PENNATl'LIDA. 



I cannot regard the grouping of Delage & Herouard as an improvement on the system of KolUker, 

 and most of the genera, found in these authors but not in those before mentioned, will have to be 



abandoned. 



Nor can I, for my part, regard the system of KoUiker as satisfactory ; the characters, by whicli 

 the division into groups has been made, hardly always make the line of separation in a natural way; 

 when, for instance the first group Pennatulcce is made to contain the forms in which the individuals 

 of the polyp-series are coalesced to form wings, this feature is seen not to be of absolute validity: in 

 the genus Virgularia, species are found in which the wings are partly dissolved into free individuals, 

 or, more properly, in which only some individuals are coalesced, and onh' to quite a slight extent 

 [Virgul. bromlcyi K. and Virg. cladisais niihi); the same is found in the genus Stylatula (in the sub- 

 genus Diibenia). On the other hand, forms are found in the group Spicatce^ in which a coalescing of 

 some individuals in an oblique series takes place, in which, accordingly, there is as distinct a beginning 

 of wings as in the mentioned Virgtilarue; this is seen in Authoptilnm grandiflorum Verr., while in 

 another species of the same genus, A. murrayi K. the polyps are quite separate, and parth- even not 

 arranged in regular oblique series. A one-sided prominence, given to this single character, will thus 

 easily separate closely related forms into two groups; but when other structures are also taken into 

 consideration, I think that the four groups of Kolliker may be kept — at least provisionally — but, 

 to be sure, with altered contents. If we knew more of the development of the colou>- in a larger 

 number of Pennatulids than is the case now, the mutual relations of the forms niiglit certainh" be 

 seen more clearly, and then, perhaps, the grouping might be different. For the present, it is exceedingly 

 difficult' to reach a sure classification ; difficulties occur on all points, especially in limiting off the species. 

 Most of the features that have generally been used as specific characters prove to be more or less 

 useless: the number, grouping, size of the polyps and zooids often change considerably during growth, 

 as also all the relative sizes of the parts of the polyps and the stem, which are moreover exceedingly 

 influenced by the degree of contraction. Nor does the anatomical structure give many holds for the 

 classification, as it is upon the whole too uniform; the .•^jncules also give but slight help: species of 

 the same genus may quite lack .spicules or be abundantly provided with them (for instance, within 

 the genus Umbcllula)\ nor is the form and size of the spicules often of much assistance; in specimens 

 of the same species both the number and the size ma\- be very \-ar>ing (for instance, in Fimiculina 

 (/iindrangularis, Kophobclrmnon stcllifrrmn). Amongst the better characters, I reckon the presence or 

 absence of a calyx on the polyp.s, and the form of the calyx. For the delimitation of genera aud 

 families, these features along with the grouping of the polyps and zooids in the developed colony will 

 for tile present be of most value. 



I am not in a position to undertake a complete reformation of the system of the Pennatulids 

 on the basis of the material I liave hitherto had for examination, but on several points 1 shall be able 

 to revise, with greater certainty than has before been possible, certain parts of the system of Kolliker 

 and the later additions to, and alterations of, this system; on other points I shall have to be content 

 with giving suggestions more or less positive. The more particular reasons for the position of the 

 forms dealt with in the following section, will be given there; in this place I shall only give some 

 conclusions with regard to these forms. 



