THE STRUCTURE OF DIATOM COMMUNITIES UNDER 

 VARYING ECOLOGICAL CONDITIONS 



Ruth Patrick 



Department of Limnology, Academy of Natural Sciences of Philadelphia, 



Philadelphia, Pa. 



During the preceding 15 years we at the Academy of Natural Sciences of 

 Philadelphia have spent a great deal of time studying the composition of diatom 

 communities in the eastern and southern United States. The findings from sec- 

 tions of streams characteristic of this area which have not been adversely 

 affected by pollution are discussed in this paper. To understand these com- 

 munities of diatoms not only the species which compose them but also the 

 sizes of the populations of these species must be known. This necessitates 

 collecting species from all types of habitats in the community. It also neces- 

 sitates counting sufiEicient specimens to determine most of the species compos- 

 ing the community. Obviously, when studying a community which has one 

 or two species with large numbers of individuals, many more specimens must 

 be counted to discover the species composed of small populations. 



From our studies it is evident that 7000 to 8000 or more specimens usually 

 must be counted before a reliable picture of a diatom community can be ob- 

 tained if one wishes to compare the quantitative characteristics of commu- 

 nities. From TABLE 1 it is evident that only a small percentage of the species 

 composing the community are seen when only 200 or 500 specimens are counted 

 and approximately 50 to 75 per cent of the number of species are seen when 

 1000 specimens are counted when compared with the number seen when several 

 thousand specimens are counted. Similarly, the percentage of the population 

 composed of specimens of dominant species in some cases varies greatly when 

 based upon counts of a few hundred specimens as compared with counts of a 

 few thousand specimens. As seen in table 2 the numbers of species compos- 

 ing the diatom community remain fairly similar when similar segments of the 

 communities are analyzed if no serious change in the environment occurs. As 

 seen in table 3 the percentage of the population composed of dominant species 

 does not vary greatly for similar environments when similar segments of the 

 community are studied. 



When the structure of these populations was plotted by representing the 

 number of species as the ordinate and the number of individuals composing 

 each species as the abscissa, the data approached the shape of a truncated 

 normal curve, figure 1. To determine what mathematical formula might 

 best express the results of these studies several formulae were tried (Patrick 

 et al., 1959) and the truncated normal curve provided a little better fit than 

 the other methods investigated. The use of a truncated normal curve to ex- 

 press the structure of communities of organisms has been supported by the 

 work of MacArthur el al. 



By using this method we objectively compared similar segments of diatom 

 populations. For example, if enough specimens are counted and enough species 

 are identified to always place the mode in approximately the same interval, 



359 



