INTRODUCTION. CXXVll 



I jet ween tin- two sections, it would only be one more proof 

 that in our classifications we may not look for sharply de- 

 fined provinces, hut must be content with the " vanishing 

 lines " of nature. 



Our real perplexities commence when we attempt to 

 frame the families and genera; for in the case of the 

 Polyzoa it is extremely difficult to seize the significant 

 characters. This remark applies especially to the Cheilo- 

 stomatu. Amongst the Cyclostome forms* there is so 

 much uniformity in the zorecia, that we are compelled to 

 base our divisions on the grouping or mode of combina- 

 tion. Throughout this suborder the cell has undergone 

 the slightest modification ; but there is an almost endless 

 variety in the zoarial form. In dealing with the Cteno- 

 stomata important characters are supplied by the polypide 

 and the structure of the zooecium, and we have compara- 

 tively little difficulty in defining well-marked groups. 

 But in the case of the Cheilostomata, and especially of the 

 Escharine forms (in the classification of Smitt), it is 

 otherwise. Here the polypide gives us no material help, 

 exhibiting few variations of any systematic importance. 

 Amongst the Aeteida it is of a very simple type, and 

 reminds us of that of the Crisiida ; in this, as in other 

 elements of structure, this group makes an approach 

 to the Ctenostomata. In the genus Bugula it exhibits 

 strongly marked features ; but such differences as generally 

 occur are devoid of any special significance. The avicu- 

 are for the most part much too inconstant to serve 

 the purposes of the systematist ; and their history shows 

 us that as indications of affinity they arc seldom to be 



* On the gemmation of the Cyclosiomata, we Barroio, ' Embryologie,' 



