« PreviousContinue »
striking than in agriculture. The farmer of the last century was content to do as his fathers had done before him, and ploughed and manured in certain ways, because it had been the custom to do so from time immemorial. If he were possessed of superior por os of observation (a far rarer gift than is generally supposed), he noted what occurred on his own or his neighbors' farms; and if it happened that any particular operation was attended by favorable results, he adopted it in future years. It will be readily understood that, so long as this method was followed, the progress of agriculture was necessarily slow, for it depended not merely on the chance of a particular result occur sing, but of its occurring under the eyes of an individual possessing the qualifications necessary to enable him to take advantage of it, and, even when thus observed, the knowledge acquired passed but slowly from man to man, and many years elapsed befom
became generally known; for there was then no agricultural press to diffuse the knowledge of it over the country, and no agricultural societies in which its good effects eould be discussed. Very striking is the contrast between this state of things and the active watchfulness of the modern farmer, ever on the alert for what is new, rapidly taking advantage of every step that is made, and anxious to do what in him lies to con bibute to the general fund of information.
Amidst the various methods by which agriculture may be promoted, the claims of er periment have not been overlooked; and the farmer has of late years devoted much time and labor to this important means of establishing the principles on which his art de pends. The introduction of experimental inquiry has indeed, in a certain sense, pra duced a revolution in agriculture; for it has raised it to the dignity of a science, and shown that it depends on general principles, which it is possible for the farmer to develop and establish on a sure and firm foundation. Nowhere have experiments been made with greater zeal and activity than in our own country; and we owe their number and accu. macy very greatly to the fostering care of the Highland and Agricultural Society. Up Tards of twenty years ago the Society commenced offering prizes for reports of experiments, and it has continued to do so ever since; and the result has been that the pagea of its “Transactions" contain a larger number of careful and accurate experiments Phan those of any similar periodical. Every year the subject receives new attentionz and the premium-book contains a series of suggestions for experiments, selected by a committee including a large number of the most experienced practical farmers, as being Those which especially merit investigation, and are likely to give results useful to agriculture. Although the number of experiments made by farmers in Scotland and else There is very considerable, there is no doubt they might be beneficially extended. Nen subjects of investigation might be undertaken; and by the adoption of a systematio plan, the value of the results obtained might be increased; and I have, therefore, thought that the whole subject might be advantageously discussed here, and the aims and objects of experiments, as well as the conditions requisite to insure accuracy, explained. I am de more induced to do this because it has been erroneously alleged that I am inclined to depreciate experimental agriculture, than which there cannot be a greater misappre hension. So far from this being the case, I have never lost an opportunity of doing all in my power to promote it. By many elaborate analysis I have sought to give completoDess and precision to the results obtained in the field, and have always expressed my sppreciation of the care and accuracy of those who have devoted themselves to such lo quiries.
It seems to me, however, that the time has now arrived when experiments might be mmdertaken with wider objects than those with which we have hitherto been satisfied, and that the progress of agriculture, while it demands more minute and extensive ide quiries, affords also the means of accomplishing them. It was not long since remarked
to me by a distinguished agriculturist, who has himself made many accurate and welldėvised experiments, that he thought farmers had got too much into a beaten track, add vent on repeating the same experiments over and over again; and in this I think them is some degree of truth; and, without undervaluing repeated experiments, the great importance of which will be afterwards pointed out, there is no doubt that new subjecta
inquiry might be opened up with the effect of interesting a great number of persons, ad inducing them to enter into this field of usefulness. The fact is, that from the very nature of things, the field of experiment increases every year. As our knowledge ada ronces new subjects suggest themselves, and it frequently happens that the result of one veries of experiments gives the first indication of the necessity for another, or they may how that the method which appeared well adapted to elucidate the required facts is not pufficient for the purpose, and render it necessary to throw overboard what had been done, and to commence again ab initio. In such a case it may seem that the labor e pended has been lost, though such is really not the case, these imperfect experiments being a necessary step towards the more perfect. It must be borne in mind also that, as wur knowledge advances, not only are more minute and accurate inquiries necessary, but the methods of investigation themselves improve; so that no experiments can be eonsidered as final, but all must be repeated and revised at more or less distant intervals . time.
Without entering for the moment more fully into this question, we may set out with the position that careful and minute experiments are one of the most indispensable means of promoting both the science and practice of agriculture; and as they are very laborious, it is most important that they should be directed into the most profitable channels, and the maximum of useful results be obtained from them. In considering the natter, it may be observed at the outset that all agricultural experiments may be divided into two great classes, one set leading to special, the other to general conclusions; the first having for its object to determine particular facts, the second to establish general prindiples.
