291 Dr. F. A. Bather on the Homologies 
This is allied in structure to H. heros, Guer. j in both 
species the sixth sternite of the male has four longitudinal 
carina?, excluding the raised lateral margins, the outer pair 
ending in a small spine and with a large tubercle at 
the base, but the seventh sternite is broadly rounded at the 
apex in titan and truncate in heros. I do not think that 
titan is a mere colour variet3% though heros varies much in 
colour, especially on the thorax ; but it may be a local 
race. 
XXXIII. — The Homologies of the Anal Plate in Antedon. 
By F. A. Bather, D.Sc, F.R.S. 
(Published by permission of the Trustees of the British Museum.) 
Of late years, since rigid distinctions were first drawn between 
the various plates in the posterior interradius of Palaeozoic 
Crinoids, the plate which appears, migrates, and disappears 
in the posterior interradius of the larval Antedon, and is 
called the anal plate, has been regarded as homologous with 
the plate generally known as anal x (the brachiaual of 
Bather, 1890) in the Crinoidea Inadunata and Flexibilia. 
Dr. Austin Hobart Clark, however (1912, Journ. Washing- 
ton Acad. Scr. ii. pp. 309-314, and 1915, Monogr. Existing 
Crinoids, vol. i. part 1, pp. 331-339 *), attempts to prove 
that the anal of Antedon is homologous with the radianal 
(Bather, 1890). A plate presumed to be the same occurs in 
the young of Promachocrinus. The representative of anal x 
is found by Dr. Clark in the posterior one of the small inter- 
radial plates occasionally observed in Antedon and other 
normal comatulid genera, while in Promachocrinus he would 
homologise it with one of the additional arm-bearing plates 
(pararadials, Bather, 1900). 
Considering the extensive use that has been made of the 
anal plates in the classification of the Palaeozoic Crinoids, 
it seems advisable to examine Dr. Clark's arguments. But 
first let us recapitulate the main characters of anal x and the 
radianal, as seen in the Inadunata and Flexibilia. 
Both of these plates are intimately connected with the 
* The references, except when otherwise stated, are to the latter 
work. 
of the Anal Plate in Antedon. 295 
right posterior radios. This is universally admitted with 
regard to the radianal, which is in its origin the right posterior 
inferradial, in other words part (if not really the whole) of 
the radial itself. The sanu- connection does not appear to be 
so generally recognized in the case of anal .r, but this con- 
nection is one of the reasons that led mo to give that plate 
the name brachianal. Facts proving the connection were 
published in the ' Geological Magazine ' for January 1899, 
and summarized in Lankester's ' Treatise on Zoology,' part iii. 
Echinoderma, pp. 120-122 (1900). 
The geological history of the radianal is briefly as follows: — 
It begins in Ordovician time3 as an inferradial *. Then, as 
the posterior interradins widens, the radianal of the Dicyclic 
Inadunata moves to the left of the right posterior radial, and, 
while retaining its contact with the lower margin of that 
plate and with the two underlying basals, comes also into 
contact with anal ,r, if that plate be present (e. g., Palaeo- 
crinus and Botryocrinus). In Silurian and Devonian times 
the widening of the interradius continues, the radianal con- 
tinues to move to the left and comes into contact with a plate 
(rt) which sinks down on the right side of the anal tube 
between .i; and r.post.R. (e. g., Euspirocrinus). This arrange- 
ment is emphasized in such Carboniferous genera as Poteriu- 
crinus. In some later forms of the same group the radianal 
may stretch upwards, as the body of the animal with its anal 
structures comes to lie on rather than in the dorsal cup; but 
while the other anal plates (.r, rt) pass up beyond the limits 
of the cup, the radianal invariably retains that connection 
with the lower slope of r.post.R. which bears witness to its 
inferradial origin. This arrangement is best shown in some 
North-American and British species of Ulocrinus, and I have 
recently discussed them in regard to this feature "j". A parallel 
* An inferradial is the lower half of a transversely bisected radial, the 
upper half being distinguished a9 superradiai. These terms were chosen 
to indicate the truly radial nature of these plates. O. Jaekel and A. H. 
