190 FLORENCE BUCHANAN. 



serve to a much greater extent for respiration, though retain- 

 ing their other functions as well. 



The ventral collar, usually taken as a characteristic of the 

 family Serpulidae, has not been actually mentioned as present 

 in the Spionidse. Judging, however, from a very incomplete and 

 imperfect figure given with the Report on Annelids by Webster 

 and Benedict in the 'Commissioner's Report of Fish and Fish- 

 eries for the United States' in the year 1881, there would appear 

 to be a collar of the same sort in Streblospio Benedicti.^ 



Large modified thoracic nephridia are found in several 

 families of sedentary annelids, e. g. the Terebellidae, Her- 

 mellidse, Serpulidae, and Cirratulidse, sometimes a single 

 pair only, sometimes two or even three pairs, and sometimes, 

 as in the Serpulidae, one pair with a common opening to the 

 exterior. They have not been described in other genera of 

 the family Spionidse. 



The sinus round the intestine has also not, as far as I am aware, 

 been described in other Spionids, but a vascular plexus occurs 

 round the intestine in some forms, e. g. Nerine cirratulus.^ 

 A sinus is found in the Serpulidae, ChaetopteridcC, Ariciidse, 

 Terebellidse, and in many of the Cirratulidae amongst others, 

 but cannot be regarded as of much classificatory importance. 



The dorsal collar of the 2nd segment is not found in other 

 Spionids,' nor in other families; but in some Spionids, e. g. 

 Spiophanes Verrilli (described in the samepaper by Webster 

 and Benedict), there is a membranous ciliated dorsal ridge con- 

 necting the bases of the opposite so-called '^ cirri " on every 

 segment from the 6th backwards, and this may be something 

 of the same kind. 



We see, therefore, that Hekaterobranchus has many charac- 

 ters in common with the Serpulidee,^ and I think there is good 

 reason to regard it as the degenerate descendant of a form 



^ See note at the end of this paper. 



2 Claparede, ' Annelides scdentaires,' 1873. 



^ It also has many characters in common with the Cirratulidae which olh 

 Spionids have not. This family is notably closely allied to the Spionids, 

 being probably an earlier and more primitive offshoot than the Serpulida;. 