As an illustration of the first of these, we may take the case of experiments made by ndarmer, in which he contrasts two or more manures with one another, for the sake of satisfying himself as to which it will be most profitable for him to use on his own farm. 1 he obtain a definite result, the immediate and tangible gain is obvious, because he bs krabled to introduce a material economy by confining himself in future years to that manure which proved most advantageous. But it by no means follows that the result shall be of any use to his neighbor, for differences in the soil may effect the results; and even were they completely identical, differences in height and exposure or in the meteof. dogical conditions may completely alter the circumstances. No doubt the case is raro in which an experiment of this kind is altogether devoid of instruction or value to other agriculturists; but very frequently an individual, when experimenting for his own Hnefit, may adopt a kind of inquiry very different from that which he would select X Bis object were the instruction of the agricultural public.
The other classes of experiments which are directed to the general advantage of agreb alture have a higher aim, for they seek to establish some general principle or principles which underlie its practice, and thus enable us to generalize, as we say in scientifte language that is, enable us to draw conclusions which are not only true in a few special vadles, but, if rightly used, are applicable under a great variety of apparently different qaditions. Such experiments, for example, would take a manure not as a whole, but would dissect it, as it were, and selecting esch of its constituents separately, would en daayor to ascertain what part of the total manurial effect it produced was due to each of miem; and if this course woro pursued with a variety of substances, we should in the
long run be enabled to see what result ought to be produced by any particular mixture under any circumstances. The great laws by which the action of manures is governed Vould thus be established—a knowledge valuable in all coming time and in every posable condition.
There is, however, a very important difference between these two classes of experi. ments, inasmuch as those made with a special object give an immediate answer of some kind or other, and decide, even when repeated only a few times, which is the most useful manure to employ in the particular case; while those made for the purpose of determin. ing general principles must be frequently repeated and varied in every possible way, and may be carried on for a great number of years, and with the expenditure of much time and labor, before any conclusions whatever can be drawn from them. With such a difference it will be no matter of surprise that the majority of experiments hitherto made should belong to the former class, and comparatively few to the latter, which at first sight appear of a more abstract cha ter, and less immediately applicable to practice. It is for this reason that the prizes offered by this society for special experiments have generally been well competed for; while those which seek to determine general laws have attracted comparatively little attention, though they are really the most important, end this is after all very natural; for every one naturally desires to see his own laborg producing an immediate practical result, and looks with far less favor or gratification on the slow process of storing up facts from which no safe conclusions can be drawn until they have been accumulated during many years. Yet this patient accumulation is indis pensable to the proper foundation of scientific principles. Its necessity is recognized in all the sciences, and it is very forcibly illustrated in the case of astronomy, where the abserver is often compelled to wait for many years until the recurring course of events enables him to observe the facts be requires, and he is content to go on year after year amassing facts by which his successors are destined to profit.
The agriculturist is in a somewhat similar predicament. His experiments necessarily extend over an entire season, and are exposed to the disturbing effects of weather, and many other causes which, ag every one knows, will tend to invalidate, or at least to effect, the accuracy of the conclusions to be drawn from them, and are often very dis. heartening to the experimentor. Perseverance should indeed be the motto of the agricultural experimentor. He should remember that his results are valuable in proportion to their number, for it is frequent repetition which en es him to eliminate the effect co soil and weather, and gives precision and certainty to his results.
What I have to say regarding the conducting of agricultural experiments may be di. vided into two parts-1st, the precautions necessary to procure accuracy, and, 2d, the mode in which experimental agriculture may be best promoted. In commencing any set of experiments, it is of primary importance that a definite plan should be laid down The experimentor should put before himself some question or questions for solution, and should then set himself to consider carefully how his experiments may be so shaped as to gre a definite answer with accuracy and precision. To secure this no precaution should be omitted, and he ought always to bear in mind that an inaccurate experiment is not only useless, but is misleading, and that the omission of some apparently unimportant piece of information may deprive an otherwise excellent experiment of half its value. Above all things, he must avoid undertaking too much. He will find no difficulty at seed-time in beginning a large number of experiments, but it is not so easy to make all the necessary weighings at harvest, when he has so many other matters to attend to. It is very desirable also that the experiments should be so arranged that they may bo na
peated during several successive years on an exactly similar plan, by which their value will be enormously increased.
Having fixed upon the particular subject to be examined, the next point is to ascertain exactly the existing state of information on it, the current opinions regarding it; and any previous experiments bearing on it should be minutely examined, by which means much labor and unnecessary repetition may be avoided. The previous experiments should be very carefully studied, and their weak points traced, so that the same errors or defects may not occur in those about to be begun. Having arranged the best mode of proceeding, it is well to consider whether the experiments about to be performed might Dot at the same time be made to throw light on some points not included in the original plan; or whether, by some slight modification, they might not embrace some further matter without detracting from their accuracy in relation to their primary object. These arrangements should, if possible, be made long before the time at which it is necessary to commence the experiments, so that the plan may be well digested, and nothing of importance be overlooked. The plan being arranged, the next matter for consideration is the field in which the experiment is to be made, and to this very great attention should , be paid. It is of primary importance that the soil should be as uniform in quality and texture as possible. Absolute uniformity can rarely be secured, but every effort should be made to come as near it as circumstances will permit; and if the experimentor wishes to give the highest degree of accuracy to his results, he will endeavor to produce some oxperimental evidence of the uniformity of his soil, or of the limits within which the produce of different parts of it differ point in regard to which most experiments are very deficient.