Clark call the inferradials " subradials." To this term there are two 
objections. First, it implies that the plates are not radial elements, but 
independent plates developed below the radials. This, however, does not 
appear to be the intention of those two authors. Secondly, the term 
subradial was used for the basals by De Koninck, and Dr. Clark seems to 
have forgotten (p. 104) that in this use De Koninck was followed by 
several American writers of repute down to the end of the nineteenth 
century. As to possible homologies of the inferradials, see Bather, 1900, 
' Treatise on Zoology,' iii. p. 112, and 1913, Bull. Mus. Ottawa, i. 
pp. 9, 14. 
t 1917, Trans. Geol. Soc. Glasgow, xvi. pp. 210-212. On p. 211, in 
line 4 of the middle paragraph, " left posterior basal " is a blip fur " right 
posterior basal." Of course, RA never touches l.post.B. in any genus. 
206 Dr. F. A. Bather on the Homologies 
course is followed by the Eupachycrinus series. In many 
allied forms (e. g., Delocrinus) it is plain that the radianai 
Lad disappeared before the migration of the other anals from 
the cup was complete. There is no evidence that the radi- 
anai ever followed those plates outside the cup limits. It 
could only do so by losing its primitive morphological con- 
nection with the lower end of r.post.R., a connection which it 
invariably retains throughout the manifold modifications of 
the anal area. The fate of the radianai, as indicated by the 
fossils from Cambrian to Permian, is to disappear by atrophy 
or resorption while still below the upper margin of the dorsal 
cup. In the fossils from the Trias to the Pleistocene no 
trace of it is found. 
In the Monocyclic Inadunata the history of the radianai 
is somewhat different. In the Pisocrinidae and their descen- 
dants it is pushed to the right of r.post.R. instead of to the 
left, and is eventually squeezed out of existence at the lower 
end of the cup ; it never rises between the radials. 
In the Palaeozoic Flexibilia Impinnata the radianai may 
assume a position abutting on anal x, similar to that in 
Botryocrinus (vide supra) } or it may remain below r.post.R., 
or even be thrust down into the basal circlet. The essential 
point in the present connection is that no part of it ever rises 
between the radials, as in the Dicyclic Inadunata. The facts 
are given by Springer (1906, Journ. Geo!, xiv. pp. 516-519). 
If any Of the later crinoids, including the comatulids, are 
descended from the older Flexibilia, and correctly classed as 
Flexibilia Pinnata, then it is important to note that the radi- 
anai has not been observed in the adult of any one of them 
from Triassic to Recent times. 
Returning to anal x, and confining our attention to the 
Dicyclic Inadunata, we find it beginning in Ordovician 
genera above, or partly above, or between the two posterior 
radials. In the last two cases it rests on the posterior basal, 
but if a radianai be present it abuts or in part rests also on 
that plate. It sinks furthest down into the cup in forms 
with a wide anal area, such as Carabocrinus and Thenaro- 
criaus, or in the peculiar Gasterocomidae, where the anus 
opens on the side of the cup itself. Further width is attained, 
especially in Poteriocrinus and its allies, by the sinking of 
right and left tube-plates (rt and It) into the cup, the former 
even meeting the radianai. The extreme of this development 
is reached in some Lower Carboniferous genera, such as 
Woodocrinus. Then begins the consolidation of the cup and 
the raising of the viscera. As the rectum passes upwards, so 
of the Anal Plate >'n Antedon. 297 
also does its proximal supporting plate ,r, which thus grad- 
ually rises above the radial circlet, until the two posterior 
radials meet beneath it (e. g., Erisocrinus) . Thenceforward, 
from the Trias onwards, the identity of anal x i.s lost in the 
general plating of the adult anal tube when such a structure 
exists. 
Apart from all hypothesis, the historical facts make it 
clear that anal x and tho radianal are structures differing in 
origin and development and ultimate fate. While anal a 
has supra-radial relations, those of the radianal are inferradial. 
While anal x finally passes up out of the cup, the radianal is 
resorbed when within the cup limits. 
If, then, in recent crinoids a plate is observed in young 
2jes between the posterior radials, and then migrating 
upwards beyond the limits of the cup, it is natural, on the 
theory of recapitulation, to regard that plate as x rather than 
as the radianal. 
Let us now consider Clark's arguments. 
