246 W. C. ALLEE. 



survey that there is always a lively interest in single factor 

 indices of associations, and at the beginning of the present 

 studies I rather expected to find such an index in the pH relations. 

 While a combination of the average pH, the extent of range and 

 the relative position of the extremes does allow one to place 

 these associations in their natural order with considerable 

 exactness, and while such data is very suggestive it does not 

 classify these associations with the precision necessary for a 

 successful single factor index. This is emphasized in the rock 

 series of associations, Tables XIV. and XV. 



The use of the oxygen content of the water is out of the question 

 as such an index of an association. Temperature is an aid but 

 needs confirmation. Salinity, whether average or range, con- 

 sidered with position of extremes, does arrange the different 

 associations in their logical order, and so qualifies for the recom- 

 mendation that was predicted for pH, as the best single index 

 when water conditions alone are considered. The readily deter- 

 mined factors of salinity, temperature and pH taken together 

 give a much stronger index than any one of them alone. 



If, however, on account of urgent haste, I should be forced to 

 make use of a single criterion to divide the communities of the 

 Woods Hole littoral, I should depend more on observation of the 

 character of the sea bottom than on any other one factor. 1 This, 

 the most obvious, the longest used, is still the least treacherous 

 single factor index of littoral distribution in this region. It 

 should be used with discretion since a rock well back on the flats 

 supports a different set of animals from one on an exposed point, 

 but the corrections are more obvious and more easily applied. 



All the data collected in the present investigation agree in 

 supporting the common sense conclusion that animal associations 

 in a region such as this under consideration are not normally 

 limited by any one factor, but by the interaction of several, and 

 when feasible all these should be analyzed and recorded. 



1 Shelf ord ('14) in writing of the suitability of water for fishes concluded that 

 the amount of clean bottom, the amount of carbon dioxide and the amount of 

 hydrogen sulfide, taken together serve as an index of availability of bays and 

 enclosures of the seas for fish life. Longley ('22) in studying the local distribution 

 of Tortugas fishes concludes that the local distribution of many species is determined 

 by the character of the bottom. This holds particularly for what he calls the 

 "sand-patch" association. 



