92 THE ANALYST. THE ANALYSIS OF SOILS. BY J. ALAN MURRAY, B.Sc. I SHALL be glad to be allowed to offer a few suggestions in regard to certain proposals contained in a valuable and timely paper-the report of a committee of the Agri- cultural Education Association-on uniformity in soil analysis, which appeared in the November issue of the ANALYST. Whatever differences of opinion there may be as to the desirability of uniformTHE ANALYST. 93 processes of analysis, it seems to me that uniformity is essential in the following points : (a) I n taking the samples in the field; ( b ) I n manipulation of the coarse samples in the laboratory ; (c) I n making the solutions ; (d) I n expressing the results ; and it is upon these points chiefly that I should like to offer some observations on the report of the committee.With regard to the method of expressing the results, the committee recommend that results should be stated in percentages of the air-dry fine earth. I have, of course, no knowledge of what considerations weighed with the committee in coming to that conclusion; but I notice that the proposal was unfavourably commented on a t the meeting, and it is quite clear that results so stated would not be mutually comparable. It seems to me, however, that there are other and more serious objections to this popular convention. First, it should be kept in view that the results may often have to be interpreted, not by chemists, but by agriculturists ; and it is of the greatest importance that they should be stated in some such plain and simple way as to afford at a glance trustworthy information as to the stock of the various substances in the whole soil.I t cannot be maintained that a statement of percentage in the fine earth, air-dry or dry, does this : on the contrary, such information can only be deduced by long and troublesome calculations, even when the other necessary data are given, and too often they are not. For example, it may happen that 3, or 30, or 90 per cent. of the total mass of the soil may be made up of stones ; the specific gravity of the fine earth itself, and still more the density of picking, varies enormously. It is apparently to be left to the discretion of each chemist whether this information shall beincluded in his report or not, and without such information a statement of percentages means little or nothing.Secondly, the figures showing the percentages of the most important ingredients -phosphoric acid, potash, etc.-in the fine earth are, as a rule, inconveniently small, and the importance of those of the third decimal piace may easily be unappreciated. Thus, the percentage of total phosphoric acid is often represented by figures of smaller value than 0.1, and that of the ‘‘ available ” portion by figures less than 0.01 ; and yet 0.001 per cent., which to many people might seem almost a negligible quantity, would, in most cases, be equivalent to about 20 pounds of phosphoric acid per acre, and would require over 19 cwt. of superphosphate to supply it as manure. Thirdly, the total amount of water present in a soil at the time of sampling is naturally extremely variable ; and though it depends largely upon circumstances beyond control, and which may for many purposes be regarded as accidental, it is by no means devoid of interest and importance, especially if the rainfall in the district for the few days immediately preceding sampling be known or can be ascertained.The fact that it cannot conveniently be included with the other substances in the total of 100 per cent. must therefore be regarded as an objection to the plan of rendering results in terms of percentages rather than as a justification for ignoring the water altoget her.94 THE ANALYST. Finally, a plant’s chances of obtaining its necessary supplies of any substance depend, not upon the amount of that substance in a unit of mass, but in a unit of volume ; and, consequently, since a given volume of one soil may in extreme cases weigh perhaps twice as much as an equal volume of another, a simple statement of percentages does not afford a reliable basis for comparison of degrees of fertility of different soils.The remedy for this last-mentioned objection appears simple-namely, to express the results of analyses in terms of the quantities per unit of volume of the soil iizsitzc. Thus, the quantities of potash, of phosphoric acid, etc., in a given volume of soil form separate statements independent of each other, of the quantities of water and of everything else, so that the quantity of water may be included or omitted without in any way affecting the figures relating to the other ingredients.A report in these terms would yield, on multiplication of the figures by some simple factor, an expression of the quantities per acre, and by choice of suitable units it might be rendered in figures of any size we please. The question of the particular units to be chosen is, of course, a mather of minor importance; but, all things considered, it would probably be hard to find any with more to recommend them than ounces per cubic foot. These are common English standards, with which every farmer is familiar ; they are easily converted to pounds and acres respectively, and they produce figures of about ten times the magnitude of those expressing the same results in terms of percentages, so that it would only be necessary to carry the figures to the third decimal place instead of the fourth. Even the fact that the measurements would in most cases be made in the units of the metric system offers no obstacle, for, as it happens, the relation of the gramme to the ounce is practically the same as that of the cubic decimetre to the cubic foot ; and hence, if results expressed in terms of grammes per cubic decimetre be simply read as ounces per cubic foot, the error involved is not greater than 1 in 700-a quantity which for practical purposes might be neglected.In other words, the exact relation between grammes per cubic decimetre and ounces per cubic foot is as 700 to 701, which gives the factor 1.00143 for converting the former to the latter in cases where In order to give effect to the suggestion, however, it would be necessary to determine the mass of some definite volume of the soil in situ by means of a specially constructed sampler, as recommended by the committee, or in some other way.The sampler, if used, need not be exactly a cubic foot or a cubic decimetre, but its volume content must be accurately known. A cylinder 9 inches deep by 3 inches in diameter holds approximately a cubic decimetre, and is a convenient instrument in many ways, though it cannot be used in every case. A wider one is necessary for very stony soils, and a longer or shorter one where the soil is t c be sampled at different depths. The depth of soil sampled is very important in this connection, as it determines the factor for converting ounces per cubic foot into poixnds per acre. Thus, if the depth be 12 inches, the factor will be 43560 x & = 2722 nearly ; but if the depth be 9 inches only, the surface area in the cubic foot will be greater, and the factor in that case would be 43560 x & x 9 = 2043 very nearly.In the course of the discussion which followed the reading of the paper, Dr. Dyer in speaking of the proposal of , great accuracy is required.THE ANALYST. 