MAlSrUAL OF CATTLE-FEEDIKG. Ill 



A preparatory period of feeding is, of course^ imeessmy, 

 as explained iii tlie previous bectiou, and tliis nuibt be long 

 enough to allow the body to come into equilibrium with 

 the foody so that the effects of the latter may have fully 

 developed themselves. The experiment proper must also 

 extend over a sufficient time to give a fair average. At 

 least twenty-four hours is necessary for this, but better 

 results are obtained when the experiment covers several 

 days. 



Finally, it should be remembered that the results ob- 

 tained show, in the first place, only the gain or loss of ni- 

 trogen^ and that the factors used Jor converting this into 

 protein or flesh are average nniBbers only, and that, while 

 they are nearly enough true for practical purposes, they 

 are not absolutely accurate in all cases. 



Produofcion of Pat. — ^As the production of flesh is es- 

 timated by a comparison of the receipts and expenditures 

 of nitrogen by the body, so the production of fat is esti- 

 mated by the gain or loss of carbon. 



For this pm-pose it is necessary to take into account the 

 gaseous i)roducts of respiration and perspiration, since the 

 larger part of the carbon excreted leaves the body through 

 these channels. 



These products can only be estimated with accuracy by 

 means of a special apparatus, first constructed in a practi- 

 cal form by Pettcnkofer, in Munich, and now widely used 

 under the name of " Pettenkofer's Eespiration Apparatus." 



The Respiration Apparatm.—'Vh.G principles of this most important 

 apparatus are well illustrated in an ordinary stove, in which the gases 

 coiniog from the fire may represent those coming from the lungs of the 

 animal 



As long as the chimney draws, no smoke escapes from the doors and 

 draughts of the stove, but, on the contrary, the air presses from all 

 sidea mto the stove, to pass out through the chimney. 



