236 BULLETIN 100, UNITED STATES NATIONAL MUSEUM. 



of rays in the lateral fins. In devising the keys these specificities 

 are used and, except for Sagitta maxima (Conant), which I still find 

 impossible to separate from. S. lyra, every species recognized as valid 

 by Bltter-Zahony (1911) and several others described subsequent to 

 his report are included. There is great need of keys for identifying 

 poorly preserved material, but the minute structure of seizing jaws 

 and other skeletal parts of the head, upon which such identification 

 depends, is still undescribed in nearly half the species. The keys 

 included in this report are therefore adapted only to the identification 

 of well-preserved specimens. 



The area covered by the expedition is too large and the hauls 

 were too scattered to yield definite information concerning the dis- 

 tribution of the various species obtained. As pointed out elsewhere 

 (1916, p. XVIII), variability in plankton distribution is enormous, 

 and hydrobiological relations are too complex to be revealed without 

 frequently repeated collections in very restricted areas and searching 

 hydrographic observations corresponding in time and place to each 

 net haul. Even though thousands of individuals of one species and 

 none of another be obtained by a single haul, no conclusion is justi- 

 fied other than that the former species was obtained and that the 

 latter was not obtained. Such data afford no adequate evidence 

 for concluding that the former species is more abundant in or more 

 typical of that particular locality than is the latter. Had 20 or 30 

 hauls been made at each station with nets of similar filtering capaci- 

 ties, sufficient evidence for such conclusions might have been obtained. 

 But rarely more than one haul was made at each station, so the data 

 3 r ield little else than records of occurrence. These records are given 

 for each species. 



In comparing the species occurring in the Philippine region with 

 those obtained from the vicinity of San Diego (the only coastal 

 region of the Pacific off either of the American continents from which 

 chaetognaths have been described), two interesting and suggestive 

 facts come to light: First, those species obtained in largest numbers 

 from the Philippines are those which, as a rule, occur rarely, if at all, 

 in the San Diego region, and the opposite. Second, of those species 

 common to both regions, the number of teeth is greater in Philippine 

 specimens. 



This is contrary to what might have been expected. For, judging 

 from the fact that chaetognatha collected under the auspices of the 

 Scripps Institution from as far south as 23 north, off Lower Cali- 

 fornia, are essentially like those within the San Diego region proper, 

 and, realizing that this is but two degrees north of the northern 

 boundary of the Philippine area from which chaetognatha were 

 obtained, one would naturally infer a close relationship between the 

 Philippine and San Diego faunas. To find it quite the reverse is 

 therefore suggestive of a fundamental and far-reaching difference in 