95 the committee that soils should be sampled to a depth of 9 inches, mentioned that Sir Henry Gilbert now considered about 8 inches would, as a rule, be more suitable, and he suggested one-fifth of a metre as a convenient international standard. I do not propose to discuss in this connection what depth would be most suitable under ordinary or any circumstances ; but it occurred to me at the time that, if the fifth of a metre were adopted, it would fall in very conveniently with the suggestion I have ventured to put forward.It will, further, be seen that the proposal to state the results in quantities per unit of volume of the soil iu situ would involve a large departure in the methods of treating the samples in the laboratory from those at present commonly adopted and recommended by the committee. Thus, a sample taken in the manner and for the purpose referred to might be accepted as representative of the soil of the field or part of the field, or it might be rejected as the reverse; but in either case it would have to be accepted or rejected as a whole, and its properties would have to be determined in the mass as it occurs-ie., without any preliminary process of picking or sifting, or any other arbitrary method of separating the fine and coarse particles.For it is evident that a statement that a cubic foot of soil i?z situ contains a certain quantity of any ingredient soluble in concentrated or dilute acids implies that a representative portion of such soil as it occurs in situ has been submitted to the action of the solvent. I t may be argued, of course, that the coarser particles of the soil do not, within a measurable time, affect its chemical properties; but this, even if it were true, is not a satisfactory reason for excluding these particles from the action of the reagents, but quite the reverse. The roots of plants are commonly found adhering very closely to the stones. I t is possible that the plants derive merely mechanical advantage in this way, though the facts, so far as they are known, point in the opposite direction; but if the roots extract nothing from the Etones, then neither will those reagents with which we endeavour to imitate their action, if they do imitate it ; and if the roots do obtain nourishment from the stones, then the stones should be treated exactly like the other particles of the soil.The arbitrary use of a sieve of certain dimensions for the purpose of discrimi- nating what is fine and what is coarse material-i.e., what is soil and what is not- probably does not commend itself greatly even to the members of the committee who recommend it, and it is certainly much easier to excuse it as necessary than to defend i t as reasonable and satisfactory.Probably few persons could be found to approve of manipulating a sample of any other kind in a similar way; but without doubt, in the case of soils, the line must be drawn somewhere. Thus, a sample might be drawn containing a stone as big as a man’s head, and perhaps only as much fine stuff as would fill a teacup ; a sample might be drawn containing the stump of a tree ; but such samples would not be representative, and would therefore be rejected altogether. It may be taken for granted, then, that some method of deciding what substances might be legitimately included as representative in a sample of soil is indispensable, but it does not follow that it should be the arbitrary one at present in general use. On the contrary, it seems to me that each case should be judged separately and on its own merits, on the principle that d l those stones, etc., of whatever size, which are96 THE ANALYST.more or less uniformly distributed throughout the mass of the soil should be regarded as essential parts of it, and that those which are exceptional in point of size or in other respects should be regarded as accidental, when, of course, no sample containing them could be accepted. It is to be observed that this principle, should it be accepted, does not leave the decision to the somewhat haphazard method of simple inspection-though even that would in most cases form a sufficiently reliable guide-because in estimating the mass of a cubic foot it is always necessary to make several trials, and, under ordinary cir- cumstances, the weights will usually be found to differ only by a few grammes ; but should an exceptionally large stone or other accidental substance be included in one of the samples so drawn, the weight of that one will be found to differ very widely from the average of the others, and the sample would therefore be rejected.The plan I have adopted in taking a portion of the soil for analysis has been to disintegrate the various masses of wet; soil taken out by the sampler with the fingers, mix them as thoroughly as possible, and bring the lot together in a heap; then with a long-bladed spatula divide off such a portion as I reckoned would constitute a fair sample of the heap, and sxeep it into a weighed beaker, and weigh again to find the quantity of soil taken. In most cases I found that about 200grammes was sufficient, but under certain circumstances it was necessary to take larger quantities.It is, of course, necessary at the same time to set aside a portion to air dry, and afterwards to dry it thoroughly, in order to determine the moisture. Any errors which might be involved by the rougher method of sampling would probably be more than compen- sated by the difference in the size of the portions taken. The exact nature of the diEculties to be encountered and the best methods of dealing with them could, how- ever, probably be most satisfactorily dealt with by discussion in committee. I have found it easier to get the solvent thoroughly mixed with the large mass of soil, and in one case I found a considerably larger proportion of '' available " phos- phoric acid extracted when the sample was treated in the undried condition. Dehy- dration of hydrated phosphates might possibly account for this ; but whatever was the cause, the point is one of great importance, and I regret that I have been so far unable to pursue my investigations in regard to it. I think the scheme of the committee would also be greatly improved if it included a recommendation that the amount of humus should be deduced from the determi- nations of the organic carbon. The loss on ignition, after drying at 100" C., can rarely correspond to the amount of humus, for it includes the combined water of clay, etc. In a soil which I examined recently I found only about 0-4 per cent. of organic carbon and no carbonates at all, and yet this sample lost nearly 5 per cent. on ignition after drying. I n the course of certain other experiments, I found that, as a rule, the organic carbon corresponds to about half the weight of the humus present ; but the relation is probably not constant, and the matter requires further investigation.