64 



MASS. EXPERIMENT STATION BULLETIN 197. 



Our own results with four sheep show some noticeable variations. Sheep 

 XII and XI seem to have digested the feed better than the two other 

 sheep. This variation is noticeable in each of the several ingredients. 

 The average results agree quite closely with those secured by Ewing and 

 Smith ^ with steers. 



Appl5dng the average coefl&cients to the chemical composition of the 

 velvet bean feed we find 2,000 pounds of the material to contain the 

 following: — 



Velvet bean feed contains about the same amount of digestible protein 

 as does wheat bran. Its content of digestible fiber, extract matter and 

 fat is somewhat in excess of that contained in the bran, and on the basis 

 of total digestible organic nutrients it is shown to have some 11.5 per cent 

 more feeding value than the latter feed. 



D. Velvet Bean Feed for Cows. 



During the fall and winter of 1918-19 two feeding experiments with vel- 

 vet bean feed were made with milch cows. In one case six and in the other 

 four animals were used. They were divided into groups in each experi- 

 ment, and were fed by the reversal method for periods of five weeks 

 (besides prehminary periods) on a ration composed of hay for roughage 

 and a grain mixture made up of 20 per cent cottonseed meal, 40 per cent 

 corn feed meal (corn meal in second experiment), and 40 per cent of either 

 velvet bean meal or wheat bran. 



Before starting the experiment the cows were carefully chosen and paired 

 off as well as possible in regard to age, breed, period of lactation, yield of 

 milk, fat, etc. 



The hay and grain rations were carefully calculated for each animal on 

 the basis of milk and maintenance requirements according to the Haecker 

 Standards. 2 The general care and management of the animals in no way 

 differed from that always used in our feeding experiments, and require no 

 discussion here. Hay and grain samples were taken at regular intervals, 

 composited and analyzed. Millc samples were taken on the first, third 

 and fifth weeks of each half of each experiment. 



1 p. V. Ewing and F. H. Smith in Journal of Agricultural Research, Vol. XIII, No. 12, p. 616. 

 Results with five different steers in eighteen single trials. 



2 See Minn. Bui. No. 140, p. 56. 



