46 



BULLETIN 150, UNITED STATES NATIONAL MUSEUM 



more than six. The following table presents the facts from which 

 the above conclusions have been drawn. 



> 5 in first dorsal. 



> 4 In first dorsal. 



The rays of the first dorsal differ from those of the second dorsal 

 not in being unsegmented but in being undivided and therefore 

 consisting of a single rod. The writer has not demonstrated that 

 the notch in all those species in which it is present marks the division 

 between these two types of rays. In the species without the dorsal 

 notch the anterior rays which represent the first dorsal are undivided. 

 This statement is based on an examination of L. dennyi, tunicatus, 

 and tanakae. 



The position of the origin of the dorsal varies somewhat but can 

 not be utilized to any extent in distinguishing species. 



In two species, Liparis rutteri and atlanticus, the first dorsal rays 

 are sometimes greatly elongated. The elevated dorsal is supposed 

 to be found only on the males and during the breeding season. 



The dorsal connection with the caudal is constant within a species. 

 The connection is typically shortest among the tide-pool series. 

 In some of these species the dorsal appears to be free from the caudal. 

 In nearly all such species the dorsal membrane connects with the 

 skin-covered base of the caudal. The connection between dorsal 

 and caudal is typically greatest among the deep-water forms. In 

 Liparis pulchellus, however, a shallow water species, the dorsal 

 connects with fully three fourths of the caudal. Of the deeper- 



