LIBRARY OF THE UNIVERSITY OF CALIFORNIA. Class LECTURES ON SOME 0V THE PHYSICAL PROPERTIES OF SOIL HENRY FROWDE, M.A. PUBLISHER TO THE UNIVERSITY OF OXFORD LONDON, EDINBURGH, AND NEW YORK LECTURES ON SOME OF THE PHYSICAL PROPERTIES OF SOIL BY ROBERT WARINGTON, M.A., F.R.S. FORMERLY SIBTHORPIAN PROFESSOR OF RURAL ECONOMY IN THE UNIVERSITY OF OXFORD; EXAMINER IN THE PRINCIPLES OF AGRICULTURE TO THE DEPARTMENT OF SCIENCE AND ART WITH A FRONTISPIECE AT THE CLARENDON PRESS 1900 PRINTED AT THE CLARENDON PRESS BY HORACE HART, M.A. PRINTER TO THE UNIVERSITY PREFACE THE contents of this volume formed the substance of a course of lectures delivered by the author as Sibthorpian Professor of Rural Economy in Michaelmas Term, 1896. The subject of the final chapter did not actually form part of these lectures, though it had been prepared for delivery if time permitted. The whole of the matter has since been carefully revised. The form of lectures has not been adhered to in the present volume ; the language has in fact been greatly condensed. The title of lectures has, however, been retained, as it best expresses the character of the work. This is not a textbook, dealing exhaus- tively with the physical properties of soils ; but lectures, discussing with some fullness particular por- tions of the subject. In these lectures the attempt has been made to treat every subject from an experimental point of view, and a considerable space will be found occupied by accounts of the investigations which appear to have thrown most light upon the subjects discussed. 164161 vi PREFACE The behaviour of actual soils under known conditions has been made as far as possible the foundation of the conclusions drawn. A great mass of results has accumulated from the investigations made in the very numerous Agricul- tural Experiment Stations in Europe and America ; with these results the agricultural teacher is too often unacquainted. His valid excuse is the scattered publication of the reports, and his want of time to correlate the several results recorded. The writer hopes that the publication of these lectures will stimulate others to labour in the abundant harvest field of Experiment Station Reports. It must ever be borne in mind that it is only on the results of experimental investigations that Agricultural Science can be safely built. The reader will probably be surprised that so little is said respecting English soils, and so much respecting the soils of America. The writer heartily wishes that this might have been otherwise. In fact, however, the physical constitution and properties of English soils have as yet not been investigated, save in a very few exceptional cases ; this has been doubtless due to the great lack of investigators and research laboratories in this country. The general properties of soils can of course be equally well illustrated by any well studied examples, but the deficiency of knowledge of our own local soils is nevertheless a very real evil, and must greatly hinder the practical application of general principles. PREFACE vii The following pages will show the author's special indebtedness to the writings of Hilgard, King, Wollny, Mayer, Schloesing, and Lawes and Gilbert. The splendid work of Hilgard upon the soils of Mississippi and California must be regarded as in many respects a typical investigation ; it needs to be repeated in every English county. The frontispiece to the present volume is the por- trait of John Sibthorp, M.D., F.R.S., Sherardian and Regius Professor of Botany in the University of Oxford ; it is copied from the oil painting in the Botanical Library. Dr. Sibthorp was the founder of the Chair of Rural Economy in the University. He died in 1 796, after a short life spent in active research. Should the present volume meet with a favourable reception, it is the author's intention to publish a continuation of the work, in which some of the chief points in the chemistry of soil may be discussed on a similar plan. R. WABINGTON. Harpenden, October, 1899. CONTENTS PAGB INTRODUCTION xi CHAPTER I PHYSICAL CONSTITUTION OF SOIL . CHAPTER II RELATIONS OF SOIL TO WATER 51 CHAPTER III RELATIONS OF SOIL TO WATER (continued) 92 CHAPTER IV RELATIONS OF SOIL TO HEAT 139 CHAPTER V MOVEMENTS OF SALTS IN THE SOIL . . . . . .189 INTRODUCTION THE physical properties of soil, and their bearing upon its fertility, is a subject which has been much neglected by the scientific investigator till quite recent years. The horticulturist and farmer have indeed from the earliest times realized the immense importance of a favourable texture of the soil, if crops were to be successfully grown. The special suitability of certain soils for certain crops, and the striking results which can be produced by skilful tillage, have always formed part of the inherited experience of agricultural art. In works dealing with practical agriculture, tillage operations have always occupied a prominent place. The only early investigation on soil physics is that of Schubler, made more than sixty years ago. His work was comprehensive in its scope, and in general very accurate. At the present day his results still furnish some of the favourite quotations of our agricultural textbooks. The rapid rise of agricultural chemistry in the middle of the present century diverted attention from the importance of such investigations. When it was realized that a chemist could analyse a crop and ascertain the elements of which it was composed, and then analyse the soil and ascertain what xii INTRODUCTION proportion of these elements it contained, it was felt that the causes determining fertility were now at last firmly grasped. The large increase in the supply of artificial manures helped forward these novel views by making their practical application easy. At the present day agricultural chemistry still to a considerable extent monopolizes the field, although a fuller knowledge of the facts of agriculture has not borne out the theory to which we have referred. The part of agricultural science which is most prominently brought before our farmers in rural classes and lectures is still the need and application of artificial manures. There can be no doubt that the neglect of the physical conditions of the soil as a subject of study, and in consequence as a subject of teaching, has done much to hinder the appre- ciation of science by practical men. The experienced farmer knows the overwhelming importance of a proper texture of the soil for the profitable culture of each crop. His scientific teacher has, however, little to say on this subject, while he freely recommends the use of expensive manures which a proper culture of the soil might render unnecessary, and which must fail to yield a profitable return if a favourable physical condition of the soil is absent. The farmer feels that this teaching is out of touch with the experience he has gained on the farm ; he also frequently finds that the plan suggested is not a financial success. He therefore charac- terizes the advice given as 'theoretical/ and concludes that science is not a safe guide for the farmer. A soil may be rich in all the elements of plant food, and yet be quite infertile. If seeds are to germinate in a soil, or if roots are to develop in a healthy and vigorous manner, there must be a suitable soil climate. The conditions as to INTRODUCTION xiii air, moisture, and temperature within the soil are quite as essential for vigorous plant growth as are the corresponding conditions in the atmosphere above. Different crops, and even different varieties of the same crop, demand quite different degrees of moisture or dryness in the soil to bring them to perfection. The most favourable proportion of water varies even in the different stages of a plant's growth. The temperature of the soil required for various purposes is also very different. A knowledge of these facts is quite essential for deciding what crops should be grown upon a particular soil, and what treatment the soil should in each case receive for their successful culture. The texture of the soil also largely determines the availa- bility of the plant food which it contains. The surface presented to the action of the roots is far greater when a soil is in a condition of fine tilth, than when the same soil is consolidated, or contains unbroken clods, The supply of air and water, and the condition as to temperature, also largely determine the intensity and character of the chemical pro- cesses which take place in the soil, and by which plant food may be either produced or destroyed. The movements of salts in a soil are to a large extent determined by physical actions, and on the extent and direction of these movements the effects produced by these salts will greatly depend. The fertilizing action of a saline manure, or the injurious effect produced by the accumulation of salts in alkali lands, is thus largely governed by physical conditions, and these conditions come more or less under the farmer's control when he is fully acquainted with their nature. In most cases a good deal may be done to improve the physical conditions of a soil, and render it more suitable for xiv . INTRODUCTION the production of the desired crop. It is possible to increase the retentive power of a soil for water, and to diminish the evap- oration of the water it already contains. It is equally possible to reduce the amount of water in a soil which is naturally too wet. The surface temperature of the soil can also be increased in spring and summer by lessening its contents of water. Both the consolidation of the soil, and the loosening of its particles till they become a fine powder, are to a con- siderable extent under the farmer's control. By skilful management the salts of alkali lands can be prevented from rising to the surface and becoming a source of mischief. In every case a knowledge of the physical properties of the soil, and of the physical actions which go on within it, places them more or less under our control. The overwhelming importance of the physical conditions of plant growth is perhaps most strikingly seen in the case of sandy soils extremely poor in plant food, which, nevertheless, from their extremely favourable physical condition, and the equally favourable climate of the locality, are soils of high agricultural value. Such an example is furnished by the narrow band of sandy soil in the State of Florida now devoted to the cultivation of pine-apples. This sand is almost entirely destitute of plant food, yet it responds so abundantly to the capital invested in it that the planted land has become worth from ^loo to ^500 per acre. It is clear then that we must not judge of the value of a soil by the result of its chemical analysis ; we must take its physical properties also largely into account. Indeed, in a majority of cases, the physical properties and climate will do more to determine its fertility than its chemical composition. INTRODUCTION xv In recent years much has been done both in Germany and in the United States of America to increase our knowledge of the physical properties of soils. Since 1878, a periodical edited by E. Wollny, Forschungen auf dem Gebiete der Agriculturphysik, has been devoted to the publication of papers on this subject. In America the study of the soils characteristic of several States and districts has been ener- getically carried on by Hilgard, and with a more limited scope by King. Since 1893, the subject has been taken up by the U. S. Department of Agriculture, at first in connexion with the Weather Bureau, and since 1895 under a separate Division of Agricultural Soils, superintended by Professor M. Whitney. The inquiry which was at first local has thus grown to be one of national application. The facts available for a discussion of the subject have thus become very numerous. Some of the investigations made, and some of the conclusions arrived at, will be found in the following pages. PHYSICAL PROPERTIES OF SOIL CHAPTER I PHYSICAL CONSTITUTION OF SOIL Physical Constitution of Soil — Methods of Mechanical Analysis — Relation of Physical Constitution to Fertility — Tenacity of Soil — Cementing Materials in Soil— Coagulation of Clay — Shrinkage on Drying — Nature and Origin of Tilth— Specific Gravity of Soil— Weight of Soil per Acre— Colour of Soil— Odour of Soil. Physical Constitution. A soil is a mass of solid particles, differing in their size, shape, and nature. In approaching the subject of the physical constitution of soil it is simplest to take, in the first instance, the case of an ideal soil, consisting of uniform particles, having all the same size, shape, and nature. In a system composed of solid spheres the number of particles in any given volume will depend firstly on the size of the particles, and secondly on their mode of packing. If the spheres are all of the same diameter, there will be a closest and a loosest mode of packing, in each of which every particle is in contact with all those surrounding it ; such arrangements are shown in Fig. I . In both these systems of spheres there is clearly a consider- able proportion of empty space between the particles, which only touch each other at certain points. With the closest packing the interspaces form 25-95 per cent, of the total *> B 2 PHYSICAL CONSTITUTION OF SOIL volume of the mass ; with the loosest packing the interspaces are 47-64 per cent, of the total volume. The proportion of the volume occupied by interspaces, though varying so much with the mode of packing, is quite independent of the size of the particles, so long as all the particles are of one size. Thus, if a single spherical marble, one inch in diameter, be placed in a square box having an internal capacity of one cubic inch, the unoccupied space will be 47*64 per cent, of a cubic inch. If the same box is now FlGUEE I. filled with one million marbles, each of T^jth inch in diameter, the packing being of the loosest description already referred to, the unoccupied space will again be 47-64 per cent, of a cubic inch x. In the second case the interspaces are indi- vidually far smaller than in the first case, but their united volume forms the same proportion of the whole. This is an important fact to bear in mind. The proportion of inter- spaces in a soil plainly determines both the volume of- air which the soil will contain when dry, and the volume of water which it will hold when fully saturated. In a coarse sand, and in a fine clay, the maximum capacity for air and water will be the same, if each is composed of uniform particles packed in the same way. 1 This, and several other illustrations in this chapter, are taken from Professor King's very interesting book The Soil. PHYSICAL CONSTITUTION OF SOIL 3 If the particles composing a soil are not all of one size, but are a mixture of large and small, then the proportion of interspaces in the whole volume is diminished, the small particles fitting in between the large ones. If in the system of solid spheres with closest packing, a second set of spheres is introduced, exactly fitting into the spaces between the larger ones, the proportion of the interspaces will be reduced from 25-95 to 6-76 per cent, of the total volume. If the process is repeated by the introduction of a third set of still smaller spheres, exactly fitting the remaining spaces, the proportion of interspace will fall to 1-76 per cent, of the total volume (Soyka, Wollnys Forsch. der Agrikulturphysik, 1885, i). Any mixture of particles of different size — and all soils consist more or less of such mixtures — tends therefore considerably to reduce the space available for air or water. If, on the other hand, the particles composing a soil are themselves porous, the volume of the interspaces may be considerably increased. It is indeed obvious, that if spheres of sponge were substituted for the solid spheres in the systems represented in Figure I, the proportion of unoccupied space in the total volume would become very much greater. The particles of quartz sand in a soil are non-porous, but particles of limestone are generally porous, and the particles of decayed vegetable matter (humus) are highly porous. A fertile soil also largely consists of compound particles, made up of fragments of various sizes, held loosely together; such particles act as large particles in the soil, but are themselves porous. The proportion of empty space in a mass of soil may also be greatly increased by appropriate tillage operations, the object of which is to bring the soil into a state of loose powder, thus increasing its bulk. The condition of a fertile B 2 4 PHYSICAL CONSTITUTION OF SOIL soil, rich in humus, and possessing a good tilth, is thus one attended with a large proportion of interspaces. We shall go into further details on these points when we treat of the capacity of soils for water. Another important fact is plainly taught by the system of solid spheres which we have taken as the simplest example of a soil. Although the size of the spheres is without influence on the proportion of the interspaces, it has an enormous influence on the extent of surface which the spheres present. In the case of the spherical particles contained in one cubic inch, we have the following variations in surface brought about by alterations in the size of the spheres. Number of spheres in Diameter of each Total surface of spheres i cubic inch. sphere. contained in i cubic inch. One ...... ... 1 inch ... 3-1416 square inches A thousand ...... TV inch ... 31416 „ A million ......... T^ inch ... 314-16 „ A thousand millions ... iifov inch ... 3141-6 ,t A million millions ... -- inch ... 31416 A solid mass of matter, reduced to particles of TtfiI>GOO00>OOO IQ -* Oi CO rH CO GO l£> >O a -3 rl r- ^O "^ C1! rH T— H qo»ocorHTHOooqoo c6rH6666666666 i i i i i i i i i i i i qiocoTHr-iqqqqqqq TH66666666666 RELATION TO FERTILITY 19 analysis shows it to contain a good deal more clay than No. i. Hilgard points out, however, that the heaviness of a soil is largely determined by the sum of the three finest constituents, the influence of which is further modified by the presence of coarse sand, lime, and ferric oxide. In the present soil 106 per cent, of ferric oxide, and 0-8 per cent, of lime were present, with some sand ; the effect of these constituents was to reduce the tenacity of the clay. The loam subsoil which stands next in the table represents a first-class upland cotton soil. It is pretty easily tilled, but suffers much from swelling in alternate frost and thaw, and from denudation during heavy rain, owing to its being not sufficiently pervious to water. The proportion of clay is here considerably reduced, but the soil is composed almost entirely of the finer constituents. The sandy loam soil and subsoil belong to the drift period ; they are characterized by the growth of the long-leaved pine. The soil is very light and easily tilled. We have come now to soils containing much less clay, and a considerable pro- portion of coarse sand. It will be remarked that the coarser constituents preponderate in the surface soil, and the finer constituents in the subsoil. This is a general fact in all soils from districts having a considerable rainfall; in the arid regions of the United States this difference is not observed. The preponderance of coarser particles in the surface soil is due to the gradual removal of the finest particles from the surface by rain water. Further illustrations on this point will be found on pp. 45-9. The river deposit is a light, very porous soil, of great fertility, recently deposited by the Mississippi. It is largely composed of silt and very fine sand, the principal portion of C 2 20 PHYSICAL CONSTITUTION OF SOIL its constituents having almost the same hydraulic value. This deposition in one place of particles of a similar size is characteristic of the soils formed by rivers. The sandy soil is from the uplands of Mississippi. It is of considerable fertility, but possesses so little cohesion that it loses its finer constituents in a high wind when left unpro- tected. Unlike the soils previously noticed, it is mainly composed of coarse sand. A further very instructive example of the results obtained by a detailed mechanical analysis of soil is furnished by the analyses of certain typical subsoil's in Maryland, published by Whitney (U.S. Weather Bureau, Bulletin 4 ; Wiley, op. cit., 249). The separation in this case is into fewer groups than those in Hilgard's analyses. TABLE II PHYSICAL ANALYSES OF MARYLAND SUBSOILS (WHITNEY) Diameter of Particles mm. 1 Early Market Garden. 2 Market Garden. 3 Tobacco Land. 4 Wheat Land. 5 Wheat and Grass. t> Grass and Wheat. Fine Gravel . . 1.0 -2.0 0.49 0.04 1-53 1.34 Coarse Sand . . 0.5 -1.0 4.96 1.97 5-67 0-40 0.23 0.33 Medium Sand . 0.25 -0-50 40.19 28-64 13.25 0.57 1.29 1.08 Fine Sand . . 0.10 -0.25 27.59 39.68 8-39 22-64 4.03 1.02 Very fine Sand . Silt .... 0.05 -0.10 0-01 -0.05 12.10 7.74 11.43 4-95 14.95 28.86 30-55 13.98 11.57 38.97 6.94 29.05 Fine Silt . . . 