6 BULLETIN 966, TJ. S. DEPARTMENT OF AGRICULTURE. 
Breeds haying a larger proportion of twin births than the South- 
downs can be expected to give larger returns from flushing. Also, 
most farm ewes should show larger increases in lamb yields as a 
result of flushing than were obtained in these experiments because 
of the fact that ordinarily they are thinner at the time of breeding. 
RELATION OF WEIGHT GAINS TO NUMBER OF TWINS. 
Individual weights of the ewes were kept, and these permit an 
analysis of the relation between actual gains and numbers of twins 
produced. 
Of 57 cases in which ewes lost from 1 to 7 pounds during the breed- 
ings season, twins were produced by 21, or 37 per cent. 
Of 178 cases in which ewes gained up to 7 pounds during the breed- 
ing season, twins were produced by 60, or 34 per cent. 
Of 117 cases in which ewes gained 7 to 30 pounds during the 
breeding season, twins were produced by 57, or 49 per cent. 
These results suggest that in order to produce the largest number 
of twins, ewes should gain 7 pounds or more. 
FEED FOR FLUSHING. 
With a view of determining the kinds of feed most efficient and 
most economical for use in flushing, the flock at Beltsville was divided 
into three lots in 1918, 1919, 1920, and 1921. One lot in each year 
was unflushed while the other two were flushed. One of the flushed 
lots was fed grain in each case while the other was kept on extra 
good pasture. In 1918 the grain-flushed lot, No. 10, gaining an 
average of 12.25 pounds per head, produced lambs at the rate of 4.5 
more per 100 ewes than Lot 11, pasture flushed, which gained 11.86 
pounds. 
In 1919, Lot 14, given extra pasture and gaining 3.76 pounds, 
produced lambs at a rate af 4.8 more per 100 ewes than Lot 13, 
which was flushed with grain and gained 10.71 pounds. 
In 1920, Lot 16, flushed with grain and gaining 5.8 pounds, pro- 
duced at a rate of 3 more lambs per 100 ewes than those flushed with 
extra pasture and gaining 10 pounds, while in 1921, Lot 20, flushed 
with extra pasture and gaining 8.15 pounds, produced 21.8 more 
lambs per 100 ewes than Lot 19, flushed with grain and gaining 3.37 
pounds per head. No other lots show so large a variation due to the 
different methods of flushing. Lot 20 has also the highest lambing 
percentage of all the lots. 
The fact that in two of the four cases the grain-flushed lots pro- 
duced the highest lambing percentage, however, would indicate that 
there is no decided advantage in the kind of feed used in flushing 
except that where pasture flushing is practiced there is the saving in 
labor and more expensive grain feed. There are times, however, 
