five acres in all; and I do not hesitate to say that there is nothing more difficult than to find that quantity of soil of the necessary uniformity, and that in the great majority of cases plots of an eighth, or even a sixteenth, of this size would give much more accurate results. It is urged, also, that the crop can be more accurately weighed when the spaces are larger; but this is only because experimentors generally are content with the rude weighings which suffice for the ordinary farm purposes, while, if they used a somewhat more refined apparatus, equal, if not greater, accuracy might be secured in the small scale. The size of the plots must, in fact, be greatly influenced by the nature of the experiments to be undertaken. Where the object is merely to make what I have called special experiments, a considerable breadth of land is best; but it need never exceed a quarter, or at most half an acre-a quantity amply sufficient to give a fair average of the whole field. Where, however, the intention of the experimentor is to contribute to our knowledge of the general principles of agricultural science, much smaller quantities of land will suffice, and a twentieth, or even a hundredth, of an acre will often be sufficient. In these cases, it must be remembered that it is comparative results that are wanted; and it does not matter whether the produce from the plots be above or below the average over the whole field, provided only they be made under precisely similar circumstances, and are fairly comparable among themselves.

The main objection which may be made to small experiments is that, when the crop is one of a kind which throws out long roots, it is likely that the mutual interpenetration of the roots of adjoining plots may occur to such an extent as to interfere to some extent with the accuracy of the results. If, for example, a plot to which nothing has been applied adjoins one which has been heavily manured, some of the roots from the outer part of the former will penetrate into the latter, so that its produce will be somewhat increased; while, in the same way, the outer plants of the manured soil will stretch into the unmanured, and the difference between the produce will consequently be less marked than it ought to be. The difficulty can, however, be easily overcome by a slight separation of the plots by a sort of outer rim. If, for example, it be resolved to make the experiments, say, in a fiftieth of an acre, the space should be measured off by fortieths for manuring, while the portions actually weighed should be in each only a fiftieth of an acre in the center of each fortieth. So that the field would be arranged somewhat as in this diagram, where the inner squares represent the quantities to be weighed. Another

objection is the difficulty which is experienced in gathering the crop without overstepping the boundaries of the plots, which, when they are small, may lead to great inaccuracies. I was myself led to direct my attention to this point some years since when

engaged in examining the progress of vegetation in the wheat crop, when, from the very nature of the inquiry, it was indispensable to work on a very small scale. The plan I then adopted was to measure out the spaces with great care, and to drive firm pins, projecting about six or eight inches above the surface, into the corners, between which slender wires were stretched, so that when the produce was required it could be cut with the greatest certainty, the plants taken being always those included within the wires of each plot.

This plan has recently been adopted at my suggestion by Mr. Thomson, of Grange, Kilmarnock, who expresses himself highly satisfied with it. I am convinced, indeed, that when such precautions are taken, small experiments will be in many cases more accurate than large ones. And in this respect agriculture would only be in the position of the other sciences. In chemistry, for example, the experimentors of the sixteenth century used pounds, and those of the eighteenth, ounces, where we use grains; while there is no comparison whatever between the accuracy of the results. I am far from supposing, however, that the use of small plots makes experiments easy; on the contrary, I am satisfied that they require more care and attention to minutiae; but a greater number can be made, and they are more rapidly performed. This latter is, indeed, a matter of great moment, because the experimentor can then superintend personally the whole operations; whereas, when the plots are large, he is almost of necessity compelled to leave a considerable part of them in the hands of ordinary farm-laborers, who cannot always be trusted to carry out with intelligence the instructions given them. There is no question at all that, if we are satisfied with small experiments, the number and variety of those performed are likely to increase very greatly. They can be carried on at a much smaller expense than large experiments; and this is no light consideration. When experiments are made on half acres or acres, the expense of the manures employed is often serious, and when a considerable breadth of land is to be taken up by them, the experimentor will naturally arrange his work so that the crop shall in all cases be remunerative, though this is not always the best condition for elucidating scientific principles. An experiment may be a failure in one sense, inasmuch as it may show that the manure used produced no practical benefit; but it may not only be no failure in a scientific sense, but may actually contribute very greatly to the establishment of some important facts. It may even be very desirable that experiments likely to give negative results should be made; but it is not to be expected that farmers will be found ready to sacrifice the produce of a whole acre, though they might not object to losing that of a smaller space.

Taking the question as a whole, I think that small experiments should be more encouraged than they have been, provided they be made with the necessary precautions, not certainly to the exclusion of large experiments-some of which it would be always advisable to conduct-but because, by giving due weight to them, we should undoubtedly induce many more persons to engage in these interesting and useful inquiries. Whether we choose to operate on a large or a small scale, it must never be forgotten that single experiments have comparatively little value, and that every result, to be trustworthy, should be made in duplicate, and it is better still if it is repeated three, or even four, times on the same soil. So great is the importance of this precaution, that it may be safely asserted that, if a person has resolved to set aside a sufficiency of land for experiments on half-acre plots with ten manures, the value of his results would be increased fifty-fold by dividing each space into four, so as to make it a four-fold experiment. When they are made in duplicate or triplicate, it is very obvious that many of the uncertainties which beset a single experiment are avoided. The risk of error from local peculiarities of soil is so far more effectual than it can be by increasing the size of the

plots, and mistakes in weighing or measuring the land and produce, which with every care must sometimes occur, are greatly diminished.