0.005-0-01 2.23 2.02 7.84 4-08 8-84 11.03 Clay .... -0.005 4.40 8.79 14-55 21.98 32.70 43.44 99.70 97.52 95-04 94.20 97.63 94.23 The early market-garden soil1 is a light yellow sand, 1 In American writings, market-garden soils are spoken of as ' truck land.' RELATION TO FERTILITY 21 having little power of retaining water ; it is therefore char- acteristically warm and dry. With abundant dressings of farmyard manure, it produces spring-sown garden vegetables about ten days earlier than any other soil in the State. It is the coarsest soil in the series, and contains nearly 73 per cent, of coarse to fine sand. The market-garden soil is of much the same character as the preceding, but it contains more clay, and the sand is of a finer description. Its power of holding water is greater, and it produces larger crops of vegetables than the previous soil, and is much superior to it for small fruits, peaches, and for autumn-sown crops ; but the produce of spring-sown crops matures later than on the first soil. The tobacco lands of Southern Maryland contain 10-20 per cent, of clay. The lighter soils yield the smallest crop, but the finest tobacco. Wheat is grown on tobacco land as part of the rotation, but the soil is too light for wheat to be made the principal crop. The wheat land represents the lightest soil on which wheat can be profitably cultivated in the climate of Maryland. The soil is too light for permanent meadow or pasture, and too heavy for the best quality of tobacco. Market-garden crops are late in coming to maturity on this soil. The wheat and grass land represents the heavier wheat soils ; it is considerably more productive than the preceding, and is sufficiently retentive of water to make good grass land. The grass and wheat land is an example of a still heavier soil lying on a limestone formation ; it possesses considerable fertility. This series of typical soils is most instructive ; their com- position ranges from soils consisting chiefly of coarse particles UNIVERSITY CF 22 PHYSICAL CONSTITUTION OF SOIL to those in which the finest groups largely preponderate. With these differences in physical constitution, the agricultural value of the soils, and their suitability for the growth of different crops are plainly connected. We could hardly have a better illustration of the great influence of physical structure, and of the extent to which this can be revealed by the methods of mechanical analysis. The alluvial soils, which cover large areas of the central and western States of America, are not represented in the above table. Ancient alluvial soils, often described as Loess l, have, like their modern representatives (see River Deposit, Table I)} a very simple constitution, the currents of water which brought them to their present resting place having deposited in one spot particles of similar hydraulic value. The Loess soils of Illinois and Nebraska contain according to Whitney (Soils, Bulletin 5, p. 13) from 50 to 70 per cent, of silt, the particles of which are mostly -01—05 mm. diameter. The so-called ' Plains Marl ' of the same district in America is a soil of apparently similar origin, but composed of rather coarser particles ; the samples examined by Whitney contained 7 1-75 per cent, of very fine sand, diameter of particles '05—10 mm. Both these kinds of soil contain but little clay, yet their water-holding power is considerable, and they make good wheat land. We shall see presently (p. 99) that it is in soils of this class that the capillary movement of water takes place to the most beneficial extent. By assuming an average diameter for the particles in each of the separated groups, and knowing their specific gravity, 1 The typical Loess of the Rhine is an extensive deposit apparently formed at the close of the glacial period, when the rivers in the northern hemisphere greatly exceeded their present bounds. The Loess of America has probably a similar origin. TENACITY OF SOIL 23 it is possible to calculate approximately the number of particles contained in any given weight of soil, and to estimate their total surface area ; this has been done by Whitney in the case of the soils mentioned in Table II. The number of parti- cles in one gram of dry soil, and their surface area were as follows. Number of Particles Surface Area, of in one gram. Particles. Soil 1 ... 1958 millions . . . 760 square centimetres „ 2 ... 3955 „ ... 1008 „ „ 3 ... 6786 „ ... 1902 „ ,, ^ ... 10229 „ ... 2493 „ „ 5 ... 14736 „ ... 3593 „ „ 6 ... 19638 „ ... 4575 „ The average diameter assumed for the particles of clay is of course very uncertain. The above calculation assumes that all the ultimate particles in the soil are free, and not arranged so as to form compound particles, the existence of which would necessarily dimmish both the number of particles and their available surface area. The figures given above are thus probably maxima, which are not actually reached in the respective instances. The calculation, however, shows in a striking manner the character- istic differences which exist between soils composed of coarse and those composed of fine particles. The Tenacity of Soil. A fertile soil must possess a sufficient solidity, one of its functions being to afford a firm support to the plant, and enable it to withstand the force of wind and rain. It must not. on the other hand, offer too great an obstacle to the spread of roots. An open texture of soil is especially needed during the early growth of a seedling plant. A good arable soil should also allow of the easy use of tillage implements, and should break down readily when they are skilfully employed. 24 PHYSICAL CONSTITUTION OF SOIL A farmer is accustomed to classify soils as light and heavy ; this language is to be understood as referring to the resistance which they offer to tillage. A heavy soil is one of great tenacity, the particles of which are held together by a strong cohesive force. When such a soil is dug, or ploughed, great strength is required to perform the work, and the soil is therefore said to behave as if it were heavy. This resistance to the application of mechanical force is due to the cohesion and not to the weight of the soil, the weight of a cubic foot of wet clay is indeed much less than the weight of a cubic foot of sand. A considerable degree of stability may be obtained when the individual particles of a soil are very large and heavy, as in the case of a coarse gravel ; in such cases the stability is due to the weight of the particles, and not to any appreci- able extent to the cohesion between them. As the particles diminish in size, they become more easily moved, and we obtain blowing sands, such as are seen on our own Norfolk coast, and which occur to a serious extent in other countries. On the other hand, a decrease in the size of particles is attended with a large increase in their total surface, and sand exhibits a considerable degree of coherence when wet, though scarcely any when in a dry state. When the particles become extremely fine, corresponding in fact to the groups described as silt in the physical analyses of soil already given, the cohesion of the mass when in a wet state is very similar to that exhibited by clay ; the wet soil is in fact a sticky mud. Soils of this character are often spoken of as clay soils ; their true nature becomes revealed when they dry, as they then easily fall to powder. It is important to note that the coherence of silt, sand, CEMENTING MATERIALS IN SOIL 25 chalk, and humus is greatest when they hold sufficient water to fill up all the finer spaces between their particles. As these substances dry, they lose to a great extent the coherence they possessed when wet. Clay, on the other hand, increases greatly in coherence as it dries, and finally becomes a hard, solid substance. These are fundamental facts to bear in mind when conducting tillage operations on a farm. Cementing Materials. We have seen that as the particles diminish in size, and their total surface increases, there is a marked increase in the cohesion of the mass. The difference in the size of the particles is not, however, the only circum- stance which determines the different degrees of cohesion observed in different soils. Soils, in fact, contain various cementing materials, the presence of which has a marked influence on their tenacity. The principal of these cementing materials is clay. We have already mentioned (p. 16) that ordinary clay consists of extremely fine particles held together by a small proportion of a colloid body. Its constitution thus resembles that of putty, in which the particles of whiting are united by means of linseed oil. In typical clay (kaolin, A12O3 2SiO2 2tL,0), derived from the decomposition of felspar, the whole substance has practically the same chemical composition, the various sediments into which it may be divided all containing the same percentages of silica and alumina. The small portion of the clay which possesses a colloid character is probably, however, more hydrated than the rest. The clays ordinarily met with in soil are not, however, chemically homo- geneous, the fine sand which they contain commonly consists of quartz particles, though it may at times have a different origin. Thus in marls we have a clay in which the sandy 26 PHYSICAL CONSTITUTION OF SOIL element is largely composed of calcium carbonate. The true colloid clay is always an aluminium silicate. The character of a natural clay is largely determined by the size of the particles which form its chief bulk. Hilgard remarks that the groups of coarser particles, numbered 1-9 in Table I, all tend to diminish the tenacity of a soil, while Groups 11-13 increase it. The same amount of true clay will thus generally produce a far more tenacious soil when associated with fine particles than when associated with coarse ones. Curious cases sometimes arise of ' putty soils/ drying to extreme hardness, which contain much coarse sand, while extremely little clay is shown by Hilgard's method of analysis. In such soils there is always sufficient silt to fit in between the coarse particles of sand. We must also bear in mind that the cementing power of clay is confined to the small amount of colloid matter present, and that the proportion of this is not shown by the analysis. Though the tenacity of heavy soils is largely due to the clay which they contain, it is by no means necessarily in proportion to the amount of clay present. Not only has the fineness or coarseness of the other soil constituents an influence on the cohesion of the mass, the nature of these constituents may also have a considerable effect. The presence of lime and of humus in a clay soil may diminish its tenacity very considerably ; ferric oxide, according to Hilgard, also acts in the same direction. For these substances to affect the tenacity of the soil they must be thoroughly distributed throughout it. Aggregations of oxide of iron or of carbonate of calcium occurring in a soil will scarcely affect its general physical texture. We must always remember that the dis- tribution or aggregation of the constituents in a soil is not CEMENTING MATERIALS IN SOIL 27 shown by a chemical analysis. The physical properties of a soil cannot safely be inferred from its chemical composition. We shall presently have to refer to a variety of other conditions which affect the character of clay soils. Besides clay, soil contains other colloid bodies which help to bind the particles together. Humic acid is well known as a colloid substance ; when combined with lime it possesses considerable cementing power. Schloesing prepares calcium humate by first extracting a soil with dilute hydrochloric acid -to remove the bases with which the humic acid is combined, and then, after washing on a filter, extracting with ammonia. To the dark-coloured solution of ammonium humate thus obtained, hydrochloric acid is cautiously added with constant stirring till a precipitate begins to form. A solution of calcium chloride is then added, and the precipitate of calcium humate collected. Schloesing determined the amount of freshly prepared calcium humate required to give a suitable coherence to a siliceous sand, a comparative experi- ment being made at the same time with a good plastic clay. He found that i per cent, of humic acid in the form of calcium humate had as great a cementing power as n per cent, of plastic clay. If however the humate is thoroughly dried, and then remoistened, it will be found to have lost its cementing power, while the cementing power of clay remains unaltered by this treatment. Humus is well known as one of the most effective materials for improving the physical condition of sandy soils ; it acts partly by giving cohesion to the particles, but still more by increasing the power of the soil to retain water. We have already mentioned that humus diminishes the tenacity of clay, an action which is plainly the reverse of 28 PHYSICAL CONSTITUTION OF SOIL that which it exhibits when mixed with sand. The humus of soils is made up of very various matters, consisting as it does of the remains of decayed plant tissue, and of its products in various stages of decomposition. Not only do the coarser parts of the humus tend to lighten a clay with which they are mixed, the colloid products already mentioned act apparently in the same direction. Schloesing mixed with clay 2, 4, and 6 per cent, of calcium hamate, and found that with each increase in the proportion of humate the clay became more pulverulent on drying. It is not difficult to conceive possible explanations of this action of humus on clay, but experimental proof is as yet wanting. Hydrated ferric oxide is another colloid substance which in sandy soils undoubtedly plays the part of a cementing agent. In rocks and soils of the Red Sandstone formations its influence is plainly marked. Perhaps however the most obvious example of the cementing action of ferric oxide is afforded by the formation of iron-pan in moor soils. In moor soils, and especially in those covered by heather, the iron has been dissolved out of the surface soil by the action of the humic acids, the sand at the surface being left remarkably white. The iron passes in solution into the subsoil, where it is reprecipitated, with the result that the sand at a certain depth is cemented together, and an iron-stone produced l. 1 The precipitation of the iron oxide may be brought about by contact with calcium carbonate contained in the subsoil ; or, possibly, the whole action of solution and precipitation is determined by alterations in the condition of the soil. In autumn and early winter, with a soil saturated with water, the iron may enter into solution as a ferrous salt, and be carried below ; and in summer time, in a dry and aerated soil, it may be precipitated as ferric oxide. The iron-pan would then be formed at the line at which oxidation chiefly occurred. CEMENTING MATERIALS IN SOIL 29 Like humus, the hydrated ferric oxide diminishes the tenacity of clay, while increasing that of sand. Calcium carbonate is one of the commonest of the cementing materials occurring in rocks. When deposited from its solution in carbonic acid by the escape of this gas, it encrusts and unites the particles on which it is precipitated. This action is familiar to us in the case of petrifying springs. Calcium carbonate is a common ingredient in soils, and is usually the most abundant of the solid constituents held in solution in soil water. Under the varying conditions of the supply of water and carbonic acid, and with alterations in temperature; calcium carbonate will at times enter into solution in the soil, and at other times be redeposited. By the action of rain and vegetation a gradual removal of carbonate of calcium from the surface soil is generally in progress. In districts in America having a deficient rainfall, a hard pan frequently forms in the subsoil at the depth to which the drainage water usually penetrates ; the cementing matter in this pan is carbonate of calcium. The lime-pan, which sometimes forms in English soils which have been dressed with quicklime, is probably due to the conversion into carbonate in the subsoil of the calcium hydrate carried down in the drainage water. The effect of calcium carbonate in increasing the coherence of surface soils is most plainly seen after a dressing of rnarl or chalk has been applied to a sandy soil. Calcium carbonate is a crystalline and not a colloid body. As already mentioned, it diminishes the tenacity of clay. Soils usually contain various hydrated silicates, and some- times hydrated alumina ; these colloid substances are doubtless 30 PHYSICAL CONSTITUTION OF SOIL not without influence on the cohesion of the soil particles, but nothing very definite can be said on the subject. Coagulation of Clay. The tenacity of a soil containing clay is greatly influenced by the condition of the colloid clay which it contains. If this jelly-like matter is in its fully swelled condition, the soil exhibits its maximum stickiness and is perhaps quite impervious to water ; while if this jelly is in a shrunk, coagulated state, the same soil may be pervious to water, and capable of successful tillage. In the case of a natural soil the facts are complicated by the circumstance that the permeability of a clay soil to water, and the production of a good tilth, largely depend on the formation and maintenance of compound particles, and the conditions under which compound particles are produced are sometimes also favourable to the coagulation of clay. The conditions which destroy compound particles are also some- times those which bring about the expansion or diffusion of the colloid clay. We may at present state, that a change in the clay from the coagulated to the diffused condition is necessarily destructive to compound particles in all cases in which the colloid clay is an essential constituent of such particles; but the converse is not necessarily true, for it is quite possible to destroy compound particles without affecting the coagulated condition of the clay. The phenomena presented by soils have thus in some cases a mixed origin ; in our attempt to understand them we must consider the causes separately. The behaviour of colloid clay is best studied when we have it diffused in water. Pure clay, when mixed with distilled water, remains permanently suspended in the same, however long the mixture may remain at rest. The addition of COAGULATION OF CLAY 31 various substances to the water will however speedily bring about the coagulation of the clay, which then completely separates from the water as a flocculent precipitate. If this precipitate is collected, and the precipitating agent separated from it, the clay will be found to have regained its power of permanent suspension in water. This behaviour of colloid clay distinguishes it sharply from the various extremely fine particles with which it is associated, and which are so difficult to separate from it. If a clay containing fine silt is diffused through water and the clay then coagulated, the whole of the suspended matter falls together, the colloid clay carrying the silt down with it. If the silt is separated, and purified from clay by the methods already described, and is then diffused in water alone, it does not exhibit the phenomena of coagulation when treated with the reagents which precipitate true clay. The substances which determine the coagulation of clay when diffused in water are very numerous. Acidification with a mineral acid is very effective, and a great number of salts produce this effect. According to Schloesing, who has made a special study of these phenomena (Chimie agricole, 62), lime, either as lime water or as salts of calcium, is especially effective, 02 gram of lime being sufficient to cause a speedy precipitation in a litre of clay water. Magnesia, according to the same authority, is nearly as effective as lime. The salts of potassium are less effective, and must be used in five times greater quantity than the corresponding calcium salts. The salts of sodium are still less active ; common salt is however a very effective precipitant of clay if used in sufficient quantity. Sachsse and Becker (Landivirth. Versuchs-Stationen, xliii. 1 5) 32 PHYSICAL CONSTITUTION OF SOIL state that carbonic acid is an effective precipitant of pure clay. Calcium hydrate they found far more effective than any other calcium salt. Pure clay, when precipitated by lime water, did not apparently carry down lime with it. Monocalcium phosphate was very effective as a precipitant. Sulphate and chloride of ammonium were fair precipitants, as also was chloride of sodium. Nitrate of sodium had very little precipitating power. Alkaline substances appear generally to favour the diffusion of clay in water, and to destroy flocculation when it has oc- curred. Ammonia and sodium carbonate possess this property. The facts now mentioned as to the coagulation of clay have a wide-reaching application. Schloesing points out that the clearness or turbidity of river water depends essentially on the proportion of lime present. For a stream to be capable of depositing the clay brought into it, the water must contain 70-80 mgrams. of lime per litre. Glacier streams are always turbid, owing to the purity of the water. The waters of the Loire and Garonne are turbid, containing respectively 27 and 36 mgrams. of lime per litre. The Rhone clarifies slowly, the lime and magnesia amounting to 68 mgrams. per litre. The Seine clarifies quickly, the lime being 104 mgrams. per litre. The immediate precipitation of river mud on contact with sea water, giving rise to the formation of banks, bars, and deltas, is well known ; the principal cause of this precipitation is doubtless the coagulation of the clay on mixing with a strong saline solution. In soils the same class of facts may be readily observed. Jn the laboratory, a clay soil may be washed on a filter with weak acid, or with water containing lime, but if distilled water is employed the washings become turbid, and the per- COAGULATION OF CLAY 33 meability of the soil to water is much diminished. If a treatment with acid has preceded the washing with distilled water, the results just mentioned will be especially apparent. One reason of the accumulation of clay in a subsoil is doubtless the washing out of the clay in the surface soil by rain water. The amelioration of clay soils by dressings of lime or chalk is a well-known practice ; the effect of such dressings is most marked, the clay losing much of its stickiness, becoming more friable, more pervious to water, and being more easily cul- tivated. A farmer once told the writer that dressing his heavy land with chalk had enabled him to plough with two horses instead of with three as formerly. Sachsse and Becker filled several wide glass tubes with powdered kaolin and with a powdered heavy loam ; some of the tubes were without lime, and some had 2 per cent, of quicklime intimately mixed with the soil. Water was then poured on the surface. Where no lime had been added percolation did not occur, though a considerable head of water lay on the surface. With lime, percolation proceeded regularly in both cases. The kaolin tube without lime burst from the excessive swelling of the clay. It may be laid down as a general rule that clay soils are friable, and permeable to water, only when the colloid clay is in the coagulated condition. The action of the various salts used in agriculture upon a clay soil demands much further study. There appears some evidence that salts which determine the coagulation of clay when added to clay diffused in water may nevertheless be capable of disintegrating the particles of a clay soil, and liberating the clay, when used as a strong solution, or allowed to crystallize in the soil. On some heavy soils a top-dressing D 34 PHYSICAL CONSTITUTION OF SOIL of nitrate of sodium produces a marked effect. After heavy rain the water is observed to stand on such soil in puddles, and after the water has disappeared the surface of the soil is seen to be white and glistening. The compound particles of the soil have in fact been disintegrated, a layer of fine sand remains on the surface, and the clay lies under it in a puddled condition. This effect of nitrate of sodium on clay soils results in a sticky condition which greatly increases the difficulty of tillage. The addition of superphosphate to the nitrate