VARIABILITIES OF WATER RELATIONS 85 



on three individuals (all those having water deficits exceeding 2% 

 of Bo, fig. 11), the rate of gain varied by C.V. ± 32. It may be con- 

 cluded that ingestive gain and urinary output are managed with 

 equal uniformity. There is here no evidence that drinking is less 

 accurately controlled by the dog than other exchanges. 



Another question is whether the variation is different if instead 

 of averaging those measurements of rate taken at one time I ascer- 

 tain the distribution of rates at one load. The variability turns out 

 to be similar. Further it would be possible to find the distribution 

 of times or of loads at which uniform rates prevail. 



Measurements of variability in rates of exchange make possible 

 the detection of unusual responses of the individual. How small a 

 rate of water excretion, following the ingestion of 3% of Bo of 

 water, shall be considered pathological? There are data of Falck 

 (E, fig. 31) which, so far as the methods and conditions of measure- 

 ment are described, duplicate those of Kingsley (fig. 2). But the 

 rates of output are much smaller and the lags much greater. It is 

 difficult to believe that three individual dogs used by Falck, con- 

 sistent among themselves, were genetically different from three 

 used by Kingsley, fairly consistent among themselves. Diversities 

 of method, race, diet, conditioning and acclimatization that will 

 produce such wide divergencies of result remain to be precisely 

 investigated in some one laboratory. 



Again, ''distilled" water by stomach in quantities of 1 and 2% 

 of Bo did not invoke diuresis, while tap water did (Hashimoto, '14). 

 All distilled water is not similar. But few investigators would have 

 considered it worthwhile to test the differences among samples of 

 water, yet to disregard a difference in water administered is physi- 

 ologically confusing. Unfortunately Hashimoto did not demon- 

 strate that the fact of being distilled water was what mattered in 

 those tests, nor did he show that the difference was a significant one. 



In general at diverse rates of water exchange, the variations are 

 relative to the mean rates. In recoveries from water deficits the 

 initial water intakes vary by C.V. ±: 32. In recoveries from water 

 excesses the initial outputs vary by C.V. ±: 24. Hence the responses 

 to negative and to positive water loads are of equal accuracies. 

 Unusual responses may be adjudged by reference to these manifest 

 variabilities. 



