SABELLIDES AND SERPULIDES 187 



genus is restricted to these two species, with the first taken as type. 

 But, as no figures have been given, we can form no definite conception 

 of the form and arrangement of the setae or of other important features, 

 showing the great need of a more careful study of these species. 



In constructing the following analytical table for the genera which 

 are related to the genus Sabella^ an attempt has been made to base it 

 on characters which can be readily seen with the aid of a good pocket 

 lens, the tables hitherto published being so complicated as to require 

 much careful microscopic work before one can arrive at the generic 

 relation of any species. 



In studying the various forms representing the numerous genera, 

 certain details in structure are found to be repeated a certain number 

 of times, forming a definite sequence or continuous evolution, as in the 

 development of the collar. 



Taking the form without a collar as the primitive type, the anterior 

 edge of the first segment becomes more or less elongated in front, form- 

 ing one or two more or less conspicuous lobes. When a collar begins 

 to develop, the entire anterior edge may be produced into a free mar- 

 gin without any openings ; or one incision or cleft may occur, forming 

 an opening on the back, the ends being in contact or meeting; or only 

 a portion along the sides and in front may be produced, forming a 

 collar open on the back with widely separated ends. The same 

 process of development taking place in the anterior margin of the first 

 segment of the two-lobed type will produce a two-lobed collar, either 

 with ends in contact or separated on the back. When additional 

 incisions or clefts develop on the sides of either of these two-lobed 

 forms, two corresponding four-lobed collars are formed, those with 

 separated ends usually having the lateral incisions toward the front 

 (ventro-lateral), while in those where the ends are in contact the 

 incisions are toward the back (dorso-lateral). It therefore seems 

 desirable to use the collar as an important character in grouping the 

 genera. Other characters also of these primitive forms are found to 

 be repeated; the setae and uncini especially, or variations of them, 

 being repeated many times in various combinations which can be ar- 

 ranged in definite groups. 



It will be found that the concise facts in regard to many of the 

 36 genera cited are much too meager to render it possible for one to 

 place each genus in its exact or correct relative position. There is 

 still much work to be accomplished before a perfect analytical table 

 can be formulated. 



