468 Journal of Agricultural Research voi in. xo. 6 



production of animals is variable, depending undoubtedly upon the particular 

 animal under investigation and possibly, also, upon the nature of the ration. Thus 

 the average percentage de\'iations for the four animals for which 10 or more obser- 

 vations are recorded, with their probable errors, are: 



Animal 1 2- io±o. 23 



Animal A , '^2. 25±o.35 



Animal A 6 i. 77 ±0. 18 



Animal B o. 7S±o. 13 



Animal D i. 32 ±0. 28 



While, according to these figures, animals I and A can not be differentiated from 

 each other, they are both clearly differentiated from animal B, tlie differences, 

 mth their probable errors, being i.35zto.26 between animals I and B and 1.02 ±0.22 

 between animals A and B. The difference between animals I and D was 0.78 ±0.36, 

 the significance of which may be questioned. Using another metliod of compari- 

 son, it may be shown tliat the odds are 124 to i that animals C and H are definitely 

 distinct as regards the percentage deviation under discussion, I should not hesi- 

 tate to conclude, therefore, that this percentage deviation is affected by the indi- 

 viduality of the experimental animals. Furthermore, there is a slight suggestion, 

 especially in the data of animals A and B, that the nature of the ration may aSect 

 the percentage deviation of ^-our determinations. 



It is of interest to note in this connection that the results of Kellner's 

 respiration experiments (uncorrected for standing or lying) likewise 

 show variations of much the same order of magnitude between indi- 

 vidual (not consecutive) days. When, therefore, comparisons are based 

 upon the average results for 48 hours, it is impossible to assert that 

 these results represent, as they should, periods of average muscular 

 activity, although it would appear that the error thus introduced is 

 usuallv not large. In Kellner's experiments it is still further reduced 

 by the fact that in most cases the results of four or five single runs are 

 averaged. 



ANALYSIS OF HE.\T PRODUCTION 



In Table VII were shown the increments of heat production per 24 

 hours in standing animals as compared with those lying. It is evident 

 that of the total corrected heat production recorded in Table IX an 

 amount equal to one-half of the corresponding increment shown by 

 Table VII is to be regarded as the effect of t-he 12 hours' standing, while 

 the remainder represents the metabolism of the animal per 24 hours 

 lying. On the basis of Zuntz's recent results it is possible to carry this 

 analysis of the heat production a little farther, at least approximately. 

 The expenditure of energy caused by standing obviously includes that 

 required for the muscular effort of rising and lying down. Von der 

 Heide, Klein, and Zuntz (20, p. 823) estimate this on the basis of 

 experiments by Klein at 9.7 Calories per 550 kg. of live weight for 

 once rising and lying down again. The same investigators (20, p. 795) 

 compute from Markoff's experiments (37, 38) that the methane fer- 

 mentation in cattle gives rise to the evolution of 4.374 Calories of heat 

 per cubic centimeter of methane, equivalent to 6.07 Calories per gram. 

 While both the foregoing figures are confessedly but approximations, 

 nevertheless they permit a partial analysis of the heat production with 



o Including questionable observation. t> Not including questionable observation. 



