To investigate the effect of replacing part of the concentrates by roots, a 16-week winter-feeding experiment was made using twelve Ayrshire cows. All the cows received 8 lb of hay per day and enough grass silage for a measurable residue to be left at each feed. Two rations were supplemented with 45 lb of fodder-beet per day but the other two contained no roots. At each level of root feeding, concentrates were fed at either 3·50 or 1·75 lb per 10 lb of milk.

All the fodder-beet was eaten (8·4 lb dry matter) and, as this caused a reduction in the intake of silage dry matter of only 3·4 lb per day, there was a net increase in drymatter intake. The level of concentrate feeding had little effect on the intake of the other feeds. Rations which included roots contained significantly less crude fibre than rations without roots.

The mean daily milk yields for the treatments with the low and high concentrate levels were 29·5 and 33·9 lb, respectively, without roots and 31·8 and 35·8 lb with roots. The fat percentages of the milk were similar on all four treatments whereas the solids-not-fat (S.N.F.) increased significantly from 8·42% on the treatment without roots and low concentrates to 8·64% on the treatment with roots and high concentrates. The response per 1 lb of extra S.E. was 1·0 lb milk and 0·02% S.N.F. between the low and high concentrate treatments and 0·7 lb milk and 0·05% S.N.F. between the two root treatments. The analysis of milk from individual quarters of all the cows during each period showed the continuous presence of some subclinical mastitis. It was estimated that this lowered the general level of S.N.F. percentage by 0·2% but did not bias the results from any one feeding treatment. For milk from healthy quarters only, the response to 1 lb extra S.E. from concentrates was 0·03% S.N.F. and from beet it was 0·08% S.N.F.