POPULATION FACTORS AND SELECTED POPULATION PROBLEMS 365 



spiders, isopods, diplopods, and insects. 

 Cole was able to distinguish both random 

 and "contagious" dispersion patterns. 



The only group examined by Cole that 

 showed a random distribution of individ- 

 uals under the boards was the spiders 

 (Table 26). 



Scytonotus granulatus (Table 27). Here, 

 the observed and the expected do diflFer 

 significantly (by chi square) from each 

 other, there being more instances of no 

 forms per board than would be expected 

 on the hypothesis of random dispersion, 

 and also more instances of 4, 5 and more 



Test of observed against expected frequen- 

 cies by chi square shows that these do 

 not differ significantly from each other 

 (P = 52.7 per cent). In other words, as 

 great a deviation between the observed 

 and Poisson distributions could be antici- 

 pated about 53 per cent of the time as a 

 result of chance alone, so that there is cer- 

 tainly no reason for assuming a nonrandom 

 dispersion of spiders under the boards. 

 Cole also found that the spiders were dis- 

 tributed at random in all the observational 

 areas for all seasons of the year. 



individuals per board than expected. Thus, 

 instead of finding frequencies of 3.4 with 

 4 diplopods per board, 0.7 with 5 diplo- 

 pods per board, and 0.1 with more than 5 

 diplopods per board (see table), Cole 

 actually found frequencies of 8 with 4 dip- 

 lopods per board, 5 with 5, and 3 with 

 more than five. This, of course, strongly 

 suggests aggregation or grouping of the 

 Scytonotus under the boards. The data of 

 Table 27 are also graphed in Figure 130 

 with number of stations on the ordinate 

 against number per station on the abscissa 



The other species of this Illinois wood- 

 land showed, not a Poisson, but a "conta- 

 gious" distribution (see P61ya, 1931; Ney- 

 man, 1939; Cole, 1946a). The distinguish- 

 ing feature of such distributions so far as 

 they apply to Cole's study is that, by al- 

 lowing for aggregating tendencies, they 

 show that the boards sheltering the larger 

 number of animals occur more frequently 

 than would be true if the dispersion were 

 random. 



The point is well illustrated by the un- 

 der-board distribution of the diplopod, 



and with the observed and Poisson curves 

 plotted within these coordinates. The con- 

 tagious character of the distribution be- 

 comes immediately apparent. The two 

 curves cross each other at about the level 

 of the frequency class. "Three diplopods per 

 board." Of this Cole says: "More detailed 

 investigation might reveal that when about 

 three Scytonotus are present under a board 

 of this size, they serve to attract other dip- 

 lopods to the board or influence wander- 

 ing individuals to remain once they find 

 that particular board. Thus any 'social in- 



