224 POPULATIONS OF THE SEA 



is an excellent opportunity for such collaboration in the study of 

 the world populations of Calanus and related genera about which 

 a lot of new information has emerged in recent years. 



In the examples I have chosen, and indeed throughout ecology, 

 there is an implied emphasis on the study of what Hedgpeth 

 called "ecological units or communities." As I have suggested, it 

 is extremely difficult to define and classify communities of pelagic 

 organisms. Indeed, Allee et al. (1949) did not confine this difficulty 

 to the pelagic zone ; they argued that all the components of marine 

 life are interdependent and they treated the w'hole of marine 

 ecological associations as a single major community. On the other 

 hand, some workers have used "community" to describe any as- 

 semblage of species found in the same place, without any implica- 

 tion of ecological relationships. 



Distribution, abundance, and species composition are to some 

 extent parts of a continuously variable system, but we all know 

 that certain combinations of species occur more frequently than 

 other combinations and that the organisms in such groups are a 

 fairly constant part of each other's biological environment. It is 

 impossible to complete the study of the ecological relationships 

 of a species without reference to the other organisms with which 

 it lives. 



The approach to the study of associations between species owes 

 much to the techniques of the psychologists who ha\'e faced the 

 problem of grouping various measures of human performance. 

 The statistical techniques known as cluster analysis or factor 

 analysis have been used. Many of the tests depend on the calcula- 

 tion of correlation coefficients of various kinds which are then as- 

 sembled into a matrix from which similar and dissimilar groups 

 can be selected, preferably by some objective test. Kontkanen 

 (1957) has described some of these techniques. The terrestrial 

 botanists have been concerned with the same problem of con- 

 tinuity or discontinuity of vegetation; for example Poore (1956) 

 and Goodall (1953, 1954). Fager (1957) devised an index of affinity 

 between species and showed how ranking methods and concorda?2ce 

 analysis could be applied to the study of various interspecific rela- 

 tionships. MargalefT (1958)' discussed the problems of temporal 

 succession and spatial homogeneity in the phytoplankton. He 



