EXPERIMENTS ON FARM.

New Zealand Herald, Volume LXII, Issue 19124, 16 September 1925, Page 18

EXPERIMENTS ON FARM.

VALtTE OF INVESTIGATIONS.

FACTORS FOR CONSIDERATION. SELECTION OF CONDITIONS. It is impossible to turn a commercial farm into an experimental station —the financial loss would be too groat. At tho same time, there is no reason why a few simple investigations should not bo carried out upon a field devoted to this purpose. No two farms are alike, and narrow enquiries which may be particularly appropriate to tho needs of an individual holding cannot always bo prosecuted in a general research establishment. Detailed conditions, such as sods, markets, climato and local breeds of livestock, vary so widely that it is an undoubted advantage for the farmer himself to plan experiments which may in many small ways load to a reduction in running costs. Tho substitution over tho whole farm of a less expensive fertiliser which, on an experimental plot, has been proved of equal efficiency to that of its predecessor, may effect a surprising economy. Again, on a given soil, or with a given crop or climate, ,it may be ascertained that a singlo harrowing will rival in economy two "strokes." In many cases tho busy practical man cannot afford to risk a radical change in his methods, because such a change, when adopted throughout, the entire farm, may the previous practice involving, perhaps, | hayo disastrous results. On tho other hand, it may thoroughly justify itself. If, these changes be attempted first on an experimental plot, an aero or so of land only is subjected to the risk, assuming that tho conditions operating on the plot closely approximate to thoso under which tho rest of tho farm is worked. Choice ol Representative Site. Tho first essential when dealing with crop experiments is to choose a representative site. This may not be so easy as it appears on casual consideration. A field can usually be found which exhibits a surface character fairly typical of the whole farm, yet—especially in the case of alluvial deposits—the lower levels of soil may not "match." Adjacent outcrops so widely differing as clay and chalk '.may be uniformly overlaid with loam, S3 that it is necessary to study the general geology of a farm before deciding upon a site for operations., Drainage is another factor which should be carefully considered. In the case of a ridge-and-furrow system, tho furrows will always be wetter. Within normal limits, then, it is along the troughs that the greatest crop yield is to be expected. For this reason experimental plots should .be laid transversely to, and not parallel with, tho ridges. This i 3 especially a necessity where plots are long and narrow in shape. Similar considerations apply to under-drains, for the soil immediately overlying these is invajrjabiy wetter than i 3 the rest of a field. In conducting a crop-experiment, the various forms of treatment under comparison should first be carefully listed; and the list will be all the better if not too ambitious. For instance, a nitrogenotia artificial experiment dealing with three forma of fertiliser will be found tpite sufficiently complex to occupy the tune and acreage which a busy man has at his dispusaL Two plots st least should be devjted to each style of treatment, and duplicates are placed as far apstrt as possible in order to ensure uniform representation. Even the most consistent of soils will often vary slightly from one end of the field to the others—- " plots should lie parallel with the direction of variation, so that each is subject to the same gradation in character as is the whole of the area occupied. Largo Plots Desirable.

The sizrt of plots is a consideration in *lrich the farmer is limited. It i 3 obvious nnrepreswilativ" re solts diminishes as the experimental area u increases, but it has been stated that butween the sizes of l-50th and f acre there E> not a .great deal of variation in the possible error. The chances of inaccuracy da® tc excessive snsallness are naturally increased in the case of root crops, ■where msuy fewer plants per acre are present than in the case of cereals. Furthermore, the perimeter of a plot receives more light and air than does its centra Al--thougb . where ploto are of eqnal size, all except the extreme members are subject to identical circumstances in this respect, the conditions are by no means representative of those which obtain in a normal field where the perimeter bfars & smaller proportion to the total area. This discrepancy is reduced as , the size of an experimental plot increases. The safest practice is to make the plots as largo as possible, having due regard to the total area, t;me and labour which are available. Plots should be rectangular in shape. Land destined for experimental purposes should be as free as possible from apparently minor factors which may lead to . inequality. The ground'' surrounding tree roots is invariably drier than is the rest of a field. The shelter afforded by a belt of trees or by a building may have abnormal effects upon the moisture-con-tent or lighting of a particular area, while a field located' on the crest of risitog ground may slope in two opposite directions—the one side being less open to evaporation by a prevailing wind than is the other. Simultaneous Cultivation, Cultivation should, where posrible, be performed simultaneously over the whole experimental area (Le., transversely to the direction of plots). This practice ensures that any unevenness . consequent upon tha suspension of Work, and its' subsequent continuation under different conditions, will be distributed uniformly over isvery plot instead of being vested in individuals. In the case of experiments ;where cultivation is itself the factor under consideration, th«i rule is, of course, en- j tirely different. Hero each plot would receive its particular treatment separately, and it is very desirable that a given operation should be completed over the whole experimental area in one day. This consideration again emphasises tho need lor modesty in the scale of investigation. 'Hie foregoing outlines the method of choosing ideal experimental conditions. On a given farm it is, of course, impossible to fulfil every such ideaL Considerations of geology, accessibility artd business routine preclude such an achievement. The best that can be done is to make a careful selection of those available conditions which approach ihe ideal from a majority of standpoints, but there is no reason why a lot of valuable information should not be gathered under eemi-ideal. circumstances. Interpretation of Results. The interpretation of results is a matter which requires care. Where yield experiments are carried ont, it is very desir able to weight the produce from an entire plot—not from an aliquot part thereof. The liability to error is thus spread ever as wide an area as possible. In tho case of root experiments, weighing should take place before storage, while produce from individual plots should be harvested and stored independently. Scales should bo carefully adjusted and tested both before jmd after weighing. The ocular observation of growth thtri fag the period of investigation will often lead to, valuable deductions. One plot may do remarkably well up to & certain stage, but may subsequently be rapidly overtaken by a second. Similarly, ono jnay succumb to a disease, carrying it thrmghout the whole growing season, pother, though attacked, may sbow 'signs of resistance. Residual effects should be noted upon rlhe succeeding ..crops Among thase would be classed the .response to unexhausted (manures, susceptibility to disease, moisture resources of the soil a id the physical *»t«rc ' a clay.