260 University of California PuhJicafions in Zoology [Vol. 18 



sequently, in hauls encountering only a very few protruding chains 

 the chance of losing the single solitary salpa of each chain would be 

 much greater than that of losing all the aggregate salpae. This, I 

 believe, is why no solitary forms were found in these three hauls. 

 Moreover, in four others (1432, 1779, 1854, and 1864) individuals of 

 the aggregate generation alone were recognized at first, but on reex- 

 amination mature solitary forms were found, which makes the above 

 explanation more plausible. The crude data also show that out of 

 thirty-six hauls made in water below 18? 6C, each of twenty-five con- 

 tained both solitary forms and aggregate forms, while each of the 

 remaining eleven failed to obtain a single individual of either genera- 

 tion. This seems explicable only on the assumption that protruding 

 chains were encountered in nearly if not all successful hauls. 



It is well known that chains of even such large species as Salpa 

 fusiformis and S. zonaria are obtained entirely intact only with the 

 greatest difficulty. Says Herdman (1889, p. 59): "Out of the 

 enormous number of aggregated Salpae collected during the Challenger 

 expedition, none were adhering together when they reached my hands. 

 In all cases the chains . . . had become broken up into their con- 

 stituent Salpae." Although this has not been the result of collections 

 made under the auspices of the Scripps Institution, chains of no 

 species have been obtained entirely intact except when collecting with 

 extreme care with a dip-net. Would it be surprising, then, if the 

 pressure and swirl of the water in a tow-net completely breaks up 

 whatever chains of the smallest and most delicate species, S. dcmo- 

 cratica, may be encountered? 



Granting this to be the explanation why protruding chains of 

 this species have never been described, it follows that the effect of the 

 swirling water in causing breakage would be less pronounced the 

 shorter the duration of the haul. Working over the collections from 

 this point of view, it was noticed that, while all the hauls herein con- 

 sidered consumed upwards of twenty minutes, a few thirteen-minute 

 hauls had been made in the vicinity of Santa Rosa Island during 

 August, 1911. Furthermore, while no portion of a chain was found 

 in any of the hauls entered in table 1, several fragments ivere dis- 

 covered in one of the above mentioned thirteen-minute hauls (2766). 

 In three instances two aggregate forms were found attached together ; 

 in another, three larger ones were attached ; and in still another, five 

 were attached together so as to form a double chain as illustrated in 

 plate 9, figure 1. There can, therefore, be no doubt that, as Agassiz 