First, as to the anal of Antedon and similar forms. He 
says this is more closely connected with r.post.B. than with 
l.post.R. But this is just what I have alwa\ d on 
as the case with anal x in the Ordovician and Silurian In- 
adunata. Jf this tendency of the Antedon anal to keep to 
the right M is a fact of the very highest importance," then I 
welcome it as confirming the views I have always expressed 
concerning the true anal x. 
nk then passes to the anal in the young of Promacho- 
crinus (p. 332). First he describes it as arising "in the 
rhombic area between the corners of the basals and orals" 
before any of the radials appear. That is to say, it lies in 
the right posterior radius. " Soon afterward the [r.post.] 
radial appears, just to the right of and in line with''' the anal 
plate. This radial grows faster than the anal and gradually 
"surrounds" it, so that the anal ''comes to lie in a deep 
concavity in the side of the radial." Later the r.post. 
"radial extends itself beneath the anal and the concavity 
becomes straightened out and disappears, the anal con- 
currently being shoved diagonally forward (to the let t) and 
disappearing by resorption." In a subsequent paragraph it 
is added that the concavity which receives the anal is in "the 
lower left hand portion of the radial," also that the anal 
migrates upwards, 
These facts, says Dr. Clark, " leave no room for doubt that 
the so-called anal of the pentacrinoid larvae is nothing more 
nor less than the radianal of the fossil forms." 
.1////. & Mag. JS. Hist. Ser. 0. Vol i. 20 
298 Dr. F. A. Bather on the Homologies 
In the absence of figures for Promachoerinus it is impossible 
to be perfectly clear as to the precise relations of the plate in 
question to the right posterior radial ; but the description is 
far from convincing me that Dr. Clark's conclusion is justified. 
In the well-known figures of the Antedon larva by W. B. 
warpenter and others, and in those of Hathrometra prolixa 
Chich Clark here reproduces from Mortensen as evidence in 
his favour, I find nothing to indicate that the anal is anything 
other than anal x. The upward migration of the plate 
entirely favours this view. If the anal of Antedon be not 
homologous with the plate in Promachoerinus , then the latter 
might possibly be the radianal, since it does not migrate 
beyond the limits of the cup. That fact, however, is scarcely 
conclusive, since there is a special reason for it in this genus, 
as will shortly appear. 
Let us, then, see what further arguments Dr. Clark has to 
offer. " Since," he writes, " the radianal is represented in 
the pentacrinoids of the comatulids we should expect also to 
find in the posterior interradius a second plate which we 
could with a reasonable degree of probability identify as the 
representative of the plate known as anal x ; and such a plate 
actually occurs." This, of course, would be almost con- 
clusive ; but the statement needs careful checking. 
Fiisf, consider the facts adduced for genera with five 
radials, as in Antedon. Wyville Thomson (1865, Phil. 
Trans, p. 540) "in one or two cases observed " in Antedon 
bifida, " about the time of the first appearance of the anal 
plate, a series of five minute rounded plates developed inter- 
radially between the lower edges of the oral plates and the 
upper edges of the basals." The fate of these plates is 
uncertain. Thomson himself identified them with certain 
perisomic interbrachials of the adult, but P. H. Carpenter 
(1884, Chall. Rep. p. 40) doubted this, and regarded them as 
true interradials, ultimately rtsorbed in Antedon, but homo- 
logous with the permanent interradials of Thaumatocritms. 
With these plates A. H. Clark (p. 335) homologizes five 
plates which appear in Comactinia (species not stated) at the 
time of formation of the fii st primibrachs (IBr,) and lie on 
the shoulders of the radials at such a height that their upper 
halves are between the IBij. In a single specimen of 
Comactinia rneridionalis (p. 317, fig. 412) each such plate was 
surmounted by two others. In Comatilia iridometriformis 
(of rather later stage, pi. ii. figs. 528, 529) a " large rounded " 
plate rests in each interradius above the interbrachial 
process of the radials. 
of the Anal Piatt in Antedon. 209 
If these plates in Comactinia and Comaiilia are houx] >. 
gous with those first observed by Thomson in Antedon bifida, 
then they Beern to support Tliomson's interpretation of those 
plates. Further, plates occupying such a position in the 
adults of Palaeozoic crinoids, or in such Meaozoic genera as 
Grttet tardier iniu and Uintacrinus, are no longer called " inter- 
radials" but " interbrachials," and are not regarded as 
homologous with the true interradials of such forms as the 
Bhodocrinidae. Consequently there does not seem to be good 
reason for regarding such plates in the ordinary comatulid 
larva as homologous with true primary interradials. Their 
late appearance in development also suggests that they do 
not represent plates of former importance in the cup. Even 
if these plates were primary interradials, they would not, in 
my opinion, have any bearing on the anal question. Anal x 
is a characteristic plate of the Inadunate Crinoids — in other 
words, of those crinoids which are devoid of true interradials 
or of any interbrachially situate cup-plates in interradii other 
than the posterior. Whatever anal x may be, it is a special 
plate developed or adapted for the widening of the anal area 
and the support of the rectum. The same is the case with 
the corresponding plate in the Adunata and the Flexibilia. 
