common both to Leche and 
his opponents, that genuine and actual successional enamel- 
germs are present. But even. supposing that such are really 
in evidence, we cannot see that their ‘‘ long persistence” is 
any argument against their vestigial character. On the con- 
trary, if they are really vestiges of teeth which have been lost, 
it is precisely these remnants of structures which formerly 
persisted throughout a considerable part of the life of the animal 
which will retain the strongest tendency towards durability. 

Again, the occasional occurrence in other mammals (e. g. 
Erinaceus and Phoca) as discovered by Leche of teeth of a 
third dentition, developed from outgrowths of a residual dental 
lamina by the lingual sides of teeth of the second dentition, 
does no doubt establish the capacity of the dental lamina— 
after partial exhaustion by the production of a first and a 
second dentition—to once more liberate itself as a free residual 
lamina, from which in rare cases there may actually be gene- 
rated members of a third and entirely new dentition. 

But just in so far as this argument is relied upon for Leche’s 
purpose his case is weakened for the establishment of the 
homology of the supposed potential enamel-germs of Mar- 
supials to the Eutherian second dentition, a case which is, at 
its best, in our opinion an exceedingly weak one. In fact, we 
may say that if the production in some mammals of teeth of a 
third dentition, from “enamel-germs” similar to those of 
Marsupials, be admitted, this may fairly be claimed as at least 
establishing, on Leche’s own premises, the possibility ofa 
true morphologicalcorrespondence between the similar structures 

VOL. 39, PART 4,.—NEW SER. PP 


548 ds T.. “WILSON AND 0, P. HELE. 


in the forms compared, and thus as tending to overthrow that 
homology of the adult marsupial teeth to the Eutherian 
milk dentition which Leche accepts as definitely fixed. 

To this deduction from Leche’s reasoning we have already 
given in our complete adherence, and we are thus after all at 
one with Leche in denying that the so-called “ enamel-germs ” 
are vestigial remains of fully developed successional teeth, 
since we deny that they are “ enamel-germs” at all. 

In further opposition to the interpretation of the residual 
laminar downgrowths as vestigial in character, we may cite the 
fact of the non-occurrence, in any single case, of occasional 
more advanced stages of tooth development, such as might 
legitimately be expected to crop up as the result of atavistic 
tendencies. That such an expectation is not without warrant 
is shown by the fact that in cases where undoubted vestigial 
teeth are found, as in the case of the so-called “ prelacteal ” 
teeth of Leche (our “milk” teeth), these frequently, if not 
generally, exhibit a prematurely advanced and “ abbreviated ” 
developmental condition. 

Denying then, as we do—with Leche—the vestigial cha- 
racter of the supposed rudimentary enamel-germs in Marsupials, 
and—against Leche—their character as successional enamel- 
germs, we have to make the attempt to homologise the mar- 
supial dentition with the typical mammalian upon totally 
different lines. Such an attempt can, we believe, be quite 
satisfactorily carried out, with the help of material in the way 
of facts and observations, of which some of the earliest and 
most significant were contributed by Leche himself. These 
facts and observations concern the existence of representatives 
of an undoubtedly vestigial tooth-series, whose imperfect de- 
velopment precedes the evolution of the persisting adult teeth. 

Concerning the So-called “Prelacteal” Teeth. 

Vestigial representatives of this earliest “ prelacteal”’ tooth 
generation have been described by Leche in Myrmecobius, 
and, as indicated in our introductory sketch, he has interpreted 
the discoveries by Woodward and Rése of similar vestigial 
teeth as demanding an identical explanation. 


DEVELOPMENT AND SUCCESSION OF TEETH IN PERAMELES. 549 


If now it be admitted, on the strength of our previous con- 
tentions, that the doctrine of the homology of the existing 
antemolar teeth of Marsupials to the Kutherian milk denti- 
tion rests upon very slender grounds, surely the discovery in 
Marsupials of vestiges of a dentition preceding the adult one 
may well give us pause before adding, with Leche, a theore- 
tical third to the two dentitions already recognised as the 
typical mammalian heritage. We do not think that due con- 
sideration has yet been given to the question whether these 
discoveries alone do not enable us to decide upon the precise 
value to be attached to the supposed rudimentary enamel- 
germs, in whose determination as such Leche has at least 
taught us to exercise a very cautious and even sceptical criti- 
cism, and which he himself after all accepts rather as poten- 
tial than as actual “ enamel-germs.” 

If these problematical germs can be dispensed with, the 
view of their successional homology proving to be ill-founded, 
the theory of the marsupial dentition may be reduced to very 
simple terms. The whole of the adult marsupial antemolar 
teeth (except dp. 3) could then simply be recognised as homo- 
logous to the permanent or successional teeth of other mam- 
mals; whilst the homologues of the Eutherian milk-teeth 
would be found in the deciduous premolar, together with any 
vestigial so-called ‘ prelacteal”’ teeth which may be recog- 
nisable. 

This now proposed explanation of the marsupial dentition 
has not been wholly ignored by Leche. Thus, in reference to 
the supposed “ prelacteals”? of Myrmecobius, he says, ‘‘ Ent- 
weder stellen die fraglichen rudimentéren Zihne die erste 
(Milch) Dentition dar, welche bis auf diese Reste ver- 
schwunden ist, wahrend die zweite Dentition, welche die erste 
wahrend der Phylogenese ganzlich ihrer Funktion enthoben 
und verdrangt hat, durch die persistirenden Zahne reprasen- 
tirt wird. Oder, die persistirenden Zahne entsprechen bei 
Myrmecobius wie bei den andern Beutelthieren der ersten 
Dentition, so dass die erwaihnten rudimentaren Zahne nichts 
anderes als Reste einer Dentition, welche der ersten Dentition 


550 J. T. WILSON AND J. P. HILL. 


vorangegangen ist, darstellen konnen” (3, p. 91). The argu- 
ments used by Leche to dispose of the first view seem to us 
exceeding weak. ‘“‘Gegen die erste alternative spricht nun 
zunachst der Umstand dass dieselbe ohne jegliche Analogie 
bei den iibrigen Beutelthieren ist, denn bei diesen entspricht 
ja, wie die neusten Untersuchungen tibereinstimmend darthun, 
das persistirende Gebiss der ersten Dentition der placentalen 
Sdugethiere. Und da gerade Myrmecobius in Bezug auf die 
Anzahl der Backenzihne die primitivste Form unter den 
lebenden Beutelthieren ist, wurde, falls wir diese Alternative 
acceptiren wollten, das Myrmecobius-Gebiss durch das Vor- 
kommen einer ganzen Reihe von Zahnen der zweiten Denti- 
tion zugleich hoher als die tibrigen Beutelthiere entwickelt 
sein-eine Annahme, welche durch ihren Mangel an wahrschein- 
lichkeit von selbst fallt ” (8, p. 91). 

It is plain that the “general acceptance in Marsupials of 
the homology of the persisting teeth to the milk-teeth of Pla- 
centals’’ simply goes for nothing in the present connection, 
seeing that it is just the validity of this “general acceptance ”’ 
that is challenged, partly on the grounds of the existence of 
the “prelacteals.” To determine the latter as morphologically 
‘‘ prelacteal”? on the grounds of the accepted “ milk ” homo- 
logy of the persisting marsupial teeth appears to us like 
reasoning in the very narrowest of circles. 

And with regard to the second argument from the condition 
in Myrmecobius we are at a loss to perceive its bearing upon 
the question at issue. Granting for the present, at least, that 
Myrmecobius is the most primitive marsupial form, are we 
in the least bound to assume that even the most primitive 
mammal should exhibit partial monophyodontism, or at least 
incompleteness of its successional series? ‘To make such an 
assumption is again to beg one of the most important questions 
at issue. Why, we ask, should it not simply be held that the 
most primitive mammals possessed both a milk and a succes- 
sional tooth series, but that the former had undergone almost 
total suppression amongst the ancestors of the modern Marsu- 
pials (? Triconodon). We have already pointed out that the 


DEVELOPMENT AND SUCCESSION OF TEETH IN PERAMELES. 951 


retention of the single tooth-change from Mesozoic times to 
the present offers neither more nor less difficulty than any 
other mode of explaining the Triconodon condition (cf. also 
Leche [28], p. 278). 

The only argument used by Leche in this relation which 
appears to us to require serious consideration is his statement 
respecting traces of true “ prelacteal’’ tooth-rudiments labially 
from the milk enamel-organs of higher mammals.) But when 
we examine the descriptions and figures upon which Leche 
bases this startling statement, we find that the indications 
referred to are of the very feeblest and most inconclusive kind 
(see his figs. 3 and 5’). We have ourselves met with struc- 
tural indications in the pig (cf. figs. 83—85 ; text-figs. 2 and 
3) which at first tend to suggest, even more strongly than 
Leche’s, an interpretation such as he has given. But they are 
really to be explained in a quite different manner, as we hope 
to show incidentally later on when dealing with the question 
of molar homology (cf. p. 567, e¢ seq.). 

It is unnecessary here to recapitulate our own observations 
upon the so-called “ prelacteal”’ vestigial teeth in Perameles, 
which, in accordance with our theory, we have constantly 
designated as “milk ” or deciduous teeth. It is sufficient to 
say that we have recognised what we believe to be vestiges of 
these teeth in connection with at least four, and perhaps all, 
of the upper incisors, at least two lower incisors, and both 
upper and lower canines. Of these, d22, di+, and d< have 
been traced in the cupped and papillated stage, while di; 
becomes strongly and d; weakly calcified. 

In our examination of these interesting rudiments we very 
early became convinced of their serial homology with the 
tooth dp. 3. As we have stated in the introductory part of 
this paper, this conviction at first compelled us to take up the 


' The argument from the topographical relations between those portions of 
the dental lamina connected with the rudimentary and persisting teeth, 
respectively, may be passed over as unimportant. The various stages we 
figure of vestigial teeth in Perameles quite sufficiently explain any pecu- 
liarity in the late stages of the vestigial teeth, alone studied by Leche. 


Bae J. T. WILSON AND J. P. HILL. 


novel and revolutionary view that the deciduous tooth of 
Perameles, and probably of other Marsupials, was a member 
of a supposed degenerate “prelacteal” series. But ere long 
we were led, first to call in question, and then confidently to 
reject the “ prelacteal” theory of this earliest tooth-series, as 
a violation of that principle of parsimony which should govern 
our procedure in the construction of explanatory hypotheses. 
For if dp. 3 and the supposed “ prelacteals”” be regarded as 
representatives of an otherwise suppressed milk series of mar- 
supial teeth, we shall have before us a workable hypothesis 
which we believe to be free from the difficulties and complica- 
tions which beset that more generally current, and which 
offers no insuperable difficulties of its own. 

In preceding pages we have dwelt upon the application of 
Leche’s criterion of the contemporaneousness of origin 
of the tooth-Anlagen from the dental lamina. We 
have shown that, judged by this test, dp. 3 belongs to a 
different category than that of the other persisting antemolar 
teeth. The identification of dp. 3 and the “ prelacteals” as 
members of one and the same tooth-series may be designated 
as a positive result of the application of the same cri- 
terion. The investigation of our Stages 11 and 111 has yielded 
the most striking evidence of the validity of such a conclu- 
sion. 

Thus in Stage 11 the enamel-organs of the future permanent 
antemolar teeth are in abeyance, or are represented only by 
localised undifferentiated thickenings of the dental lamina, 
But already dp? and di+ show well-marked cupping (figs. 8, 9, 
14, and 12), while d¢3 shows a slighter degree of it (figs. 2, 
8, and 4), and the enamel-organ of d£ is, at least, in process of 
differentiation from the main mass of the canine Anlage 
(fig. 13). 

