512 
HUBERT LYMAX CLARK 
derived from the Dendrochirota at a comparatively recent date 
and there is good reason for thinking Eup3Tgus has had a differ- 
ent origin from the rest of the order and has simply converged 
into the group. 
The relationship of the Holothurioidea to other Echinoderms 
If the proposition made above, that the five interradial ten- 
tacles of the synaptid pentactula are characteristic organs of the 
ancestral holothurian, be accepted, the first question which natur- 
ally arises, when we consider what relation the holothurians bear 
to other Echinoderms, is this: What interradial outgrowths of 
of the hydrocoel are there in other Echinoderms with which these 
tentacles can be homologized? At first approach the problem 
seems a very difficult one, for so far as our knowledge of echino- 
derm embryology goes, the only direct outgrowths of the hydro- 
coel in the other classes are the radial canals, pore-canal and 
polian vessels, all of which are also present in our hypothetical 
holothurian ancestor. A careful study of recent work of Grave 
('02) and AlacBride ('03) on the development of echini, has led 
me to the rather surprising conclusion that the 'primary tentacles 
of synaptids are homologous with the dental sacs of the echini. This 
idea at first seems preposterous because the dental sacs of echini 
have no connection whatever with the hydrocoel. More careful 
consideration shows, however, that this objection is far from 
conclusive. Both Grave and AlacBride show that the dental- 
sacs of echini arise as five interradial outgrowths of the left coe- 
lomic pouch, which lies below for on the inner side of) the hydro- 
coel and from which the hydrocoel itself arose. These outgrowths 
push out between the budding radial canals, and thus these two 
series of outgrowths come to occupy the same position with rela- 
tion to each other which the primary- and secondary outgrowths 
of the hydrocoel occupy in the pentactula. There is, of course, 
this important difference that in the young echinoid, the lumens 
of the dental sacs have no connection with the water-vascular 
system, while in the pentactula the homologous lumens are con- 
tinuous with that of the circular water-canal. If, however, the 
essential identity of the left hydrocoel and the left posterior coelom 