There is no reason for regarding it as one of five primary 
interradials, retained while the other four have disappeared. 
Consequently the existence of plates, whether interbrachials 
or true interradials, in all five interradii of certain comatulids, 
does not prevent us from regarding the specially developed 
anal plate as the homologue of anal x. 
Dr. Clark's final argument, on which he lays most stress, 
is drawn from Promachocrinus and Thaumatocrinus. Ever 
since P. H. Carpenter described Thaumatocrinus renovatus in 
1884 there has been a tendency to regard the plate in the 
posterior interradms, which supports a short somewhat arm- 
like process, as an anal plate. It is, however, one of five 
similar plates, each separating the adjacent radials, and 
therefore, so long as attention was confined to the original 
specimen, all to be regarded as true interradials. Dr. Clark 
still so regards them, and at the same time homologizes the 
posterior one with anal x. That homology is open to the 
same objections as have just been raised in the case of the 
supposed interradials of the ordinary comatulids. 
Dr. Clark's own work, however, by putting a new com- 
plexion on T. renovatus, has made his homology even moie 
difficult of acceptance. He has shown, in the first place, 
that T. renovatus is the young of the species later described 
20* 
300 Dr. F. A. Bather on the Homologies 
as Promaclwcrinus abyssorum. The process borne by the 
posterior plate is indeed a developing - arm, and Clark suspects 
" that smaller arms borne on the other interradials have been 
lost. . . . During growth the posterior interradial arm of 
Thaumatocrinus becomes reduplicated on all the other inter- 
radial plates, and all of the five interradial arms gradually 
increase to the size of the five primary arms so that the 
10-armed Promachocrinus abyssorum results" (p. 338). It 
follows from this that the supposed interradials of Thaumato- 
crinus, including the posterior one or supposed anal, are of 
precisely the same nature as the five arm-bearing plates 
which in Promachocrinus have been added to the five normal 
radials of the ordinary comatulid. Dr. Clark calls all these 
plates "interradials," a term which suits his argument, but 
which scarcely seems justified. 
In 1900 (Lankester's 'Treatise on Zoology/ iii. p. 150) I 
suggested that these " interradial radials " were of the same 
nature as the arm-bearing plates in the cup of the Monocyclic 
Inadunata Calycanthocrinus and the Catillocrinidae, for which 
plates Jaekel in 1895 had devised the excellent term " para- 
radials." Dr. Clark now tells us how the pararadials of the 
Promachocrinidae develop. They "arise very early in life 
and are from the first equal in height to the radials. They 
are probably . . . best interpreted as a sort of lateral budding 
from, or a delayed reduplication of, the radial to the left. 
As the radials move apart [the pararadials] continue to 
broaden, and their development in all ways is proportionate 
to their breadth as compared with the breadth of the normal 
primary radials" (p. 337). The development of the arms 
which they support bears a similar relation to the arms borne 
by the normal radials. The opinion that each pararadial is 
in a sense derived from the radial to the left of it is confirmed 
by various facts. Thus the posterior pararadial always 
maintains " a closer relation with the" left posterior radial 
than with the right posterior (p. 336). In some thirty 
6-rayed specimens of Promachocrinus studied by Dr. Clark 
the supernumerary ray is in all cases but two inserted to the 
right of the left posterior radial, and receives its food-groove 
from the groove-trunk leading to that radial (p. 338). 
The posterior pararadial appears to originate slightly before 
the others ; in the original specimen of Thaumatocrinus 
renovatus its arm was more developed, and,- as just stated, it 
is sometimes the only one to be formed. These facts are 
very simply explained as due to the relatively greater widening 
of the posterior interradius by the pressure of the rectum. 
of the Anal Plate in Antedon. 301 
ft is not necessary to suppose, as Dr. Clark does, that the 
posterior pararadial represents a phylogenetically persistent 
anal ./•, which subsequently is reduplicated in the other 
interradii. 