But it is in connection with Stage 111 that the most strik- 
ing pictures are obtained. The horizontal series from this 
stage was of especial use to us during our period of transition. 
Specially would we refer to our fig. 22, as affording a most 
instructive comparison between the canine ‘ prelacteal ” 


DEVELOPMENT AND SUCCESSION OF TEETH IN PERAMELES. 553 


enamel-organ, so-called (our de.), and the deciduous third pre- 
molar, dp. 3. 

The position of the latter with regard to the neighbouring 
dental lamina is seen to correspond—not to that of the other 
adult teeth—but most accurately with that of the canine 
** prelacteal ” enamel-organ, which is well seen in the section ; 
dp* and d? are both seen to project labially so as to lie ina 
sagittal plane distinct from that occupied by the developing 
enamel-germs of the permanent teeth. This relation is also 
very evident, if slightly less striking, in coronal sections 
through the same Anlagen ; we have only to imagine the bulk 
proportions of the respective parts in figs. 20 and 24 to be 
somewhat altered in order to realise the most complete corre- 
spondence between the two. 

It is of the greatest interest to compare Rése’s fig. 1, 
representing the dental condition of a foetal Phascolomys 
(11, p. 752) with that just described. There, in the wombat, 
it appears that dc. is comparatively large and calcified, but is 
evidently of precisely the same order as the smaller calcified 
cheek-tooth marked D, whose homology with dp. 3 cannot, in 
our opinion, be doubted (cf. also his fig. 3). Here, therefore, 
the bulk proportions of de. and dp. 38 are altered from what 
we find in Perameles, but otherwise the condition presents a 
remarkable similarity. 

The very marked labial situation occupied by dp3 in all the 
entire series of our stages is most striking, and must not: be 
confounded with such a lateral displacement of. enamel-organs 
as may ensue as a result of over-crowding of the jaw. This 
latter occurs more especially in the incisor region, and its 
effects are well illustrated in the figures of Rése’s models of 
Didelphys already referred to. But the distinction in the 
orientation of the enamel-organ of dp. 3 is established from 
its very first appearance. It is most striking when the germs 
of the anterior premolars begin to appear, and long before any 
displacement through encroachment on their part could be 
imagined. ‘Thus both in serial position and in period of 
differentiation the enamel-organ of dp. 3 relates itself to the 


554 J. T. WILSON AND J. P. HILL. 


so-called ‘ prelacteals,” 


teeth. 

Along with these considerations may be borne in mind the 
tendency, so often illustrated amongst Marsupials, for this 
same tooth to share the fate of its fellows and to become 
vestigial. We have referred in the introduction to its con- 
dition in Dasyurus and Phascologale, and here we need 
do no more than refer to our figures of the tooth in the first- 
named of these forms (figs. 80 and 81). 

On the grounds, therefore, both here and elsewhere detailed, 
we can have no hesitation in affirming that dp. 3 ought to be 
regarded as a member of the same series which includes the 
vestigial canines and incisors, and that this series corresponds 
to nothing else than the normal mammalian milk-series, which 
in Marsupials has been profoundly modified in the way of 
suppression. 


and not to the permanent antemolar 


Primitive Diphyodontism and “Suppression.” 


Along such lines of reasoning we thus reach the standpoint 
of the primitive diphyodontism of mammals—a_ position 
already occupied, though, as we believe, on erroneous grounds, 
by Kikenthal, Rose, and others. We believe that we are 
justified in seeking for the cause of the almost total suppres- 
sion of the milk-teeth in front of the last premolar, in the 
modified condition of the mouth in marsupial young in con- 
sequence of its peculiar adaptation to the sucking function. 
This would seem to be a more natural employment of this 
factor than that of Leche, who has already suggested it. For 
he seeks by it to explain the non-appearance of an entirely 
new series of replacing teeth. Surely such an organic 
modification would more immediately tell upon an existing 
milk dentition in the way of suppression, rather than— 
merely through delay of such dentition—affect the subsequent 
development of successors. 

In fact, Leche himself, in opposing the view that the second 
dentition has actually once been present and has since dis- 
appeared, contends in so many words that, according to the 


DEVELOPMENT AND SUCCESSION OF TEETH IN PERAMELES. 905 


principle of ‘‘ abbreviation,” the suppression should affect the 
first rather than the second dentition. “‘ Ferner falls wirklich 
jemals eine vollstandige zweite Dentition ausgebildet gewesen 
ware, ist es kaum zu erklaren, wesshalb z.B. bei Didelphys, 
wo durchaus keine Ursache zu einer Reduction oder retro- 
graden Entwicklung des Zahnsystems als Ganzen vorliegt und 
actisch auch keine Reduction eintritt, die zweite und nicht 
vielmehr die erste, im allgemeinen schwachere Dentition un- 
terdriickt wurde—etwas, das ja auch nach dem Princip der 
Abkiirzung der Entwicklung zu erwarten gewesen ware” 
(3, p. 105. See also 28, p. 372 [2)]). 

We note also in this connection the admission of the same 
author, ‘‘ Wo aber sonst innerhalb der Klasse der Siugethiere 
Monophyodontismus auftritt, spricht der zur Zeit vorliegende 
Thatsachenbestand zu Gunsten der Annahme, dass die erste 
Dentition verschwunden ist und die zweite persistirt ” (3, p.143). 
Leche finds in Erinaceus an illustration of this view. He 
has come to the conclusion—we believe justly—that in this 
form those antemolars which are subject to no tooth-change 
are members of the second dentition which have hastened their 
time of development, and that Krinaceus represents “a transi- 
tion form from the diphyodont to the monophyodont stage.” 
Thus suppression in Krinaceus is interpreted in a totally 
different way from that occurring in Marsupials, so that three 
different mammalian conditions may be recognised: (a) the 
lowest or marsupial being characterised by an almost complete 
monophyodontism, from failure to develop successional teeth ; 
(6) the typical mammalian condition, where complete duplicate 
sets of antemolar teeth—milk and successional—are developed ; 
and (c) a still more advanced condition, attained in Erina- 
ceus, Bradypus, and Pinnipedia, which again approaches 
monophyodontism, but this time through defect of the first 
(milk) or historically older dentition. An elaborate hypothesis 
like this, to explain what are essentially very similar con- 
ditions as regards suppression in Hrinaceus, &c., and Mar- 
supials, seems to us to bear its condemnation upon its face. 
We think the proofs are overwhelming that in Marsupials, as 


556 J. T. WILSON AND J. P. HILM. 


in the other forms referred to, but to an even greater extent, 
the suppression has affected the milk dentition and not the 
successional, and that we may extend to Marsupials generally 
the conclusion of Rose, that “‘ Phascolomys Wombat ganz 
ahnlich wie die placentalen Sauger zwei typische gesonderte 
Dentitionen besitzt 7 (11, p. 745). 

In a paper upon the milk dentition of the Rodentia (18) 
Woodward has discussed the occurrence of vestigial teeth 
in certain members of that order, and has shown that, in the 
forms dealt with, these vestigial teeth are representatives of 
the milk dentition. It may be noted that this conclusion is 
reached in the case of the mouse in spite of the fact that 
‘there is present at one stage in the development of the large 
lower incisors a slight downward prolongation of the dental 
lamina on the posterior side of the enamel-organ of that tooth.” 
This is, of course, simply a repetition of the ordinary mar- 
supial condition. Upon it he comments :—‘‘ This might be 
regarded as presenting an indication of a permanent tooth, the 
large incisor being referred to the milk dentition, while the 
vestigial incisor in front might represent the pre-milk dentition. 
On the other hand, if this very slight prolongation of the 
dental lamina has any morphological value at all, I should be 
inclined rather to consider that it represented that supposed 
fourth dentition, as described by Kiikenthal and Leche for the 
Seals” (p. 626). Here, then, we find Woodward applying 
to the vestigial teeth of the mouse precisely that mode of 
interpretation which is claimed by us to be alone capable of 
affording a satisfactory solution of the marsupial problem. 

In Kikenthal’s latest utterances (in reply to Leche) upon 
the problems of the mammalian dentition (27) we have a con- 
venient summary of his views, which have undergone little if 
any change from those of his earlier papers. 

He adheres firmly to the view that both the ordinary denti- 
tions of mammals were inherited from sub-mammalian ances- 
tors, but apparently admits as established the assertion that 
“traces of two other extinct dentitions are occasionally still 
present embryonically.” 


DEVELOPMENT AND SUCCESSION OF TEETH IN PERAMELES, 557 


But it is extremely difficult to follow Kikenthal in his ex- 
planation of the manner in which suppression has_ been 
brought about in different groups. He first formulates a 
“law of reduction,” stipulating that the tendency to sup- 
pression has constantly affected the first dentition most 
strongly. Thus he explains the suppression of the supposed 
“ prelacteal” series. So far all is intelligible. But when he 
comes to explain suppression as witnessed within the mamma- 
lian class itself the case is much ess clear. ‘ Bei den nie- 
deren Sdugetieren dominirt die erste Dentition, die zweite 
kann wohl ausgebildet sein (Edentaten),! oder aber sie fallt 
aus besonderen Griinden-—secundaren Anpassungena—dem 
Rudimentarwerden anheim (Beutler, Cetaceen).? Bei den 
hdheren Saugetieren verliert die erste Dentition an Bedeutung, 
und die zweite kommt mehr zur Geltung. Wenn jetzt gewisse 
Lebensbedingungen einen Zahnwechsel ungiinstig sind, wird 
nicht die zweite Dentition unterdritickt werden, sondern die 
erste eine raschere Entwickelung und ein fritheres Ende 
nehmen.” (The last sentence expresses the corollary from the 
previously enunciated “law of reduction,” and formulates 
precisely what we believe to have happened in the case of 
Marsupials.) 

In view of this startling exception of Marsupials from the 
ordinary operation of the law of reduction, it is rather 
strange to find Kukenthal going on at once to say that “ein 
einheitliches Gesetz beherrscht also die Dentitionen aller 
Wirbeltiere ” (p. 657). But why should Kikenthal ask us to 
make a special exception of the Marsupials from the operation 
of his own “law”? Why should he expect us to assent to 
the idea that while ‘‘ conditions of life unfavorable to tooth- 
change,” amongst mammals generally, should tend towards 
suppression of the milk dentition, yet nevertheless the 


? Upon this statement with regard to the Edentata, as well as upon the 
entire passage, Leche’s rejoinder (28, p. 275) may be consulted. 

2 We are constrained to avoid special discussion of the Cetacean condition 
because we are disposed to hold with Leche that the homology of the persis- 
tent teeth of Cetacea must still be left as an open question (28, pp. 274-6). 


558 J; Tl WILSON AND J}; BP: HIGE. 


‘secondary adaptations ” which have operated amongst the 
marsupial group have resulted in the well-nigh complete ex- 
tinction of the successional dentition ? 

We have little doubt that the answer to these queries ulti- 
mately resolves itself into this,—that Kiikenthal has not been 
able to bring himself to question the validity of the prevalent 
judgment which has stamped mere secondary or residual laminar 
downgrowths—sometimes thickened marginally—as veritable 
enamel-germs of successional teeth. 

We are quite unable to sympathise with Kiikenthal’s final 
protest against Leche’s views as calculated “to disturb anew 
the now clear conception” of the dentitional problem. In 
our opinion Leche’s systematic criticism has been most timely, 
and if he has not gone the whole length in carrying out his 
criticisms to what we believe to be their logical conclusions, 
yet to him belongs the credit of having first introduced syste- 
matic critical criteria into the discussion of the developmental 
aspects of the problems of dentition. 


