1884.] 139 [Hyatt. 



in their development. The absence of the type-larval stage in an 

 animal is not, therefore, evidence against its association with a 

 branch to which its structure brings it in undeniable affinity, pro- 

 vided the omission can be explained upon the basis of concentra- 

 tion and specialization ; but on the other hand the presence of this 

 stage is absolutely determinative. 



The term Radiata established by Cuvier has been banished 

 from the literature of Zoology under the impression that the 

 general homologies which might be made between the types 

 of Porifera, Hydrozoa, Actinozoa and Echinodermata would have 

 no taxonomic value. While recognizing that the distinctions which 

 have been shown between these types are sufficient to entitle 

 them to the rank of different branches, we have never felt dis- 

 posed to slight the homologies founded upon the prevalence of 

 the tendency to vertical division of the body and the predomi- 

 nantly radiatory symmetry in the arrangement of parts. This 

 appears now to receive additional confirmation since the radiate 

 symmetry evidently correlates with similarities in the modes of 

 growth of the diverticula of the archenteron as was shown by 

 Alexander Agassiz (Embryol. Starfish Agass. Contrib., vol. v, p. 

 65) before the views of the Hertwigs with regard to the coelom 

 enabled him to use the general homologies of the body cavity in 

 support of these views. The Hydrozoa, Actinozoa and Echino- 

 dermata are the higher terms of this grand series of animal types, 

 and to this list we add the Porifera as the first and lowest term, 

 thinking that we are amply justified by the facts adduced above 

 in estimating them as true Metazoa radiata. 



We can now, therefore, proceed with a certain confidence to 

 point out still more definitely the primitive character of the struc- 

 tures of Porifera. These animals with their thick middle layer 

 filling up the space between the diverticula, and with these cavi- 

 ties in direct connection with the exterior by means of tubes, 

 present us with a structure, which in horizontal section would not 

 be very remote from Prof. Sedgwick's ideal diagram of the sup- 

 posed ancestral type of the segmented animals (op. cit., pi. 3, f. 

 7). If we imagine this figure altered so as to obliterate the 

 peripheral tube connecting the coelomatic cavities, it would 

 become an approximate diagram of a horizontal section through 

 an adult sycon or the ampullinula stage among Carneospongiae. 