The facts given by Dr. Clark serm to me to show that 
flu' pararadials of the Promachocrinidae are really of radial 
origin. There is no more reason to regard them as inter- 
radiala than there would he in the case of the Oatillocrinidae. 
A similar proliferation of arm-bearing, or, rather, brachiole- 
bearinu', elements is seen in many Oystidea Rhombifera, 
e.g., Chdrocrinus (Bather, 1913, Trans. R. Sue. Edinburgh, 
xlix. p. 446, 6gs. 52-55). 
Dr. Clark, it is true, has some ingenious remarks, appa- 
rently intended to show that there is no real difference between 
a radial and an interradial. To quote from his paper of 
1912 (p. 312) : " while a plate if situated below the ventral 
edge of the perisomic surface may give rise to a simple series 
of more or less similar plates running up to the edge of the 
ventral surface, and possibly continued further along the anal 
tube, the same plate if situated just at the ventral surface will 
give rise to an arm or a group of arms exactly like those 
arising from the radials. The character of the ossicles 
following a plate is not determined so much by the character 
of the plate itself as by its position in reference to the 
boundary between the dorsal and ventral surfaces of the 
animal." This seems to me hopelessly inconsistent with 
"well-known facts. On the one hand, there are plenty of* 
genera in which the anal is on a line with the radials and yet 
does not give rise to an arm (e. g., Hexacrinus) ; on the 
other, there are genera in which some radials, though on a 
level with the others, cease to bear arms (e. g., Tribrachio- 
crinuSf Sycocrinus). 
We have, then, found no reason for accepting Dr. Clark's 
statement that a " representative of anal x actually occurs " 
along with the plate which he calls the radianal. Conse- 
quently that argument for the radianal nature of the latter 
plate falls to the ground. The facts do, however, throw 
light on the non-migration of the anal in Promachoainus, the 
peculiarity which, it was admitted, might conceivably en- 
courage one to regard it as the radianal (anted, p. 296). 
Dr. Clark .-ays (p. 337) : " I have examined pentacrinoids of 
Promachocrinus kerguelensu in which both the radianal and 
anal x are present, the former dwindling, the latter increasing 
in size. They are situated side by side between the two 
posterior radials-" Obviously the lateral growth of the 
302 Mr. O. Thomas on the 
posterior pararadial (Clark's " anal x ") prevents the upward 
migration of the anal (Clark's " radianal ") and inhibits its 
further growth, partly by drawing on its stereom for its own 
supply of calcium carbonate. 
We may, therefore, continue to regard the anal in the 
Promachocrinidae as homologous with that of the other 
comatulid larvae, and, in all, as the representative of anal as. 
XXXLV. — On the Arrangement of the small Tenrecidae 
hitherto referred to Oryzorictes and Microgale. By Old- 
field Thomas. 
(Published by permission of the Trustees of the British Museum.) 
There has long been some doubt as to the distinction from 
each other of the two genera Oryzorictes and Microgale, the 
latter of which I described in 1882, twelve years after 
Grandidier had described the former, and in consequence of 
this doubt the generic allocations used by Forsyth Major in 
describing the many new forms of Tenrec-shrews which he 
discovered during his successful expedition of 1894-96 have 
never been revised or confirmed. 
I have now had an opportunity of going over the Museum 
material of the group with a view to putting its generic 
arrangement more in order than it was left by Dr. Major, 
who never completed the admirable work he began on it. 
No additional specimens have been received since his 
collection came, but the fine series he obtained, combined 
with those previously collected by Mr. Deans Cowan and 
worked out by me, have enabled me to obtain some idea of 
the natural arrangement of the group. 
I find that it may be divided into five genera, whose chief 
characteristics are set out in the following synopsis : — 
A. Claws not markedly fossorial, the anterior 
not or little longer than the posterior. 
Canines not dominant, commonly low and 
bifid, and never surpassing the anterior 
incisors. 
a. Molars with marked internal lobe. In- 
cisors diminishing backwards, the canine 
considerably longer than P. Muzzle 
little elongated, the teeth touching each 
other. Fore-claws not longer than 
bind.