III. Serial Homology of the Molars. 


The question concerning the serial homology of the molars 
is not altogether separable from the fourth question, which 
deals with the origin of multicuspidate teeth in general. 

For a connected historical sketch of the state of opinion on 
these points we may simply refer the reader to Leche’s mono- 
graph (3, pp. 145-8). A succinct account may also be found 
in Woodward’s paper (14), and a more extended discussion of 
the subject in Schwalbe’s address (16). The principal views 
held may be shortly summarised as follows: : 

(a) The molars belong to the first or milk dentition (Owen, 
Beauregard, Kiikenthal [earlier view], Rose [earlier view], 
Osborn, Hoffmann, Leche). 

(2) The molars belong to the second or “replacing” denti- 
tion (Lataste, Magitot, Woodward). 

(c) The molars are not homologous to any single series, but 
to two or more dentitional series, either by actual fusion 
of the tooth-germs of the latter, or of the material from 


DEVELOPMENT AND SUCCESSION OF TEETH IN PERAMELES. 509 


which such tooth-germs would have been formed (Schwalbe, 
Kikenthal). 

(d) The molars are the lateral end-members of separate 
dentitions: thus, in man, the dentition of which m. 1 is the 
end-member lies between the first and second dentitions (Rése 
[later view]). 

The arguments mainly relied on by Leche for the establish- 
ment of the “ milk” affinities of the molars largely hinge 
upon that interpretation of the development of the antemolar 
teeth with which we are now sufficiently familiar. 

The presence of a residual dental lamina—when its distal 
portion is somewhat swollen, and especially when the sur- 
rounding connective tissue is somewhat condensed—has been 
liberally drawn upon by him to furnish proofs of the milk 
homology of the antemolar teeth. And, having been accepted 
as practically sufficient for that purpose, it is now utilised in 
establishing a like homology for the molars. 

There can be no doubt that the molar segment of the dental 
lamina in Marsupials does very constantly develop a well- 
marked and often thickened residual laminar downgrowth 
Opposite each tooth as a consequence of its differentiation from 
the parent lamina. Here, as elsewhere, however, the residual 
lamina is in each case directly and serially continuous, fore and 
aft, with the intermediate undifferentiated parent lamina; in 
other words, the residual laminar downgrowth does not ex- 
hibit discrete swollen portions, but is an elongated continuous 
band, more or less uniformly thickened towards its free distal 
margin. 


Value of the Molar Residual Lamina. 


No secure basis for the establishment of the dentitional value 
of the molars is afforded by the presence of this residual 
lamina. It might be hastily concluded that, with our definite 
opinion of the “successional” character of the persisting 
antemolar teeth, we are bound to attach the same value to the 
molars. But we are by no means compelled todo so. If 
the molars are to be regarded as homologous with any one 


560 J. T. WILSON AND J. P. HILL. 


dentition, as is likely enough, they may very well be equivalent 
to the milk series, as represented, e.g., by dp.3. For the pro- 
duction of a residual dental lamina, during the differentiation 
of the enamel-organ of a tooth from the parent lamina, for us 
neither involves nor excludes the belief that successors 
may or do arise from that residual lamina. In other words, 
the production of a residual dental lamina by the lingual side 
of a milk tooth-germ in itself does no more than guarantee the 
possibility of the origin from it of a true successional tooth, 
just as its presence by the side of the Anlage of a successional 
tooth involves the potentiality of the origin from it of a tooth of 
a post-successional dentition. . 

We may, perhaps, advantageously refer once again to the 
condition of the first molar in the early stages (11 and 111) of 
Perameles. It will be remembered that in Stage 11 that 
enamel-organ is already recognisable as such, and its stage of 
development, though slightly less advanced than dp. 3, is fairly 
on a par with that of the anterior milk-teeth (‘ prelacteals”’). 
Its papilla is just indicated. 

In Stage 111 the resemblance between dp. 3 and m.1, both 
in their stage and manner of development, is worthy of note. 
At the very least it may be urged that there is no prima facie 
case for assigning m. 1 to a different series from dp. 3. 

Woodwaid at first (2, p. 460) seems to have based his 
belief in the successional character of the molars upon his 
inability to discover in Didelphys, in the Macropodide, 
and some other mammals “rudimentary enamel-germs” or 
“HKrsatzleisten” at their lingual sides. And he attempted 
to explain away what he regarded as a “suggestion ” of such 
structures in Lepus and Talpa by a critical distinction 
which, thoroughly carried out in reference to marsupial develop- 
mental features, leads directly up to the main thesis of this 
work. Woodward in a subsequent paper (18) has fully re- 
cognised the existence of such lingual growths of the molar 
dental lamina in certain Eutherian forms, and in that connec- 
tion reconsiders the question of molar homology. He still 
inclines strongly to the theory of their successional character, 


DEVELOPMENT AND SUCCESSION OF TEBTH IN PERAMELES. 561 


though, of course, no longer upon the grounds of the ab- 
sence of what he regards as rudiments of possible successors. 
But the latter he would prefer to relegate to the category of 
Leche’s fourth (post-sucvessional) dentition. His reasons for 
such a decision are largely of a speculative character. 

But there is one feature of his case which is by no means 
merely theoretical, but is based directly upon observation of 
structure. We refer to the presence of alleged rudiments of 
a molar dentition older than the existing molar series. We 
shall recur to this subject presently. 

We have examined several series of sections of early Ma- 
cropus embryos, and, like Woodward, we there find that the 
large and deeply cupped molar enamel-organs are as yet indis- 
tinguishable from the dental lamina, the latter appearing for 
the time being to have been wholly converted into the cup-like 
enamel-organs. But this is precisely what we find to be the 
case with the milk enamel-organs of the incisor and canine 
teeth in embryo pigs of 20 mm. head-length, where of course 
successional teeth arise later on. It is entirely a question 
of the stage of advancement of the developmental processes. 
And however it may be in Macropodidx, there is not the 
slightest doubt that lingually placed “ residual” prolongations 
of the dental lamina (and with swollen free distal margin) do 
occur constantly in Pe rameles during the development of every 
one of the molars, as shown in our descriptions and figures. 
Thus fig, 27 shows the tolerably early appearance (Stage 111) 
of the residual lamina by the side of m4. Here there can be 
no possible confusion with the Anlage of m2 which is seen in 
the horizontal sections (figs. 25 and 26) to be just developing 
further back. That the evidence derived from these horizontal 
sections is not to be gainsaid becomes apparent in the exami- 
nation of Stage 1v. There m2 is a well-developed cupped 
enamel-organ, and has itself developed a definite residual 
lamina with a swollen fundus (fig. 52), while beside m+ 
the residual dental lamina is seen persisting as a typical 
“ Ersatzleiste,” and appears on cross-section as a swollen 
and “ bud-like” (“‘knospenférmig ”) downgrowth (fig. 51). 


562 J. T.- WILSON AND J.-P. HILL. 


And with the development of each successive molar a similar 
condition is brought about. 

It would certainly appear that the evolution of the macropod 
molar proceeds less rapidly than it does in a polyprotodont 
form like Perameles.! In any case it becomes quite certain, 
from examination of our consecutive stages, that a residual 
dental lamina with thickened margin comes to project freely 
by the lingual side of the molars, just as in the case of the 
antemolar teeth. 


Concerning the Significance of Labial Outgrowths 
of the Lamina in the Molar Region. 


Holding such views as we do of the comparatively trifling 
significance to be attached to the occurrence of a residual 
dental lamina (even when “knob-like” or ‘ bud-like ” 
[“kolbig” or “knospenférmig™”] on cross-section), it be- 
comes the more important for us to come to a decision with 
regard to the real character of those labially directed out- 
growths of the dental lamina in the molar region noticed by 
Woodward. These, as has been mentioned above, have been 
interpreted by him as possible or even probable vestigial 
remains of predecessors of the existing molar teeth. 

In such vestiges he is inclined to believe (2, pp. 460 and 
470, and cf. 18, p. 630) we have the true serial homologues of 
the antemolar milk-teeth. 

It is plain that, from our point of view, a discovery of 
vestigial remains of teeth preceding the functional molars 
would tend to place the latter in the same category with, say, 
the permanent canine of Perameles, where, as has been 
shown, a pretty large residual dental lamina is differentiated 
alongside the Anlage of the permanent canine at a time when 
a mere vestige of the milk canine (dc.) alone remains (fig. 47). 

Woodward describes as “ present on the outer side of the 
enamel-organ (in Petrogale) a conspicuous outgrowth of 
its cells, extending down into the gum at right angles to the 


' Probably the future large size of the macropod molar has something to 
do with the relative delay in the process of its differentiation. 


DEVELOPMENT AND SUCCESSION OF TEETH IN PERAMELES, 563 


swollen portion of the dental lamina.” He goes on to remark 
that “‘ we can only interpret these appearances on the grounds 
that the functional molars belong to the second or replacing 
dentition, and that this downgrowth on their outer sides repre- 
sents the rudimentary, or rather vestigial, milk or first 
dentition.” 

We can fully confirm the observations of Woodward in 
regard to the existence of such labial projections as he has 
thus described (and figured in his figs. 25, @ and 8, and 26). 
The like are constantly seen in Perameles in certain regions, 
and at first we were inclined to interpret them in the same 
manner as Woodward has done. 

We have, however, been led to entertain grave doubts as to 
whether such a construction can legitimately be placed upon 
the facts observed by Woodward and ourselves. In the first 
place it may be observed that the whole labial surface of the 
dental lamina often presents considerable irregularity as com- 
pared with the lingual surface, and the irregularities not 
infrequently constitute actual outgrowths or processes. This 
condition is specially noticeable in those parts of the lamina in 
front of or behind the enamel-organ of a molar tooth (cf. fig. 
30, “d”). And even opposite the tooth-germ cellular projec- 
tions may be observed upon the labial aspect of the “ neck ” 
of the dental lamina, proximal to the place of attachment of 
the enamel-organ, as in figs. 28 and 29 (“a”). This kind of 
process appears to consist of cellular material, originally form- 
ing part of the enamel-organ, which has been left attached to 
the main stem or “neck ” of the lamina during the process of 
constriction off by which the enamel-organ is liberated from 
the lamina; and it is very probably comparable to those 
villus-like projections which not infrequently occur upon 
the outer surface of the developing enamel-organ itself. Indi- 
cations are shown in fig. 29, “c,” and especially in fig. 28, “c,” 
of such villous irregularities over the general surface of the 
enamel-organ of m-t towards its hinder end. In some sections, 
several (as many as three) labial projections from the neck of 
the dental lamina of the character indicated in figs. 27, 28, and 

VOL. 39, PART 4,—NEW SER, QQ 


564 J. T. WILSON AND J. P. HILL. 


29 (a”) may be recognised, one above the other. It is plain 
that, whatever be the real meaning of these, it is impossible to 
regard them as vestigial representatives of teeth. This form 
of labial projection must therefore be excluded. Again, it may 
be found that a projection on the labial side of the dental 
lamina is traceable to the point of severance of the lamina 
from the oral epithelium. 

Thus in the section shown in fig. 27, an outgrowth of this 
character (“ b”) appears to be the direct prolongation of the 
proximal margin of the dental lamina, which has just been 
separated from the oral epithelium though it is still in contact 
with it. If a comparison of this figure be made with that of 
dp2. in fig. 24, it will become evident that the condition in the 
former is easily derived from that of the latter. We have only 
to imagine that the freed margin of the lamina in the region 
of m+ has become shifted, or has actually grown, labially from 
its late point of attachment (or later point of severance). Such 
an explanation may possibly apply not only to the case quoted, 
but to that represented in Woodward’s fig. 26, m’ (cf. also his 
fig. 20, d. 1.). 

