No. IX.] , REMARKS UPON PRECEDING EXPERIMENTS. 79 



crops the portions dressed with sulphate of soda are described as pale in colour 

 and yet the excess of produce over the undressed parts was as follows : — 

 In the oats . , li cwt. straw. } -txri .u i i . v j 



In the hay . . 2 cwt. per acre. \ ^"^^'^'^ ^^^ ^^^P^^^e was applied. 

 The increase in neither case would bs deser-ving of much attention except 

 as showing satisfactorily thaj, wrong conclusions may be drawn in regard to 

 the efficacy of manures and top-dressings by those who judge only by the eye 

 — and that safe reliaiice can be placed on those comparative results oidtj which have 

 been tested by weight and measure. I know, indeed, that practical farmers who 

 have applied nitrate of soda to grass land, and have been delighted by the beauti- 

 ful green colour which followed, have occasionally been disappointed by find- 

 ing that after all this promise the weight of hay obtained was no greater than 

 upon the undressed parts of their fields. As to the feeding qualities of the two 

 kinds of hay no experiments have yet been made, though it is known that cat- 

 tle prefer that which has been dressed. 



Suggestion XI. — I put down, therefore, as a distinct suggestion for the pur- 

 pose of drawing attention to the subject, that this plan of specially noting the 

 appearance of the crops at stated, say monthly periods, should be adopted in all 

 future experiments. This will serve, not merely to show us more clearly what 

 kind of appearances are to be trusted, and how far, as indications of an increase 

 of crop — but may hereafter prove of further importance when experiments shall 

 begin to be instituted upon the feeding properties of crops reaped under dif- 

 ferent circumstances, and raised under different kinds of management. 



3°. Importance of having iv;o or more experimental plots similarly treated. — 

 The experiments upon hay above-mentioned exhibit another illustration of the 

 fact adverted to in page 59 of this Appendix under the head of limits of error. 

 I there drew the attention of experimenters to the difference in the produce ob- 

 tained on two equal patches of the same field of turnips, to neither of which 

 any dressing had been applied. At Erskine two equal plots of grass in the 

 same field gave a similar difference of produce. 1 pi-esent both results here for 

 the sake of clearness. The produce per imperial acre was — 



Hay at Erskine. Turnips at Barochan. 



tons, cwt tons, cwt 



1st plot 4 5 12 17 



2d plot 3 3 11 8 



Difference 12 19 



In my remarks upon the difference between the two plots of turnips (Appen- 

 dix, p. 50). I expressed an opinion that differences equally great, depending not 

 at all upon the substance applied, might be expected on equal portions of those 

 fields upon which our different saline manures may have been applied ; — and 

 that very erroneous conclusions might thence be drawn in regard to the abso- 

 lute and comparative effects of the substances with which our experiments are 

 made upon the crops' to which they are applied. 



I have since met with a confirmation of this view in a record of two pairs of 

 experiments made with equal quantities of rape cake upon equal plots of red 

 whea^, in the same season, and upon adjoining parts of the same field, ^British 

 Husbandry, I., p. 112.) The results of two experiments Vith different quem- 

 tities of rape dust were as follows : — 



Produce of Light 



Rape dust applied, marliet corn. Weight per bushel. com. 



stones. bush. lbs. oz. lbs. 



1st plot 59i ... 26 ... 52 10 ... 46 



2nd plot 59i ... 21 ... 50 8 ... 67 



1st plot 86 ... 28 ... 53 4 ... 35 



2nd plot 86 ... 22 ... 51 2 ... 91 



The differences both in the quantity and in the weight of the grain reaped, ir 



30 