would probably do much to prevent this evil, as the lime salts of the superphosphate would tend to preserve the coagulated condition of the clay. The action of the salts present in the surface soil of alkali- lands is also very marked. The principal salts in the alkali- lands of India and California are sodium sulphate and carbonate ; the presence of the latter is particularly harmful. Where the carbonate comes to the surface the wetting of the soil by rain occasions a depression in the land ; the clay becomes puddled and impervious to water, and the soil finally dries into sheets of hard pan. This evil admits of cure by treatment with gypsum, which converts the sodium carbonate into the less injurious sodium sulphate. Sodium carbonate has been already mentioned as a salt favouring the diffusion of clay. Some agriculturists in England speak of the formation of a pan in the subsoil in certain cases in which common salt has been long used as a manure. In the absence of accurate information, it seems probable that these so-called ' salt-pans ' had their origin in the washing downwards of fine clay into the subsoil due to the disintegration of the surface particles by the salt. SHRINKAGE ON DRYING 35 Frost is an effective agent in bringing about the shrinkage of colloid clay. If water containing colloid clay in suspension is slowly frozen, the water separates as clear ice, and the clay is concentrated in the unfrozen liquid, or finally separated as a solid film. The same action takes place in a soil when it freezes ; the drainage waters obtained during a thaw are always remarkable for their brightness. Shrinkage on Drying. The different constituents of soil behave very differently on drying. It is a well-known fact in the arts that wet sand suffers little change of volume when dried ; sand is thus always employed when making moulds for casting metals. The clay used by the potter is equally well known to undergo a considerable contraction on drying ; all articles of pottery have therefore to be made considerably larger than the size finally in- tended. Schiibler carefully determined the amount of contraction suffered by various soil constituents during drying. Cubes made of wet siliceous or calcareous sand did not change in volume on drying. The purest clay available, when treated in the same way, showed a contraction of 18-3 per cent, of its volume ; a sandy clay showed a contraction of 6 per cent. The greatest contraction on drying was exhibited by humus, the amount being 20 per cent, of its volume; the humus employed was apparently obtained from the centre of a decayed tree. The various soils he tried varied in their rate of contraction according to the proportions of humus and clay which they contained. An arable soil shrank to the extent of 12 per cent., and a garden soil 14-9 per cent, on drying. From these results we see that contraction on drying is largely determined by the presence of colloid matter in D 3 36 PHYSICAL CONSTITUTION OF SOIL the soil1; that a jelly-like substance should greatly shrink on drying is of course what we should naturally expect. The facts just mentioned are well known in agriculture. The shrinking and cracking exhibited by clay soils in dry weather, and by moor lands rich in humus, are familiar occurrences. The result of this shrinking of the soil is at first harmful, for the roots of plants are torn, and the cracks in the soil allow of a speedier drying of the subsoil. The after results are, in the case of a clay soil, decidedly beneficial. The fissures established in a time of drought afford in future easy lines of drainage. The texture of the clay is also improved to a considerable depth, the drying and remoistening being favourable to the formation of compound particles. Air is also admitted to the subsoil in exceptional quantity, and oxidation in the subsoil is consequently promoted. The swelling of dry clay, or of peat, under the influence of rain is also familiar to us. The overflow of peat bogs in an exceptional rainfall has often led to serious results. Hilgard mentions a 'dry bog' soil which increased 30 per cent, in volume when saturated with water. He also mentions that the alkali soils of America, containing sodium carbonate, contract strongly when the powdered soil is wetted. This interesting fact still requires explanation. Nature and Origin of Tilth. We are now sufficiently acquainted with the constitution and properties of soil to consider the important question of tilth. By tilth we doubtless primarily understand the pulverulent condition of the soil which results from successful tillage. We shall here, however, 1 Schiibler found that precipitated magnesium carbonate lost 15.4 per cent, of its volume on drying. This substance is not usually reckoned as a colloid ; it is however exceptionally light and voluminous. NATURE AND ORIGIN OF TILTH 37 take a general view of the subject, and consider all the various circumstances which serve to bring about this condition. For land to be in a high degree of fertility it is necessary that the surface soil should exhibit a certain friable, crumbly condition, allowing it to fall into powder on the application of gentle tillage, when containing a medium amount of water. The importance of this favourable texture of soil can hardly be overrated. The success or failure of a crop often depends on the character of the seed-bed at the time of sowing. Not only is a good tilth very favourable to the extension of the delicate roots of the seedling plant ; it is also most suitable for ensuring the best conditions of moisture, temperature, and chemical action in the surface soil during the lifetime of the plant, and is thus of the highest importance to fertility. On clay soils the production of a good tilth is especially important; without this there can be no profitable arable culture. The difficulties and delays in attaining this object are the chief obstacle to the effective working of clay land. Tilth is only partially or indirectly the result of tillage operations. In the case of a stiff soil, tillage frequently does nothing more than place the soil in a condition in which the natural forces producing tilth shall exercise their greatest influence. When tillage directly produces tilth, it is simply due to the fact that the soil has already acquired the friable, crumbly condition which is the essential point in the pro- duction of tilth. It is important at starting to grasp this fact. A heavy soil is not reduced to powder by mechanical force brought to bear upon it by means of horses and imple- ments; such a task would indeed be far beyond a farmer's 38 PHYSICAL CONSTITUTION OF SOIL power, and if accomplished would be but a poor imitation of what actually takes place in nature. The favourable texture of soil of which we are speaking may be brought about without any use of implements — it is due to the formation of compound particles in the soil. In an ordinary fertile surface soil the particles are to a large extent associated in groups ; the atoms are, so to speak, built up into molecules ; the fine constituents do not behave as separate entities, they form part of larger compound particles. This condition is highly favourable to fertility. In a soil consisting solely of uniform particles, all of one size and substance, the formation of compound particles could not arise, the force of cohesion being in every direction the same. Compound particles also do not arise in soils composed wholly of coarse sand, consisting of particles with a diameter not less than 0-2 mm., the cohesion being in this case too small. Compound particles are produced in soils composed of mixed particles, including a considerable proportion of the finer sizes, and the formation is greatly aided by the presence of colloid constituents capable of acting the part of cement. A soil may, however, have a constitution very favourable to the formation of compound particles, and yet none may exist ; they are formed under some conditions, and destroyed by others. Let us take as an example a good friable loam. If a spadeful of this is taken when very wet, placed on a hearth or paving stone, and thoroughly beaten, we obtain a cake of very stiff mud, utterly unsuitable for the purposes of vegetation, and which will dry into a hard stony mass. By this treatment the compound particles have been shattered and destroyed, and the soil transformed into an aggregation NATURE AND ORIGIN OF TILTH 39 of ultimate particles, a state in which it exhibits its maximum degree of cohesion. The action of mechanical force in increasing the plasticity and cohesion of clay is well known to the brickmaker and potter ; no clay is used by them till it has gone through the pug-mill, and been ground and beaten in a most thorough manner. If an engineer desires to make clay impervious to water, the same system of mechanical working while it is in a wet state is pursued ; the clay is then said to be puddled. The farmer is equally well acquainted with the effects of mechanical force on wet clay. Ploughing a clay soil in wet weather is known to be disastrous to its condition. The effect of very heavy rain on clay land is also to destroy the surface tilth. In all these cases the work done consists in the destruction of compound particles, and the resolution of the clay into a mass of ultimate particles. Under what conditions are compound particles formed? They arise spontaneously under certain natural conditions. If the lump of beaten loam mentioned above is thrown out upon a garden bed, and left exposed to variable weather for a few months, it will be found completely altered. The lump will be found to have increased considerably in bulk, and, at last, when partially dry, it will fall to pieces when touched. It is now said to have become ' mellow ' ; the condition of tilth has been re-established; it is now once more a mass of large compound particles having only a moderate adhesion to each other. This change has been brought about by natural agents without the aid of tillage. The conversion of ultimate into compound particles will not take place in a dry soil, nor in a very wet one. The soil is in the most favourable condition for producing 4o PHYSICAL CONSTITUTION OF SOIL compound particles when it is rather less than half saturated with water. In this happy condition of moisture there is sufficient freedom of internal movement, without the cohesion of the particles being unduly diminished. When this favour- able condition exists, the formation of compound particles is brought about by the alternate expansion and contraction of the mass. The expansion and contraction of the soil may be simply due to the difference between the day and night temperature, the small alteration in volume being made effectual by its frequent repetition. Much larger changes in volume may, however, be brought about by alternate frost and thaw, and by alternate drying and wetting; and these larger changes produce naturally more speedy results. In freezing, water expands to Jfths of its previous volume, and the expansion of a dry loam after rain will easily exceed this proportion. In each case in which a moist soil is exposed to the conditions assumed above, we have an unequal expansion and contraction taking place in different parts of the mass. The soil consists of particles of very various size and nature, packed in various ways, coated by films of water of different thickness, and with colloid matter irregularly distributed throughout it. In such a mass, the cohesive force being different in different parts, and the internal strains and pressures also unequal, separations take place along the lines of least resistance, and the mass becomes divided into groups of particles, which as the operation progresses become more and more isolated from each other 1. During both frost and 1 It is to be recollected that compound particles are, as already stated, only ormed in soils of mixed constitution. Frost will, of course, disintegrate a uniform moist sand, but in such a case no compound particles, but a powder composed of individual grains, will be produced. SPECIFIC GRAVITY 41 drought, a precipitation of the colloid matter must take place on the solid particles of the soil, and tend greatly to increase the stability of existing groups. The striking effect of winter frosts upon clay land roughly ploughed in autumn is well known to every farmer. The precious tilth thus acquired must be carefully preserved. The following spring cultivation must be done with grubbers and harrows ; to plough the land again would be to bury the fine surface soil obtained by winter exposure. When a farmer ploughs heavy land in summer time he has generally to wait till the clods have thoroughly dried, and been again moistened by rain, before he can obtain the required tilth. These striking changes in the texture of a soil are easily observed, the slower changes are not less important. A perfect cure for an unworkable clay soil is to lay it down to grass. The sod protects the soil from the injurious puddling of the clay by heavy rain. The mischief done by tillage operations ceases. The soil is left to itself, and to the natural influences of the changing seasons. When the land is again ploughed up, it is found that an excellent friable condition of soil has been established. On old grass land the action of worms in increasing the depth of friable soil is important. The worm-casts consist of soil which has passed through the worm's body ; this soil when ejected readily breaks down to a coarse powder. Worms do little to increase the surface soil on arable land, but on grass land the benefits of their action are very con- siderable. Specific Gravity. The comparative weights of equal volumes of various soil constituents are shown in the following table. 42 PHYSICAL CONSTITUTION OF SOIL TABLE III SPECIFIC GRAVITY OF SOIL CONSTITUENTS Water ... 1-00 Dolomite 2-8 - 3-0 Humus 1-2-1-5 Mica 2.8 - 3.2 Clay ... Quartz ... Felspar Talc ... 2.50 2-62 2-5-2.8 2-6-2.7 Hornblende . . . Augite ... Limonite Hematite 2.9-34 3-2 - 3-5 34-4-0 5.1-5-2 Calcite .. 2.75 Of all the solid constituents of soil, hunius is by much the lightest. Clay is a little lighter than quartz sand, and crystallized calcium carbonate (calcite) a little heavier. These are the most common constituents of soil. Dolomite (calcium and magnesium carbonate) is distinctly heavier than calcite. Of the common siliceous minerals, felspar and talc have specific gravities quite similar to quartz and calcite ; mica is distinctly heavier. Hornblende and augite may contain a good deal of iron, and then show a decided increase in specific gravity. Limonite is a natural hydrated ferric oxide; hematite is the same oxide in its anhydrous state. These oxides of iron are the constituents of soil which possess the highest specific gravity. It is clear, from what has been stated, that humus soils will be the lightest, and soils rich in iron the heaviest, if we have regard to their true specific gravity. The true specific gravity of ordinary arable soil is usually about 2-5. We have already seen that a soil is composed of particles which touch each other only in certain points, spaces filled with water or air lying between them ; a cubic foot of dry soil has not therefore the weight which we might assume SPECIFIC GRAVITY 43 from the specific gravity of its constituents, but a weight very much smaller, only a portion of the cubic foot being occupied by solid matter. The weight of a given volume of dry soil, divided by the weight of the same volume of water, is called its * apparent specific gravity.' According to Wollny, the apparent specific gravity of powdered quartz is 1-449, °f clay i -on, of hunius 0-335 ; all these were weighed in an air-dried condition. As one cubic foot of water weighs 62-32 lb., these specific gravities are equivalent to weights of 90-3 lb., 63-0 lb., and 20-9 lb. per cubic foot. Siliceous sand is thus the heaviest constituent of ordinary soils, clay much lighter, and humus far lighter still. According to Wiley, an ordinary arable soil, in good tilth, will have an apparent specific gravity of about J-2, and will consequently weigh 74-8 lb. per cubic foot.. The difference between the apparent specific gravities of quartz sand and clay is seen to be considerably greater than the difference in their true specific gravities ; this arises from the fact that the extremely fine particles of clay lie more loosely, and are much more difficult to pack than the large heavy particles of sand. The difference between the apparent and true specific gravities is still more marked in the case of humus, the proportion of empty space in dry humus being far greater than it is in the case of either clay or sand. The whole of the apparent specific gravities and weights per cubic foot given above relate to dry powdered soil, shaken or pressed together in a vessel of known capacity ; they represent therefore the weight of soils when in a condition of fine tilth, and not their weight when in a consolidated condition in the field. King has, by boring, cut out successive cylinders of the 44 PHYSICAL CONSTITUTION OF SOIL Wisconsin soil to a depth of 6 ft., and determined the weight of dry soil in each foot. One set of determinations will be found in Table IX, p. 70. The first foot of soil, a loam, is seen to weigh 76-8 Ib. per cubic foot; while the fifth foot, a fine sand, weighs 116-7 lb. We have here the lighter weight of the loam, and the heavier weight of the sand, both exag- gerated by the position in which they occur ; the loam being at the surface, and lightened by tillage and vegetation, while the sand is consolidated by the weight of the four feet of soil which lie above it. In a later set of determinations (Wis- consin 8th Rep., 107), the weights vary from 79 Ib. per cubic foot for the surface foot of loam, to 1 1 1 Ib. for the fifth and sixth foot of pure sand. The results furnished by experi- ments at Rothamsted and Woburn will be found on p. 47. Anything which tends to diminish the proportion of empty space in a soil, as the presence of stones, or a considerable variety in the size of the particles, leading to closer packing, will tend also to increase the volume weight ; while anything tending to loosen the texture of the soil will diminish it. Weight of Soil per Acre. It is important for many pur- poses of calculation to be acquainted with the weight of dry soil in a given depth per acre ; to ascertain this fact it is necessary to remove definite volumes of the soil, and to dry and weigh them. The accurate information on this head is not very extensive. The best mode of operation is to drive into the land a short, wide steel tube, sharpened at its lower edge, till the top is level with the surface ; the contents of the tube then furnishes a sample of the soil to the depth represented by the length of the tube. The tube should not have a diameter of less than six inches ; with narrow tubes, the resistance to the WEIGHT OF SOIL PER ACRE 45 passage upward of the soil within the tube is so great that the contents of the tube becomes less than that proper to the depth reached. When the sample of the surface soil has been taken in the manner described, the earth can be dug away around the tube, and when the tube has been emptied it can be again driven down, and a sample of the succeeding depth obtained. At Rothamsted, the soil of the experimental fields has been sampled in this way to a depth of nine feet. The iron or steel frame employed at Rothamsted is square in sec- tion, the sides of the square being six inches, and the depth nine inches. Table IV shows the average weights of soil per acre obtained at Rothamsted, Herts, and at the experimental farm at Woburn, Beds, by the method just described. The weights here given represent soils in their natural con- dition of consolidation. The pasture soil is of course quite undisturbed by tillage, but in the case of the arable soils a part of the upper nine inches has been so disturbed. The Rotham- sted soil is a heavy loam, containing many partially rolled flints ; it rests on a variable subsoil of loam or clay, beneath which is the chalk. The Woburn soil is a light sand. The stones mentioned in the Table were in every case separated by a sieve having quarter-inch meshes. Looking first at the total weight of dry soil per acre, we see that the soil is in every case lightest at the surface, and increases gradually in weight as we descend into the subsoil. This is in part due to the increasing pressure to which each succeeding stratum of soil is subjected, and in part to the result of particular actions tending to make the surface soil loose and porous. The action of rain is to wash out of the surface soil its finest particles, and to carry them into the subsoil ; the surface soil is thus made up of coarser particles, 46 PHYSICAL CONSTITUTION OF SOIL TABLE IV WEIGHT OF SOIL PEK ACRE i. Old Pasture, Rothamsted, Loam with Clay Subsoil Original Dry Soil. Wet Soil. Total. Stones. Fine Soil. Eoots. Ib. Ib. Ib. Ib. Ib. First 9 inches . . 3,294,380 2,328,973 174,091 2,144,470 10,412 Second 9 inches . 3,867,780 3,098,939 353,322 2,744,715 902 Third 9 inches . . 4,091,620 3,273,324 217,515 3,055,501 308 Fourth 9 inches . 4,139,420 3,343,787 280,730 3,063,057 2. Arable Land, Rothamsted } Loam with Clay Subsoil First 9 inches . . 3,288,553 2,919,689 340,656 2,578,634 399 Second 9 inches . 3,668,115 3,044,615 141,861 2,902,682 72 Third 9 inches . . 3,882,285 3,215,285 213,190 3,002,095 Fourth 9 inches . 3,995,723 3,313,563 197,400 3,116,163 3. Arable Land, Woburn, Sandy Soil First 9 inches . . 3,835,104 3,157,448 93,763 3,063,074 611 Second 9 inches . 3,947,640 3,381,804 201,527 3,180,277 Third 9 inches . . 4,046,364 3,462,498 170,443 3,292,055 Fourth 9 inches . 