On the other hand, neither of the explanations suggested 
above appears sufficient to account for the appearance shown 
in his figs. 25 a, and 25 6. These represent a form of out- 
growth or projection with which also we are familiar in Pe- 
rameles. Thus in figs. 30 and 31, passing through the lamina 
just behind m+ in Stage 111, and where the last trace of the 
enamel-organ of that tooth is disappearing from the sections, 
we find that a labial process (/. 0.) of very distinct and definite 
character manifests itself, springing from the neck of the 
dental lamina. This becomes established, not as a mere local- 
ised projection, but as a continuous secondary lamina 
extending right back to the place where the entire dental 
lamina ends abruptly in the Anlage of m2 (figs. 25-6). 

The structure described cannot, we believe, be explained by 
either of the modes of interpretation already put forward as 
applicable to some other cases. It cannot in any sense be a 
consequence of disintegration or severance, and its constancy 


DEVELOPMENT AND SUCCESSION OF TEETH IN PERAMELES. 565 


and persistent continuity in the hindmost molar region forbid 
our regarding it as a more or less accidental or capriciously 
developed villous outgrowth or process. But at the same time 
the fact that it does not constitute any mere knob-like excres- 
cence, but that it really extends longitudinally back- 
wards asa flattened laminar or ridge-like outgrowth, 
seemed to us upon closer consideration to be unfavorable to 
its claims to be regarded as of rudimentary or vestigial dental 
character. 

On examining the next succeeding stage (1v), where m2 is 
a well-developed enamel-organ, and m3 has attained a degree 
of development comparable to that of m2 in Stage 111, we 
find that the secondary labial ridge or lamina aforesaid has 
now disappeared opposite m+ and up to near the hinder part 
of the now developed m2. Its disappearance is associated 
with the disappearance in most of the sections of the connec- 
tion between the dental lamina and the oral epithelium, i.e. 
with the onset of disintegration of the proximal portion of the 
dental lamina. 

But towards the hinder end of the now well-developed m2 
the secondary labial laminar outgrowth again comes into view 
(figs. 52 and 53, /.0.), this time extending backwards from 
near the hinder end of m2 to the posterior limit of the dental 
lamina with its thickened Anlage of m3 (fig. 54, /.0.). 

We see, therefore, that the conditions which formerly 
obtained further forwards are exactly repeated here at the pro- 
gressively growing segment of the dental lamina. Here 
again, however, there is nothing to suggest a vestigial dental 
character. The continuous laminar character, indeed, is an 
indication to the contrary. 

We may sum up these and further observations in this 
direction by affirming that, towards the hinder end of the 
dental lamina (which in the molar region is the growing end, 
and is thus ontogenetically younger than the parts of the 
lamina in front of it), there is constantly to be found, in addi- 
tion to the main stem or axis, as it were, of the dental lamina 
a secondary offshoot or lateral leaf-like outgrowth, directed 


566 J. T. WILSON AND J. P. HILL. 


labially, nearly at right angles with the stem, as shown in figs. 
30 and 81 (/. 0.) and in Woodward’s figs. 25 aand 25 6. 

The definiteness, constancy, and longitudinal extension of 
this form of labial projection markedly distinguish it from 
those other kinds of outgrowths on the labial side of the dental 
lamina to which we have alluded above. 

No satisfactory explanation of the meaning of this structure 
seemed forthcoming until, after some incidental study of serial 
sections of pig embryos, the conclusion was forced upon us 
that this secondary labial laminar projection really represented 
the continuation of the labio-dental, or, better, labio-alveolar 
epithelial lamina (precursor of the lip-furrow or labio-alveolar 
groove) which the researches of Rose (9) in the human teeth 
have shown to arise from a common Anlage with the dental 
lamina proper. 

This identification seems to us so important and so unex- 
pected, as applied to structures so far back in the molar region, 
and especially in marsupial animals (in which, at least ante- 
riorly, the lip-gum furrow is long suppressed), that we find it 
necessary to exhibit in some detail the facts upon which our 
conclusion is based. 

It must be admitted that the explanation offered seems far- 
fetched and improbable, nor do figs. 830 and 81 seem to lend 
much countenance to it. Nevertheless we hope to show good 
reason for the belief that the labial laminar outgrowth is, at 
least, an epithelial lamina which is largely independent of the 
dental lamina proper. 

In the first place we must state that in respect of the mode 
of extension backwards of the dental lamina in Marsupials, 
our observations are so far at variance with those of Rése 
upon human molar development. There can be no doubt that 
in the animals examined by us the greater part of the molar 
lamina grows posteriorly by extending backwards its con- 
tinuity with the oral epithelium, and not merely by extending 
back freely into the mesoderm beneath the epithelium as Rése 
has described it. The anterior, greater part of it appears to 
be as much the product of direct ingrowth from the oral 


DEVELOPMENT AND SUCCESSION OF TEETH IN PERAMELES. 567 


epithelium as is the dental lamina in the antemolar region. 
That this is the case as far as the entire molar lamina in Stage 11 
is concerned we have already pointed out, but we may here 
at once state that we have observed the extension backwards 
of the dental lamina in continuity with the oral epithelium in 
other stages as far as the region of the third molar inclusive 
in the upper jaw, and just short of this region in the lower. 
In fact, it may be stated that this continuity with the oral 
epithelium exists just as far back as the oral epithelium 
overlies the molar Anlagen. The most posterior molars 
appear to originate precociously from a backward free pro- 
longation of the lamina before the lengthening of the jaw 
permits a co-extensive prolongation backwards of the oral epi- 
thelium. 

In pig embryos of 20 mm. head-length, we also find the 
connection of the dental lamina with the oral epithelium 
maintained right up to the hinder end of the molar region, or 
within a couple of sections of it. 

We are, therefore, little surprised to find that in the pig the 
labio-dental or labio-alveolar lamina (i. e. the epithelial Anlage 
of the lip-furrow which is formed anteriorly) is continued back 
into the molar region as the Anlage of what we may here call 
the gum-cheek-furrow. And at the extreme posterior end of 
the proper dental lamina the latter is seen to be fused near its 
basal or attached margin with the labio-alveolar lamina, the 
two structures forming parts of acommon Anlage, as Rése has 
shown to be the case at first in the anterior region of the jaw. 
Fig. 83 illustrates the condition met with close to the pos- 
terior end of the molar lamina in the upper jaw of such a pig 
embryo. The lamina stops rather abruptly a few sections 
further back. 

The section figured passes close behind the posterior end of 
the molar Anlage,—in fact, partly shavesit. Projecting labially 
from the neck of the lamina near the oral epithelium there is 
seen in cross-section a laminar projection (d.0.) of the kind 
whose significance we are discussing, and which is proved, by 
examination of the series, to be neither more nor less than a 


568 J; & “WILSON AND: J: BP. RI, 


continuation backwards of the labio-alveolar lamina or Anlage 


of the “ lip-furrow.” 
The accompanying figures in the text are outlines of the 


identical epithelial structures in the lower jaws of the same 


D 


Fic. 2, a—z.—Pig embryo of 20 mm. head-length. Four serial coronal 
sections through the molar region of the lower jaw. a. Some distance 
behind m;. 8B. Two sections in front of a. c. Through posterior part 
of mz. D. Through middle region of mz. /.0,= labial laminar out- 
growth. /.gr.=lip groove. x 45. 


series. (The lower jaw was chosen as affording a better series 
of gradations of otherwise exactly similar conditions.) 


DEVELOPMENT AND SUCCESSION OF TEETH IN PERAMELES. 569 


Text-fig. 2 (aA) shows the common laminar Anlage attached 
to the oral epithelium (which exhibits on its oral surface a 
trace of the actual furrow). This section corresponds exactly 


Fic. 3, a—c.—Pig embryo of 20 mm. head-length. Three serial coronal 
sections, upper canine region. a. Just in front of © 3B. Through 
middle region of ¢. oc. Through posterior region of * x 45. 


in its relation to the molar lamina with that shown in fig. 83 
from the upper jaw, and shows only a very slight variation 
therefrom in form. The next three outlines (text-fig. 2, 
B, C, D,) are from sections taken at intervals of a few 
sections further forwards. Of these, p shows the definitive 
separation of the labio-alveolar Anlage from the proper dental 
lamina by an increase of the connective-tissue basis of the 
alveolus. 


570 J. T. WILSON AND J. P. HILL. 


The remaining figures (text-fig. 3, a, B, and c, and figs. 
84 and 85) from the pig are taken, not from the molar 
region, but from much more anterior parts of the jaws, in 
order to show the striking appearances in the way of labial 
processes which are produced at certain stages of the differ- 
entiation of the common Anlage of labio-alveolar groove and 
proper dental lamina. Attention may be especially directed 
to the transitional forms of the labial projections in the upper 
and lower jaws figured in figs. 84 and 85. It must be men- 
tioned that every one of the labial projections figured has 
actually been traced into continuity with the definite lip- 
furrow Anlage as clearly as in the case of the series shown in 
the first four outline figures (text-fig. 2) taken from the molar 
region. With regard to the latter it may be remarked that, 
as might be expected from the fact that the more immature 
portion of the molar lamina is the posterior, the primitive 
connection of the Anlagen of labio-alveolar and proper dental 
lamina is recognisable behind, and in proceeding forwards in 
the molar region a transition toa state of mutual independence 
of the two structures occurs. The fact that again, anteriorly, 
there is areturn to the status of a common Anlage is doubtless 
conditioned by other developmental factors. 

Comparison of fig. 83 and text-fig. 2 of the molar lamina in 
the pig, with figs. 30 and 31 of Stage 111, and 52—54 of 
Stage 1v of Perameles, and also with Woodward’s figs. 25a 
and 25 6 of Petrogale, will hardly fail to carry conviction of 
the essential identity of the structures therein represented. 

From a careful investigatiou of these labial appendages in 
the molar region we therefore conclude that while such ap- 
pendages may not be uniformly homologous, yet those which 
alone are definite, and irreducible to the results of unimportant 
incidents of development, require a different mode of inter- 
pretation from that suggested by Woodward. We must accord- 
ingly dismiss the idea that any structural features occur in the 
development of the marsupial molars which can fairly be 
interpreted as vestiges of degenerate “‘ milk” predecessors. 

We are thus thrown back upon the already noted resem- 


DEVELOPMENT AND SUCCESSION OF TEETH IN PERAMELES. 571 


blance between the development of dp? and m+ in the early 
stages. Such resemblance as has been adverted to cannot, of 
course, by itself prove that an actual serial homology exists 
between the two teeth. It may indeed establish a prima facie 
case for such an homology. But, however we may decide the 
special problems of the distinctively marsupial dentition, the 
serial homology of the molars must remain as a general question 
of mammalian dentition, whose decision must yet be regarded 
as a separate and largely independent issue. 

The theories of molar homology set forth under (ce) and (d) on 
page 558 still remain to be dealt with. Of these the latter 
has not as yet been supported save by its author, and may, we 
believe, be regarded as a mere speculation. The former, which 
is supported by both Kiikenthal and Schwalbe, may best be 
discussed in connection with the next question, which primarily 
concerns the mode of origin of multicuspidate teeth, but which 
raises the general question of fusion of originally distinct 
dental units. 


lV. TuHeoriEs oF DENTAL FUSION AND THE ORIGIN OF 
MULTICUSPIDATE TEETH. 


