Of course all three factors are considered as only 

 approximately measured in any case, they arc t*ken as true 

 only within the limits of variation due to unknown or neglect- 

 ed conditions, within the limits of what are commonly termed 

 the unavoidable errors of ooservations, measurement or experi- 

 mentation. A3', that was necessary in the present study is 

 that fluctuation in the approximately ifcx maintained tempera- 

 ture, and uncorrected errors in the measurement of the length 

 of the exposure period should be of no greater relative magni- 

 tude than were the unknown and neglected fluctuations in the 

 approximately controlled material and cf the nutrient medium* 

 A decision in such matters, for physiological studies, still 

 depends largely on general judgment rather than upon mathe- 

 matical calculation. 



It has he en emphasized, by lehenhauer (1914) and others, 

 that the optimal temperature for a given physiological pro- 

 cess, with a giver complex of controlling conditions, can he 

 logically stated without reference to the length of the expo- 

 sure period only ^nen it is assumed that the process rate is 

 constant, thrcug'-cut that period. If the rate in question 

 changes in any '.vay wit> the lapse cf time the irdex of optimal 

 temperature can have no -rery precise meaninr unless the le 

 of the period considered is definitely stated. 1 is :s a 

 consideration frequently neglected in experimental detenu ra- 

 tions cf temperature optimal for physiological processes. 

 Of course this whole matter of the reictiori of length of expo- 

 sure period to trie three cardinal points of the temperature re- 

 lation specifically involves the consideration of progress: 



