754 BECKETT 



2Sab 



cc = 



Sa + Sb 



where S^ is the number of species in sample A; S^ is the number of 

 species in sample B; and Sab is the number of species found in 

 sample A which are also found in sample B (Mcintosh, 1967; 

 Whittaker, 1975). Values of this index can therefore range from 0.0 

 (perfect dissimilarity) to 1.0 (perfect similarity). This index has been 

 used in determining similarities among plant communities (see review 

 by Mcintosh, 1967) and among animal communities (e.g., spiders, 

 Uetz, 1976;benthic macroinvertebrates, Burlington, 1962). 



I analyzed my data for each collection date. Using the CC index, 

 I compared each station to each other station, thereby forming a 

 similarity matrix (called a Q matrix or secondary matrix by some 

 authors) for each of the collection periods. I then formulated 

 similarity matrix diagrams (Fig. 4) along an upstream— downstream 

 gradient directly from the similarity matrix. In these diagrams, 

 ccmmunities showing similarities equal to or greater than 0.65 are 

 associated by connecting lines. In cases where a number of stations 

 aU showed marked similarities (>0.65) to each other, the entire 

 grouping of similar stations was connected to a single similarity line. 

 Those stations not exhibiting marked similarity to any other station 

 were left as single entities. The use of 0.65 as a cutoff value for 

 marked similarity was somewhat arbitrary. However, Hanson (1955) 

 and Hurd (1961) also found 0.65 to be indicative of high biotic 

 similarity in their studies of plant communities. 



The similarity matrices were also used as a basis for ordination of 

 data. Ordination provides a graphical summary of the simileirity 

 between stations or communities. The ordination technique I used 

 was introduced by Bray and Curtis (1957) and is commonly called 

 Wisconsin comparative or polar ordination. The computational 

 procedure that I used in ordinating the data is presented by 

 Whittaker (1973). In addition, I used ordination as a form of indirect 

 gradient analysis (Whittaker, 1973). 



RESULTS AND DISCUSSION 



Diversity values (H') showed a number of interesting trends 

 (Fig. 2). Upstream stations (stations 1, 2, and 3) on the Mad, 

 Stillwater, and Great Miami rivers had moderate to high values on all 

 dates. Station 4 (near the F. M. Tait power-plant intake) exhibited a 

 high macroinvertebrate diversity (above 3.0) for all three sampling 

 periods; for two of the three sampling periods the highest value of H' 



