04 fi:vi:k. 



The oK'Vatioii of the UMiipcTiiluri- of tlie air during its passage, in tlio exp«'rimcnt 

 uudtT discussion, is t = l°..V2 in tin- forniida, Q = W X t X *>[>• 1». Hie- hptritic heat 

 of air is O/i^H, and the fuiiuula bttonK-s 



Q = '6.'W9 X 1.5'2 X 0.'2:374 = 2.2()'»4. 



The licat "ivi-n to the air hv tlie animal is therefore 'i/itJlU units." 



Ihe animal in tlie box is constantly giving off moisture, and heat is becoming in 

 this way insensible; condensation of moisture also may or may not be goin^c *"' in 

 the box. In order to estimate the disturl)ance of heat in tiiis way, as already ex- 

 plained, calcium tubes or sulphuric acid bulbs were employed. The gain of weight 

 of the sample in calcium tubes, as shown in its appropriate column, is (1.1 1,S4 granunc. 

 In order to determine liow much moisture is in the air enming fmm tlie box, the 

 amount of the sample (().2;J51 ) is divided into the whole aiiiouul of air passi-d 

 through the box, and the moisture in the sample is multiplied by the resultant, 

 which represents the proportion between air and samjile, and is known in this 



mciisuro as the niuAitiit /<ir t/ie l/ox. Thus '" '"^ = 350 = Quotient for the box. 



().'i351 



Thus t). 1 l.^l grannne X 355 = 4'2.03*2 = wlmle amount of moisture coming from 

 the box. By a similar process the amount of moisture entering the box is cal- 

 culated. The '■'(jiiiitleiil J'or dir" is obtained by dividing the whole amount of air 

 passed through the box by the amount of the air analyzed in the outside apparatus. 

 Then this is multiplied by the amount of moisture found in the analyzed air. 



Thus '^'^■^'^^='272.5= quotient for air. ^T^.S X 0.057*) = 15.778 = moisture 

 0.3063 



entering the box. Moisture leaving box, 4*2.032 grammes. 



Moisture entering box, 15.778 grammes. 



Moisture evaporated in box, 26. '254 gramm<s. 



' Strict nccorary would of course require tbat allowance be made for baromctricol TarialionB, aiici 

 tbnt tlu- |iri')i.suri- be reduced to the slandnrd jirussure 20.92 incbcs. Tliislia.s not been done in anr of 

 my expvrinient.s, lH.'cnuse llie rcstillniit error is so .small ns to be of no ini|>i>rt»ncc wliatcvcr. The 

 experiincnts were mostly performed in tlic course of a few sucec-jfiivc hours, and elinnpes in the barw- 

 meter lu such a )ierio<l rarely amount to 0'*.5. A^isuminf^ 30 inches as the standard, and that in llie 

 present instance the barometrical pressure was 30'''.5, the calculation would be 



^. V' •;.4:ox30.5 g^^^ y,_ 84.^8T^_,9 2 AVV'x 0.08073- C.3W. 



Q = 6.394 X l.'ia xO.2374: 2.3072. 



Ii will be seen that the dilTerence U'tween thi.'i result and that arrived at is onl.v 0.0378. When it 

 is borne in mind ihnl the supposed barometrical rnriation is cxirenie, that the e.xperinients are all 

 n'lutiTc, and that in fnml comparisons of results no decimals are of any importance, it is plain thai no 

 injury is done by tlii' omission of banimrtrical allowances. Even in the fever experiments, Bllowinir 

 for a pontiblc barometrical variation of one inch, only the second decimal would be affectetl and that to 

 a Tort sliirht extent. It must be borne in mind that absolute accuracy cannot be reachi-*! in any of 

 tl' ircmonts of heat production, and that the trnlh of n fact can only lie established bj tbo 



*■' ' hi-at priKlnction beinp so Inrpc am] so conslntit as to remove the danirer of error. KrP'r* 



f' 'n**" preponderate in one direction, sometimes in another, and if results an* alwnya toe 



HauW-, lite error rauat lie nniinportant. 