The results of the present research do not, any more than 
Woodward’s, lend countenance to the view once advocated by 
Kikenthal, that molars are a result of the fusion of germs of 
more than one dentition, in the sense in which fusion has been 
supposed to occur between tooth-germs of the same series. 
But Kikenthal’s present position in regard to this question is 
rather less simply expressed. He holds “dass die echten 
Molaren im Wesentlichen zur ersten Dentition gehéren, dass 
sie aber ein Verschmelzungsproduct der Anlagen erster Denti- 
tion mit dem Material, aus dem sonst die zweite Dentition 
entsteht, darstellen ” (20, p. 112, and 27, p. 659). 

Against this view both Woodward (14, p. 447) and Leche 
(3, p. 146) have pointed out that it ignores the fact, upon 
which Kukenthal had himself insisted, that a residual dental 
lamina does normally develop by the lingual side of the first 


972 J. Us WILSON “AND oJ RP. HUGE. 


molar. ‘To this criticism Kikenthal has omitted to reply 
during the course of his criticism of Leche, embodied in his 
latest publication on this subject—‘ Zur Dentitionenfrage ’ (27). 
And we find him towards the close of this last paper simply 
reiterating the opinion in the course of his remarks in reply to 
Hoffmann. 

We confess that Kiikenthal’s view is still very far from 
being clear to us. When he speaks of the molars as repre- 
senting “ein Verschmelzungsproduct der Anlagen erster Den- 
tition mit dem Material, aus dem sonst die zweite Dentition 
entsteht,” are we to understand that the enamel germs of a 
first dentition are first of all differentiated from certain residual 
“material ” representing enamel-germs of a second dentition, 
and that, having been differentiated from one another, these 
are subsequently fused together? If this be what is meant, 
then certainly the facts of marsupial development give no 
countenance to such anidea. Indeed, the mode of origin, pre- 
servation, and gradual disintegration of the molar residual 
lamina prove that certain “ material”? remains over which is 
not fused or taken up during molar development. 

But if Kikenthal’s words are not to be understood in the 
sense referred to, and if we are to believe that he does not 
conceive of a morphological separation or form-differentia- 
tion ever having been effected between the material of the 
first dentition and the material representing the second (as, 
indeed, his employment of the term material would seem to 
imply), then we can only say that, so far as we can see, his 
view differs but slightly, if at all, from Hoffmann’s idea of 
“ physiological ” fusion. 

The opinions of the last-named author are thus approvingly 
summed up by Leche in a passage which deserves special 
comment :—‘* Will man aber den Begriff der Verschmelzung 
umbedingt beibehalten, so kann man, wie Hoffmann richtig 
bemerkt, und wie auch ich bereits oben (p. 142) hervorge- 
hoben habe, sich vorstellen, dass das Schmelzleistenmaterial, 
welches bei den niederen Wirbelthieren zur Ausbildung einer 
Ganzen Anzahl von Zahnserien verwendet wird, bei den Sauge- 


DEVELOPMENT AND SUCCESSION OF TEETH IN PERAMELES. 5738 


thieren zur Ausbildung von bedeutend wenigeren, dafiir aber 
komplicierteren Zahnen benutzt wird” (3, p. 155). 

Concerning such a view we would remark that it amounts 
simply to a surrender of the idea of the fusion of morphological 
structures as such. The aggrandisement of one morphological 
structure at the expense of the material required for the deve- 
lopment of another, whereby the survivor is enabled to launch 
out into a further structural differentiation, may be a justifiable 
conception, but it ought not to be confused with a case of 
morphological concrescence. True morphological homo- 
logy deals with the identity of organic forms, not with that of 
the mere “ material” organised under such forms. In fact, 
it may be questioned whether the identity of organic 
material, as distinguished from that of the form 
under which it is organised, is at all a fruitful or even 
valid biological concept. 

In any case, until Kikenthal has disposed of the objection 
based upon the presence by the side of the molars (e.g. of 
Marsupials) of a residual lamina entirely comparable to that 
found beside other admittedly simple teeth, his theory of 
homology of the molars must be regarded as untenable. 

A still more fundamental question underlies the preceding 
discussion, i.e. whether the multicuspidate form of the modern 
mammalian molar has originated by the concrescence of several 
originally distinct and simple tooth-germs, or by the gradual 
evolution of cusps as outgrowths of a single simple conical 
tooth. 

In favour of the latter view are ranged the majority of 
paleontologists (cf. especially Cope [10] and Osborn [84]), 
who have elaborated a most plausible scheme showing the 
steps in the supposed evolution of the complex mammalian 
molars from a primitive conical type of tooth, through a very 
early tri-tubercular condition. 

The fusion theory, on the other hand, has taken shape chiefly 
at the hands of a few embryological investigators—notably 
Kikenthal and Rose. 

Rose has described the appearance, at a very early period 


574 J. T. WILSON AND J. P. HILL. 


of the development of a molar enamel-organ, of several meso- 
dermal papillary upgrowths prior to any marked differentiation 
of the enamel epithelial germ. These upgrowths he confi- 
dently interprets as the homologues of the originally distinct 
denticles of which, in his view, a complex tooth is really built 
up (cf. 29, fig. 1). 

Woodward has remarked upon Rése’s attempt to prove 
his theory in the case of the chameleon, “ where the back teeth 
are each composed of three cones, which, according to him 
(Rése), arise independently of one another,” to the following 
effect :—this is true to a certain extent, but these cusps develop 
under a common enamel-organ, and there is no indication of 
their ever having possessed independent organs, as would have 
been the case if they were distinct teeth, there being merely a 
differentiation of the cylindrical enamel-epithelium over each 
cusp, which Rése considers sufficient evidence in favour of his 
view’? (14, p. 446). 

We have occasionally come across appearances which at 
first seemed to suggest multiple papillary upgrowths in con- 
nection with the developing molar lamina. Thus in fig. 69, 
though the whole of the surrounding connective tissue has not 
been represented, it is plain that the irregularities of the labial 
surface of the dental lamina might (if the connective tissue 
were all filled in) suggest multiple papillation. But, as we 
have already shown in our previous references to this figure, 
this view would be quite erroneous. The true and only papilla 
of m2 is marked “mp.” Fig. 69 may be compared with figs. 
68a and 70. So far as our observation and experience go a 
truly multiple papillary character is never found in the 
earliest stages of development of a marsupial molar. At its 
first differentiation as such the molar enamel-organ is as 
simple, single, and uncomplicated in form as any incisor 
enamel-organ. And as certainly we find that we can watch 
the ontogenetic evolution of the marsupial molar cusps by a 
gradual differentiation of the primitively simple cup-shaped or 
bell-shaped enamel-organ. 

It would appear, however, that such a statement as the 


DEVELOPMENT AND SUCCESSION OF TEETH IN PERAMELES. 575 


latter is by no means regarded, at least by Kiikenthal, as in- 
consistent with the doctrine of phylogenetic concrescence of 
dental germs. Replying to Hoffmann, he observes, “ Der 
Backzahn eines heutigen Saugers is ein so complicirtes plas- 
tisches Gebilde, welches im Laufe seiner Stammesgeschichte 
einer solchen Summe von Verinderungen unterworfen worden 
ist, dass ich nie daran gedacht habe, dass seine Entwickelung 
uns tiber die ersten Stammesgeschichtlichen Vorgange, die in 
die Uebergangsperiode von Reptilien zu Saugern fallen, wiirde 
Aufschluss geben kénnen, und ich habe schon friiher.... 
die wenig glickliche Idee Rése’s, dass die Ausbildung einzelner 
kegelf6rmigen Schmelz- und Dentinkappen an der spitze von 
Backenzahnanlagen fiir die Verschmelzung spricht, als nicht 
beweiskraftig zuriickgewiesen” (27, p. 658). 

Certainly the crude theory here disclaimed will not stand 
even a very superficial criticism. But the differentiation of 
the molar dental lamina in Marsupials does not in the least 
suggest the occurrence of a fusion-process of any kind. And 
we are in a position to add that even the teeth of Ornitho- 
rhynchus! are to be found, at a stage considerably anterior to 
that described by Poulton (25), still in the condition of quite 
simple deeply-cupped enamel-organs, from which the advanc- 
ing complication of form shown in Poulton’s figures, and well 
known in the fully developed teeth, is derived by differentiation 
in the way of mere inequality of growth. 

There can be no doubt that the advocates of the fusion- 
hypothesis, instead of being aided by ontogenetic considera- 
tions, will have to face the fact—perhaps not necessarily fatal, 
but at least disconcerting—that, in a series of comparatively 
primitive mammalian forms, the cusp-development of a compli- 
cated molar can be traced out as a mere growth-differentiation 
of a primitively smooth and simple bowl-shaped enamel-organ, 
just as, according to Osborn (84, p. 206), we can be “in at 
the birth of every successive cusp” in phylogenetic develop- 
ment. 


1 We hope ere long to make our observations in this direction the subject 
of a special paper. 


576 J. T. WILSON AND J. P.. HILL. 


It may be conceived, of course, that prior to the simple cup- 
shaped enamel-organ, there is a stage like that alleged to exist 
by Rése, in which several mesodermal papille beneath a 
common enamel-germ represent the future tooth. But the 
evidence for the existence of such a condition is as yet so 
extremely meagre and inconclusive that we cannot admit that 
a serious case for it has been made out. 

Schwalbe’s observations in favour of fusion, referred to by 
Kikenthal, cannot be regarded as necessarily possessing any 
great phylogenetic significance. The fusion referred to (16, 
p- 21) of the upper milk incisor with the precociously develop- 
ing crown of the successional tooth need not be taken as 
possessing either more or less morphological importance than, 
e.g., syndactyly in the human subject. Like the latter, the 
condition might even be to some extent hereditary, but its 
interest and importance are probably pathological rather than 
morphological. Were it not for the undeniable importance of 
Kikenthal’s observations on the process of division of the 
cheek-teeth of whales into single conical teeth (19),! and of 
the further statement of the same author that he has been able 
actually to detect a fusion as taking place between molar 
enamel-germs in the case of the walrus (20), we should be 
disposed to regard the fusion theory as a very shadowy hypo- 
thesis indeed. And even with the support alluded to we 
cannot regard its claim to acceptance as at all a strong one. 
There is plenty of room for sceptical criticism. Thus, concerning 
Kiikenthal’s fig. 89 (19) Woodward has remarked (14, p. 447) 
that the condition illustrated and described (19, p. 411) appears 
“rather indicative of the formation of a cusp by outgrowth 
from a simple conical tooth” than of the fusion process 
supposed. 


1 Oshorn has pointed out (84, p. 199) that “even by Kikenthal’s hypo- 
thesis the typical Mesozoic mammal could not furnish as many teeth as are 
found in some of the dolphins;” and he suggests as a likelier explanation 
that “‘as the jaws were elongated the dental fold was carried back and the 
dental caps were multiplied.” 


DEVELOPMENT AND SUCCESSION OF TEETH IN PERAMELES. 577 


PART IV. 
ConcLupING REMARKS. 


Before concluding this discussion it may not be wholly out 
of place to ask the question whether our observations and the 
conclusions deducible therefrom can be made to throw any 
light upon the more general problem of the affinities of Mar- 
supials with other mammalian groups. 

Hasty generalisation is certainly to be deprecated, and it is 
only with the greatest diffidence that we even approach the 
discussion of such a weighty question with the equipment of 
our own partial and imperfect knowledge, more especially of 
the paleontological aspects of the problem. 

There are, however, one or two more or less obvious deduc- 
tions from the views which have been advocated in the fore- 
going pages, and partially summarised in the introductory 
section of the paper (p. 441). 