The disturbing causes which are not eliminated in this way are the effects due to the kind of soil, and to the meteorological conditions. These can only be ascertained by extending the experiments, so as to include different descriptions of soil (which may be selected in the immediate neighborhood of one another), and localities differing in their temperature and rainfall, which must almost of necessity be at some distance apart. When these matters are to be taken into account, the scope of the experiments is so greatly extended as to make them far beyond the power of an individual; and such a plan could only be accomplished by several persons in different parts of the country associating themselves together for the purpose of repeating the same experiments. To the value of such an arrangement I shall afterwards advert. Of other precautions to be taken, several must be referred to. Among others, I would particularly refer to the importance of leaving plots entirely without manure-a matter to which I should not have considered it necessary to refer had I not seen many series of experiments otherwise well performed rendered comparatively useless by this omission; and I have felt great regret that good experiments, which had clearly involved much labor, should have had their value diminished by an omission which might have been so easily supplied.

The application of the manures involves some attention. Those which are to be employed should be procured from dealers of reputation, so that their quality may be good. Before each is applied, it should be turned out of the bags on a clean wooden or stone floor, and carefully mixed with the spade, any lumps contained in it being broken down. It should then be passed through a sieve and again mixed, and a sample taken for analysis from six or eight handfuls of the stuff. The quantity required should then be weighed and mixed with once or twice its bulk of damp sand, and the whole being again passed through a sieve, it should be put into a bag ready for use. The object of mixing with sand is to admit of its uniform application, and to prevent the more dry and dusty manures being blown about by the wind during application. It is very necessary, when precise results are required, that the actual manures used should be analyzed, in order to see that they come up to the proper standard; and this precaution should never be omitted, for mistakes will occur, in proof of which I may just mention that I not long since examined a manure supplied for experimental purposes by a manufacturer of the highest standing, and stated to contain 38 per cent. of soluble phosphates, in which only 28 per cent. was found. It is scarcely necessary to observe that, had the manure been used without analysis and on the faith of its containing 38 per cent., any conclusions drawn from the experimental results would have been completely fallacious. For still greater completeness the soil should also be examined, or at all events a sample should be taken just before the experiment is begun, which may be afterwards analyzed should it be deemed necessary. For this purpose a hole should be dug to the depth of the soil, say ten inches, and from its perpendicular side a slice two or three inches in thickness should be taken. This should be repeated at five or six different parts of the field, and the whole of the soil thus taken should be carefully mixed with the spade, passed through a coarse sieve, and spread out in a thin layer on large sheets of paper for some days to dry in the air; and 5 or 6 lbs. fairly selected from this should be preserved in a tin box. It is scarcely necessary to observe that during the growth of the crop attention should be paid to everything bearing on its progress. The time at which it braids in each plot, when it comes into ear, if a cereal, and all similar events, should be carefully noted.

A knowledge of the rainfall and temperature during the time of the experiments is also of much importance, and the necessary information on these points, obtained from

any meteorological observer in the neighborhood, will generally suffice. The most perfect plan would be to have a rain guage on the field itself, and if the strength of the solar rays were also observed, the results would doubtless be most interesting; but these are refinements which we can scarcely expect to see adopted in the present state of agriculture. When the crop approaches maturity, the proper time for harvesting becomes a matter demanding much careful consideration. The general system is to collect the produce from all the crops on the same day. As far as root crops are concerned, this, is probably right, because there is no definite point at which they can be said to be ripe; but it is quite otherwise with the cereals. In them the period of exact maturity should be carefully attended to; and, as the effect of some manures is to accelerate this point, it would be wrong to reap all the produce on the same day. If this be done, it is necessary: either to wait till the latest plot is ripe, in which case some of the grains of the earlier ones are liable to be shed, particularly if the weather is windy; or to begin when the first is ripe, in which place the others are placed at a disadvantage.