Preference should always be given to a field which is perfectly flat. Undulations on the surface are generally connected with differences in the soil or subsoil, and are often due to the existence of rocks beneath the heights, which, of course, modify to a great extent the nature and quality of the superincumbent soil. On a considerable slope the Boil can never be uniform, as any one may easily convince himself when the state of a ploughed field, on even a very moderate inclination, is examined after a heavy fall of rain. The furrows at the lower part of the field will then be often seen filled with matters washed down from the higher level; and what is there deposited is sometimes the coarser sandy part of the soil, the finely divided clay, which is generally its most valu. able portion, having been carried off to the ditches or drains; while at other times, and more particularly when the slope is gentle, the finer particles are deposited at the bote tom, and the coarser left behind. Whichever of these be the case, it is very manifest that the effect is to create a difference between the two portions of the field, which must in the course of time become considerable. When it is unavoidable to make the experiments on such a soil, the plots for the different substances should be arranged in a single row along the middle of the slope, the upper part of each being at the same distance from the top. In this way the chances are that all of them will be fairly comparable with one another; but it must never be forgotten that, though experiments in such a soil are admissible, they are open to some risk of fallacy, which can only be avoided by the greatest possible care.
The importance of uniformity of soil is so obvious that there is scarcely an experiment on record in which it is not referred to, and a statement made that the portion selected for the experiments was suitable in this respect; but it is genearlly a mere statement of a fact which we are led to infer has been determined merely by ocular inspection. But as this is liable to be fallacious, it would be infinitely preferable if some more precise method of ascertaining it were adopted. This could be readily done by deciding on the Dature of the experiments and the position of the plots a year beforehand. The plots should then be at once staked out, and the crop being raised in the usual way, the produce of each should be harvested and weighed separately; and if it proved equal in all cases, the soil would be perfectly uniform. I am not acquainted with any experiments in which this precaution has been taken-indeed, I believe it is now suggested for the first time; but there can be no doubt that the year's delay and greatly creased labor would be amply repaid by the additional value and precision of the results.
The condition of the soil is also a matter of great importance. As a rule, it may be stated that experiments are most satisfactory when made on soils which have been for some time under careful cultivation, for by this means equality of texture is most readily secured; but is advisable to avoid those which are in the highest condition, or which have been recently and heavily manured. In general, a soil which has been well worked but not highly manured is best, because on it the differences between the manures employed will be most marked.
Not less important is the consideration of the size of the plots on which the experiments should be made—a subject of some difficulty, and in regard to which very great difference of epinion exists. The general impression is that the larger they are the better; and it has been held by some that no experiment should be made on less than an acre, while others kave made half an acre their standard; and still smaller quantities, down in some instances to a very small fraction of an acre, have been used. It seems to me that no good general rule can be laid down for our guidance in this respect, and that much must depend on the object of the experiments, and the circumstances under which . each individual investigator is placed. If, for example, the object be to grow a crop by , means of different manures, and then ascertain the relative feeding values of the produce by feeding expriments on cattle, it is very obvious that no results worth having can be got except by working on a considerable scale; and even an acre may be too small & quantity to give good and trustworthy results. On the other hand, where it is . merely intended to compare the weight of the produce, that quantity is unquestionably far toe large.
Without undervaluing the advantages of large plots, some of which are sufficiently obvious, I have no hesitation in saying that experiments on a small scale offer great con: reniences; and when made with the necessary care, are quite as accurate, if not more : so, than those made on larger plots. The arguments which have been brought forward in support of large experiments are chiefly that the crop can be raised more in accord. ance with the ordinary-methods of cultivation; that local inequalities of soil are to a great extent avoided; and that a fairer average is obtained. When small experiments are made it is argued that, from some unexpected or unobserved peculiarity of the soil, one or more plots may be greatly superior or inferior to the others; and the result of the experiments would be to place the manures used on them in a position above or below that which they ought to occupy. It will be observed, however, that this argument proceeds or the assumption that no method other than the fallacious one of occular inspection has been used to ascertain the uniformity of the soil. If the mode of testing this point by weighing the produce of the ordinary crop in previous years, already adverted to, had been used, this difficulty would have been avoided. The fact is that, though large plots may have the advantage of eliminating any local peculiarity of the soil, they are no guarantee for its uniformity in respect to gradual changes. It frequently happens that a field is heavy at one end, and gradually shades off into a lighter soil at the other; and when this is the case, large plotf are very disadvantageous, because such differences are peculiarly apt to escape detection, though they may greatly affect the results of the experiments.
Supposing it be resolved to make only ive experiments on an acre each, this requires