I9O W. C. ALLEE. 



each year we have broken previous records with monotonous 

 regularity, for numbers of species from most of the localities we 

 visit. This could not have continued so long had the animals 

 been becoming less abundant. 



The number of animals present in a given locality must 

 depend more on the availability of suitable breeding places and 

 abundance of food than upon such disturbing influences as 

 summer collecting, particularly when the collecting does not 

 reach all the breeding habitats of a region and there is adequate 

 means of distributing young stages. This conclusion is empha- 

 sized by the rapid recovery in numbers of Arbacia after their 

 almost complete disappearance following the winter of 19 17-18 

 (Allee, '19) and in the face of their destruction by the thousands 

 in the research work carried on in the Woods Hole laboratories. 



VI. Summary. 



1. Analysis of distribution records in the four major types of 

 habitats of the Woods Hole littoral, viz., wharf pilings, rocks 

 and rockweeds, flats, and the sea bottom in deeper water show 

 that mere records of species present in the different habitats 

 fail to indicate any relationship between the different types 

 of associations. 



2. By eliminating species known to be approximately equally 

 distributed throughout and records for one year, only, and 

 classifying the remaining species in terms of places where they 

 are most abundant and next most abundant one finds : 



(a) The association of the wharf pilings is closely related to 

 that of the rocks. 



(b) Species taken in dredging on clean hard bottom are found 

 in next abundance on the rocks. 



(c) The associations of the flats are highly independent of the 

 others in the region but continue in the mud and sand under and 

 around rocks. 



(d) That some degree of quantitative work- is necessary in 

 order to determine the relationships of animal associations. 



3. Preliminary collecting in a region tends to give the obvious 

 forms and gives similar results in analysis to the type of quasi- 

 qualitative work described in this report. 