4,014,432 3,501,466 274,239 3,227,227 with wider interspaces, than the subsoil. In the case of arable land, the mechanical action of rain just described is much aided by the loosening of the surface soil by tillage, and this is itself a direct agent in producing a looser and lighter soil at the surface. In heavy soils under arable culture, a so-called ' hard pan ' or ' clay pan ' is often formed immediately under the surface WEIGHT OF SOIL PER ACRE 47 soil by the treading of horses and men in the furrow, and the pressure of the sole of the plough. The formation of this consolidated layer is clearly aided by the accumulation of fine clay immediately beneath the ploughed soil. Such an impervious layer is of course most injurious to the growth of crops, and if formed requires to be broken up by a subsoil plough following the ordinary plough in the same furrow. The difference in the weight of the surface soil and subsoil is most marked in the case of the old pasture land at Rotham- sted ; here the surface nine inches is only three-quarters the weight of the second nine inches. This is an excellent illustra- tion of the lightening of a soil laid down to grass brought about chiefly by the accumulation of vegetable residues in the surface soil. In cases in which the sampling of the subsoil was carried out at Rothamsted to a greater depth than 36 inches, it appeared that below this depth there was little further in- crease in the consolidation and weight of the soil. The sandy soil at Woburn is seen to be, at every depth, of greater weight than the heavy loam at Rothamsted, thus supplying a further illustration of the comparative lightness of clay soils. The apparent specific gravity, and the weight per cubic foot of the dry soils at the surface, and at the depth of 27 to 36 inches were as follows: — Apparent Weight Specific Gravity. per Cubic Foot. Ib. Rothamsted Old Pasture, first 9 inches ... 1-144 ... 71-3 „ „ „ fourth „ ... 1-642 ... 102-3 Rothamsted Arable Land, first 9 inches ... 1-434 ... 89-4 „ „ „ fourth „ ... 1-627 ... 101-4 Woburn Arable Land, first 9 inches ... 1-550 ... 96-6 fourth 1-715 106-9 48 PHYSICAL CONSTITUTION OF SOIL These weights supply excellent examples of the consolida- tion of a soil below the surface. The distribution of the stones in the Rothamsted soils deserves some attention. Their distribution in the subsoil is extremely irregular in different places, but in the surface soil two points are invariably noticed, namely the deficiency of stones at the surface of the old pasture land, and their marked accumulation at the surface of the arable land. These facts admit of explanation. The sod of a pasture is to a considerable extent a new formation, consisting of living and dead vegetable matter which has been added to and has covered the previous soil. On pastures also the burying action of worms is carried on to its fullest extent, the worm-casts of fine earth, brought from the subsoil, covering the stones and causing them apparently to sink beneath the surface. In the case of arable land the circumstances are quite different. When a stony piece of land is dug or ploughed in the autumn, and left exposed to the weather, it will be found in spring time covered with stones. During the winter frosts the surface soil has swollen and become disintegrated, the stones have partially protruded, and have been left surrounded by pulverulent matter. When rain comes, the fine soil is washed away from the stones, which are left exposed on the land, often very much to the farmer's astonishment, who not unfrequently asserts his belief that ' stones grow.' The stones thus accumulated on the surface may be buried again at the next ploughing, but the kind of action here noticed, continued for centuries, tends to a gradual removal of the finest particles from the surface soil, and consequently to an increase in the proportion of stones, where these, as at Rothamsted, consist of flint or other minerals little affected UNIVERSITY OF 'F- '_ r ~ '' *^ COLOUR OF SOIL 49 by weather. On the surface soil the full action of wind and rain, and the disintegration by frost are experienced. It is the finer particles which are first removed, both by mechanical and chemical agents, and the large stones thus gradually become a larger proportion of the whole. The Woburn arable land does not show this accumulation of stones at the surface ; this land was in fact in pasture at no very distant date. In the case of soils derived from an underlying soft rock, say an oolitic limestone, the distribution of stones will be the reverse of that experienced at Rothamsted. The stones, in this case, are easily disintegrated by atmospheric influences, and will therefore be fewer and smaller in the surface soil, and will increase in size and number as we proceed down- wards. Colour of Soil. Soil owes its colour chiefly to two con- stituents— humus, and ferric oxide. Humus alone causes a soil to be grey when dry, but nearly black when wet. Ferric oxide is the colouring matter of all red soils ; the tint varies with the condition of the oxide, and in some cases may be brown or yellowish. The admixture of humus of course modifies the colour. The intensity of the colour is not a certain indication of the proportion of iron or humus present. If the soil consists of large particles, as a coarse sand, a little ferric oxide or humus may strongly affect the colour, owing to the small extent of surface to be coloured ; while a soil consisting of fine particles will need a much larger amount of colouring matter to produce the same tint. The colour of * blue lias/ and probably of other blue clays, is due to finely divided iron sulphide, FeS2. When lias is treated with strong hydrochloric acid the pyrites remain as 50 PHYSICAL CONSTITUTION OF SOIL a heavy black powder, which yields a sublimate of sulphur when dried and heated in a glass tube (Jour. Chem. Soc. 1864,379). Odour of Soil. According to Berthelot and Andre* (Compt. rend., cxii. 598) the characteristic odour of soil is due to a volatile organic substance, which can be distilled in the presence of water at the temperature of 60° C. The chemical nature of this body has not been satisfactorily ascertained ; it is said to belong probably to the aromatic family. CHAPTER II RELATIONS OF SOIL TO WATER Water required by Crops — Deposition of Dew — Hygroscopic Water -^ Maximum Water Capacity of Soils — Optimum Proportion of Water — Power of retaining Water — Percolation. To supply water to the plant is one of the proper functions of the soil. A plant is not constructed with a view to its ab- sorbing water by means of its leaves, or indeed by means of any of its above-ground parts ; water is only absorbed from dew or rain by these parts when the plant is in a wilted condition. With this limited exception the whole of the water required by a plant is taken up from the soil through the roots. The amount of water supplied by the soil is one of the most important factors in determining the luxuriance of plant growth. It is seldom that there is a sufficient supply of water during the whole time of the growth of a crop. In localities receiving an ample amount of sunshine the supply of water from the soil becomes the circumstance which more than any other determines the quantity of the produce. The large increase of produce which results from artificial irriga- tion is well known. Water required by Crops. A large supply of water is 52 RELATIONS OF SOIL TO WATER needful from the soil partly because a living plant contains so large a proportion of water — seldom less than three- quarters of its weight ; but still more because the passage of water through the plant is one of the most important means of plant nutrition. Evaporation from the surface of a living plant, chiefly from the leaves, takes place energetic- ally whenever the amount of heat received from the sun, or the state of dryness of the atmosphere, admits of water being converted into vapour. The evaporation of water from the leaf surface produces an upward stream of water from the roots, by means of which the solution of plant food held by the soil is carried to the leaves, and the substances which it contains made use of for the production of organic matter. Thus the greater the evaporation, the greater is the trans- ference of plant food from the soil to the plant. So plain is the connexion between plant evaporation and plant nutrition, that it has been thought probable that a definite relation exists between the quantity of water evapo- rated and the quantity of organic matter produced. The relation in question appears to be fairly definite so long as certain conditions remain constant, but to vary considerably under a wider range of circumstances. The following table gives the results arrived at by various investigators em- ploying different methods of experiment (Lawes and Gilbert, Jour. Hort. Soc. 1 850 ; Hellriegel, Exp. Station Record, iv. 532 ; King, Rep. Wisconsin Exp. Station, 1 894, 248 ; Wollny, Exp. Station Record, iv. 532). That the conditions of the experiment have had a con- siderable influence on the result is evident from a glance at these figures. Barley, for example, has been experimented with by each investigator, and has under various conditions WATER REQUIRED BY CROPS 53 i required 2,62, 310, 393, and 774 parts of water to produce one part of dry matter. TABLE V WATER EVAPORATED BY GROWING PLANTS FOR ONE PART OF DRY MATTER PRODUCED Results obtained by Lawes and Gilbert. Hellriegel. King. Wollny. Beans . . . 214 Beans . . 262 Maize . . 272 Maize . . 233 Wheat . . Peas . . . Red Clover . 225 235 249 Peas . . Barley . . Red Clover 292 310 330 Barley . . Potatoes . Red Clover 393 423 453 Millet . . 416 Peas . . 447 Sunflower 490 Barley . . 262 Wheat . . Buckwheat 359 371 Peas . . Oats . . 477 557 Buckwheat 646 Oats . . 665 Lupine Rye . . . Oats . . 373 377 402 Barley . .774 Mustard . 843 Rape . . 912 Two circumstances seem especially to influence the relation of water supply to produce — i . The amount of water supplied ; 2. Its richness or poverty in plant food. These two circum- stances are naturally connected. When the soil contains little water, the solution in the soil is of a comparatively con- centrated character. When much rain has fallen, the supply of water in the soil may be ample, but it usually contains very little plant food. The evaporating organs of the plant are constructed so as to act to the greatest advantage under both circumstances. When much water enters the plant, the cells controlling the stomata in the leaves become turgid, and their swelling determines the opening of the stomata as widely as possible, thus increasing the evaporation. When, on the other hand, the supply of water falls short, the cells in question shrink, the stomata are closed, and evaporation is diminished. 54 RELATIONS OF SOIL TO WATER Striking illustrations of the sufficiency of a small supply of water under specially favourable conditions are to be found in the case of some of the soils rich in soluble salts which occur in the semi-arid regions of the western States of North America. Good crops of wheat are here grown with an annual rainfall of 13-18 inches, most of which falls in winter time before the growth of the crop has commenced. The water level in the subsoil is 20-30 feet below the surface. Such a supply of water would prove quite inadequate on the soils of the eastern States. The richness of the saline soils of the western States in soluble plant food will be referred to later (p. 215). The presence of much soluble saline matter in the soil is probably in itself a check to the transpiration of water by the plant. A plant may succeed in reaching perfect maturity with a scanty supply of water, and in this case there will be a relatively large produce for the quantity of water consumed, but a maximum crop will not be obtained in this way. A luxuriant growth demands permanent turgidity of the cells, and in an ordinary climate this condition can only be at- tained by a large supply, and a large evaporation of water. The largest crops can thus only be grown with a luxurious or wasteful consumption of water. King's experiments at Wisconsin were made in barrels sunk in pits, the surface of the soil in the barrel being on the level of the field. The crop grown in the field was the same as that grown in the barrel, so that the experimental crop was surrounded by a similar growth in the same manner as it would be under ordinary cultivation. The water was supplied through a pipe passing to the bottom of the barrel, and the water level in the barrel was permanently maintained at six DEPOSITION OF DEW 55 inches above the bottom. The depth of soil in the barrel was 3 ft. 6 in. With this large and permanent supply of water very luxuriant crops were obtained, far exceeding those pro- duced in the surrounding field. With oats and barley a yield of 10,000 Ib. of dry matter per acre was obtained ; with potatoes 1 2,000 Ib. ; with maize over 20,000 Ib. The water consumed would average about 24 inches per acre ; a quantity of course far exceeding the summer rainfall. Average English crops, yielding about 4,000 Ib. of dry matter per acre, will probably evaporate five or six inches of water during their growth. There is not enough evidence to point out any particular crop as especially wasteful or economical in its consumption of water ; the figures in the table however certainly suggest that the cruciferae should be ranked with those requiring most water, and maize with those requiring least. The sub- ject is -a promising one for further investigation. The principal source of soil water lies of course in the rain, hail, and snow precipitated from the atmosphere ; the amount of this precipitation in any place, and its distribution through the different seasons of the year, are facts which fundamentally affect the fertility of the land. The amount and distribution of the rainfall is however a subject lying beyond the scope of the present lectures. Of the underground water supply, which in some localities is of very considerable agricultural importance, we shall have something to say by-and-by. It will be convenient before proceeding further to notice two sources of water supply to the soil which are of minor im- portance, but which fairly come under the physical properties of soil. Deposition of Dew. If air containing water-vapour is 56 RELATIONS OF SOIL TO WATER cooled till it becomes supersaturated it will deposit water on the solid bodies with which it may be in contact. If a cold body is introduced into moist air the temperature of the air may be so reduced on the surface of the body that the air becomes supersaturated, and water condenses upon the cold surface. In any case the deposition of dew ie greatly favoured by the presence of dust particles on the solid body. A soil must thus gain water from the air whenever the air is sufficiently moist, and the surface of the soil sufficiently cold to occasion precipitation. It more frequently happens that the deposition of dew or hoar-frost takes place on the herbage occupying the surface rather than on the soil itself. Occasions however do arise in our variable climate in which rather considerable amounts of water are condensed upon the surface of arable soil. These occasions occur most frequently towards the end of winter. The soil is still not far above its minimum temperature when the advancing season brings with it a mild moist wind ; under such circumstances a considerable deposi- tion of water may occur on the surface of the cold soil. We obtain evidence of this fact by studying the results furnished by drain-gauges filled with bare soil. In the case of Mr. Greaves' drain-gauge, containing fine gravel, kept free from weeds, the monthly drainage through the soil has exceeded the monthly rainfall twice in December, seven times in January, seven times in February, and three times in March, during fourteen years. As in each of these months some evaporation must have taken place from the gravel, the amount of water con- densed from the air must have been considerably greater than the differences between rainfall and drainage would indicate. The gravel having little power of retaining water, and being an excellent conductor of heat, was especially suited to exhibit HYGROSCOPIC WATER 57 the results due to rapid changes of temperature. The same class of results is shown, but much less frequently, by the drain-gauges filled with loam at Rothamsted. The persistent wetness of land in February is doubtless connected with the frequent condensation of moisture upon the cold soil. Besides the deposition of water upon the surface from the atmosphere, there may be deposition on the surface from vapour rising from the subsoil. This will happen chiefly in the case of well-drained soils, made up of coarse particles, in which a free movement of gases can take place. Move- ments of the soil air will be determined by variations in its temperature and pressure, but the condensation of water- vapour at the surface will of course only happen when the surface is colder than the subsoil. When this is the case a veritable distillation may occur. An action of this kind may possibly explain in part the maintenance of the moist condition of the surface soil during a nearly rainless winter, which is observed in the case of the early market-garden lands of Florida. These soils consist of coarse sand, and hold on an average only 3 per cent, of water, yet early vegetables are successfully grown on them without irrigation during the winter months. The water level is 15-20 ft. below the surface (Soils of Florida, Soils Bull. 13, p. 9). Hygroscopic Water. Quite apart from the power of a cold body to condense water from a moist atmosphere is the power which bodies possess in a varying degree of condensing water on their surface even when at the temperature of the atmo- sphere, and even when the air around them is not saturated with water. Every body in a moist atmosphere has a moist surface, the thickness of the covering film of water depending on the temperature, and on the degree of saturation of the 58 RELATIONS OF SOIL TO WATER atmosphere. Such water is known as hygroscopic water. It is parted with slowly in a perfectly dry atmosphere, or on a sufficient rise in temperature. The amount of hygroscopic water belonging to any given weight of a substance depends on its extent of surface ; a mass of glass becomes far more hygroscopic when reduced to powder. Porous bodies are thus especially hygroscopic. Wood charcoal recently burnt will increase about 20 per cent, in weight in a moist atmosphere. In comparing the power of different bodies to absorb water- vapour, the question is often complicated by the existence of affinities for water which are quite distinct from the purely surface action we have just described. Water- vapour may be absorbed in large quantities by bodies which have a chemical affinity for water, and form hydrates with it, as in the case of calcium chloride, potassium carbonate, and'* many other salts. It may also be absorbed to a considerable extent by colloid bodies, as silk or gelatine ; the water in this case is not held simply at the surface, but penetrates the whole mass. Bemmelen (Rec. Trav. Chim., vii. 37) has made a long inves- tigation on the amount of water held by the dried gelatinous precipitates of silica, ferric oxide, alumina, and other similar colloid bodies, and on the gain or loss of water which they suffer at various temperatures, both in dry and moist air. These precipitates when dried by long exposure to ordinary air contain a large quantity of water. Placed in perfectly dry air (over strong sulphuric acid) a portion of this water is lost l. 1 The precipitated ferric oxide used in some of my own experiments still contained 15-5 per cent, of water after drying over sulphuric acid, and precipitated alumina 33-1 per cent. HYGROSCOPIC WATER 59 Heated to 100° C. still more water is lost. Heating to 300° C. fails however to remove the whole of the water, which can only be certainly expelled at a red heat. If these substances have not been over-heated, a considerable part of the water- lost is regained when they are placed in a moist atmosphere. In the case of an ordinary soil of mixed constitution, the power of absorbing water- vapour depends in part on the extent of surface which it presents, that is, on the fineness of its particles ; but it depends also to a still greater extent on the amount and character of its colloid constituents. Chemists are at present unable to separate with certainty the portions of water held in these various ways ; all we can say is that the water lost in perfectly dry air is that most feebly held, and that the water expelled by heat is more firmly combined in proportion as it requires a higher tempera- ture for its expulsion. We may indeed be certain that water requiring a higher temperature than 100° C. for its expulsion is not hygroscopic water held on the surface of a non-colloid body ; but, on the other hand, we may be equally certain that the water expelled below that temperature has been derived in part from the colloids present. It is important to bear in mind that soil dried at 100° C. is not dry, but that the colloids still hold water; any further loss obtained on ignition is thus not simply due to the combustion of organic matter, but is occasioned in part by a further loss of water by the cla}^ and other silicates, and by the colloid ferric oxide and alumina which may be present. Loss on ignition is thus no measure of the amount of organic matter present. The methods used for determining the hygroscopic water in soils have varied a good deal, and the results obtained 6o RELATIONS OF SOIL TO WATER by different investigators are often not comparable. In some methods the powdered soil is dried over sulphuric acid, or at ioo°C.j and then placed in an atmosphere saturated with water, and the gain in weight determined. Such a method affords low results. In Hilgard's method the air-dried soil is exposed to a saturated atmosphere at about 15° C., weighed, and then dried at 200° C. This method will yield maximum results, more of the colloid water being included. The following table will give a general idea of the absorp- tive power of Mississippi and California soils for water- vapour as determined by Hilgard and Loughridge (Portland Meeting, American Ass. Advanc. Science, 1873; Rep. Agri. Expt. Stations California, 1892-3-4, 70). In the rough classifica- tion here adopted, a sandy soil contains less than 5 per cent, of clay ; a sandy loam 5-10 per cent. ; a loam 10-15 per cent. ; a clay loam 15-20 per cent. ; and a clay soil over 20 per cent, of clay, as determined by Hilgard's method of mechanical analysis (see p. 12). TABLE VI WATER EXPELLED AT 2OO° C. FROM SOILS SATURATED WITH WATER-VAPOUR, PER IOO OF DRY SOIL (HILGARD) Minimum. Maximum. Average. Sandy Soils . . Sandy Loams . . Loams .... 0-79 1-84 2.30 4-18 6.12 9-18 2-59 3-39 5-19 Clay Loams . . Clays 5-06 4-20 10.26 18-60 6-49 10-83 This rough classification plainly shows that the power of absorbing water- vapour is least in the case of sandy soils ; HYGROSCOPIC WATER 61 the coarsest sand was the one absorbing only 0-79 per cent, of water. As the proportion of clay increases the absorptive power for water- vapour rises. In one case the whole soil, and some of the separated portions from its mechanical analysis, were individually tested as to their power of absorbing wa ter- vapour ; the percentages of water expelled at 200° C. were as follows. Whole Soil. Separated Portions. Clay. Hydraulic Value in millimetres per second. < 0-25 ... 