HYDROIDA 



sequeiitly the system brought about b\- the Auierican investigators, picking up heterogeneous char- 

 acters, is indeed, as I have shown in an earHer treatise (1909), to a great extent one of arbitrariness, 

 placing the species now in one, now in another genus, according to the character it is thought desir- 

 able to emphasize at the moment. The system has, in tliis way, become a matter of chance. 



Before proceeding to the special explanation of the several species of hydroids and their occur- 

 rence in the seas of the northern Atlantic, we, therefore, must try to realize, from a systematic and 

 phylogenetic point of view, the value of each single character. In the first instance the question must 

 be answered what part the gonophores play as to the phylogeny of the hydroids or, in other words, 

 what importance has to be attached to the gonophores and their conditions for the purpose of 

 classification. 



Tiie application of the characters of the gonophores as distinguishing marks of higher system- 

 atic unities, of families, or of subfamilies, has been abandoned by all modern investigators of the 

 hydroids. On the other hand, it is held by several investigators that they are of importance as to the 

 limitation of genera. The last significant publications maintaining this view, are those of Kiihn (1913) 

 and Stechow (1913). Kiihn, however, applies the characters of the gonophores with much caution 

 and discretion, while Stechow, as a glance at his tables (1913, p. 36 and the following pages) suffices 

 to show, undiscerningly recurs to the principle of laying the main stress on the "medusa" as contrasted 

 with the "sporosac". In fact, as long as the limits between these designations are not more precisely 

 defined, they will be subject to much arbitrariness. An interesting instance is afforded by the point 

 in dispute, where the limits are to be drawn between the genera Curync and Syncorync. The old 

 criterion, accepted by Allman and other investigators, was the free medusa as contrasted with the 

 fixed gonophore or sporosac, and, to all appearance, it is the same limitation Stechow tries to 

 maintain in his table. Indeed, this principle of limitation was only supportable on the ground of the 

 deficient knowledge of the organisation and development of the hydroid gonophores attained to in 

 AUmau's days. Kiihn, therefore, (1913, p. 229) is seen to take quite another departure, defining the 

 Syncorync as including all species having "medusa:", while the Coryi/c is distinguished by styloid 

 gonophores. It is evident from Kiiiin's previous argument (I.e. p. 174) that under "medusae" he also 

 includes the eumedusoids which normally do not break loose, and that accordingly Coryne, in his 

 opinion, exclusively embraces species having styloid gonophores. 



Parallel to these genera Stechow also treats of Podocorync (with medusae) and oi Hydractinia 

 (with "sporosacs"). Kiihn (1913, p. 227) groups them in the following vague way: 



o ocorync 1 ^^^^^^^^^^ (Margelinen), Eumedusoide, Cryptomedusoide, Styloide. 

 Hydractinia ) 

 From his premises (1. c. p. 226) it appears that he agrees with most modern investigators, think- 

 ing it right that the two old genera should be merged into one, Hydractinia; for he states that ni 

 fact we have here before us a series of closely allied forms, in which "zwischen Arten mit Vollmedusen 

 und einfachen Medusoiden die verschiedensten Ubergange bestehen". Thus Kiihn, elsewhere considering 

 the characters of the gonophores as significant generic criteria, in this place actually reduces them 

 to mere specific characters. 



In another connection Kiihn (1913 p. 197) gives an instance of the fact that the sexes in 