It is plain that, if our view of the “ milk” homology of the 
so-called “ prelacteal” teeth be admitted, we are bound to 
believe that the marsupial order as a whole—if not derived 
from truly Eutherian ancestors, as seems unlikely from the 
general type or organisation exemplified—is at least an off- 
shoot from a diphyodont stock common to both Metatheria 
and Eutheria. 

Following upon the publication of Kiikenthal’s and Rése’s 
researches upon marsupial teeth, Osborn (34, p. 204) has 
remarked that “the discovery of the complete double series 
seems to have removed the last straw from the theory of the 
marsupial “ancestry of the Placentals.’ And the adoption 
of the conclusions of the present paper in no whit weakens the 
general purport of this criticism, since it still leaves un- 
touched the important fact that one of the two typical mamma- 
lian dentitions has been lost during the evolution of the Mar- 
supialia. In respect of their tooth-equipment then, the 
Metatheria are degenerate mammals. 

With this view it is interesting to correlate such specula- 


578 J. T. WILSON AND J. P. HILL. 


tions as inevitably arise in consequence of the discovery by 
one of us (H., 30) of a true allantoic placenta in Perameles. 
It is obvious that this discovery admits of only a limited 
number of alternative explanations. It implies either (@) that 
an allantoic placenta has been quite independently evolved 
within the limits of the Metatherian and Eutherian groups, 
thus affording a quite remarkable instance of parallelism in 
development ;! or (4) that the Eutheria have inherited their 
allantoic placenta actually from polyprotodont marsupial 
ancestors closely allied to the modern Peramelida; or (c) 
that both the Metatheria and the Eutheria have sprung 
from an earlier mammalian proto-placental stock. 

If the latter alternative were adopted (and we believe that it 
alone will in the end commend itself to the judgment of 
morphologists), then we should be in a position to define the 
common mammalian ancestors of the Metatheria and Eu- 
theria as both placental and diphyodont. 

The much-debated question as to whether or not this early 
mammalian stem was also marsupial in its organisation may, 
in view of developmental researches upon the mammary organs 
during recent years, admit of a not unsatisfactory reply. For 
the researches of Gegenbaur (86), Klaatsch (82), and others 
appear to have established the fact of the possession of rudi- 
ments of ‘‘mammary pouches” (“ Mammartaschen”’) by 
embryo Kutherians. Such mammary pouches (to be distin- 
guished from a true Marsupium) are, in ruminants, accord- 
ing to Klaatsch, converted into the tubular teat-cavity.? Bonnet 
expresses Klaatsch’s view in this connection as follows :—“ Wo 
aber die Mammartasche unverdndert bestehen bleibt, wie bei 
den Wiederkaiiern und den Equiden macht er den Anschluss 
dieser Formen an ganz niedere Zustande, die gar kein Mar- 
supium ausgebildet haben, wahrscheinlich ” (81, p. 628). 


1 But in such case Perameles, contrary to all previous belief, would 
exhibit a more advanced or specialised condition than any other known 
Marsupial. 

2 The contrary opinions of Rein, though accepted by Minot in his ‘ Human 
Embryology,’ have been subjected to severe criticism by Klaatsch, and the 
views of the latter observer have been adopted by Bonnet (31). 


DEVELOPMENT AND SUCCESSION OF TEETH IN PERAMELES. 579 


It would appear probable that we must be prepared to en- 
tertain favorably Klaatsch’s suggestion that the “ Placentals ” 
have originally all passed through a kind of marsupial stage, 
“ein Form, die jederseitz zwar eine Mammartasche, aber 
keinen Beutel besass.” He denominates the animals which 
exhibited this form “ Taschentiere ” or “ Bursalia.” 

Such a rudimentary or primitive condition must have been 
a characteristic of the common stock of both Meta- and 
Eutheria, and was only a marsupial one in the sense in 
which Echidna is now marsupial. 

Klaatsch has recently pointed out (83) what strong reason 
there is for the surmise that the evolution of mammary pouches, 
such as Echidna possesses, is probably an even older trait of 
mammalian development than the evolution of the mammary 
glands themselves. Its existence in the oviparous Echidna 
proves that it isa much older mammalian characteristic than any 
form of placental connection can possibly be. We cannot follow 
Hubrecht (35) in his assumption that the Protamniote an- 
cestors of the Promammalia (and of the Sauropsida) must 
have been viviparous. 

In our view the Monotremes must be taken to represent a 
quite distinct offshoot of a common Promammalian (Hypo- 
therian) stem, which was originally both oviparous and “bursal”’ 
(in the sense of Klaatsch), and in which both hairs and mam- 
mary glands were gradually developed. We do not believe that 
in the long run it will be found possible to maintain the essential 
dissimilarity of the mammary glandular organ in Monotremes 
(36, and cf. Minot, 1. c., pp. 565-6). 

After the separation of the Monotreme phylum we must 
suppose the placenta to have appeared, constituting the steck 
protoplacental. At all events, the existence of a placenta in 
the marsupial order seems to indicate that the primitive pouched 
common ancestors (Bursalia) of the Meta- and Eutheria, 
long after losing their oviparous character, developed an 
allantoic placental connection between the embryo and the 
uterine wall. At the same time the mammary function, 
already evoked during the oviparous period for the nourish- 

VOL. 39, PART 4,—NEW SER. RR 


580 J. T. WILSON AND J. P. HILL. 


ment of the immature young as in Monotremes, became further 
and more perfectly developed. We may further suppose 
that now one branch of the stock thus circumstanced, i.e. the 
Eutherian, went on to further exhaust the nutritive possibilities 
implied in the acquirement of a placental connection; retain- 
ing, however, and at the same time elaborating, the function 
of lactation in relation to more advanced stages of development. 
Another branch,—the Metatherian,—neglecting, from some 
cause or other, the opportunities offered by the original for- 
mation of a placental connection, may be supposed to have gone 
on developing the latent capabilities of the mammary function 
for the provision of adequate means of nutrition for the imma- 
ture young. These capabilities included the replacement of 
the primitive mammary pouches by a true Marsupium, partly 
at least derived from the latter. The original pouched condi- 
tion is thus to be carefully distinguished from the derivative or 
truly marsupliated condition of the Marsupialia, which is 
as much a secondary development along one line as, e. g., the 
ruminant mammary condition is along another. 

And the character of the marsupial dentition as we interpret 
it renders it very highly probable that the secondary evolution 
of a true Marsupium has all along been accompanied by a 
retrogressive development or degeneration from the normal 
early mammalian diphyodont dentition. Indeed, as Leche has 
pointed out, the functional adaptation of the mouth of the 
marsupial young to the peculiar suckling conditions prevalent 
in the order has no doubt conditioned the almost entire sup- 
pression of one of the two dentitions. The dentition sup- 
pressed is, according to Leche, the second (his third) ; but in 
our view it is, beyond all doubt, the first or milk dentition 
which has degenerated. 


University oF Sypvey, N.S.W.; 
July 4th, 1896. 


DEVELOPMENT AND SUCOESSION OF TEETH IN PERAMELES. 581 


ADDENDUM. 


Reference has been made in a foot-note in the body of this 
paper to a contribution by Dr. Marett Tims (87) of ‘‘ Notes on 
the Dentition of the Dog,” published in February in the 
‘ Anatomischer Anzeiger.’ Since our research was completed 
and the bulk of our work written up before Dr. Tims’ paper 
came into our hands, we have found it easier to leave its con- 
sideration for this special note. 

From his observations upon tooth-development in the dog, 
Tims has arrived at certain conclusions with regard to the 
mammalian dentition with which our work will be found in 
general agreement. And although this writer has apparently 
not carried on special researches on the teeth of Marsupials, 
his own work upon the dog has led him to foreshadow in some 
degree the views independently arrived at by the present 
writers on the grounds of a somewhat extensive and prolonged 
series of observations. 

Dr. Tims’ further confirmation of the common occurrence of 
a residual dental lamina beside the permanent teeth of 
Eutheria is a very welcome one. 

The conclusion from his paper with which we are most parti- 
cularly concerned is the following :—‘“‘ That there is not suffi- 
cient evidence for believing in the existence of a pre-milk 
dentition even in a rudimentary condition in the Mammalia 
of the present day.” And a corollary, drawn in part from 
this proposition, is that if Tims’ view be accepted, “ the teeth 
of Marsupials would once more appear to belong to the per- 
manent series.” 


582 J. T. WILSON AND J. P. HILL. 


J 


BIBLIOGRAPHY. 


. Rost.—* Ueber die Zahnentwickelung der Beutelthiere,” ‘ Anatomischer 


Anzeiger,’ Bd. vii, 1892. 


. Woopwarp.—“ Contributions to the Study of Mammalian Dentition. 


Part I. On the Development of the Teeth of the Macropodide.” 
‘Proc. Zool. Soc.,’ London, 1893. 


. Lecue.—“ Zur Entwickelungsgeschichte des Zahnsystems der Sauge- 


thiere,” ‘ Bibliotheca Zoologica,’ Heft 17, Stuttgart, 1895. 


. FtowEer.—“‘ On the Development and Succession of the Teeth in the 


Marsupialia,” ‘ Phil. Trans.,’ London, vol. clvii, 1867. 


. THomas.—“‘ On the Homologies and Succession of the Teeth in the 


Dasyuride, with an Attempt to trace the History of the Evolution of 
the Mammalian Teeth in General,”’ ‘ Phil. Trans.,’ London, vol. elxxviii, 
1887. 


. Winez.—‘ Om Pattedyrenes Tandskifte isaer med Hensyn til Taendernes 


Former,” ‘ Vidensk. Meddel. fra d. Naturhistoriske Forenung i 
Kjébenhavn,’ 1882. 


. KixentHaL.—“ Das Gebiss von Didelphys,” ‘ Anat. Anz.,’ Bd. vi, 1891. 
. Tuomas.— Notes on Dr. Kiikenthal’s Discoveries in Mammalian Denti- 


tion,” ‘ Ann. and Mag. Nat. Hist.,’ series vi, vol. ix, 1892. 


. Roést.—‘‘ Ueber die Entwicklung der Zalne des Menschen,” ‘ Archiv 


f. mikros. Anat.,’ Bd. xxxviii, 1891. 


. Cope.—“ The Mechanical Causes of the Development of the Hard Parts 


of the Mammalia,” ‘ Journ. of Morphology,’ vol. iii, 1889. 


. RosE.— Ueber die Zahnentwicklung von Phascolomys _Wombat,” 


‘Sitzungsber. der k. preuss. Akad. d. Wissensch. zu Berlin,’ Bd. 
xxxvili, 1893. 


. Lecns.— Nachtrage zu Studien tiber die Entwicklung des Zahnsystems 


bei den Saugethieren,” ‘ Morph. Jahrbuch,’ Bd. xx, 1893. 


. Lecne.—“ Studien iiber die Entwicklung des Zahnsystems bei den 


Saugethieren,” ‘ Morph. Jahrbuch,’ Bd. xix, 1892. 


. Woopwarp.—* On the Succession and Genesis of Mammalian Teeth,” 


‘Science Progress,’ July, 1894. 


. OwEn.—* Odontography,”’ London, 1840-45, ‘ British Association Re- 


ports,’ 1848, p. 93. 


. SchwaLBe.— Ueber Theorien der Dentition,” ‘Verhandlungen der 


Anatomischen Gesellschaft auf der 8te Versammlung in Strassburg,’ 
1894. 


DEVELOPMENT AND SUCCESSION OF TEETH IN PERAMELES, 583 


17. 


