362 PHYSIOLOGICAL REGULATIONS 



quantities of no visible mechanical inertia such as respiratory 

 quotient. 



Data concerning variation are capable of representing what the 

 organism does to exert control of content, whether or not this con- 

 trol employs recognizably special forces or organs. It may be that 

 the rectal temperature of a snake would be, under like conditions, 

 just as constant as was that of dog A; the snake perhaps "does" 

 nothing internally to preserve its temperature, while the dog does ; 

 instead the snake may ' ' select ' ' its environment to accomplish the 

 same end. Further knowledge of that is a matter for observation. 

 What is now in hand is the study of many components, preferably 

 measured simultaneously; both the organism and the conditions 

 available entering into the control of each. The dog or man is to 

 be described as found, leaving aside the dog put under selected 

 loads, and the dog placed successively under two or more recog- 

 nizably different conditions. 



Among the components in any organism, those found to vary 

 little while exchanging continuously, may be defined thereby as 

 components whose constancy is usually of more consequence to the 

 individual than others. Generally the organism is exerting a type 

 of behavior or of process that diminishes any positive or negative 

 load, restoring the content toward the mean. 



Another method of stating this general conclusion is that of 

 Gasnier and Mayer ('38, p. 119). Each property or component 

 occurs with such values (contents) as are arrayed in a frequency 

 curve. Those values which are rare (of low frequency) are evi- 

 dently prevented from occurring by a resultant of whatever proc- 

 esses operate to avoid them. Possibly the "effort" made to avoid 

 these values is inverse to their frequency. Therefore the fre- 

 quency curve that is epitomized by the coefficient of variation (and 

 by the coefficient of difference) becomes by definition the measure 

 of all that discourages extreme contents. 



Variabilities apply to rates of exchange as well as to loads. As 

 in loads, variability between successive periods may depend on the 

 length of the period. This is the case in rates of urinary water 

 loss in dogs under control conditions; the mean difference is less 

 for 0.25-hour periods than for 0.5-hour periods or for 1.0-hour 

 periods (fig. 45). This fact indicates that something prevents 

 large short-time fluctuations; it is evidence of an "inertial" gov- 



