12 FEEDING EXPERIMENTS WITH ISOLATED FOOD SUBSTANCES. 



The best hope of success if such be possible rests at present 

 on the method of trial and error in which each variable is gradually 

 eliminated by successive comparative experiments. This is the 

 scheme which we have in large measure pursued. Our efforts have 

 at first been directed toward devising a simple ration of isolated 

 food components which should satisfy the numerous requirements 

 set more adequately than any yet proposed. This established, 

 substitutions could gradually be instituted in respect to the protein 

 constituents. We learned before long that a diet which might be 

 adequate for maintenance was by no means necessarily suited to the 

 requirement of a growing animal. Hence our attention has become 

 directed to some of the features of growth as well as those of the 

 maintenance ration of the adult. 



METHODS EMPLOYED. 



The rats were kept in metabolism cages similar to that described 

 by Henriques and Hansen.* A small door permits the introduction 

 of food and water cups, such as are used in bird cages, through the 

 side of the cage. 



Figure A shows the essential features. Instead of weighing the 

 food (which was always fed in the form of a homogeneous paste) in 

 the food cups, we devised the following very simple plan to avoid 

 frequent weighings. The food is introduced into a glass cylinder 

 about 25 cm. in length and 3 cm. in diameter. A rubber stopper in- 

 serted into one end can be moved forward like a piston head and the 

 food expelled from the other end of the cylinder into the food recep- 

 tacle. The exit end of the cylinder is kept stopped when the food 

 is not being expelled and the entire apparatus with its food content 

 can be preserved in an ice-box for long periods without deterioration 

 of the diet. The food eaten can thus be renewed at intervals and 

 the quantity fed determined, when desired, by ascertaining the loss 

 of weight of each food tube. 



Figure B illustrates our feeding-tube device. 



The urine and faeces were separately collected, the former in a 

 receptacle containing boric acid and chloroform, and analyzed at 

 intervals as indicated in the protocols. Control trials made by 

 trickling known volumes of analyzed human urine over the cage 

 bottom, and, after a suitable interval, washing with boric acid solu- 

 tion, collecting the urine, etc. , just as in the rat experiments, indicated 

 losses of 10 per cent or more. This must be borne in mind in consid- 

 ering our results and presumably those of other investigators. 



We devoted great care to maintaining suitable environmental 

 conditions (temperature, etc.), since the rats are sensitive to marked 

 changes. With our diets they consumed large quantities of water. 



*Hcnriques and Hansen: Zeitschrift fur physiologische Chemie, 1904-5, xliii, p. 418. 



