220 COMPOSITION' OF SOME INDIGENOUS GRASSES. 



True Protein. 



Any relative values assigned to the grasses investigated 

 should be based on their digestible constituents. Here, however, 

 the wi-iter was faced with the difficulty that there are no 

 experimental data on the digestibility of South African grasses. 

 Henry and Morrison^" gives a very valuable compilation of 

 American analyses of feeding stuffs and results of digestion 

 experiments, and on this basis have calculated the percentage of 

 digestible nutrients in a great variety of feeding stuffs. It would 

 be feasible to use tzheir factors for average grasses, and so calcu- 

 late approximate figures for the digestible nutrients in them, but 

 this would not appreciably affect the relative feeding values 

 calcvilated. for the grasses under consideration, and so would not 

 repay the extra calculations required. Hence the calorific value 

 based on the total percentages has been taken as the basis of 

 comparison. For the heats of combustion of the three chief 

 classes of food the figures of Eubner" have been used. 



1 gram of fat yields 9,300 calories. 



1 gram of protein yields 4,100 calories. 



1 gram of carbohydrate yields 4,100 calories. 



The crude fat or ether extract, containing ether-soluble matter of 

 less calorific value than pure fat is pi'obably overrated by using 

 these figures. In accordance with the usage of Henry and 

 MoiTison and other writers, the crude protein has been used in 

 the calculation instead of pure protein. Armsby^^ considers 

 that for maintenance non-protein may be of equal value to true 

 protein. Later investigators (Atwater") give a higher value 

 for protein (4,400), but as the value of asparagine is given by 

 Ingle* as 3,500 calories, the ultimate result does not differ 

 very materially from the figures used. Crude fibre has been 

 neglected, in calculating the calorific value, although it does 

 certainly have some value. 



Finally, all results have been calculated on the absolutely 

 dry .material, thus giving a basis for comparison. The average 

 air-dried sample contained about 11 per cent, of moisture. It 

 may be added that all analyses were made in duplicate, and in 

 cases when agreement of results was not as close as was desired 

 the analyses were repeated to ensure a fair average. 



Classification and Relative Values. 



Table I gives the average analyses of the grasses under 

 review, the number of samples of each species dealt with being 

 shown in the third column. In the second column an attempt 

 has been made to allocate the species to four main groups : — 

 R. Ruderal : Such species as Elcusine iudica, Panicuvi prolt- 



ferum, which are commonly found on waste lands. 

 Veld: Average climax veld species — for example, AniJii^liria 



imberbis. 



