NO. 5 BARNARD: AMPHIPODA 9 



the high side of the scatter diagram. To some extent the low number of 

 crustaceans in canyon heads (Table 3) may be the result of sampling 

 errors, although a large proportion of the samples was taken with the 

 Campbell grab which presumably does not suffer much loss of small 

 motile organisms. 



The very low recoveries of mollusks in the shallow parts of the 

 insular canyons seem significant. The low recovery of crustaceans in all 

 of the canyons, plus their division into so many orders, results in such 

 scanty material that general statements about crustacean community 

 ecology cannot be made. Depth zonation is apparent however. 



COMMUNITY ASSEMBLAGES 



Canyon samples are dominated by the following organisms: Pecti- 

 naria calif orniensis, Maldane sarsi, Capitella capitata, Chloeia pinnate, 

 Pista disjuncta, Dentalium rectius, Heteromastus filobranchus, Ancis- 

 trosyllis tentaculata, Spiophanes fimbriata and Nothria iridescens and 

 Lysippe annectens. All are polychaetes except the scaphopod Dentalium. 

 Large and conspicuous, but not in great abundance, are the following: 

 Brissopsis pacifica and Brisaster townsendi (echinoids) ; Arynchite sp. 

 and Listriolobus pelodes (echiuroids) ; Cerebratulus sp. (nemertean) ; 

 Solemya sp., Yoldia sp. (clams) ; Asychis disparidentata, Glycera amer- 

 icana, G. robusta, Onuphis vexillaria, Lumbrineris sp. and Praxillella 

 pacifica (polychaetes). The depth zonation of some of these and of other 

 important species is depicted in Graph 6. The eurybathicity of these spe- 

 cies is striking. 



Table 5 lists the communities found on the coastal shelf of southern 

 California, in order of their importance. The occurrence of these com- 

 munity types in the canyons is insignificant except for Capitella, a genus 

 that is more frequently abundant in canyons than on the coastal shelf. 

 The most important canyon species is Pectinaria calif orniensis, which on 

 the shelf is a subdominant of the Amphiodia urtica and Amphiodia- 

 Cardita communities. 



■ The codominant frequency (Graph 5) suggests those samples to be 

 tested for community appellations and they have been selected by in- 

 spection of the lists published by Hartman (1963). No group of stations 

 is large enough to ensure statistical uniformity, but as assembled they 

 show considerable differences in the frequency of the various species 

 (Table 6). Of 109 stations of the inshore (coastal, non-insular) can- 

 yons, 78 can be assigned to one of the nine different associations. Most of 



