Report on the Bacteriology of Water. 321 



growth once in full swing is retarded, during the short period I am 

 able to measure it, by a low intensity of light independent of tem- 

 perature. 



Criticism of the foregoing Results, and Experiments on Temperatures. 



The more I became familiarised with the methods which had given 

 the foregoing results, the more evident it became that, instructive 

 and valuable as the facts may be and there can be no question as to 

 their decisive confirmation of the conclusions of previous observers, 

 as well as myself, that direct sunshine kills so long as the blue rays 

 are not filtered out they are not altogether satisfactory. The varia- 

 tions due to internal causes are so minute that, real though they are, 

 they do not seriously affect the measurements of growth, provided I 

 compare similar rods or long filaments during the same periods, and 

 it is obvious that if I use cultures in the same food-material, and 

 prepared alike, there can be no serious drawback in the method so 

 far as comparative results go. 



But over and over again I was impressed by the dangers possible 

 as regards the screens employed. It is obvious that with an organism, 

 like this, very sensitive* to changes of temperature during its growth, 

 differences of a few degrees at the critical periods might easily lead 

 one into errors which, if neglected, would vitiate many of the con- 

 clusions entirely. 



The matter is a serious one, for it must be remembered that 

 biologists, and especially botanists, have long been in the habit of 

 using coloured screens, and if these agents are responsible for changes 

 of temperature not recorded in the observations, we are in danger of 

 making fundamental errors. 



The suspicion that the temperatures, as recorded merely by thermo- 

 meters in the air beside the microscopes, might be less trustworthy 

 guides to the changes of temperature going on in the hanging drop- 

 cultures in the glass cells than had been hitherto assumed, now 

 forced itself upon me ; and with that, of course, the idea that I 

 might be running risks of confusing the effects of high or low tem- 

 peratures with those of light, became painfully disturbing. 



In any case, it was obviously necessary to institute a series of 

 experiments to test these notions, and it was well I did so, for the 

 results, although not wholly unexpected, were more startling than I 

 care to dwell on now.f 



* It was not till a much later period that I found how extremely sensitive it is to 

 temperature changes, as will be seen subsequently. 



t It scarcely needs mention that the question here concerned is not the physical 

 one whether two different screens transmit heat-rays in different proportions, but 

 the physiological one how far the differences affect the organism experimented on. 

 It can obviously be decided only by patient comparative investigations. 



