SABELLIDES AND SERPULIDES 219 



of brown. In life " white or yellowish with brown purple branchiae." 



Like all the species, there are the longer and shorter forms, but all 

 taper gradually backward from the first segment, and have long, well- 

 marked, biannular segments, which vary in number from 70 to 100, of 

 ^ which but 3 belong to the thorax. 



There are 14 pairs of branchiae, each with a rather short and broad 

 terminal portion reaching beyond the web ; the long, well-separated 

 pinnae are sometimes much curled and twisted. 



The triangular ventral lobe of the first segment is well developed ; 

 the lateral puckerings are not always noticeable, and the distinction 

 between the thoracic and abdominal regions is not clearly defined by 

 a groove or furrow. 



The hooked setae, 4 in number, were found on the second and third 

 segments and the uncial plates on the fourth (first abdominal) seg- 

 ment, and form a complete circle around the body commencing at about 

 the twelfth segment, passing posterior to the fascicle of setaa. 



The largest specimen is about 80 mm. long and 3.5 mm. broad at 

 the first segment; branchite about 13 mm. long. The smallest speci- 

 men, of about 50 segments, with 9 pairs of branchiae, is about 17 mm. 

 long and 2.5 mm. broad, with the branchiae 5 mm. long. 



Dutch Harbor, Unalaska Island, July 8 and 17, thirty specimens 

 embedded in much mucus under and between stones on shelly sand. 



Tribe SERPULIDES. 

 Family SERPULID^. 



Comparatively few authors have attempted any systematic work on 

 this difficult group. Philippi in 1844 gave results of his study of the 

 Mediterranean forms ; Morch in 1863 reviewed all the then known 

 species and gave fine figures of the operculum of many of them ; 

 Levinsen in 1883 added to the northern forms, but, as in the case of 

 the Sabellides, Saint- Joseph in 1894 gave an extensive analytical table 

 of the known genera, proposing many new ones, based on the different 

 forms and arrangement of the setae. 



In studying many species, however, one soon finds it impossible to 

 adopt all of his changes, especially in the genus Spii-orbis (see p. 252), 

 and that, although so many new names appear, there are still many 

 interesting and peculiar forms which require to be separated under 

 new genera ; no attempt, however, has been made to find the correct 

 generic relation of all the species hitherto published. 



As similarly stated under the Sabellides, the following analytical 

 table for the genera which are related to the genus Serpula is based 



