AGRICULTURAL EXPERIMENTS. 281 



where we find an animal has made a very great loss at one weigh- 

 ing, it will frequently make the best gain at the next Weighing. 

 Ii' this loss in weight was owing to a delect in the animal, then 

 we would suspect the animal was sick, or there was some de- 

 rangement of the organs, and would not expect it to make such a 

 rapid gain at the next weighing. 



To obviate these difficulties, and make our experiment satisfac- 

 tory — that is, in order to accomplish the object of experimenting — 

 it seems to me we should pay especial attention to several partic- 

 ulars which I shall enumerate. In the first place, the experiment 

 should be as simple as possible. A large proportion of the 

 experiments made thus far have been of no value, for the reason 

 that too much was attempted. For instance, a person wishing to 

 test the real value of potatoes of different sizes for seed, plants his 

 sm;Jl potatoes in drills two feet apart, and the large potatoes in 

 rows three feet apart. Here, you see, is a double variation. 

 There was no coudition the same, and no chance to compare such 

 experiments ; and a very large proportion of experiments have 

 been vitiated in that way. It arises from attempting to deter- 

 mine two things at once : that is, the effect of variation in the size 

 of the seed, and the effect of variation in the rows. You have 

 two elements, and you may try such experiments as long as you 

 please without any valuable results. It would be better to try 

 one experiment to settle the matter in regard to size, having all 

 the conditions precisely the same, and then take, as a separate 

 experiment, one in which the different distances of the rows was 

 tlie object of investigation. In reference to this difficulty in 

 •regard to the variation of the soil, it would be necessary to have a 

 large number of unmanured plats for comparison. The increase 

 of the manured plats over the unmanured should be where the 

 plats are treated precisely alike. 



1 think it would be desirable, also, to put crops for two or three 

 years in succession upon the same land, before using it for experi- 

 mental purposes. For instance, if you wish to test the value of 

 manure as applied to soil, it would be better to mark out your 

 plats accurately, and crop them for several years without any 

 manure until you ascertain the peculiarities of this crop, and then 

 continue the same crops with the addition^of manure afterward; 

 but in this case at least one half of the plats should be left 

 unmanured for comparison, so that you may compare results 

 alongside of the manured plats, and compare the plats with them- 



