204 H. JOEL CONN 



In making a comparison of these ratios it seemed reasonable 

 to assume that those falHng between the hmits of 1.2: 1 and 

 1:1.2 were so near imity as to indicate no real difference in 

 numbers between the sod and cultivated samples. In the 

 sixth column, giving the ratios of the actual numbers per gram, 

 sixteen cases are hsted that fell within these limits. There 

 were only two cases (May 19, 1914 and the third on November 

 4, 1915) when the numbers in cultivated soil were enough greater 

 than in sod to give a ratio outside these limits; and of these the 

 greatest ratio was only 1: 1.8. There were nineteen cases, 

 however, in which the numbers in sod were sufficiently greater 

 than in cultivated soil to give a ratio exceding 1.2: 1; and of 

 them the maximum ratio was 6.4: 1. The average ratio of 

 all thirty-eight cases was 2.15: 1, which is larger than the ratio 

 between the general averages of columns three and four. The 

 figures which show what percentage of the total flora consisted 

 of Actinomycetes are somewhat more striking. The average 

 ratio, it is true, (as shown in the last column of the table) was 

 2.1: 1 or practically the same as the ratio between the actual 

 numbers per gram; but there were only four cases that fell 

 between the hmits 1.2: 1 and 1: 1.2 and only two (May 21, 1914 

 and the third on November 4, 1915) when the numbers in cul- 

 tivated soil were enough greater than in sod to give a ratio out- 

 side these limits. These two cases both showed a ratio of 1 : 1.3 

 which is hardly to be compared with the maximum ratio, 7.2: 1, 

 in favor of sod soil. 



The conclusion to be drawn from this comparison is that the 

 few exceptional cases in which there were more Actinomycetes 

 in the cultivated soil a,re completely overbalanced by the numer- 

 ous cases in which there were more in the sod soil. In some of 

 the border-line cases, moreover, the number of lower bacteria 

 was greater in the cultivated soil than in the corresponding 

 sod sample, with the result that the percentage of Actinomycetes 

 was sometimes greater in the sod sample even though the actual 

 number was the same in both samples. 



The last four cases in the table are of special interest because 

 they were analyses of the same samples collected on November 



