INPUT AS BELATED TO OUTPUT 
21 
for each of the droves was predicted 10 from the actual inputs, using 
the relations shown in Table 19, these estimated gains came exceed- 
ingly close to the gains actually made. 11 
Figures 7, 8, and- 9 show graphically some of the relations given 
in Table 20. These results show that the fact of diminishing returns 
must be taken into account in setting up any sort of unit re- 
quirements. 12 It takes 8.7 pounds of grain to cause a pound of gain 
in a steer fed only 10 pounds per day, but it takes 14.3 pounds of 
grain — 64 per cent more — when fed at the rate of 30 pounds per day. 
whereas the pork by-product, which was 1 pound for each 67 pounds 
of grain at the lower input, is only 1 pound for each 112 pounds of 
grain at the higher rate of feeding. Similar wide variations in feed 
requirements are caused by variations in other feeding practices, as 
shown. 
Table 20. 
-Net relation of various feeding practices to feed requirements for 
beef prodiiction, and for pork by-product 
A. DIFFERENCES IN THE DAILY INPUT OF GRAIN 
Pounds per head per day 
Daily gain due to 
grain 
Average grain re- 
quired per pound 
of gain of 
Beef 
Pork by- 
product 
Beef 
Pork by- 
product 
10 . . .- - -- --- 
Pounds 
1.15 
1.50 
1.81 
1.95 
2.10 
Pounds 
0.149 
.205 
.238 
.258 
.268 
Pounds 
8.7 
10.0 
11.0 
12.8 
14.3 
Pounds 
67 
15 . . .. .. .- 
73 
20 - 
84 
25 . . - - - -- --- 
97 
30 
112 
B. DIFFERENCES IN THE DAILY INPUT OF ROUGHAGE 
Pounds per head per day 
Daily 
gain due 
to 
roughage 
Average 
roughage 
required 
per 
pound 
of gain 
8 . -. -- -- -- -- --- 
Pounds 
0.46 
.66 
.83 
.98 
1.11 
Pounds 
17.4 
12 
. 
18.2 
16 .. -- -- -- -- 
19.3 
20 -- ------ .- 
20.4 
24 . - - --- 
21.6 
10 Gains were " predicted " by determining- from the table hew much gain should have 
been made for the grain and roughage fed and then adjusting these gains according to the 
length of time on feed and the initial weight. 
[Gain due to corn (Effect of 1 ) Effect of 
Estimated gain =\ + gain due to roughage >X< length of J-XJ-initial 
[ gain due to protein, concentratesj (.period J J weight 
Taking the " relative gains per unit of feed " to measure the effect of the last two factors. 
11 The correlation between the actual and the predicted gains was r= + 0.82O±0.027, 
and the average error in predicting gains was 12.7 per cent of the actual gains. Separat- 
ing roughage into its components — leguminous and nonleguminous hay, straw, and fod- 
der — would have made the predictions still more accurate. 
12 The practically linear increase in grain requirement per pound of gain, as more is 
fed per day. is not due to the fact that this was determined as a linear relation. The out- 
put resulting from the various inputs was first determined, and it merely happened that 
the reciprocals of the curves (fig. 7) gave the nearly linear relation (fig. 8). Straight 
lines have been drawn into Figure 7 to show the extent to which the curves diverged from 
linear relations. 
