NO. 5 BARNARD: AMPHIPODA 13 



that canyon samples would reveal a haphazard occurrence of species in- 

 dicating unstability of bottom, with each sample representing a dififerent 

 starting point in a myriad of successional regimes. This impression was 

 entertained after observing the occurrence of Listriolobus pelodes, Dor- 

 villea articulate, and various species of Diopatra in canyon samples in 

 depths and in combinations with other species that were not normal for 

 the coastal shelf. The strength of that expectation is not fully dissipated 

 by the discovery of identifiable assemblages that indicate either a degree 

 of stability in the substrate or a rapid repopulation of substrates after 

 their consolidation. 



Hartman (1955) presented partial analyses of numerous slope sam- 

 ples (50 - 300 fms) in which Pectinaria is often mentioned, but in which 

 Chloeia pinnata is more abundant and with Maldane may represent the 

 principal community dominant of the upper bathyal outside of canyons. 



Capitella capitata has been suggested by Hartman (1963) as an in- 

 dicator of undersea leakage from emergent sweet- water aquifers in can- 

 yons. It is tolerant not only of brackish waters but of polluted conditions 

 in waters of normal salinity and may thus be an indicator of natural 

 putrefaction. It lives in high densities in the inner harbor of Los Angeles 

 in waters of normal salinity but low dissolved oxygen (see Reish, 1959 

 and Reish and Barnard, 1960). Its occurrence in some canyon samples 

 may be related to high contents of organic matter in the sediments that 

 are restrictive to other metazoans. Capitella appears to tolerate wide 

 ranges of physical conditions that are restrictive to most organisms but 

 apparently is seldom found with other animals. In the depth-sediment 

 scheme (Graph 11), the Capitella samples are grouped in the coarse 

 sediment range, indicating the presence of percolating water that must 

 leak through coarse sediments. In bays and harbors Capitella inhabits 

 fine-grained sediments (Reish, 1959). 



SAMPLE ASSOCIATIONS, METHODS 



Despite their faults, trellis diagrams of association between pairs of 

 samples, based on the percentage composition in each sample of each 

 species, matching species in samples and summing the minimum per- 

 centages for an index of association, have been used to examine the in- 

 terrelationships of canyon samples. Samples of low diversity more often 

 match as pairs with high association indices than samples of low di- 

 versity matched with samples of high diversity. Thus, sample 4851 in 

 Graph 8, with 66 species and 657 specimens, is a poor matching partner 

 with 6779, having 18 species and 116 specimens. Both are dominated 

 by Maldane sarsi and, indeed, 4851 has nearly twice as many specimens 