0-25 ... 0-50 5-34 ... 17-60 ... 7-96 ... 2-91 ... L73 The various fine constituents and colloids classed as clay had thus by far the greatest power of absorbing water, and the coarsest constituents the least. The connexion between the percentage of true clay in a soil and its absorptive power is not a very close one ; clay, in fact, is not the constituent of soil possessing the highest absorptive power ; the purest clay experimented with (the pipe-clay of Table I) absorbed in fact only 9-09 per cent, of water. The constituents of soil having the highest absorptive power are humus, ferric oxide, and the hydrated silicates decomposable by acids ; if either of these is present in con- siderable" quantity, the soil will have a high absorptive power irrespective of the proportion of clay it may contain. In cultivated soils the hygroscopic power is determined most frequently by the proportion of humus present. The soils examined by Loughridge having the highest absorptive power for water- vapour were some from the Sand- wich Islands; these soils were derived from the decomposition of lava, and were extremely rich in hydrated silicates, in hydrated alumina and ferric oxide, and in some cases humus. 62 RELATIONS OF SOIL TO WATER The percentage of water expelled at 200° C. from the vapour- saturated soils amounted in one case to 37-4 per cent, of the weight of the dry soil. When a substance condenses water from the atmosphere its temperature rises, the water necessarily evolving heat when passing from the gaseous to the liquid state. Stellwaag (Wollny, Forsch. der Agrikulturphysik, v. 210) determined the rise in temperature of various soil constituents, first dried at 1 05° C. and then placed in a saturated atmosphere at 30° C. The changes in temperature observed give an excellent idea of the relative hygroscopic power of the materials experi- mented with. Rise in Temperature in Water-vapour at 30° Quartz Sand r. ... 0°-88 C. Calcium Carbonate (precipitated) ... 1°47 Kaolin 2°.63 Hydrated Ferric Oxide 9°-30 Peat 12^.25 The moistening of a dry body with liquid water, and especially the combination of a colloid body with water, also produces a rise in temperature. At I o° C. the results obtained were as follows. Rise in Temperature when moistened with Liquid Water Quartz Sand 0°.10 C. Calcium Carbonate (precipitated) ... 0°-28 Kaolin 0°-83 Hydrated Ferric Oxide 6°.60 Sachs has taught, and his statement is current in modern textbooks, that plants are able to make use of the hygroscopic water in soils ; and that, in fact, a supply of water- vapour in the air is sufficient to preserve such an amount of water in the HYGROSCOPIC WATER 63 soil as will maintain the plant cells in a turgid condition. It is of course possible that if changes of temperature are allowed, the amount of water condensed by the soil may suffice for the maintenance of a plant ; but it is not true that a soil containing only hygroscopic water is capable of main- taining plant life, and this has been abundantly proved by the investigations of Heinrich, A. Mayer, Liebenberg, and Hellriegel (Jahresb. Agrik. CJiem. 1875-6, 368). Heinrich grew plants in very small boxes till fully developed, and then placed them under conditions of very little evaporation till they began to wilt ; the soil in the box was then mixed, and the proportion of water it contained determined. A variety of soils were employed. A weighed quantity of each soil was also placed in a dry state in a saturated atmosphere till it ceased to gain weight, and the amount of hygroscopic water which the soil could absorb was thus determined1. It was found in every experiment, that when the plants wilted the percentage of water in the soil was still somewhat higher than that proper to hygroscopic water only. The figures given in Table VII show the average results obtained with oats and maize. Experiments made with grasses, with various leguminosae, and with potatoes, gave similar results. The potatoes grown in peat required 41-4 per cent, of water in the moist peat, or 70-8 per 100 of dry peat, to avoid wilting 2. The leguminous 1 The hygroscopic water found would have been higher had the soils been dried at 200° C., but this fact will not alter the conclusion drawn from the experiment, as the amount of water remaining in the soil after the growth.of the plant would in that case have also to be determined at 200° C. 2 In experiments on the growth of crops on peaty land it has been noticed that the peat must contain more than 60 per cent, its weight of water to yield productive crops (Biedermann's Central-Blatt fur Agrikulturchemie, 1885, 297). 64 RELATIONS OF SOIL TO WATER plants required slightly more water than the grasses and cereals. TABLE VII RELATION OF PLANTS TO HYGROSCOPIC WATER (HEINRICH) Water per 100 of Dry Soil. When Plants wilted. Absorbed from Moist Air. Coarse Sandy Soil . Snndy Garden Soil . Fine Humus Sand . 1-5 4-6 6-2 1-15 3-00 3-98 Sandy Loam . . . Calcareous Soil . . Peat 7-8 9-8 49-7 5-74 5-20 42-30 From what has now been said it seems clear that if we desire to know how much water available for plant use a soil contains, we shall arrive at the result best by ascertaining how much water the soil loses when exposed to ordinary air till it ceases to lose weight. Water which can only be expelled by heat is incapable of assimilation by ordinary crops. Although a soil containing only hygroscopic water cannot support a crop, the possession of hygroscopic power is under some conditions of distinct advantage to a soil. Hilgard has observed that in the arid districts of California it is only on soils capable of absorbing 4-8 per cent, of hygroscopic water that crops can successfully resist drought. He thinks that the surface soil is cooled during the extreme heat of the day by the evaporation of a portion of the hygroscopic water, this water being regained during the night either from the atmosphere, or from the vapour rising from the subsoil. Water Capacity of Soils. A soil contains the largest pos- WATER CAPACITY OF SOILS 65 sible quantity of water when all the interspaces between the particles are filled by that liquid, the soil is then perfectly saturated. The quantity of water which a soil may contain when in a saturated condition is thus determined by the volume of its interspaces. The circumstances which determine the proportion of interspaces in a soil have been already men- tioned (p. i), and should be referred to once more before pursuing the subject any further ; the experimental results we are about to describe will serve to illustrate the principles there laid down. The proportion of interspaces in any volume of soil may be ascertained by determining the quantity of water which it holds when perfectly saturated. It may also be ascertained by calculation, if both the apparent specific gravity of the dry soil and the real specific gravity of its mixed constituents are known. Thus, if we deduct the weight of i litre of the dry soil from the weight of i litre of the solid soil constituents, and divide the difference by the specific gravity of the soil constituents, we obtain the volume of the interstices in i litre expressed in terms of water. Numerous determinations in natural gravel and sand of various degrees of fineness, and in artificial mineral powders, show that so long as the material itself is non-porous, the interstices are generally about 40 per cent, of the whole volume. The volume of the interstices tends to rise when the particles are very small, owing to the looser packing under these conditions ; it tends also to rise when the particles are of uniform dimensions. It falls when the variation in size is considerable, owing to the closer packing which then takes place from the insertion of small particles in the spaces between larger ones. F 66 RELATIONS OF SOIL TO WATER Flugge's experiments with gravel, sand, and a mixture of equal parts of both, furnish a good illustration of the very similar results obtained with masses composed of very different sized particles, and of the diminution in the volume of the interstices when large and small particles are mixed. Volume of Interstices per cent, of Total Volume Gravel 384-40.1 Sand 35-6-40.8 Gravel and Sand . . . 23-1 - 28-9 In ordinary soils, the volume of the interstices will generally be somewhat greater than in the case of sand or gravel ; this is owing to the presence of porous or compound particles in the soil. Particles of chalk or limestone are porous ; particles of humus are highly porous. Compound particles occur abundantly, as we have already seen, in all soils in a con- dition of good tilth. As the proportion of porous particles in the soil increases, so also will its capacity for containing water. A soil abounding in porous or compound particles has its capacity for water diminished if the soil is reduced to a fine powder in the laboratory. Zenger found that the soil from a peaty meadow, TOO parts by weight of which were capable of holding 178 parts of water, had its capacity for water reduced to 103 when finely powdered. There is one factor which influences the water capacity of soil, which, however, has no relation to the volume of the interspaces ; this is the action of the colloid constituents. Colloid bodies placed in contact with water take up a con- siderable quantity, and swell considerably. We have already noticed (p. 35) the increase in volume in peat and clay which WATER CAPACITY OF SOILS 67 takes place when they are wetted. When soils swell on being wetted, it is obvious that the quantity of water they will hold when saturated must be in excess of that calculated from their interstices when dry. Hilgard and Loughridge regard the hydrated silicates in the California soils as having a distinct influence on their capacity for holding water. In ordinary soils, humus is the constituent which most powerfully influences the water-holding capacity. It acts in a double way ; first, by increasing the volume of the interspaces through its great porosity ; and second, by its absorption of water due to the colloid nature of some of its constituents. As examples of the proportion of water held by various soils when fully saturated we may quote the results obtained by Meister (Jahresb. Agrik. Chem. 1859-60, 36), by Schwarz, and by Hilgard and Loughridge. TABLE VIII WATER IN FULLY SATURATED SOILS (MEISTER) Volume of Water Water bj ' Weight. per 100 Volumes of Soil. In 100 Wet Soil. Per 100 Dry Soil. Sandy Soil . . Chalk Soil . . Clay 454 49-5 50-0 23-3 28-2 27-8 304 39-2 38-5 Loam .... Garden Earth . 60-1 69-0 31-2 434 454 76-8 WATER IN FULLY SATURATED SOILS (SCHWARZ) Coarse Sand . . 394 19-8 24-7 Loam .... 45-1 24-3 32-2 Clav . 52-7 30-8 44.5 Peat Subsoil . . 84.0 78-2 359-0 F 2 68 RELATIONS OF SOIL TO WATER Looking first at the volume of water in a given volume of soil, we see in each case that the proportion is lowest in a sandy soil, rises somewhat when the soil is calcareous, or is a loam or clay, and reaches its highest point when humus is present. The figure obtained by Schwarz for peat is doubtless below the truth ; as his figures were obtained by calculation from specific gravities, the result of the swelling of the peat on wetting is thus excluded. Hilgard and Loughridge (Rep. Agri. Exp. Stat. California, 1892-3-4, 80) found that the California soils were usually saturated by 44-60 per cent, their volume of water ; a very coarse sand was saturated with 37*5 per cent, its volume. The soils which distinctly swelled or shrank when wetted are excluded from this statement, as their final volume was not accurately known. The proportion by weight varied from 23-68 of water per 100 of air- dried soil. Hilgard's method for determining the water capacity of a soil is to use a circular brass box having a sieve bottom, 10 mm. in depth, and having a capacity of 25 cc. ; the weight of the box is known. This box is filled with the air-dried soil in powder, the soil settled by tapping the box on the table, the surface struck level with a thread of silk, and the box then weighed. The box is then supported on a triangle in water, so that the bottom of the box is just beneath the surface of the water. The box is left thus for an hour or more, till the soil is fully saturated ; the box is then rapidly wiped and weighed. The result can be calculated both by volume and weight. In analyses of soil the proportions of water, and other con- stituents, are usually stated by weight ; this mode of state- ment may, however, lead to a serious misinterpretation of the WATER CAPACITY OF SOILS 69 results when soils of different volume weights are compared. It is the quantity of plant food in a given volume of soil which determines its poverty or richness. The roots are dis- tributed through a certain space : it is the quantity of water and of plant food in that space which is important. We have already seen (pp. 42 and 47) that the volume weights, or apparent specific gravities of different soils may vary greatly, the extreme differences being shown by sand and humus. In Table VIII the proportion of water has been calculated both in respect of the volume and of the weight of the soil; the results by weight are also expressed both as percentages of the wet soil, which is the usual English mode, and also as per hundred of dry soi], which is a mode frequently adopted by American and foreign writers. It will be seen at once that the relative value of the different soils as storehouses of water for the plant appears wholly different according to the particular mode of statement we regard. Thus, reckoned by volume, the peat subsoil is seen to supply rather more than twice as much water as the coarse sand, and this is the relation between them perceived by a plant. If, however, we look at the percentage by weight reckoned on the wet soil, we should conclude that the peat supplied four times as much water as the sand ; and looking at the figures per ico dry soil, we should conclude that it supplied fifteen times as much water as the sand. The unfairness of these com- parisons by weight is at once apparent if we recollect that in the last case we are really comparing the water in seven volumes of wet peat with the water in one volume of wet sand. Experiments were made by King ( Wisconsin 6th Rep., 196) on the amount of water required to saturate the soil and sub- soil of his station when these were in their natural con- 7o RELATIONS OF SOIL TO WATER dition of consolidation. Metal cylinders, one foot in depth and six inches in diameter, were driven into the soil to their upper edge, and then removed full of soil for experiment. Successive cylinders were thus filled down to five feet from the surface. The five cylinders were then placed in a tank of water till completely saturated. They were weighed in two con- ditions : (i) immediately after leaving the water ; (2) after draining four days in a saturated atmosphere. The results were as shown in Table IX. TABLE IX WATER IN SATURATED WISCONSIN SOILS (KING) Weight of Fully Saturated. Drained Four Days. Dry Soil per Cubic Foot. Water per 100 Dry Soil. Water Inches. Water per 100 Dry Soil. Water Inches. Ib. 1. Marly Loam 76-8 41.3 5-88 32-2 4-59 2. Reddish Clay . 964 28-1 5-03 23-8 4-26 3. Reddish Clay . 96-2 28-4 5-07 24-5 4-37 4. Sandy Clay . . 101-5 24-8 4-67 22-6 4-25 5. Fine Sand . . 116-7 17-4 3-76 17-5 3-77 In five feet . . 24-41 21-24 These results are interesting in several ways. They illus- trate the different capacity for water of sand and clay, which is shown in an exaggerated manner by the percentage by weight, and truthfully by the number of inches of water per unit of area. They show also the maximum amount of water which might be contained in five feet of soil T. 1 If the dry weight of a cubic foot of natural soil is known, and the specific gravity of the soil is accurately ascertained, it is easy to calculate the pro- portion of interstices in 100 volumes of the soil, and consequently the WATER CAPACITY OF SOILS 71 The quantities of water, ascertained by a laboratory ex- periment, as capable of being held by any soil are but seldom realized under natural conditions in the field. It is indeed a difficult task thoroughly to saturate a soil with water. The interstices of a dry soil are full of air, and unless the whole of this air is allowed to escape the soil cannot become fully saturated. In the laboratory saturation is best effected by allowing the water to rise into the soil from beneath ; the air then easily escapes through the dry soil above. In nature this proceeding is reversed, the rain falling on the surface and hindering the escape of air. It is thus only after long continued rain that soils are found in a fully saturated con- dition. Illustrations of the amounts of water held by field soils in their wettest condition will be found in Table X. Nos. 1-5 are Wisconsin soils, examined by King thirty- two hours after a rainfall of over three inches ( Wisconsin yth Rep., 152). Nos. 6-8 are soils, from Broadbalk wheat field, Kothamsted, examined by Lawes and Gilbert (J. Roy. Agri. Soc. 1871, no). These figures are on the whole somewhat lower than the percentages by weight given in Table IX as the result of laboratory experiments. maximum amount of water it is capable of holding. This mode of work will lead to more accurate results than experiments made in the laboratory on powdered soils, not only because the soil is taken in its natural state of consolidation, but also because, if the cubic foot is measured in a moist condition, the considerable errors which sometimes arise from the change in volume of the soil on wetting are entirely avoided. If a cubic foot of moist sand weighs when perfectly dry no lb., and the sand has a specific gravity of 2-62, then the sand contains 32-6 per cent, of its volume of inter- spaces, and would hold when saturated that proportion of water. On the other hand, a cubic foot of clay, weighing when dry 75 lb., and of a specific gravity 2-50, will contain 51-9 per cent, its volume of interspaces. These are nearly extreme cases ; ordinary soils (peat of course excluded) will fall between these limits. 72 RELATIONS OF SOIL TO WATER TABLE X WATER IN NATURALLY SATURATED SOILS Weight of Water Per 100 Wet Soil. Per 100 Dry Soil. 1. Quartz Sand (at 2 Clay Loam water level) 184 224 24-9 24-1 25-7 23-0 24-7 37-6 22-5 28-9 33-2 31-8 34.7 29-9 82-8 60.2 3. Ditto . . . 4 Brick Clay . 5 Black Marsh 6. Loam, unmanur 7. Loam, artificial 8. Loam, farmyard ed 26 years manure 26 years .... manure 26 years . . . The Eroadbalk soils afford a good illustration of the influence of manures on the water capacity of a soil. These soils had grown wheat continuously for twenty-five years ; they were sampled in January 1869 after long continued rain. The figures show the quantities of water held in the first six inches of soil. The soil which had been continuously unmanured holds the least water. The soil which had grown large crops of wheat with artificial manures holds distinctly more water : here the soil contains more humus, the residue of the roots and stubble of the larger wheat crop. The soil which had received 14 tons of farmyard manure annually for twenty-six years far exceeds all the others in the amount of water which it contains, owing to the large accumulations of humic matter within it. The accumulation of humus, and the increased water-holding power associated with it, is however practically limited to the first nine inches from the surface. Illustrations of the amount of water held by a field soil in a natural state of consolidation, and in its wettest condition, are furnished both by the Hotharnsted and Wisconsin ex1 OPTIMUM PROPORTION OF WATER 73 periments. Water was determined in the Broadbalk soils down to 3 ft. below the surface in January 1 869 ; the highest mean percentage of water found was 26-71, the lowest 23-17. Taking the weight of dry fine soil down to 3 ft. as n-6 million Ib. per acre (see Table IV), the water contents to that depth become respectively 18-68 and 15-47 inches per acre. In the case of the Wisconsin soils, in a natural state of consolidation, but saturated in the laboratory (Table IX), the upper three feet of soil contained 15-98 inches when fully saturated, and 13-12 inches after draining for four days. Both the soils may be generally described as clay loams. The quantities of water per acre just mentioned, supplemented by the summer rainfall, would be fully equal to the demands of the largest crop that could be grown (p. 55). The quantity of water contained in a soil saturated by rain is not, however, permanently held, but rapidly diminishes by percolation and evaporation. The diminution of the quantity of water in the soil below the point of saturation is indeed essential for healthy plant growth. Optimum Proportion of Water. When a soil is fully sat- urated with water air is of course entirely excluded ; this con- dition is most injurious to the health of plants. Many plants may indeed be grown with their roots immersed in water, if this water is freely exposed to air ; but the water in a subsoil is exposed to air only when the interstices of the soil are but partly occupied with water. Experiments have been made by Hellriegel, and by Wollny, in which agricultural plants were grown in jars of soil in which certain proportions of water were constantly maintained. It appeared that when the soil contained 80 per cent, of the water required to saturate it, the proportion was too high, and that when the water amounted 74 RELATIONS OF SOIL TO WATER to only 30 per cent, of the saturation quantity, the proportion was too low for the production of a maximum crop. The largest crops were obtained when the proportion of water lay between 40 and 60 per cent, of that required for full saturation. When a soil is half saturated with water it of course implies that the interspaces are half filled with air, and this is apparently the condition to be aimed at. Power of Retaining Water. The utmost capacity of a soil for water is a subject of comparatively little practical import- ance, as most soils are fully saturated only when the level of standing water is quite near the surface, or immediately after long continued heavy rain. Soils usually occur in nature in a more or less drained condition, and it is the quantity of water which they retain when fully drained which determines the supply which they are able to furnish to a crop. The proportion of water held by a soil in a fully drained condition is termed by Mayer its absolute water capacity. If a wide tube of sufficient length is filled with coarse sand, consisting of particles of uniform size, the lower end