211 



formation that has been given here of its larva is sufficient to show that 

 it differs very markedly from the true Strongylocentrotus-lsLTYSS in its body 

 skeleton in the first stage forming a basket-structure; there is then very 

 good reason to expect that also in its second larval stage it will prove to 

 agree with the Toxopneustid larval type. At least the facts known are 

 decidedly not in favour of Clark's views. 



Regarding the genera referred by Clark to the family Echinometridse 

 there can be no doubt that they really form a natural group, Parasalenia 

 alone being doubtful. It must then be claimed that the larvae of these 

 forms should be in conformity with one another in their essential features. 

 It is a pity that we do not know much about these larvae, only one species, 

 Echinometra lucunter, having been reared to its full larval shape and through 

 metamorphosis. But it is known that the larvae of Echinometra oblonga and 

 Colobocentrotus atratus agree with the Ech. lucunler-larva in the very char- 

 acteristic feature of the recurrent rod being double, and it is highly prob- 

 able that the same character applies to the larvae of Echinometra Matheei 

 and Heterocentrotus mamillatus. Thus the facts hitherto known regarding 

 the larvae of the Echinometridae agree very well with the results derived 

 from the study of the adult forms as to their natural affinities. 



Summarizing now the preceding discussion it must be stated that the 

 study of the larvae most decidedly lends support to the author's 

 views as to the classification of theCamarodonta, and thus also 

 gives proof of the correctness of ascribing comparatively great 

 importance to the minor microscopical characters of pedicel- 

 lariae and spicules in the classification of this group. The classi- 

 ficatory results reached on using these characters, combined with the 

 characters of the test, have been splendidly confirmed through the study 

 of the larvae. Disregarding these characters, Clark was led e. g. to such 

 an absurdity as to include Sphserechinus granularis in the genus Slrongylo- 

 centrotm, otherwise so naturally circumscribed through the peculiar char- 

 acter of its globiferous pedicellariae. There is no reason to enter here on 

 a discussion of the objections raised especially by Clark against using 

 these microscopical characters in classification, as it may now be regarded 

 as an estabhshed fact that they are really of eminent classificatory value. 

 Thereby I do not mean to maintain that my classification was correct in 

 all details. Thus e. g. the position assigned by me to the genus Strongylo- 

 centrotus, mainly on account of the structure of its globiferous pedicel- 

 lariae, as being probably related to the Toxopneustidae, was evidently in- 

 correct. The study of the larvae shows that it is most likely to be an offshoot 

 from the Echinid^ s. str. Upon the whole I would emphasize that the true 

 position especially of the forms belonging to the Camarodonta, (excepting 



27* 