18. 


19. 


20. 


21. 


22. 


23. 


24. 


25. 


26. 


27. 
28. 
29. 


30. 


31. 


32. 


33. 


34. 


Osporn.—‘‘The Rise of the Mammalia in North America,” ‘Proc. 
American Assoc.,’ vol. xlii, 1898. 

Woopwarp.—* On the Milk Dentition of the Rodentia, with a Descrip- 
tion of a Vestigial Milk Incisor in the Mouse (Mus musculus),” 
* Anat. Anz.,’ Bd. ix, 1894. 

KUKentHat.— Vergleichend-anatomische und Entwicklungsgeschicht- 
liche Untersuchungen an Walthieren,” ii Thiel, ‘ Denkschriften der 
medic.-naturw. Gesellschaft zu Jena,” Bad. iii, 1893. 

KixentHat —“ Entwicklungsgeschichtliche Untersuchungen am Pinni- 
pediergebisse,” ‘ Jenaische Zeitschr. f. Naturwissensch.,’ Bd. xxviii, 
1893. 

BatEeson.—‘ Materials for the Study of Variation,’ London, 1894. 

Tuomas.—‘ Catalogue of the Marsupialia and Monotremata,’ 
Brit. Mus., London, 1888. 

Batpwin Spencer.—‘ Report of the Horn Expedition to Central 
Australia,’ part ii, Zoology, February, 1896. 

Kixentuat.— Hinige Bemerkungen tiber die Saugetierbezahnung,”’ 
* Anat. Anz.,’ Bd. vi, 1891. 

Povtton.—‘ The True Teeth and Horny Plates of Ornithorhyuchus,” 
‘Quart. Journ. Micros. Sce.,’ vol. xxix, 1889. 

Baume.—‘ Versuch einer Entwicklungsgeschichte des Gebisses,’ Leipzig, 
1882 (quoted by Leche and others). 

Ktxentuat.— Zur Dentitionenfrage,” ‘ Anat. Anz.,’ Bd. x, 1895. 

Lecur.—“ Zur Dentitionenfrage,”’ ‘ Anat. Anz.,’ Bd. xi, 1895. 

Roéss.— Ueber die Entstehung und Formabanderungen der menschlichen 
Molaren,” ‘ Anat. Anz.,’ Bd. vii, 1892. 


Hitt.—* Preliminary Note on the Occurrence of a Placental Connection 
in Perameles obesula and on the Foetal Membranes of certain 
Macropods,” ‘ Proc. Linn. Soc.,’ New South Wales, vol. x (2nd series), 
part 4, 1895. 

Boynet.—“ Die Mammarorgane im Lichte der Ontogenie und Phylo- 
genie,” ‘Ergeb. d. Anat. u. Entwick.,’ Bd. ii, 1892. 

Kuaatscu.—‘“ Ueber die Beziehungen zwischen Mammartasche und 
Marsupium,” ‘ Morph. Jahrb.,’ Bd. xvii, 1891. 

Kiaatsco.—“ Studien zur Geschichte der Mammarorgane. 1 Thiel: Die 
Taschen und Beutelbildungen am Driisenfeld der Monotremen,” 
‘Zoologische Forschungsreisen in Australien,’ Bd. ii, Lieferung ii, 
Jena, 1895. 

Rosz.—“ Beitrage zur Zahnentwicklung der Hdentaten,” ‘Anat. Anz.,’ 
Bd. vii, 1892. 


584, J. T. WILSON AND J. P. HILL. 


35. Husrecut.— Die Phylogenese des Amnions und die Bedeutung des 
Trophoblastes,” ‘Verhand. der K. Akad. v. Wetenschappen te 
Amsterdam,’ Tweede Sectie Dl. iv, n. 5, 1895. 

36. GecenBAUR.—‘ Bemerkungen iiber die Milchdriisen-papillen der Sauge- 
tiere,” ‘ Jenaische Zeitschr.,’ Bd. vii, 1873. ‘Zur genaueren Kenntniss 
der Zitzen der Saugetiere,” ‘Morph. Jahrb.,’ Bd. i, 1876. ‘ Zur 
Kenntniss der Mammarorgane der Monotremen,’ Leipzig, 1886. 

37. Maret? Tims.—‘ Notes on the Dentition of the Dog,” ‘Anat. Anz., 
Bd. xi, 1896. 


EXPLANATION OF PLATES 25—82, 


Illustrating Messrs. J. T. Wilson’s and J. P. Hill’s paper, 
“Observations upon the Development and Succession of 
the Teeth in Perameles ; together with a Contribution to 
the Discussion of the Homologies of the Teeth in Mar- 
supial Animals.” 


All sections drawn were outlined by means of Zeiss’s camera lucida. 
List oF Common REFERENCE LETTERS. 


C. S. Coronal Section. H.§. Horizontal Section. %. S. Sagittal Section. 
U. J. Upper Jaw. LZ. J. Lower Jaw. 


cb. Connecting bridge. d/. Dental lamina. mp. Dermal papilla. dl. 
Residual dental lamina. cm. Meckel’s cartilage. 7%. 0. Labial laminar out- 
growth. o.¢. Oralepithelium. ¢. Tongue. 7, *, p?, &c. First upper incisor, 
upper canine, second upper premolar, &. 73, z pg, &e. Second lower incisor, 
lower canine, third lower premolar, &c. diz, d*, dp, &e. First lower milk 
incisor, upper milk canine, third upper milk premolar, &c. 


The lingual side of the figure of each coronal and horizontal section is 
indicated by a cross. 


Stace II. 


Fics. 1—5.—Serial coronal sections, U. J. 
Fig. 1. In front of third incisor region. 
Figs. 2 and 3. Next succeeding sections through enamel-organ of di3, 
attached to “neck” of dental lamina, and showing earliest formation 
of dermal papilla. 


DEVELOPMENT AND SUCCESSION OF TEETH IN PERAMELES, 585 


Fig. 4. Third section behind Fig. 3, showing di merging posteriorly in 
common dental lamina. 
Fig. 5. Third section behind Fig. 4, through Anlage of 73. 
All x 210 diameters. 


Fic. 6.—S. §. Cutting the anterior portions of the upper and lower dental 
lamine transversely, not far from the mesial plane. a. Alinasal cartilage. 
ch. Choanal passage. cm. Meckel’s cartilage. J.O. Organ of Jacobson. 
J. c. Cartilaginous capsule of Jacobson’s organ. ma. Maxillary palate. x. e. 
Epithelium of nasal cavity. pmz. Premaxilla. rf. Olfactory bulb. 8. d. 
Stenson’s duct. ¢. Tongue. x 25. 


Fic. 7.—U. J., C. §. Canine Anlage. x 210. 

Fig. 8.—U. J., C. S. Milk premolar (dp3). x 210. 

Fic. 9.—U. J., C. S. Milk premolar (dp?) in another embryo from same 
pouch, with commencing liberation of the residual dental lamina. x 130. 

Fic. 10.—S. §. dp® and molar segment of dental lamina (d/.). x 50. 

Fie. 11.—U. J., C. S. Anlage of m+ with commencing papilla-formation 
(mp.m+). xX 180. 

Fie. 12.—L. J., C. S. Anlage of diz. x 230. 

Fie. 13.—Z. J., C. 8. Canine Anlage. x 210. 

Fig. 14.—8S. S. dpz and molar dental lamina (d/.). x 50. 

Fie. 15.—Z. J., C. 8. First molar Anlage. x 180. 


Stace ITI.—P. obesula. 


Fic. 16.—U. J., H. S. Anterior end of dental lamina, showing the labial 
portion of the Anlage of 2+ as a lobe projecting at right angles from the 
main dental lamina (d/.), and the continuation backwards of the latter (d/’.) 
towards the region of 22. x 140. 

Fie. 17.—U. J., C. S. Second incisor mass, showing the labial portion of 
the Anlage of 22 projecting from the swollen lamina, and the representative 
of di2. x 180. 

Fic. 18.—U. J., H. S. Second and third incisor masses (II and III), with 
rudiments of di® and di3. The level of the section is indicated by the dotted 
linesa ....6in Figs.17 and 19. x 180. 


Fic. 19.—U. J., C. 8. Third incisor swelling, showing the narrow ‘ neck” 
attaching the mass to the oral epithelium and the remains of the enamel-organ 
of a3. x 180. 

Fic. 20.—U. J., C. S. Canine mass, showing the as yet undifferentiated 
Anlage of the permanent canine, with the cup-shaped enamel-organ of the 
milk canine (da), c.dl. The morphologically free edge of dental lamina. 
c. The probable equivalent of the labial lobe of the incisor masses. x 210. 


586 J. T. WILSON AND J. P. HILL. 


Fie. 21.—U. J., H. S. Cutting the oral epithelium at the point marked 
0.:6.- X90: 

Fic, 22.—U. J., H. S. At a somewhat higher level than Fig. 21. x 90. 

Fie. 23.—U. J., C. S. Through the third incisor (2) of an early fcetus of 
Dasyurus viverrinus. x 330. 

Fic. 24.—U. J., C. §. dp. x 180. 

Fie. 25.—U. J., H. S. Molar region at a higher level than Fig. 22. x 90. 

Fic. 26.—U. J., H. S. m+ at a higher plane than Fig. 25. x 90. 

Fie. 27.—U. J., C. S. Middle region m2. x 130. 

Fie. 28.—U. J., C. S. m+ in the third section behind Fig. 27. x 180. 

Fic. 29.—U. J., C. §. Posterior region of m1. x 120. 

Fig. 30.—U. J., C. S. Dental lamina just behind m2. x 120. 

Fie. 31.—U. J., C. S. Dental lamina five sections behind Fig. 30. x 120. 

Fie. 32.—U. J., C. 8. Dental lamina in region of future m2, and just in 
front of its termination. x 120. 

Fic. 33.—Z. J., C. 8. iz and diz. x 280. 

Fie. 34.—L. J., C. §. Anlagen of tg andiz. x 120. 

Fic. 35.—Z. J., H. 8. Anlage of tz and zz. x 90. 

Fie. 36.—Z. J., H. S. Nearer the oral epithelium than Fig. 35. x 90. 

Fic. 37.—L. J., C. S. Anlage of permanent <, with deeply cupped enamel- 
organ of dz. x 210. 

Fig. 38.—Z. J., C. 8. Anlage of pz. x 150. 

Fie, 39.—L. J., C. 8. dpz. X 280. 

Fie. 40.—Z. J., C. 8S. dpz, seven sections behind Fig. 39, showing the 
differentiation of cusps. x 180. 

Fic. 41.—L. J, C. 8S. mz. x 120. 

Fic. 42.—Z, J., C. §. Dental lamina behind enamel-organ of mz. x 1920. 


Stace IV. 

Fic. 43.—U. J., €. S. Dental lamina in front of 22. ca. Common Anlage 
of lip-groove and dental lamina proper. xX 120. 

Fie. 44.—U. J., C. 8. i+. ep. Epithelial ingrowths (see text). x 120. 

Fic. 45.—-U. J., C. S. i2 three sections behind Fig. 44. x 120. 

Fic. 46.—U. J., C. 8. i+ five sections behind Fig. 45. x 120. 

Fic. 47.—U. J., C. S. ¢ with remains of d. x 100. 

Fic. 48.—U. J., C. 8. dp® passing through the protocone. x 130. 

Fie. 49.—U. J., H. 8. Through the region between p2 and m2. For ex- 


DEVELOPMENT AND SUCCESSION OF TEETH IN PERAMELES. 587 


planation of the projections from the dental lamina marked “a” and “db” 
see text. x 100. 