of the tube being closed by a piece of linen, and water then poured on the top of the sand till it is fully saturated, and the whole then allowed to stand till dropping ceases, there will be found in the tube two distinct layers of wet sand, a short column at the bottom fully saturated, and a long column above it fully drained, and containing throughout a nearly uniform propor- tion of water. In the lower layer the interstices are completely filled with water. In the upper layer the water coats the surfaces of the particles, and is held around their points of contact, but the main interspaces are empty. If the particles of the sand are not uniform in size, but consist of a mixture of large and small, as in a natural soil, POWER OF RETAINING WATER 75 then, when dropping has ceased, three layers may be dis- tinguished, but the divisions are not marked with the sharpness that appeared in the former case. The lowest layer as before is fully saturated, and the highest layer is fully drained, and contains a uniform proportion of water, but there lies between them an intermediate layer, often of considerable length, in which the proportion of water is not uniform but increases from above downwards till it merges into the full water contents of the lowest layer. In this case, the water in the fully drained layer not only coats the particles, but fills the finest of the interspaces. In the intermediate layer, more and more of the interspaces are occupied with water as the sand gets nearer the bottom, till at last the largest are occupied, and the sand is found completely saturated. If, in a third case, the tube is filled with an extremely fine powder, firmly packed together, and then saturated with water, this powder may be found to exhibit no loss by drainage, but the tube remains filled throughout with matter of one uniform degree of wetness. In this case the interspaces are so fine that the water filling them is held too firmly to obey the force of gravity. The cause of these various results will be better understood when we have discussed the subject of capillary action. Schlibler made many experiments on the power of various soils to retain water, but his results, and those of other early investigators, are generally too high, the experiments being made in short tubes or funnels in which the soils were never thoroughly drained. A. Mayer has made use of a tube i metre long, composed of two pieces joined by caoutchouc, the upper piece 25 cm., the lower 75 cm. in length. The tube is filled with powdered soil, which is then saturated with water. When 76 RELATIONS OF SOIL TO WATER drainage has ceased, the upper portion of the tube is dis- connected, and the amount of water held by the drained soil which it contains is then determined. Wollny's apparatus follows the same principle, and is still more complete (Forschungen der Agrikulturphysik, 1885, 177). It is only by methods such as these that the true amount of water retained by a soil can be ascertained. We have previously pointed out that the coarseness or fineness of the particles has no direct influence on the quantity of water that will be held by a mass when fully saturated ; when however we have to deal with the amount of water retained after thorough draining, the size of the particles, or — to speak more accurately — their extent of surface, becomes the factor having the preponderating influence on the result. The larger are the particles, or the less the internal surface of the mass, the smaller will be the proportion of water retained after draining. Mayer separated powdered quartz by sifting into three degrees of fineness : when fully saturated, each of these powders contained more than 40 per cent, of its volume of water ; when thoroughly drained they retained as follows : — Diameter of Volume of Water retained Quartz Particles. per cent, of Total Volume. 0-9 -2-7 mm 7-0 0-3-0.9 mm 13-7 below 0-3 mm 44-6 The coarsest powder has thus lost by draining about five- sixths of its water, while the finest powder retains after draining about the same quantity of water which it held when fully saturated. Schloesing points out that the very different relation of fine and coarse particles to water may be shown by sifting a sample of moist sand, and then determining the POWER OF RETAINING WATER 77 proportion of water in the coarser and finer parts ; the portion passing through the sieve will be found to contain much the most water. An excellent illustration of the manner in which water is distributed in columns of sand after thorough draining is afforded by an experiment made by King (Wisconsin icth Report, 176). He filled five tubes, 10 ft. long and 6 inches in diameter, with sand of different degrees of fineness, prepared by sifting successively through sieves having 100, 80, 60, 40 and 20 meshes to the inch. The columns of sand were saturated with water, and then allowed to drain, protected from evaporation, for 1 1 1 days ; the water in each six inches of every column was then determined. The results given by three of the columns are shown in Table XI. TABLE XI WEIGHT OF WATEE PER IOO OF DEY SAND AFTEE DRAINING in DAYS (KING) Meshes per inch of Sieves used to prepare Sands. 20-40 60-80 100 -a: per cent. per cent. per cent. First Foot . . . 1.92 240 3-35 Second ,, . . . 2-34 2-72 3-53 Third , . . . 2-36 2.79 4-03 Fourth , . . . 2-36 2-93 5-16 Fifth , . . . 245 2-98 6-99 Sixth , . . . 2-62 3-12 9-87 Seventh , . . . 2.74 3-11 10-98 Eighth „ . . . 3-04 3-54 15-88 Ninth „ . . . 3-81 13-50 18-90 Tenth „ . . . 14-04 20-51 19-99 Mean . . . 3-77 5-76 9.87 78 RELATIONS OF SOIL TO WATER We at once see that the finer the sand, the larger is the proportion of water retained after draining. In» the case of the two coarser sands it is evident that we nearly approach the condition of a column composed of uniform particles, the proportion of water retained throughout each column vary- ing very little till the bottom is approached. The finest sand is, on the contrary, clearly a mixture of particles of different size, as it is only in the first two feet that we find a uniform minimum contents; from this point down- wards the proportion of water rapidly increases. We have already stated that the reason why a mass of fine particles retains more water after draining than a mass of coarse particles is simply due to the far greater surface for holding water existing in the first instance, and to the much narrower interspaces between the particles. It follows of course that porous particles will have a much greater power of retaining water than solid particles of the same dimensions. Mayer prepared a fine powder from quartz, calcspar, clay- stone, and wood ; the particles were in each case as far as possible of the same size (0-3-0-9 mm. diam.). He then determined the volume of water retained by 100 volumes of each powder after draining. The results were as follows : — Calcspar. Quartz. Claystone. Wood. 11.7 ... 13-7 ... 24-5 ... 45-6 The porous particles thus retained far more water than the solid particles, though all were of approximately the same diameter. Schloesing has given some determinations of the weight of water held by fully drained soils. His results were as follows : — POWER OF RETAINING WATER 79 Weight of Water in 100 Drained Soil Coarse Sand 3-0 Fine Sand 7-3 Calcareous Sand ... 32-0 Clay Soil 35-0 Forest Soil 42-0 We have here again examples of the increase of the water- holding power as the particles become finer, or more porous (e.g. calcareous sand). The forest soil consisted chiefly of extremely fine sand, probably with some humus. Some further illustrations of the amount of water held by drained soils will be found in the next section. The state of consolidation of the soil, or in other words the closeness of the packing of its particles, has a great influence upon its power of retaining water. The operations of tillage may thus supply a means of ameliorating the ex- cessive dryness or wetness of a soil. Referring once more to the examples of the loose and tight packing of soil particles already given on p. I, it will be evident that in the system of close packing the points of contact between the particles are about twice as numerous as in the system of loose packing (Fig. i), and the interstitial spaces are also much reduced in size. The closely packed particles will in fact retain when drained at least twice as much water per unit of volume as the loosely packed particles. The reason of this will become clearer when we have discussed the subject of capillary action. In practice, the water-holding power of a coarse sandy soil may be increased by consolidation with a heavy roller, or by the treading of animals on the land. On the other hand the water-holding power of a heavy soil may be greatly reduced if the soil can be pulverized, and brought into a loose state of aggregation. 8o RELATIONS OF SOIL TO WATER The amount of water retained by a soil after rain is one of the factors which, more than any other, determines its suitability for different kinds of agricultural crops. The typical American soils described on p. 20 owe their suitability for their respective cultures chiefly to the varying percentages of water which they retain. The requirements of a plant for water vary a good deal in the various stages of its life. In the earlier stage of leafy growth, when the production of veget- able tissue is proceeding with the greatest vigour, the demand for water is greatest, and luxuriant growth at this period is largely determined by the quantity of water supplied. But in the later stage of seed production, when the transference of matter rather than its new formation is the great business of the plant, the presence of an excess of water is for many plants highly injurious, and greatly diminishes the proportion of seed yielded by the plant. For seed production, therefore, dry conditions are desirable. The general idea of the relation of water supply to plant function we have just presented serves to explain why different crops, or different styles of culture, require different proportions of water in the soil. Wheat land must be drier than grass land, if both crops are to develop to the best advantage. A soil for the production of a fine sample of malting barley must be drier than one yielding maximum wheat crops. A soil that is to supply early market-garden crops must be a dry one, for the object is to obtain early and not heavy crops ; and to obtain early maturity the crop must be hastened through its preliminary stage of tissue formation and brought as quickly as possible to the completion of its career. Little information exists as to the proportion of water actually held by the soils most suitable for the production POWER OF RETAINING WATER 81 of various crops ; a commencement of an investigation of this kind has however been made in America. Certain typical soils in various places have been left for the purpose of experiment without a crop, and samples of the soil have been obtained every day by boring to the depth of I ft., and in these samples the proportion of water has been determined. Whitney (Boils, Bulletin 3) gives the results of the mechanical analysis of these soils, and also the amounts of water which they contained in the months of June and July 1895. TABLE XII PHYSICAL ANALYSIS OF TYPICAL SOILS (WHITNEY) Diameter of Particles. Market- garden Soil. Bright Tobacco Soil. Shipping Tobacco Soil. Burl ey Tobacco and Grass Soil. Fine Gravel Coarse Sand . Medium Sand . Fine Sand . . Very Fine Sand Silt .... mm. . 1-0 -2.0 . 0-5 -1-0 . 0-25 -0-50 . 0-10 -0.25 . 0-05 -0-10 . 0-01 0-05 0-06 046 7-08 4843 26-20 8-52 3-20 4-55 0.71 1.12 7-37 27-90 24-26 22-77 4-20 8.30 2-07 0-15 0-05 0-18 0-11 0-34 5-13 63-28 5-19 20-55 1-76 1-63 1.24 0-58 1-59 46-36 9-56 30-20 Fine Silt . . Clay . . 0-005-0-01 Loss at 100°. . 0-15 1.10 2-10 3-06 4-29 5-32 Loss on ignition Total .... 99-75 98-85 99-99 102-53 Percentage of Water in ift. of Soil, June and July 1895 Market- garden Soil. Bright Tobacco Soil. Shipping Tobacco Soil. Burley Tobacco and Grass Soil. per cent. per cent. per cent. per cent. Minimum .... 6-2 4-0 12-1 18-0 Maximum . . . 10.5 11.7 17.9 23-1 Average .... 8-7 7-2 154 20-1 82 RELATIONS OF SOIL TO WATER These soils were in different localities ; the results do not therefore show how different soils behave with the same rainfall, but rather the amount of water found in soils specially suited to certain crops in a season of fair average production. The soils best suited for market-garden purposes, or for the growth of the bright yellow tobacco used for cigarettes, thus only held from 5-11 per cent, of water; the shipping tobacco soil 1 2-1 8 per cent. ; and the pasture soil (used also for the coarsest tobacco) 18-23 per cent. Each of these conditions was especially suited for the purpose of the particular crop cultivated. A detailed account of the physical texture and water- holding power of soils producing distinct varieties of tobacco is given by Whitney in a later Bulletin (Soils, No. n). The delicate, elastic leaf, used for cigar wrappers, and the bright yellow tobacco already referred to, are only produced on sandy soils holding but little water. The thicker, coarser leaf, with which is associated a much larger return per acre, is grown on soils containing a more or less considerable proportion of clay, and holding a much larger quantity of water. - According to Whitney's observations, the water contents may rise to one-quarter more than the normal amount, or fall to one-quarter below it, without seriously disturbing the characteristic quality of the soil. With a greater diminution of water drought will be felt, and with a permanently greater increase in the water contents the crop will be injured. An attempt is being made in the United States to obtain daily records of the water contents of typical soils by measur- ing the resistance to an electric current passing between two electrodes sunk in the soil. POWER OF RETAINING WATER 83 Before leaving this section we may sum up the chief conclusions by saying that extreme fineness of the particles is by itself capable of 'giving to a soil a maximum power of retaining water, this condition alone sufficing to keep all interstices full of water after percolation has ceased. The quantity of water retained is thus in proportion to the internal surface, and is doubtless increased by the presence of colloid bodies in the soil. Of all soils peat has the greatest capacity for retaining water, its porosity supplying an enormous internal surface, the effect of which is heightened by the affinity for water of its colloid constituents. At the other end of the scale we have gravels and coarse sands, which have hardly any power of retaining water. For swamp-loving crops, as rice, soils retaining a maximum amount of water are desir- able ; soils having a high retentive power may also have a special value in climates of small rainfall. For ordinary farm crops it has been already pointed out that a very large proportion of water in the soil is distinctly injurious, and that the most vigorous growth is obtained when the soil contains about one- half of its saturation quantity. Sandy soils are by no means so inferior as suppliers of water as is often supposed. We must recollect that the quantity of water retained by a given volume of sand is greater than would appear from the usual percentage deter- minations made by weight. A moist sand containing 7 per cent, of water by weight, contains 11-2 vols. of water per 100 vols. ; and this difference between weight and volume is greater in the case of sand than in the case of a clay soil, and still more than in the case of a soil containing humus, owing to the higher weight per cubic foot of the sand. The soils just named are in fact not so unequal when their G 2 84 RELATIONS OF SOIL TO WATER water contents is expressed in pounds per cubic foot as when it is expressed as per cents, by weight. We have further to bear in mind the fact already noticed (Table VII), that a sand gives up its water far more completely to a crop than will clay or humus. This is in great measure due to the smaller internal surface of the sandy soil. In the case of a sandy and clay soil, both containing 7 per cent, of water, the water in the sand occurs as a much thicker film, being spread over a much smaller surface ; it is thus in a condition better suited for absorption by the roots of plants. The same weight of water in the clay soil is held as an extremely thin film, being spread over an enormous surface. The thinner is the water film the more firmly is it held by the solid particles of the soil, till it finally becomes quite incapable of assimilation by plants. The greater availability of the water in a sandy soil will also be partly due to the absence of the colloid bodies which occur in clay and humus, as these, as we have already seen, have the property of firmly retaining water. We have further to take into account the wider distribution of the roots in a sandy soil, and the greater freedom with which water moves within it. The facts just mentioned will not unfrequently turn the balance in favour of a sandy soil as a supplier of water to the crop. King (The Soil, 161) mentions the case of a sandy and clay soil at Wisconsin having water capacities of 1 8 and 26 per cent. Maize grown on these soils reduced the water in the sand to 4-17, and in the clay to 11-79 Per cent. Calculating from these data, he tells us that the sandy soil had yielded the crop 13-83 Ib. of water per cubic foot of soil, while the same volume of the clay soil had yielded only 12-5 Ib. Experience in the United States supplies abundant PERCOLATION 85 examples of the efficient supply of water to crops in semi- arid regions by soils destitute of clay, but consisting of fine particles of silt and sand. Th. Schloesing junior (Compt. rend., cxxv. 824) has cajled attention to the greater speed with which sulphate of ammo- nium nitrifies in a sandy soil than in one containing much clay, when both contain a similar moderate proportion of water. His experiments were made with artificial mixtures of sand and clay in various proportions, including a small quantity of chalk ; the same percentage of water was added to each mixture. All the soils were in a loose condition, and abundance of air was thus provided. He concludes that the different rate of nitrification is mainly due to the different thickness, and therefore availability, of the water film coating the particles of the various mixtures. By increasing the per- centage of water in the soils containing most clay the rate of nitrification was raised to that observed in sandy soils. A soil of coarse sand will show to best advantage in a season of frequent slight showers. These small supplies of water may be as thoroughly retained by the sand as by clay, while they will penetrate the sand to a much greater depth, and more effectually supply the needs of plant roots. Percolation. The conditions which affect the passage of water through the soil require some remark. Percolation is of course greatest where the retention of water is least ; the characters of the soil which produce little retention are thus favourable to a large percolation, and vice versa. King (The Soil, 171) determined the rate at which water would pass through columns of sand of different degrees of fineness, columns of clay loam, and black marsh soil. The columns were one-tenth of a square foot in cross section, and 86 RELATIONS OF SOIL TO WATER fourteen inches high. A head of water two inches in height was maintained on the top of each column throughout the experiment. The results were as follows : — Inches of Water passing in twenty-four hours Meshes per inch of Sieves used to prepare Sands. Clay Loam. Black Marsh. 40-60 ... 60-80 ... 80-100 ... 100- Inches. Inches. Inches. Inches. Inches. Inches. 301 ... 160 ... 73-2 ... 39-7 ... 1-6 ... 0-7 This series of results serves excellently to illustrate the fact, that the finer are the particles of a soil the slower will be the rate of percolation through it. King remarks that the whole of the rates of percolation observed are far above what the same soils would yield in the field ; this is owing to the shortness of the columns used, the absence of air in the inter- stices, and the considerable head of water maintained. King (nth Wisconsin Report, 285) has also determined in great detail the rate at which water drains from saturated sands of various degrees of fineness. The sands, prepared as before, were filled into tubes 8ft. long and 5 inches in diameter ; each column was completely saturated with water from below, and drainage was then allowed to commence. All the sands contained nearly the same proportion of water when saturated. Air was allowed to enter at the top of each column, but precautions were taken to prev.ent evaporation. The amount of percolation in the first hour shows in the most striking manner the different behaviour of the coarsest and finest sand. The finer sands, retaining so much more water at first, discharge after the first hour a little more than the coarsest. At the end of nine days, regular percolation had ceased in all the columns ; but from time to time slight per- PERCOLATION 87 eolation recommenced and was duly recorded, the whole experiment lasting for 268 days. TABLE XIII WATER DRAINING FROM EIGHT FEET OF SATURATED SANDS, PER 100 OF DRY SAND (KING) Meshes per inch of Sieves used to prepare Sands. 20-40 60-80 100- Water per cent. Water per cent. Water per cent. One Hour 9.6 6-6 14 One Day 13-8 11-8 6-3 Three Days .... U-5 12.5 7-5 Nine Days .... 