ERRORS OP THE EXPEBIMENTS. 123 



either from the marl or the putrescent manure. These causes will 

 be briefly stated. 



1st. The quantity of marl (800 bushels) m q r and r p is nearly 

 double the amount that ought to have beon used; and this error 

 has not only increased the expense uselessly but has served to pre- 

 vent the increase of product that would otherwise have taken place. 

 This loss is proved by the gradual increase, aad %t *ast the greater 

 product of w p, marled at only 450 bushels. 



2d. The comparative superiority of all the marled ground to s 

 q, not marled, is lessened by this circumstance : most of the large 

 logs, as well as all the small branches, were burnt upon the land, 

 when it was cleared in 1824, before the experiment was com- 

 menced ] and the ashes have durably improved a spot where each 

 of these large fires was made on s q, but have done no good, and 

 perhaps have been injurious, to the marled pieces that were made 

 sufficiently calcareous without the addition of ashes. At least, the 

 good effect of ashes, on spots, is very evident in s q, and has helped 

 somewhat to increase all its measured products, and no such benefit 

 has been visible on the marled parts. 



3d. The quantity of putrescent manure applied to s t (900 to 

 1100 bushels) was much too great both for fair experiment and 

 profit; and the excess of quantity, together with the imperfectly 

 rotted state of the stable manure, has given more durability to the 

 effect, than is to be expected from a more judicious and economical 

 rate of manuring on such land when not marled. For these several 

 seasons, it is evident that far more satisfactory results than even 

 these would have been obtained, especially in the amounts of nett 

 products, if only half as much of either marl or manure had been 

 applied. 



There are other circumstances to be considered, which, if not 

 attended to, will cause the comparative increase or decrease of pro- 

 duct in this experiment to be misunderstood. It is well known 

 that poor land put under tillage immediately after being cleared, 

 as this was in 1824, will not yield near as much as on the next 

 succeeding course of crops. This increase, which depends merely 

 on the effects of time, operates independently of all other means 

 for improvement that the land may possess ; and its rate, in this 

 experiment, may be fairly estimated by the increase on the piece 

 s q from 1824 to 1828. The increase here, where time only acted, 

 was from 12 to 21 i bushels. But as the corn gathered here was 

 always much th.9 most imperfectly ripened, and would therefore 

 lose the most by shinking, I will suppose eight bushels to be the 

 rate of increase from time, and that so much of the product of all 

 the pieces should be attributed to that cause. Then, to estimate 

 alone the increased or diminished effects of marl or manure on the 



