SPORE-FORMING BACTERIA 189 



cubated at 18°C. for seven days. This length of incubation 

 was necessary in order to allow the late colonies (particularly 

 of B. megatherium) to appear. The chief disadvantage of such 

 a long incubation was that B. mycoides and B. cereus often had 

 time to liquefy the plate completely unless high dilutions were 

 used. Dilutions of 1-20,000 and sometimes even 1-100,000 

 or 1-200,000 were necessary in order to avoid this trouble. 

 At such dilutions the numbers of colonies of the spore-forming 

 bacteria were so few that a long series of plates had to be made 

 in order to obtain a reliable count; and even then no signifi- 

 cance could be attached to variations in the count unless they 

 were quite large. 



In the first of these tests a temperature of 85° was used; but 

 later it was learned that at temperatures only about 10° higher 

 than this large numbers of the spores were killed and it was 

 suspected that even 85° might destroy some of them. For this 

 reason 80° was used instead for a while, and in the last tests 75° 

 was used. To test the efficiency of this last temperature the 

 bacteria developing on the plates after heating the infusion 

 were studied, and it was found that nothing but spore-bearing 

 bacteria had survived (leaving out of account an occasional 

 colony of some non-spore-forming type that might easily be 

 due to air contamination). 



The greatest source of error in this method, which could not 

 well be avoided, is the possibility that the bacteria investigated 

 may occur in soil in clumps or chains instead of as isolated indi- 

 viduals. There is some evidence that clumps of bacterial spores 

 can be broken up by the action of heat, which would tend to 

 increase the count in the heated infusion provided clumps do 

 occur in the soil. No increase of any appreciable amount has 

 ever been observed, however; and indeed, so far as microscopical 

 examinations of soil have been made, no evidence has been ob- 

 tained of chains or clumps of organisms of this type. For this 

 reason this possibility of error did not seem great enough to 

 invalidate the conclusions. 



A series of twenty-six tests was made. The results are given 

 in the following table. The most striking fact to be observed 



