36 Report of the Director of the 
If we consider our experiment as one trial, we have : 
■ 
No. of plants 
per acre. 
MBECHANTABLE. 
No. of ears, 
acre. 
Yield per 
acre, bush. 
B 2. Not hoed 
12,430 
12,430 
12,580 
10,290 
11,460 
12,290 
36 
43 
51 
B 3. Hoed three times 
If we consider our experiment as in duplicate, we have : 
No. of 
plants per 
acre. 
MERCHA 
No. of ears 
per acre. 
NTABLE. 
Yield per 
acre, bush. 
1. 
B 3 upper. Hoed three times 
B 4 upper. Hoed seven times 
2. 
B 2 lower. Not hoed 
B 3 lower. Hoed three times 
12,760 
12,820 
12,520 
12,100 
12,040 
12,640 
12,500 
12,720 
11,280 
8,080 
10,200 
13,300 
51 
52 
44 
21 
34 
58 
If we compare B 2 upper, not hoed, 51 bushels; B 3 lower, hoed 
three times, 34 bushels, and B 4 upper, hoed seven times, 44 bushels, 
we have again different figures, and so on through other changes. 
This trial is additional illustration of the indecisive evidence of plat 
work. 
General Considerations. ~ 
"Is there any hope for agricultural experiment?" some may ask. I 
reply confidently that there was never better justification for high 
expectation than the present, but the hope must come through the 
acknowledgment of errors, and their abandonment. If we follow up 
successes, much can be accomplished. By working from the simple to 
the complex, and by investigating the separate factors which go 
towards making crop, by using logic boldly, and by interpreting justly, 
we may anticipate the obtaining of practical conclusions which shall 
possess the value of accuracy and certainty. We workers must openly 
claim that in agricultural experiment only exact scientific methods can 
avail, and must welcome all criticism which shows the inadequacy of 
our methods, and which suggests improved systems. In this my final 
rep< >rt, I must freely acknowledge that I have yielded to public pressure 