The experimentor must, therefore, examine the crop day by day, and gather each when in the same condition as to ripeness. The crop has then to be weighed, the bulbs and tops of roots being taken separately, and the gross weight of the cereals being noted in the meantime, while the separate weights of grain and straw is left until it is convenient to thrash them. Here the experiments terminate for the year, unless it be thought necessary to ascertain the nutritive value of the produce by analysis. Where the experiments have no other object than the determination of the utility of different kinds of manures under special circumstances, this is not absolutely necessary; but where they have been made with the intention of contribating to our knowledge of the general principles of agriculture, it forms an important part of the inquiry. Suppose, for example, a set of experiments to be made by applying different quantities of ammonia to the soil, it may be a question whether a larger dose may not only increase the crop, but also yield a produce richer in nitrogenous matters. In this case, analysis is indispensable. Although the series of experiments may be considered to end here, they would still leave an important question regarding the effects of the manure unsolved. Ex perience has shown that in no case does the increase on the produce contain more than a veryi small proportion of the valuable matters contained in the manures; and if the experiments are to be exhaustive, they ought to ascertain what has become of the remainder, and what effect it will produce on subsequent crops. The determination of this point renders it neces sary to weigh the produce from the same plots-to which no manure of any kind has been applied in the meantime in the subsequent year, or, better still, during the remainder of the rotation. It is very difficult to accomplish this in a satisfactory manner, particularly with small plots; and hence the cases in which it has been attempted are very few, although the question is one of great importance, and calls loudly for investigation.

I have now discussed the precautions necessary for draining, complete and exhaustive results, in one great step of agricultural investigation; but there is another, in which the great› object is to determine the nutritive value of different substances by feeding experiments on stock. Experiments of this kind are more difficult and troublesome than those into the discussion of which I have gone so fully; and time will not permit me to consider all the precautions they require. They are, however, of great importance, and offer a wide and fertile field for the agricultural experimentor. It was my intention to have entered here into some details regarding the exact kind of experiments which it is desirable should be undertaken, but I have already occupied so much time that I fear I should exhaust the patience of my audience by doing so; and I must reserve its full consideration for some future opportunity, and con- › tent myself with a very few remarks on the present occasion.

The fact is, there is scarcely a department of scientific agriculture in which experiments, embracing all the precautions and details already given, would not be valuable. Among sub

jects particularly requiring examination, I may mention the effect of phosphates in different states. In bones the phosphates exist in an insoluble form, but they nevertheless operate very powerfully as a manure. If, however, the bones be burned, the phosphates left appear, so far as the few experiments made with them have gone, to be without manurial effect, unless they are treated with acid, and so rendered soluble. Repeated and varied experiments are neces sarily due. A comparison between the action of nitrogen in the state of ammonia, and as gelatine, in which form it exists in the bones, would also be of interest. A series of experiments on the manurial effects of the salts of potash is also much required. Some years since, when discussing the alleged exhaustion of the soil of the British islands by the modern system of agriculture, I expressed the opinion that, so long as the present supplies existed, there was no probability of the ammonia or phosphates being exhausted, but that a diminution in the available supplies of potash was a possibility, though a distant one. Up to the present time we have very little information regarding the effects of potash; and what there is is conflicting. A few years ago experiments were made at my instigation with muriate of potash on the potato-a crop requiring a large amount of that alkali-and the results were very favorable; and in one case so remarkable that the gentleman who made the experiments procured a sup ply, and used it on a considerable scale in the subsequent year, but the results were then entirely negatives. The cause of this remarkable difference, and, indeed, everything relating to the manurial action of potash, is still to learn. The same applies to the salts of magnesia, which are sometimes used as manures for the cereals, under the impression that they must be useful because the ash of these plants contains a large quantity of magnesia.

As regards experiments in cattle, I would just observe that we are still greatly in want of a complete and exhaustive series of experiments on the relative nutritive value of swedes and common turnips. An effect of water, in promoting the fattening of stock, also merits inquiry. It has been recently expressed by a French experimentor, that when a horse was liberally sup plied with water it gained weight rapidly, even when its daily rations of solid food were ma terially diminished. Should this be confirmed by further experiment, and found to be applica ble to cattle and sheep, it would obviously be a matter of great importance to the farmer. I have now further to consider what may be done to promote the pursuit of experiment, and here it may, perhaps, be thought that the description of all that is necessary for a perfect e periment, and the numerous precautions it requires, may have rather the effect of deterring those who might otherwise wish to engage in it-and it must be admitted that good experi ments do require both time and patience; but it is by no means so difficult to secure accuracy, and to attend to all these precautions, as it may appear. Any one inclined to take up experiments will find that, as in everything else the hand gets into it, and that what at first occasions a good deal of trouble becomes much simpler, and can with experience be accomplished without the effort it required at the outset. Neither should he be deterred from the good work by being in a position in which he cannot carry out all the details, for he may rest assured that experiments in which some of the precautions are omitted may still be useful, although neces sarily not so valuable as those in which they are observed. This much, however, may be said, that wherever the choice lies between many experiments with few precautions and few exper ments with many precautions, the latter alternative should unhesitatingly be adopted. The true plan is to undertake no more than can be well and effectually accomplished; and where a variety of results is required, it is best that several individuals should associate themselves together for the purpose. Where this is done, the results are sure to be most valuable, particn

e the same experiments are repeated on different soils and districts for several successive years, so as to elucidate the effect of disturbing causes. But when several persons unite for this purpose, it should be made a strict regulation that the experiments should be made exactly in the same manner, and that no one should be allowed to modify any of the details. I have seen instances in which the value of experiments was materially diminished in this way.