Notes on the Embryology of Starfishes. 245 



jecting beyond the general outline as in the adult Brachiolaria, 

 but remain always, in the most advanced specimens observed, in 

 an embryonic condition (e e', Figs. 6, 8), as rudimentary 

 arms at the extremities of the horseshoe-shaped groove, which 

 divides the anal and oral plastrons. We find in older larvae quite 

 well marked epaulettes ie' e\ Figs. 6, 8) which are, as I have 

 shown already for Toxopneustes, in my paper on the Embryo- 

 logy of Echinoderms,* only dilatations of the vibratile chord, 

 ■and not special organs of locomotion as Miiller had been in- 

 duced to consider them. 



The prominent characters of Tornaria can be summed up in 

 the permanence of the embryonic features of Brachiolaria, and 

 it will be a curious point to ascertain whether this embryonic 

 type gives rise to what I have been induced from embryological 

 data to consider the lower types of Starfishes, such as Luidia, 

 Ctenodiscus, and Astropecten. It will be most interesting to 

 observe also how far the larvre of types, which are shown tore- 

 tain embryonic features of some adult, have themselves embry- 

 onic features of their larvee. Comparisons of this kind have not as 

 yet been attempted, and promise to afford valuable aid fur clas- 

 sification. The presence of a single cavity of the water system 

 at the oral extremity of the Tornaria throws additional light on 

 the nature of the circulating cavity observed between the rudi- 

 mentary arms of Echinaster embryos. It requires but very slight 

 modifications to transform our Tornaria into a larva similar to 

 the Echinaster embryo with its three club-shaped arms ; imagine 

 the whole of the anal extremity of the Tornaria occupied by a 

 small pentagonal Echinoderm, as in Echinaster, and we have 

 short rudimentary appendages left, inclosing a cavity in which 

 a circulation could easily be perceived ; the opening, b, of this 

 cavity being placed on the edge of the young Echinoderm, 

 as in our Asteracanthion embryo, would readily escape notice. 

 We have thus an additional link to show that the development 



* See Memoirs Am. Acad. Vol. IX. 18G-A. 

 APRIL, 1866. 1 9 Akn. Lye. Nat. Hist. Vol. VIII. 



