1. An experiment is described which assessed the effects on the production obtained from milking cows, of feeding a concentrate with a starch equivalent of 64 at four levels (0, 2, 4 and 6 lb./gal. of milk produced), in addition to a basal ration with limited amounts of hay and dried grass and with silage fed ad lib.
2. The total daily dry-matter intake of the cows increased from 2·2 to 2·7% of the live weight as the rate of concentrate feeding was increased from 0 to 6 lb./gal., despite a reduction in the amount of roughages consumed.
3. For each increase of 1 unit in silage dry-matter percentage, there were increases of from 1·18 to 0·42 lb. of silage bry-matter intake per cow per day as the level of concentrate feeding increased. For each increase of 1 lb. in silage dry-matter intake there was an increase of l lb. in total dry-matter intake per cow per day.
4. The average daily milk yield per cow increased from 22·2 lb. when no concentrates were fed to 27·6 lb. when 6 lb. concentrates were fed per gallon of milk produced. This is an increase of 0·9 lb. milk/lb. increase in starch equivalent intake. A significant difference in response was found between higher yielding and lower yielding groups of cows. The higher yielding group gave an increase of 1·1 lb. and the lower yielding group only 0·6 lb. milk/lb. increase in starch-equivalent intake.
5. At the general level of feeding and of milk yields obtianed in this experiment, no evidence of diminishing returns in milk production was detected when the rate of concentrate feeding was increased.
6. The content of solids-not-fat in the milk rapidly increased from 8·3% when no concentrates were fed to 8·6% when 4 or 6 lb. of the concentrates were fed per gallon of milk produced.
7. When concentrates were fed at the rate of 0 and 2 lb./gal. declines in live weight as well as in the yield and solids-not-fat content of the milk were noted, while at 6 lb./gal. there was an unnecessary gain in live weight. It is concluded, therefore, that with the particular animals and roughages used in this experiment, the optimal level of concentrate feeding was 4 lb./gal.