15-3 12-9 84 268 Days 164 13-6 9-3 During the intermittent percolation of the last 259 days the sands lost from 6-56-9-15 Ib. of water per square foot, or considerably more than one inch of water l. This intermittent percolation deserves attention ; it is doubtless due to the vary- ing temperature and pressure of the atmosphere. A rise in temperature will act in several ways to start percolation in a soil in which the water had previously reached a state of equilibrium. As the temperature rises water becomes less viscous and its surface tension diminishes ; drainage therefore recommences in the sand column, the water films coating the upper fully drained portion becoming thinner and some of the water passing downwards. The amount capable of being held in the capillary passages at the foot of the column is also diminished. A rapid expansion of the air within the column will also cause the expulsion of a part of the water collected One inch of water on a square foot is roughly 5.2 Ib. 88 RELATIONS OF SOIL TO WATER at the lower end (see Table XI). A fall in temperature will cause this new percolation to cease and will tend to bring- about a redistribution of water towards the surface. A sudden fall in the barometer may act on the air contained in the column in the same manner as a rise of temperature. These changes all occur in a natural soil, though the results from them are not generally large ; we shall have to refer to them again when we speak of the movements of underground water (p. 129). Percolation, without a constant supply of water above, is only possible if air can enter the soil to take the place of the water leaving it. Percolation may be stopped for a time by a slight rain falling on the surface and closing the air passages. This has been observed in the case of the Rotham- sted drain-gauges. King has made some experiments upon the influence of temperature upon the rate of percolation through sand, a constant water supply being provided. He found that a rise of temperature from 9° to 24° C. increased the rate of per- colation 50 per cent. This result he thinks may possibly be above the truth. According, however, to Briggs' calcula- tion, the experimental results obtained by King are precisely those which might have been predicted from the known diminution of the viscosity of water with a rise in temperature. It is obvious therefore that soils will drain much more freely in summer than in winter. The resistance of clay to the percolation of water is a fact with which all are familiar ; the resistance is absolute when the clay is in a puddled condition, that is when it has been reduced to a mass of single particles. This resistance may be greatly modified by the coagulation of the clay with lime PERCOLATION 89 (Sachsse and Becker's experiment, p. 33), and by the forma- tion of compound particles. The resistance of puddled clay to the passage of water is due in great part to the extreme fineness of the particles, and to the great resistance which has to be overcome in passing between them ; but there can be little doubt that a considerable part of the difficulty is due to the colloid constituent of the clay, which occupies the interspaces with a jelly-like substance, and thus immensely increases the resistance offered to the passage of water. The coagulation of this colloid profoundly alters the character of the clay. In the case of a heavy loam or clay soil, under natural conditions, percolation is much facilitated by the presence of channels formed by worms, or by the roots of plants ; and by the occurrence of fissures, either originating in times of drought or natural joints in the formation. Bain may pass down these passages before the soil is saturated ; summer drainage on such soils is often chiefly of this character. The amount of water passing through a soil is measured by means of drain-gauges or lysimeters ; an instrument of this kind was constructed by Dalton in 1796, and the same method of investigation has since been employed by many observers. Most drain-gauges consist of cylinders or square frames artificially filled with soil, and usually 3 ft. deep, with an arrangement below for collecting and measuring the water which passes through. The three drain-gauges at Rothamsted, constructed in 1870 (Jour. Roy. Agri. Soc. 1881, 269), con- sist of rectangular blocks of undisturbed soil, isolated by walls of brick set in cement, and supported below on perforated iron plates, below which is placed a metal funnel, so that all drainage water can be collected and measured. The three RELATIONS OF SOIL TO WATER blocks of soil have the respective depths of 20, 40, and 60 inches; the surface area of the soil is in all cases T^Voth. of an acre. The soil is a heavy loam containing many flints, having a clay subsoil. The surface of the soil is kept free from weeds, and undergoes no tillage. The average amounts of monthly percolation during twenty years, as shown by the shallowest and deepest soil of the Rothamsted drain-gauges, will be found in Table XIV. Further results from these and other drain-gauges will be found on pp. 109, 122. TABLE XIV AVERAGE MONTHLY PEKCOLATION THROUGH BARE SOILS TWENTY INCHES AND SIXTY INCHES DEEP, ROTHAMSTED Rainfall. Percolation. Percolation per cent. Rain. 20-inch Gauge. 60-inch Gauge. 20-inch Gauge. 60-inch Gauge. inches. inches. inches. January . . February . . March . . . 2-51 2-04 1.74 1-96 144 0-80 2.06 1.44 0-86 78-1 70-6 46-0 82-1 70-6 494 April . . . May ... June . . . 2-21 2.28 2-52 0-67 0-60 0-63 0-68 0-59 0-60 30-3 26-3 25-0 30-8 25-9 23-8 July . . . August . . September . October . . 3-03 245 2-86 3-20 0-84 0-56 0.96 1.78 0-77 0-50 0-75 1-50 27.7 22.9 33-6 55-6 254 20-4 26-2 46-9 November . 3-03 2-24 2-05 73-9 67-7 December 2-42 1-90 1.81 78-5 74-8 Whole Year . 30.29 14.38 13-61 47-5 44-9 As the amount of evaporation from the soil is far greater in summer than in winter, the amount of drainage naturally PERCOLATION 91 varies in a contrary manner. The season of active drainage through a bare uncropped loam is shown by the figures in the table to commence in October and to continue till the end of February. In a climate having a severe winter, as Canada, this heavy winter drainage will not occur, but in its place there will be a large amount of drainage in April when a thaw occurs. During the spring thaw in such climates large quantities of snow water will however flow away over the surface, when the inclination allows of it, the frozen con- dition of the soil hindering percolation. The amount of annual percolation is thus considerably diminished in a severe winter climate. It will be noticed that the deep and shallow soils both deliver the same amounts of drainage at the end of winter, the deepest soil continuing draining longest. In summer and autumn the shallow soil yields the most drainage, the supply of water to its surface by capillary action, and con- sequently the amount lost by evaporation, being somewhat smaller than in the case of the deeper soil. The amount of water passing through a soil is so largely influenced by the rate of evaporation from its surface, that it will be necessary to return to the subject again when the conditions of evaporation have been considered (p. 1 25). CHAPTER III KELATIONS OP SOIL TO WATER (continued) Capillary Action — Evaporation from Bare Soil — Influence of Crop on Evap- oration— Underground Water — Wet and Dry Soils —Amelioration of the Physical Properties of Soil. Capillary Action. The rapid movements of water and other liquids in porous bodies is a fact with which all are familiar ; the rise of water in a lump of sugar, the spreading of a drop of water in blotting paper, the rise of oil in a wick, are all examples of capillary action. The typical instance, which supplies the name for the whole of the phenomena, is supplied by the rise of liquids in narrow tubes. The height to which water will rise in glass tubes of various diameters is as follows : — Height to which Water at o° C. rises in glass tubes Tube 1 mm. diameter, water rises to 15-336 mm. „ 0-1 „ „ „ „ „ 153-36 it 0-01 „ „ „ „ „ 1533.6 „ The height is thus greater the narrower is the tube. A reduction of the diameter to one-tenth causes the water to rise to ten times the previous height. The height to which a liquid rises is somewhat diminished by an increase in temperature. The rise of a liquid in a narrow tube or passage is only a particular manifestation of the familiar adhesion between a solid and liquid which is seen when a stick or clean stone CAPILLARY ACTION 93 withdrawn from water comes out with a wet surface. Water rises in a glass tube, or through the spaces existing in a mass of sand, simply because the attraction of the surface particles of the glass or sand for the particles of the water is at first greater than the attraction of gravity ; the rise of water ceases when the mass of water in the column reaches such dimensions that the attraction of gravity balances the surface attraction of the glass or sand. The surface attraction is greater, and the quantity of water raised larger, the wider the tube ; but the height to which the water is raised is greater the narrower the tube ; because while the attracting surface simply diminishes in the same ratio as the diameter of the tube, the volume of water within the tube (and thus the weight to be raised) diminishes as the square of the diameter, and the less weight is thus carried to a greater height. The same forces which occasion the rise of water in a tube will determine the distribution of water over moist surfaces, and this aspect of the subject is of considerable importance for the correct understanding of the movements of water in a soil. It is only when the interspaces of a soil are filled or nearly filled with water, that an uninterrupted passage of water through tubes of varying size, shape, and direction can take place. In the case of a fully drained soil of open texture, or consisting of coarse particles, the particles are merely covered with a water film, and it is only at the points of contact between the particles that anything of the nature of a tube is to be found. The distribution of the water over the surfaces of the moist particles is however governed by the same laws which control its behaviour in soil tubes. Briggs (Mechanics of Soil Moisture, Soils, Bull. 10) has 94 RELATIONS OF SOIL TO WATER given a very clear account of the condition and behaviour of water in a drained soil. The retention of water on the surface of soil particles, in spite of the opposing force of gravitation, is due to what is called * surface tension.' The surface of the film of water encircling a soil particle is really in the condition of an elastic membrane exerting a very considerable pressure ; the water is in consequence firmly held FIGURE 3. against the soil particle. In a fully drained soil there exists a condition of equilibrium between the force exerted by surface tension and the force exerted by gravity. If the films of water became thicker and heavier, a part of the water would gradually pass downwards in obedience to the attraction of gravity. If the films became thinner they would acquire the power of absorbing and retaining fresh supplies of neighbouring water. In Figure 3 will be found the diagram used by Briggs to illustrate the transference of water from a wetter to a drier portion of the soil, a transference which in this illustration depends entirely on the surface tension of the water, and not upon the existence of connecting tubes. The three particles of soil, which lie in contact, are each CAPILLARY ACTION 95 surrounded by an elastic film of water exerting a considerable pressure towards the centre of the particle. The result of this state of tension is that an outward pressure exists on the surface of the thicker portion of the film lying at the points of contact between the particles. This outward pressure is greater the thinner is the film on the adjoining particles. If therefore moister and drier particles are in contact, water will pass from the first to the last till an equilibrium of surface tension is established throughout the system. The figure shows this transference of water in action ; the arrows indicate the direction of the pressure in the different parts of the system. Water thus tends to distribute itself in a soil, either through capillary passages, or by the slower process of surface distribution. When these operations are assisted by gravitation, as when rain falls on a dry soil, the move- ment of the water becomes rapid. When these operations are opposed by gravitation, as when a soil dries at the surface and is still wet below, the movement is retarded, and the amount of possible work is limited, as in the case of the rise of water in a glass tube, by the final neutralization of the attractive forces by the increasing weight of the column of water lifted. The height to which water will rise in a soil by capillary action depends primarily on the size of the soil particles, and the closeness of their packing. When a soil is composed of very coarse particles, as in the case of gravel, the inter- spaces are so wide and the points of contact so few, that capillary action has little influence on the distribution of water throughout the mass. In the case of sands and loams the interspaces become sufficiently narrow for the action to 96 RELATIONS OF SOIL TO WATER assume practical importance. Capital illustrations of the influence of various conditions on the height to which water is raised, and the quantity lifted, are furnished by the experi- ments of Johnson and Armsby with loam and emery (Connecticut Exp. Stat. Rep. 1878, 83); the latter substance was selected because it can be procured in commerce graded in various degrees of fineness. A loam was separated by sifting into particles of three grades. Tubes fourteen inches long and two inches in diameter were filled with these soils ; the lower ends of these tubes were then placed in water. The quantity of water evaporated by each soil in 150 days per square inch of surface was as follows : — Average Diameter Wafer Evaporated of Soil Particles. per square inch. 3-0 mm 49-2 grams 1-0 „ 73.8 „ less than 0-25 „ 153-8 „ We see that as the size of the soil particles diminished there was a large increase in the quantity of water brought to the surface. In the case of the two coarser materials the amount of water raised was insufficient to preserve a moist condition at the surface ; the action in these cases was thus mainly due to a mere surface distribution of water. The next experiment we quote shows the influence of varying height on the quantity of water raised. The trials were all made with emery of one kind, the average diameter of the particles being 0-229 mm. The experiment lasted twenty-five days. Relative Heights Water Evaporated, of the Columns. 1 133-9 grams 2 118.2 „ 3 112-0 „ CAPILLARY ACTION 97 The highest column (13! inches) was apparently saturated with water to the top ; notwithstanding less water was evap- orated than by the shorter columns, the water rising more slowly the greater the length of the column. In the next experiment the height of the column was in all cases fourteen inches, but the tubes were filled with emery of various degrees of fineness. The trials lasted eighty- four days. Average Diameter Water Evaporated of Particles. per square inch. 0-443 mm 40-3 grams. 0-356 ,, 1384 „ 0-229 „ 155-1 „ 0-140 „ 150-6 „ 0-076 „ 144-7 „ It will be seen that the largest quantity of water is brought to the surface by the emery of intermediate fineness (diam. 0-229 mm.), and that as the particles become either coarser or finer the quantity is diminished. The coarser particles failed because the surface was never saturated with water ; the coarsest emery could saturate at 8 inches, the next coarsest at n inches. The finer particles failed because of the slower movement of the water in their passages. Had the column been shortened, one of the coarser emeries would have evaporated most water; had it been lengthened one of the finer would have appeared the most effective. This is an important lesson. For every distance of the surface from the water supply there is a particular size of soil particles which will bring up the largest quantity of water, and in the case of moderate distances it is by no means the finest particles which are the most effective. The laboratory experiments on the capillary power of some Californian soils made by Loughridge (California Exp. Stat. H 98 RELATIONS OF SOIL TO WATER Rep. 1892-3-4, 91) are especially instructive as the mechanical analysis of the soils is also given. The experiments were made in copper tubes i inch in diameter; the tubes were in i ft. lengths, fitting into each other. One side of the tube was glass so that the contents might be observed. The bottom tube was closed at its lower end with muslin. The tubes were filled with air-dried soil, stirred in the tube with a wire, and made firm by a slight tapping on the table. The composition of the soils, and the details of the rise of water in them, are shown in Tables XV and XVI. TABLE XV MECHANICAL ANALYSES OF FOUR CALIFORNIAN SOILS (LOUGHRIDQE) Clay. Fine Silt. Coarse Silt. Fine Sand. 1. Sandy Soil . . 2. Alluvial Soil . 2.82 3-21 3-03 5-53 349 1542 89-25 72-05 3.-SiltySoil . . 4. Adobe Soil . . 15-02 44-27 15-24 25-35 25-84 13-47 4541 13-37 TABLE XVI RISE OF WATER IN FOUR CALTFORNTAN SOILS (LOUGHRIDQE) No. 1 Hour. 6 Hours. 1 Day. 2 Days. 6 Days. 12 Days. 26 Days. 125 Days. 195 Days. in. in. in. in. in. in. in. in. in. 1 8 12* 14 15 16* 2 9* 19 27 30* 35 38 41 47 3 2 9 13 17 20* 25 31* 41 50 4 1* 6 10* 14* 23 26* 46 CAPILLARY ACTION 99 The first thing to note is the extremely rapid rise of water in the two coarser soils, a height of 8-9 inches being reached during the first hour, while the water in the two stiffer soils had only reached ii-2 inches. The first soil,has a very simple constitution, very little clay or silt being mixed with the sand ; the rise here is nearly completed in the first day, and is quite finished in six days. The second soil contains hardly any more clay than the first, but there is a good deal of silt, which, occupying the interspaces between the sand, entirely alters the behaviour of the soil to water. At the end of the first day the water has risen to nearly double the height reached in the first soil. After six days the rise becomes very slow, and getting slower and slower, a height of 47 inches is finally reached in 125 days, 42 days having been consumed in accomplishing the last inch. The rise of water in the silty soil is at first much slower than in the two previously mentioned, but the rate of progress is well maintained, and 50 inches is finally reached in 195 days. The heavy adobe soil is for a long time far behind the others, but like No. 3, it goes on when the others have stopped, and the water at last reaches 46 inches in 195 days. The rise of water in clay soils is thus very slow, and the considerable height finally reached is no proof of energetic capillary action. The colloid ingredient of the clay is doubt- less a hindrance to the rapid passage of water, though perhaps helping finally to carry the water to a greater height. Capillary action is most active in the case of the alluvial soil, made up of fine sand and silt ; in soils of this class, more than in any other, will this action be of substantial benefit to a crop. At the close of the experiment the percentage of water H 2 ioo RELATIONS OF SOIL TO WATER contained in the soils at different heights in the columns was determined with the following results : — Water per joo of Soil at different heights 1 in. 6 in. 12 in. 24 in. 36 in. 47 in. 1. Sandy Soil ... 24-3 ... 14-2 ... 3-9 2. Alluvial Soil ... 36-6 ... 35-0 ... 32-5 ... 21-4 ... 12-0 ... 4-3 Thus the greater the height in the column, the smaller is the quantity of water found. As we ascend, the wider interspaces would be found unfilled, and at last the particles will be merely coated with a thin film of water. The supply of water at extreme heights is thus very feeble, and it may well be doubted whether the gain of water at these elevations is not due rather to the condensation of vapour rather than to an actual ascent of liquid water. In all experiments made with dry soils the upward move- ment of the water is retarded by the fact that air has to be expelled from all the passages. Some dry soils are also difficult to wet, the particles remaining obstinately coated with a film of air ; this is frequently observed in the case of dried marsh soils. The question whether the quantity of water raised by capillary action in ordinary soils is sufficient to furnish a substantial supply to field crops has been greatly elucidated by the experiments made by King at Wisconsin. In the first experiment (Wisconsin 6th Rep., 203) the con- ditions were made especially favourable to capillary action. A cylinder 4 ft. in height and i ft. in diameter, which could be supplied with water from below, was first partly filled with water, and the fine sand from the Wisconsin subsoil was then dropped in, each addition of sand being well stirred in the water. A column of sand 4 ft. high was CAPILLARY ACTION 101 finally obtained free from air and perfectly saturated with water. The water level was then lowered till it was i ft. below the surface, and was then maintained at this point. The surface of the cylinder was then exposed to a strong current of air (the velocity of which was measured) for eight days, and the quantity of water evapo- rated from the surface of the sand was ascertained. The water-level was then successively reduced to 2, 3, and 4 ft. below the surface, and at each stage the quantity of water evaporated at the surface was ascertained ; each experiment lasted from ten to twenty-four days. The trials were after- wards repeated, using in place of sand a surface loam (Wisconsin jth Rep.