25 



number of flies. Whereas in London, 1908 and 1909, following 

 shortly upon the colder weather, both curves drop in a 

 remarkably similar fashion, in the small epidemic of 1907 

 the number of flies still remains up and even increasing, 

 whilst the epidemic is rapidly disappearing. The same 

 general facts emerge from the Manchester observations. We 

 might invoke the assistance of coincident exhaustion of 

 susceptible material, or diminished activity of flies, although 

 still numerically sufficient, in explanation of this. They 

 may, indeed, be operative, but I think that a completely 

 satisfactory interpretation of the above facts and also of the 

 fact that, with a given number of flies, the epidemic wave is 

 ascending during the early weeks of an epidemic, and sub- 

 siding a month or six weeks later, is afforded by taking into 

 account the influence of the variations in atmospheric tem- 

 perature upon the rate of multiplication of the infective 

 agent. Regard for a moment Fig. 8, the chart expressing 

 the results of Dr. Hamer's observations in London, 1908. 

 Take the weeks ending August 25th and Sept. 12th, in which 

 approximately the same number of flies were caught. The 

 mean temperature for the former week was 62- 6 F., for 

 the latter 54* 8. In the meantime it had been down 

 to 53. 



The mean atmospheric temperature is a factor to which 

 insufficient attention has been given, since the origin of the 

 epidemic showed but small relation to it. It is capable, 

 however, of explaining why the epidemic wave rises in 

 August and falls in September, although the number of flies 

 may be the same. Assuming, for the sake of argument, 

 that some form of bacterial infection is the etiological factor 

 in infantile diarrhoea, it would be in accordance with general 

 experience to suppose that the chance of infection is deter- 

 mined by the quantity of the infecting agent imbibed. This 

 will depend upon the multiplication which has taken place 

 on the infected medium prior to ingestion ; this in turn is a 

 function of the temperature at which the milk or other 

 foodstuff has been kept. 



The temperature coefficient of bacterial multiplication is 

 not higher than that governing ordinary chemical processes 

 viz. , two to three times for 10 C. , but as bacterial multipli- 

 cation proceeds logarithmically a comparatively small change 

 in rate is sufficient to produce a large difference in the 

 number of organisms contained in, say, an infected milk, if 

 an interval of some hours elapse between inoculation and 

 drinking. Thus if n bacilli coli be inoculated with milk 

 at 60 and 70 F. respectively, the number after 12 hours 

 will be 126n and 4000>* respectively. 3 



3 For the observations on which this calculation is based see Dr. Janet 

 E. Lane-Claypon's paper Multiplication of Bacteria and the Influence of 

 Temperature Thereon, Journal of Hygiene, vol. ix., 1909, p. 239. 