Fig. 50.—(Stace V.) U.J., H.S. Region between p? and m2. xX 100. 
Fie. 51.—U. J., C. S. m+ passing through its protocone. x 80. 
Fie. 52.—U. J, C. 8. m2. x 80. 


Fig. 53.—U. J., C. S. m2 in another series, and nearer its posterior end 
than Fig. 52. x 80. 


Fic. 54.—U. J., C. 8. Anlage of m3. x 80. 

Fie, 55.—Z.J., 0. 8. iz. x 100. 

Fic. 56.—L. J., C. 8. iz some distance behind Fig. 55, showing diz. x 100. 
Fie. 57.—L. J., C. 8. tg andiz. x 50. 

Fie. 58.—Z. J., C. S. Dental lamina in front of ; with d&. x 210. 


Stace V. 
Fie. 59.—U. J., C. 8S. £. x 50. 
Fic. 59a.—(Stace VII.) U.S,C. 8.5. 50. 
Fie. 598.—(Stace VIII.) U.d.,C.8.£ x 50. 
Fie. 60.—U. J., C. S. dp® and dental lamina in region of future p2. x 165. 
Fie. 61.—U. J., C. 8. Anlage of m3. x 220. 


Stace VI. 
Fie. 62.—U. J, 0. 8. *. =x 50. 
Fic. 63.—U. J., C. S. dp® with Anlage of p®. x 50. 
Fic. 64.—L. J., C. 8. dpz with Anlage of pz. x 120. 


Stace VII. 


Fic. 65.—ZL. J., C. §. Passing through the forwardly directed apex of #;, 
and showing the enamel-germ of an incisor (r. e. 0.) which arises from the 
dental lamina in front of 7; x 210. 

Fie. 66.—L. J., C. S. Through iz, showing diz independently attached to 
the oral epithelium. x 150. 

Fie. 664.—(Stace V.) JZ. J., C. §. Through 15, showing di; becoming 
attached to the oral epithelium labially from the neck of the dental lamina. 
x 150. 

Fic. 67.—(Stacz IX.) U.J., C0. 8. dp® with Anlage of p?. x 30. 

Fie. 68.—(Stace VI.) U.J., C. S. Posterior portion of m2. x 120. 

Fie. 68a.—(Stace VI.) U.J., C. 8. Some distance behind Fig. 68, show- 
ing Anlage of m&. x 120. 

Fic. 69.—(Stace VIIL) U.J., C. 8S. Enamel-germ of m2. x 120. 


588 Ji Di! WILSON: AND! J: P. HELE. 


Fie. 70.—(Stace IX.) U.S, C0. 8. m&. x 120. 

Fig. 71.—(Stacre X.) U. J, C. S. Enamel-organ of m® with residual 
dental lamina, surrounded by concentrically arranged connective tissue. x 80. 

Fie. 72.—(Stace X.) U.J., C. 8. Passing through £, showing one of the 
enlargements of the residual dental lamina surrounded by a connective-tissue 
capsule. x 230. 

Fic. 73.—(Stace X.) U.J., C. S. Hnamel-germ of p&. x 120. 

Fic. 74.—(Stace XIII.) U.J.,C. 8. Enamel-organ of pz, with com- 
mencing formation of a residual dental lamina. x 90. 

Fic. 75.—(Stace XIII.) U.J. From below. x 23. 

Figs. 76 and 77.—Coronal sections through the third premolars (lower and 
upper) of Phascolarctus cinereus, showing residual dental lamine by 
the sides of p2. S/’. Residual dental lamina by side of pz. x 30. 

(After Leche, Taf. xvi, figs. 140 and 142.) 

Fic. 78.—(Stace XIV.) U.J. Region of premolars and molars from 
below. xX 33. 

Fie. 79.—(Stace XIV.) JZ. J. Region of premolars and molars from 
above. X 33. 

Fie. 80.—Dasyurus viverrinus. ZL. J., C.S. dpz. X 180. 

Fic. 81.—Dasyurus viverrinus, JZ. J.C. 8. dpz in an older fetus. 
x 1380. 

Fic. 82.—(Stace V.) JU. J. Combination-drawing from several hori- 
zontal sections. x 55. 

Fic. 83.—U. J. Pig embryo of 20 mm. head-length. @C. 8. Dental lamina 
in posterior molar region (see text). X 75. 

Fic. 84.—Z. J. Pig embryo of 20 mm. head-length. C. 8. Passing just 
posterior to ¢z, showing labial laminar outgrowth (/. 0.) and lip-groove 
(1. gr.). xX 30. 

Fic. $5.—Z. J. Pig embryo of 20 mm. head-length. C. 8. passing through 
iz, showing labial laminar outgrowth (/. 0.) and lip-groove (/. gr.). x 30. 


INDEX TO VOL. 39, 


NEW SERIES. 


Acanthocephala, a new genus (Arhyn- 
chus) of, by Shipley, 207 

Allen, 
nervous system of crustacea, 33 

Amphioxus, a new species of, by 
Willey, 219 

Anguilla vulgaris, metamorphosis 
of, by Grassi, 371 

Arhynehus hemignathi, a new 
Acanthocephalon, by Shipley, 207 

Arthropod’s heart and vascular systein 
compared with enlarged vessels of 
Magelona, by Benham, 15 

Ascidian, new form from New Guinea, 
by Willey, 161 

Asymmetron, new species of, by 
Willey, 219 , ~ 


Benham, W. B., on fission in Nemer- 
tines, 19 


Benham, W. B., on the blood of | 


Magelona, 1 

Bernard, an attempt to deduce the 
Vertebrate eyes from the skin, 343 

Blandford, W. F. H., translation of 
Grassi on Termites, 245 

Blood colouring-matter of Magelona, 
by Benham, 4 

Blood-corpuscles of Magelona, and 
effect of reagents thereon, by 
Benham, 5 


Callotermes, Grassi on, 249 

Carinella linearis, Mont., identi- 
fied, by Benham, 21 

Carinella (Nemertine), histology of 
the nerve-cord and muscular coats, 
by Benham, 28 

Cheetz (= sete) of common earth- 
worm chiefly composed of chitin, 
those of smaller Oligocheta not 
so, Goodrich on, 66 


Edgar J., studies on the | 


Chitin and chitinoid substances pre- 
sent in the cceelomic corpuscles and 
epidermic products of Enchytreus 
and other Oligocheta (micro-chem- 
ical examination of), by Goodrich, 
62 

Chlamydomyxa montana, a new 
Protozoon, by Lankester, 233 

Ceelomic corpuscles of Euchytreus 
hortensis described, by Goodrich, 
55 

Ceelomic corpuscles of Pachydrilus, 
sp., described, by Goodrich, 61 


| Ctenophora, systematic relations of, 


by Willey, 3238 

Ctenoplana, by Willey, 323 

Cuticle of the common earthworm, 
soluble in distilled water, Goodrich 
on, 65 


Degeneration of polypides and em- 
bryos in Lichenopora verru- 
caria, by Harmer, 97 

Dentition of Perameles, and of Mar- 
supials in general, by Wilson and 
Hill, 427 

Development of the teeth of Pera- 
meles, by Wilson and Hill, 427 

Digestion, changes during, in gland- 
cells of Drosera, by Huie, 387 

Drosera, changes in the gland-cells 
of, during digestion, by Huie, 387 


Kel, common, metamorphosis of, by 
Grassi, 371 

Egg, formation of, in Lichenopora, 
by Harmer, 104 

Enehytreus hortensis, n. 
described, by Goodrich, 51 

Eyes of Vertebrates, deduced from 
the skin, by Bernard, 343 


sp. 


590 


Fission of embryos observed in 
Lichenopora, and compared with 
that in Crisia, by Harmer, 119 

Fission of posterior segments con- 
taining ripe gonads in Carinella, 
and histology of the process, by 
Benham, 22 


Gland-celis of Drosera, changes in, 
during digestion, by Huie, 387 
Goodrich, E.S., notes on Oligochetes, 

with the description of a new 
species, 51 
Grassi on the reproduction and meta- 
morphosis of the common eel, 371 
Grassi on the Termites, 245 


Harmer, 8S. F., on the development 
of Lichenopora verrucaria, 
Fabr., 71 

Hill, see Wilson. 

Huie, L., on the changes in the gland- 
cells of Drosera during digestion, 
387 


Lankester, Prof., on Chlamydo- 
myxa montana, 233 

Leptocephalus, various forms of, by 
Grassi, 371 

Lichenopora verrucaria, Fabr. 
(Polyzoon), right-handed and left- 
handed colonies described, by Har- 
mer, 76 


Magelona, a marine Chetopod, its 
mode of occurrence, characters of 
its blood and vascular system, by 
Benham, 1 

Marsupials, dentition of, by Wilson 
and Hill, 427 

Methylene blue used for demon- 
strating cells and fibres in the 
nerve-ganglia of embryo lobster 
(Homarus), by Allen, 33 


Nautilus, nepionic shell of, by Willey, 
222 

Nautilus, photograph from life, by 
Willey, 179 

Nautilus, varieties of, by Willey, 227 

Nautilus, Willey on, 145 


INDEX. 


Nephridia of Enchytreus hor- 
tensis described, by Goodrich, 51 

Nepionic shell of Nautilus, by Willey, 
229 


Nerve-fibres, course of, in nerve- 
centres of embryo lobster, 33 

Nerve-ganglion cells of lobster (Ho- 
marus) described and classified, by 
Allen, 35 


Ornithorhynchus, brain of, by Elliot 
Smith, 181 


Perameles, developmental history of 
the teeth, by Wilson and Hill, 427 

Polyclades of New Guinea, Willey 
on, 153 

Protozoon, a new species of (Chla- 
mydomyxa montana), by Lan- 
kester, 233 


Shipley on Arhynchus, a new genus 
of Acanthocephala, 207 

Siphuncle of Nautilus, Willey on, 168 

Smith, Elhot, on brain of a foetal 
Ornithorhynchus, 181 

Spectrum of blood colouring-matter 
of Magelona, 4 

Styeloides eviscerans, an As- 
aWuen from New Guinea, by Willey, 

1 


Teeth of Perameles and Marsupials, 
by Wilson and Hill, 427 

Termes lucifugus, Grassi on, 249 

Termites, Grassi on, 245 

Thread-cells of Myxine compared 
with the cclomic ‘“ thread-cor- 
puscles” of Enchytreus hor- 
tensis, by Goodrich, 59 


Vertebrates, eyes of, deduced from 
the skin, by Bernard, 343 


Willey, letters from New Guinea, 145 
»» On anew Amphioxus from the 
Louisiade Archipelago, 219 
», on Ctenoplana, 323 
», On the nepionic shell of Nau- 
tilus, 222 
»» On varieties of Nautilus, 227 
Wilson and Hill on the dentition of 
aria especially of Perameles, 


PRINTED BY ADLARD AND SON, 
BARTHOLOMEW CLOSE, E.C., AND 20, HANOVER SQUARE, W. 


or 2 ft ny fu! : 
Aas Aaa : 


pes 
% ie , 
7‘) 
* 
7 i 
ys 
Ef 
‘J 7 
a 
oh 
iy 
ve) f = 
: yy ar 
i : 7 ia) A? a 
: : ay 
od 


Li 

f ‘ i 7 ue IF 
: a hon | 
7 Vu pe 


f 
i, 


7 As 
rer, 


ut ey 
(CP ab ao 


yy 
walt 


ERNST MAYR LIBRARY 


IAEA OAT 


3 2044 110 319 829 


Date Due 


APR 1970