1884.] 95 [Hyatt. 



occurs after the endoblast is ciliated and the two membranes in 

 consequence are less distinguishable from each other. Among Cal- 

 cispongiae on the other hand it occurs when the ectoblastic cells 

 are amoeboidal and the endoblast ciliated; and as the inner layer 

 is in this case formed by the ectoblast there is here also no diffi- 

 culty in detecting them. 



The process of forming this layer by delamination occurs 

 before the permanent invagination of the endoblast in the Calcis- 

 pongiae as well as in the Carneospongiae and the parenchymula 

 stage is present even in Sycandra raphanus, thongh it is much 

 abbreviated as above stated. Schultze (Zeitschr. Wissen. Zool., 

 vol. xxxi, p. 269) considers that the middle layer arises from the 

 ectoblast and it became visible only after invagination. It may, 

 however, have existed and probably did exist previously, as in 

 MetschnikofPs larva of Sycandra raphanus (op. cit., vol. xxxn, 

 p. 368, pi. 21) which is described and figured as having an inter- 

 nal layer in the free amphiblastula stage, probably derived from 

 the endoblast. 



The later larval stages of development also show concentration 

 but not the extent of abbreviating the proximal ancestral charac- 

 teristics. The ascula has a simple cavity with a single cloacal 

 aperture, and then follows the adolescent larval stage usually 

 considered as the olynthus predicted by Haeckel ; but which, 

 with Vosmaer, we must consider as really an ascon stage. It is 

 undoubtedly a great triumph for Haeckel and the law of biogen- 

 esis that it is so closely comparable with his olynthus, but the 

 spicular skeleton, and characteristics are those of the adult 

 Ascones, and it may even be referred to the genus Asculmis. It 

 has four-rayed skeletal or supporting spicules and single acerate 

 defensive spicules arranged as in that genus. We think this close 

 correlation probably means that the distinctions between 

 Ascones, Sy cones and Leucones are artificial, and that the genus 

 Asculmis, for example, is a low or radical genus of a sub-ordinal 

 or family group including Sycandra among its higher forms. 

 Such an arrangment, cutting across the lines of classification 

 already laid down and possibly including in the same groups gen- 

 era of Ascones, Sycones and Leucones, is plainly indicated by the 

 embryology of Sycandra to one practically familiar with phylo- 

 genesis. Such a change cannot now be inaugurated, and we, 



