No. VJIL] • EXPERIMENTS UPON TURNfPS. 69 



fluence to prevent the exact results from being obtained upon a repetition of the 

 experiment. 



5°. Above all, it must be borne in mind that we are yet in the first infancy of 

 accurate experimental agriculture — that it will take many careful repetitions of 

 our experiments before we can ehminate the effects of the seasons — of the alti- 

 tude of our farms, their distance from the sea, the falls of rain to which they 

 are subject, and the kind of soil of which they consist. In the mean time our 

 most careful deductions must be considered as partial only, and as open to doubt 

 —as facts by the combination and comparison of which we are hereafter to ar- 

 rive at more general truths. 



With these preliminary observations, I turn to the experiments themselves — 



A. — The experiments upon turnips. 



The first series, those of Lord Blantyre — except the general answer that saline 

 substances cannot replace farm-yard raanure — afford no very satisfactory results. 

 They exhibit, indeed, some striking circumstances — such as 



1°. That 100 lbs. of salt per acre may, in a diy season, reduce the natural or 

 unaided produce of ti^rnips one-half— and that the same weight of nitrate of soda 

 may reduce it one-fourth! » 



2°. That in such a season as much as 16 cwt. of rape-dust per acre may be 

 applied, one-half drilled in, and one-half as atop-dressing, without producing any 

 sensible benefit. 



3°. That the same may be the case, if eight cwt, of rape-dust be drilled in, 

 and half a cwt. of nitrate of soda be afterwards applied as a top-dressing — 

 while if the same weight of common salt be used as a top-dressing "instead, 

 the crop will be increased one-half 



These results are too anomalous to be considered for the present as more tha* 

 accidental. They may possibly be explained either by the different degrees (A 

 moisture of the several parts of the field in which the mixtures were applied-— 

 or on the supposition, which is very probable, that in the concentrated state 

 some of these saline substances are more hurtful to the groacin/r plant than others. 

 It is to be regretted that the season was so unpropitious to this series of experi- 

 ments, for though the following experiments of Mr. P'leming afford some valuable 

 information, further knowledge still is wanted in regard to the relative effects of 

 different saline substaiices upon the growth of turnips, where no fermentible ma-, 

 nure is applied. 



4°. In these experiments, a striking contrast is presented between the effects 

 of rape-dust and those of guano, 16 cwt. per acre of the former gave only 3 J 

 tons of turnip bulbs, while 2 cwt. per acre of the latter gave 5 tons. It appears, 

 therefore, that rape-dust requires moist weather or occasional rain, while guano, 

 even in very dry seasons, will prodAtce a considerable effect. This is consistent with 

 what we know of the employment of the latter substance as a manure on the 

 arid plains of Peru. 



II. The next four series of experiments, those of Mr. Fleming, are rich in re- 

 sults and suggestions, 



1°, Limits of error. — The first observation which a careful examination of 

 them will lead the reader to make — and it appears to me to be a very important 

 one in reference to all future experiments of this kind — is suggested by the se- 

 cond series — those upon early yellow turnips, p. 44. 



In this series there are included two plots (Nos. 5 and 18), upon which no 

 manure was used. Upon one of these the produce amounted to 12 tons 17 cwt., 

 upon the other to 11 tons 8 cwt. only — being a difference of U tons, or one- 

 eighth of the whole. This difference between two equal portions of the same 

 field, apparently similar in soil, could scarcely, I think, have been anticipated, 

 and it shews that — where the produce obtained by the application of two tinlike 

 w^nures, to this turnip crop, does not differ 7nore than 1 i tons per acre, the effects 



