A CRITICAL KE-STATEMENT OF THE BIOGENETIC LAW. 95 
Ordovician. — Dendrocrinoid modification. Arms still non-pinnulate, but 
dichotomonsly branched ; the disk now flexible by interpolation of 
small plates between Orals and Radials, but Ambulacra remain supra- 
tegminal. Anals as before, supporting a great anal turret (xxvi.). 
Madreporite lost. 
Silurian. — Dendrocrinoid arms forked and pinnulate (Dendrocrimis, iii.). 
T-, . -V Dendrocrinoids lose RA, and then X from cup (Graphio- 
Devonian. f . i .1 • . , , . . . 
/', 1 -n Y crinus aaa li,risocrinus,c) ; their arms become biserial ; 
J Infrabasals reduced and covered by stem. 
Trias. — From Graphiocrinidse arise Pentacrinidse : cup (patina) shallow, 
the disk bulging up between the arms ; no persistent proximal 
columnal as yet ; stem cirriferous, its ossicles changing from round 
to pentagonal, with petaloid furrows and radiating ridges ; no Anal ; 
arms (again ?) uniserial. Various members of the family swim about, 
and re-anchor themselves by distal cirri of stem. Obscurely leading 
to Flexibilia Pinnata, but exact links missing (cf. Bathycrmus, with 
bi-fasciate stem-ossicles, like Bourgueticrinidas and AntedonidaB, 
figs, xlix., cxv.). 
Jui'assic. — The first " Pinnata," with persistent proximal columnal and 
reduced Basals ; stem-ossicles of modified Pentacrinid or Bathycrinid 
types. Millericrinus (no cirri) broke away from bottom of its stalk 
for swimming, the stem being slowly absorbed (Iii.). ThioUiericrinus, 
the first Antedonid, with cirriferous compound Centrodorsal, fairly 
stout stem, and bifasciate joints of stem-ossicles. 
Lias. — Antedon and later types break away from top of stalk early in life. 
With this sketch of the adult ancestry o{ Antedoii before us, let us now 
see how it is recapitulated in the ontogeny. Certain sequences of form- 
change take place in the same order. The larva on fixation exhibits 
successively an armless (" Cystid ") stage, and stages with simple arms, 
forked arms, and pinnulate arms respectively ; the Oral plates at first are 
co-extensive with the disk, as in Blastoids * and Cyathocrinoids, and then a 
peripheral growth-zone (perisome) is established between them and the 
Hadials as in Ordovician Dendrocrinoids ; an Anal plate appears within 
the circlet of Radials, rises above it and disappears, as in Carboniferous 
Dendrocrinoids ; the Infra-Basals arise separately and then fuse with one 
another and the proximal columnal, marking the change from Dendro- 
crinoidea to Flexibilia Pinnata. 
* Another Blastoid or pre-Crinoid relic appears to be involved in the abortive attempt 
■of tire 5 primary tentacles to branch and grow as radial canals in the vestibulate stag-e, each 
-peristomial tentacle (in spite of its lack of food-grooves and skeleton) representii)g a Blastoid 
brachiole. If Blastoid brachioles are homologous with Crinoid pinnules, the view that 
simple unbranched arms are primitive in Crinoids becomes untenable. 
