A CRITICAL RE-STATEMENT OF THE BIOGENETIC LAW. 97 
which rested on the Basals ; nor, if vve overlook this developmental modifi- 
cation, did one of those ancestors in its adult condition ever possess Orals 
which were hinged to the Radials on the edge of the cup, capable of opening 
and shutting over the entire disk like the valves, of a trap-door (Bather, 
fig. xxxiii. ; MacBride, 191-J-, figs. 408-410). Where does the snre, frail 
"Ariadne-thread" conduct us now? If no fossil evidence were avail- 
able, anyone who should attempt to reconstruct the ancestral Crinoid on 
the common embryological assumption that the stalked larva of Antedon 
represented an ancestral adult stage would go inevitably astray, as many 
have done already in spite of palpeontological knowledge *. For the 
remarkable thing is that several existing Crinoids possess an arrangement of 
oral valves in the adult precisely or closely similar to that of the larval 
Antedon (Uolopus, 1. c. xxxiv. ; Hi/ocrinus, Sedgwick, 1909, fig. 209 ; Thau- 
matocrinus. Carpenter, 1884, pis. iii., vi., Ivi.). The relations of the oral 
plates in Antedon to the vestibular roof of the larva, as well as the 
temporary suppression of Radials, are cleai-iy "cenogenetic" features. 
If a vestibule was a feature in the development of the earliest Crinoids, 
the oral plates must have been deposited beneath its floor, and not in 
its roof. Their relation to the roof (which alone enables them to split 
apart and function as valves) is an embryonic mutation. It is, therefore, 
scarcely open to doubt that the condition of the oral valves in the adult 
Holopus (and Hyocrinus ?) is due to the retention of a feature that was 
purely embryonic, not adult, in origin, and that, in this respect, these 
interesting Crinoids are as " predomorphic " as any Perennibranchiate 
Amphibian. 
This brings me to my last point. When the common argument is tirged 
that the stalked larva of Antedon "recapitulates" the adult stage of its 
stalked ancestors, it seems to be forgotten that every type of Pelmatozoan, 
from prse-Oambrian Cystids to the present time, must also have possessed a 
tiny fixed stage of simple structure following a free-swimming larval life, 
and that the main features of the skeleton must have been laid down in that, 
or a still earlier, stnge of its ontogeny. It follows from what has been said 
that the modern Pentacrinoid larva of Antedon is a modification of the 
corresponding stage of the ancestral ontogenj', not of the adult stage, and 
that the adult Antedon is not an addition to the ontogeny of any preceding 
Crinoid, but just a modification of the adult phase of the same ontogeny — 
partly by loss (e. g. anal turret and plates, Oral plates, stalk, &c.), partly by 
* P. H. Carpenter (1884, p. 145) compared tlie vestibulate condition of Antedon with 
tlie Camerate condition of Haplocrinus (Bather, xxxv.) and the Platycrinidce (/. c. xL). It 
is a tempting suggestion, especially as the Carboniferous Platycrinidfe possessed bifasciate 
oval stem-ossicles. But these types are Monocyclic; the relations of the "Orals" are only 
superficially similar, and it is very doubtful if these plates are other than enlarged " proxi- 
mal ambulacrals " (see Bather, pp. 127-129, and fig. xli.). 
LINN. JOURN. ZOOLOGY, VOL. XXXV. 7 