^ 151). The results were as follows :— Water evaporated daily per Square Foot Water Level I ft. below Surface. Ib. Water Level 2ft. below Surface. Ib. Water Level 3/#. below Surface. Ib. Water Level 4 ft. below Surface. Ib. 2-37 2-07 1-23 0-91 2-05 1.62 1-00 0-90 Fine Sand Clay Loam The amount evaporated in each case . thus diminished as the distance of the water-level from the surface increased. Towards the close of the trials a slight white crust formed on the surface of both sand and loam, this was removed in the case of the loam, the rate of evaporation then rose to 1-27 Ib. per day. The quantity of water raised daily a distance of 4 ft. by capillary action was thus at least J Ib. per square foot, equal to a supply of about i inch of rain in five days, a quantity quite sufficient for the most luxuriant growth *. -1 The surface of a soil may be treated both so as to favour or retard evaporation. When the surface of the wet sand was cut across with a knife 102 RELATIONS OF SOIL TO WATER The experience at Wisconsin shows3 however, that with the same sand and loam in their natural condition in the field no such large rise of the subsoil water occurs. The experimental ground of the station is near Lake Mendota, and the water level in the subsoil is, at different seasons of the year, between four and seven feet below the surface. In seasons of drought crops often suffer considerably on this land from deficiency of water, although at the time the water level in the subsoil may be only five feet below the surface. The much smaller results from capillary action in natural soils are doubtless due to their more irregular texture as compared with the artificially prepared columns employed in laboratory experiments. In the natural soil the capillary passages are less uniform in size, and are always more or less filled with air. The percentages of water found by Loughridge in the alluvial soil (p. 100) wetted by a rise of water from below, show that the soil was less than two- thirds saturated at a height of 2 ft., and one-third saturated at a height of 3 ft. above the water level, more than 125 days after the commencement of the experiment. It is in a saturated soil that water moves with the greatest freedom, the largest passages forming the most effective channels when the quantity of water moved is regarded. In a well drained soil only the finest passages remain full of water, and these will be often interrupted by wider spaces full of air. The movement of water is thus limited in many places to the depth of two inches, the rate of evaporation rose from 0-95 Ib. per square foot to 1-76 Ib. When, on\he other hand, the uppermost two inches of the sand were removed, and then replaced in a loose condition, the rate of evaporation fell to 0-63 Ib. per square foot. For a further discussion of this point see p. 113. CAPILLARY ACTION 103 to the redistribution of the water coating the particles, and the movement then becomes very slow. Another reason for the slow movement of water in natural soils lies in the fact that the water has to be drawn from moist soil and not from free water. A fine passage tries to fill itself at the expense of a wider one, a thin film grows at the cost of a thicker one, and the result in each case is merely the difference in their respective powers. This is one chief reason of the great falling off in the rate of rise as the column of soil lengthens; we have already seen that an alluvial soil (p. 99) took forty-two days in accomplishing the 47th inch. It is evident that when rain and percolation have ceased, the movement of water from a wetter to a drier part of the soil must be greatly hindered by the reluctance of the wetter soil to part with its water. Even when, as at Wisconsin, the water level in the subsoil is fairly near the surface, the rise of water in the soil is by no means free from this hindrance, the so-called water level beiug merely the surface of a mass of saturated soil. An excellent example of the slow movement of water in a dry natural soil is afforded by another of King's experiments (Wisconsin jth Rep., 143). After the dry summer of 1889 a soil was sampled on Oct. 28, to a depth of 5 ft., and the percentage of water at different depths determined. A portion of the ground was then effectually protected from rain and snow and left in this condition till April 14 in the following year, when the soil was again sampled as before, and the water present determined. For results obtained see Table XVII. The covered soil had apparently gained no water from below during the winter months, but had on the contrary actually lost some water by evaporation. The soil open to 104 RELATIONS OF SOIL TO WATER the weather had gained much water, but clearly from above. In this case the water level was about 30 ft. below the surface. TABLE XVII WATER PER IOO DRY SOIL, COVERED AND UNCOVERED, AT DIFFERENT DATES (KING) Original Soil. Soil Covered. Soil not Covered. Oct. 28, 1889. Apr. 14, 1890. Apr. 14, 1890. 1st ft. Sandy Clay . . . 2nd ft. Red Clay . . . 3rd ft. Clay and Sand . . 4th ft. Sand and Gravel . 4-03 10-07 9-11 4.35 3-32 6-68 6-32 3.71 20-23 20-01 8-32 8-63 5th ft. Sand and Gravel . 4-53 5-08 6-07 Mean 642 5-02 12-65 A comparison of the results given by the deepest and shallowest of the Rothamsted drain-gauges also affords an example of the small influence of capillary action in bringing water to the surface. Each drain-gauge consists of a rect- angular mass of heavy loam, with flints, of the area of iirijir °f an acre; the depth of the shallowest mass of soil is 20 inches, of the deepest 60 inches. The deepest soil has thus a subsoil of 40 inches to draw upon, which is wanting in the case of the shallowest soil. On an average of twenty- four years the annual evaporation from the deepest soil has only exceeded that from the shallowest by 0-6 inch ; this probably represents the quantity of water brought to the surface from below a depth of 20 inches. According to various published experiments the presence of certain salts increases the rapidity of movement in capillary CAPILLARY ACTION 105 tubes, and the height finally reached, while the soluble organic matters present in soils have a contrary effect. The results at present obtained are not, however, sufficiently definite to justify any practical conclusions on the subject. The practical effect of capillary action in raising water to the surface of the soil, or to the level occupied by plant roots, has apparently been a good deal exaggerated ; its. influence on the distribution of water in the soil is never- theless very large. We must recollect that capillary action is by no means confined to the raising of water, its effects are indeed most limited in this direction as it is then opposed by the force of gravity. The greatest manifestation of capillary action is seen in the distribution of water in a dry soil after a shower of rain. It is the surface attraction of the particles of the soil for water which causes the rain to be sucked down, with the energy with which we all are familiar, and carried into the finest passages and remotest portions of the soil. When the percolation produced by the attraction of gravity has ceased, the system of soil and water is in a condition of equilibrium, the nature of which in the case of sandy soils is well shown by the upper half of Table XI. If water is now removed from any portion of this system by root-action, or by evaporation at the surface, the equilibrium is upset, and the water coating the particles is induced by the local alterations in its tension to redistribute itself, and regain once more the state of equilibrium. In the case of coarse sands, this redistribution consists mainly in the movement of the film of water coating the particles, and such movement will be extremely slow; it will however persist at a diminishing rate till the amount of water in the soil is reduced nearly to the proportion of so-called io6 RELATIONS OF SOIL TO WATER hygroscopic water which it is capable of holding. This redis- tribution of water will take place more easily when the loss is occasioned by means of roots, because in this case the whole action is beneath the surface, and the soil particles never become perfectly dry. When perfectly dried and coated with air, the renewal of a film of water becomes more difficult. It is evident that this redistribution of water in the soil will be effected to a greater extent (because more easily) by bringing a fresh supply of water from above, or from the side, than from below ; indeed in the case of coarse sands, the supply from bolow must be extremely small. The case of a silt or loam is quite different ; here, when percolation has ceased, the soil may remain nearly saturated with water throughout its whole depth. When this condition of equilibrium is disturbed, the movements of soil-water re- establishing it will at first be far more vigorous than in the previous case, the passages in the soil being in this case more or less filled with water. The facilities for procuring a con- siderable supply of water from below, will in this case be largely increased. The advantages mentioned do not however continue to increase as the size of the soil particles diminishes ; we have in fact already pointed out (pp. 86, 97) that excessive fineness of particles, while increasing the water-holding power of a soil, greatly diminishes the freedom of movement of the soil water. Stiff clay soils, when in a puddled condition, notwithstanding the large amount of water they may contain, are quite incapable of efficiently supplying the wants of a plant ; the roots take water from the soil they are in con- tact with, but this water is replaced so slowly from the surrounding soil that the plant may die of drought. The maximum advantages of capillary action are apparently to be EVAPORATION FROM BARE SOIL 107 found in fine-grained soils, such as constitute alluvial deposits, rather than in those rich in clay. Soils which have a permanent supply of water four feet below the surface are naturally in a position to secure exceptional advantages from capillary action. Such a cir- cumstance is of course uncommon ; it should result in a high condition of fertility. Evaporation from Bare Soil. When water is converted into vapour, there is always a disappearance of heat, which. performs the work of separating the molecules of water, and lifting them as vapour. Without the presence of available heat, no evaporation can take place. By heat we do not mean temperature. On a cloudy hot summer's day there may possibly be no evaporation, while on a cold winter's day evaporation may be active. If the atmosphere surrounding the moist surface is saturated with water vapour, no evapora- tion will occur without a further increment of heat, such in fact as might be afforded by sunshine. If, however, the atmosphere is only partially saturated, evaporation of water will take place however low the temperature ; and the tem- perature of the water and air will fall till the atmosphere becomes saturated, when evaporation will stop. If the un- saturated atmosphere is continually renewed, as in the case of a drying wind, evaporation will continue ; the heat demanded for the formation of vapour being supplied by the cooling of the air and water. The popular statement that evaporation produces cold, is thus quite true. If one pint of water is evaporated from 97 pints, the remaining 96 pints will have fallen 10° F. in temperature, or an equivalent amount of heat must have been supplied by surrounding bodies. The best way of approaching the subject of evaporation io8 RELATIONS OF SOIL TO WATER from soil is to consider in the first place the simpler case of the evaporation from a water-surface. Greaves (Proc. fnstit. Civil Engineers, Feb. 29, 1876) has determined for many years the evaporation from a tank having one square yard of sur- face ; the tank was kept floating in a stream of water, the temperature of the water in the tank was thus similar to that of the bulk of water surrounding it. The average amounts of evaporation he observed were as follows : — TABLE XVIH EVAPORATION FKOM A WATER SURFACE NEAR LONDON. AVERAGE OF FOURTEEN YEARS (GREAVES) Rainfall. Evaporation Evaporation Plus or Minus Rainfall. January .... February .... March April .... o May . inches.* 2-87 1.60 1.94 143 2-06 inches. 076 0.60 1-07 2-10 2-75 inches. -2-11 -1.00 -0.87 + 0.67 + 0-69 June 2-21 3-14 + 0-93 July .... 1-77 344 + 1-67 August September .... October . . 2.33 2-35 2-73 2-85 1.61 1-06 + 0.52 -0.74 -1.67 November .... December .... 2-02 242 0-71 0-57 -1-31 -1-85 Whole Year . . . 25.73 20-66 -5-07 We see that the average amount of evaporation from the water-surface has been 20-66 inches in the year. The rate of evaporation is very different in the different seasons. From November to February the evaporation is only 0-6-0-7 inch per month. After February a rapid rise sets in, the largest EVAPORATION FROM BARE SOIL 109 evaporation being reached in June and July, namely 3-14 and 3-44 inches. After July the evaporation again diminishes till the winter minimum is reached. In five months of the year the evaporation exceeds the rainfall, but in the whole year it is less than the rainfall by 5 inches. The rate of evaporation from a perfectly saturated soil destitute of vegetation, is somewhat greater than from a water-surface, the rough soil exposing a larger area of surface than the level surface of the water. The average evapora- tion from the bare loam in the Rothamsted drain-gauges (Table XXII) during six winter months (4-8-5-2 inches) is quite similar to that from a water-surface (4-8 inches), but in the six summer months the evaporation from the Rothamsted loam (11-1-11-5 inches) is distinctly less than that from water (15-9 inches), as the soil is at this season seldom saturated. The annual evaporation from the Rothamsted bare loam is singularly unaffected by the amount of the rainfall. The following table shows the results obtained during the first TABLE XIX EVAPORATION AND PERCOLATION FROM BARE LOAM, ROTHAMSTED, DURING NINE YEARS Season Oct.-Sept. Rainfall. Soil 20 inches deep. Soil 60 inches deep. Evaporation. Percolation. Evaporation. Percolation. inches. inches. inches. inches. inches. 1873-4 22-9 17-3 5-6 18-9 4-0 1871-2 26.3 18.4 7-9 194 7-2 1870-1 29-3 18.1 11.2 22.4 6.9 1874-5 30-8 18-3 12-5 20-0 10.8 1872-3 31-6 16-6 15-0 19-4 12-2 1877-8 32.6 18-0 14-6 17-9 14-7 1875-6 34.2 18.0 16-2 17-4 16-8 1876-7 35-8 18-3 17-5 17-4 18-4 1878-9 42-7 17.2 25-5 17-5 25-2 no RELATIONS OF SOIL TO WATER nine years after the establishment of the gauges. The years are arranged in the order of their rainfall. The rainfall in the seasons mentioned shows a large range of variation, 2^-9-42-7 inches; but the amount of the evaporation, though varying somewhat from year to year, seems quite independent of the amount of rain. The large evaporation credited to the deeper soil in 1870-1, is doubtless above the truth. The gauges were constructed in the dry summer of 1870, and the blocks of soil were during the construction suffering evaporation from the side as well as from the surface. The comparative constancy of the rate of evaporation from the bare soil is doubtless due to the fact that the two factors which tend to produce a large evaporation do not occur to- gether. In a wet season there is an ample supply of water to be evaporated, but evaporation is hindered because the sky is cloudy and the temperature low ; while in a fine hot season evaporation is checked as soon as the surface of the soil has dried, and its amount is controlled by the scantiness of the rainfall1. All descriptions of soil will evaporate a similar amount of water when they are in a perfectly saturated condition. In the case of very permeable soils, consisting of coarse particles, this condition continues only a short time after rain has ceased ; when this point is passed, the capacity of a soil to evaporate water will largely depend on the quantity of water it is capable of retaining near the surface. 1 The amounts of evaporation shown inTableXIX have not been maintained in subsequent years (compare Table XXII). This diminution of evaporation, with its consequent increase in percolation, is apparently due to the fact that the soils have been left undisturbed though kept free from weeds, and their surface is now more occupied by stones than at first ; a slight growth of moss has also taken place. EVAPORATION FROM BARE SOIL m An extreme instance of the behaviour of a very permeable soil is furnished by the drain-gauge filled by Greaves with fine gravel, prepared by sifting through a screen with eleven wires in two inches (Table XXII). In this very coarse soil, with an extremely free percolation, the winter evaporation is only i -2 inches, and that for the whole year 4-2 inches. The amount of evaporation is thus largely diminished when only a small quantity of water is retained near the surface. The speed with which a soil dries depends greatly on its mechanical condition. When a soil is in an open, loose condition, and is thus readily permeable to air, evaporation may go on within its mass as well as on its external surface. A soil thus dries quicker at the surface when in the crumbly pulverulent state known as ' good tilth,' and the general effect of tillage is in the same direction. The practice of ploughing clays in autumn, and leaving the land in ridges through the winter, not only yields a better tilth, but also a drier seed-bed when the land is harrowed in spring. Soils composed of coarse particles not only retain but little water, they also dry quickly at the surface. Evaporation from the soil is greatly diminished when it is shaded from the sun's rays, and protected from wind. The effect of protection from sun and wind in diminishing evapor- ation, is seen most strikingly in the case of a forest soil. Ebermayer (Lehre der Waldstreu, 182) describes experiments made during five years, at six forest stations, on the com- parative amount of evaporation from artificially saturated soils within the forest, and freely exposed without. During the six months, April to September, the evaporation within the forest was on an average only 47 per cent, of that observed from similar soil in the open. Land covered by a 112 RELATIONS OF SOIL TO WATER growing crop is generally moist on the surface. This main- tenance of a moist surface is of importance to fertility in many ways ; it is only under this condition that the nitrifying and other organisms can discharge their functions actively. Protection from wind by means of hedges diminishes the rate of evaporation. King (Wisconsin nth Rep.> 309) deter- mined the rate of evaporation from a known surface of filter paper, constantly supplied with water (Piche's evapororneter), at various distances from a hedgerow, and other forms of shelter. The evaporometer was placed i foot above the ground. The trials were made on different days, mostly in bright sunshiny weather. The following figures show the number of cubic centimetres of water evaporated in the same time at various distances from shelter. Water evaporated at various distances from shelter 20ft. 100 ft. 150 ft. 200 ft. 300 ft. 400 ft. c.c. c.c. c.c. c.c. c.c. c.c. Oak Grove 11.1 H-3 15-7 18-5 18.5 Hedge 10-3 ., .. 12-5 ., 134 Clover Field 9.3 12-1 13.0 The effect of the Oak Grove was thus felt at a distance of more than 200 ft., but beyond 300 ft. its influence ceased. The influence of the scanty hedge, and of the clover field, was clearly felt at distances exceeding 150 ft. The effects of shelter, even of so low a kind as that afforded by growing clover, are clearly considerable ; they are due partly to the diminished velocity of the wind at the surface of the soil, and partly to the fact that the wind is less dry after passing through the shrubs and plants forming the shelter. On both these points King has made experiments. The evaporation from the soil may be considerably dimin- EVAPORATION FROM BARE SOIL ished by protective coverings. Stones are effective in this way ; on turning over a large stone in summer time the ground will generally be found moist underneath1. Mulching — or covering the soil with a layer of farmyard manure, straw, dead leaves, or cocoa-nut fibre — is extremely effective, the surface being in this way thoroughly protected from both sun and wind. In the field, valuable results may be obtained by repeated shallow cultivation, by which a few inches of loose soil are permanently maintained at the surface during summer time ; this plan is largely followed in hot climates, and on its use the success of the crop often depends. King has made many experiments on this point, one of which we will describe (Wisconsin 8th Rep., 105). A field which had been ploughed and harrowed in the spring was divided into alternating strips, each 12 ft. wide. One set of strips was rolled on May 14, the intermediate ones were cultivated frequently to a depth of three inches. The percentages of water found at different depths during the summer were as follows : — TABLE XX WATER PER 100 OF DRY SOIL IN SOILS ROLLED OR CULTIVATED (KING) First Foot. Second Foot. Third Foot. Rolled. Cultivated. Rolled. Cultivated. Rolled. Cultivated. May 29 June 9 per cent. 154 13-6 per cent. 17-2 16.9 per cent. 16-6 13-7 per cent. 16.4 15.8 per cent. 14-5 14-3 per cent. 14-3 14.1 June 17 12-0 16-0 14.1 16-1 14.5 14-4 June 20 .15-0 19-0 14-8 16-8 14-6 . 14-0 July 17 11.8 14-1 14-2 15-9 13-9 14.6 1 The beneficial effect of stones in diminishing evaporation from the soil is greatest when the stone, as in the case of flint, is impermeable to water. H4 RELATIONS OF SOIL TO WATER On June 17 and 20, the excess of water in the first 2 feet of the cultivated soil amounted to i inch. The influence of cultivation did not apparently extend to the third foot till after this date. For this method of preserving water to yield its best results it should be commenced early. The soil at the end of winter is generally nearly saturated with water ; if this water is to be preserved for the use of a crop the cultivation of the surface should be commenced as soon as a loose pulverulent layer can be obtained. The most effective plan according to King is the inversion of a thin layer of the surface soil ; but as this loose surface layer requires to be renewed after each heavy rain, the use of a cultivator, horse-hoe, or grubber, is probably the most practical proceeding. In the United States it is usual to cultivate the soil in this manner between the rows of maize till the height of the crop forbids further action. Evaporation from the soil will be more effectually checked by covering it with a mulching of manure or straw than by the method just described, but the system of mulching is one for the garden rather than the field. Trials made at the Experiment Station, New York (New York >jth Rep., 1 80), will illustrate this part of the subject. An uncropped loam, kept free from weeds, was divided into plots ; each plot was in duplicate. One pair of plots remained untouched ; three others were cultivated to various depths ; a fifth was covered by a i -inch mulch of short oat straw. The experiment began in May, and continued till the end of September. Samples of soil i ft. in depth were taken every week by boring, and the Stones of porous material, as many limestones, are less effective ; indeed slight showers may be held in such cases by the stones, and the water after- wards evaporated without benefiting the soil. EVAPORATION FROM BARE SOIL 115 water present determined. If we look at the mean water contents of the soils in the five weeks in the summer when the untouched soil was driest, we shall perceive most clearly the effect of these various attempts to preserve the water in the soil. The average water contents in the driest portion of the season was as follows : — Water per 100 of Dry Soil Untouched Soil. Surface k%)t stirred to : Oat Straiv. half an inch, two inches. four inches. one inch. 16-9 ... 19-0 ... 19.2 ... 20.3 ... 22-8 Thus keeping only half an inch of the surface stirred had a very distinct effect in preventing evaporation. The effect is increased, but not very greatly, by a deeper stirring. The maximum result is gained by the mulching with straw, one inch of straw proving far more effective than four inches of loose soil. The excess of water in one foot of the mulched soil was about equal to one inch of rain, and the influence of the mulch would doubtless be felt in both the second and third foot. Mulching should not be too deep, else slight showers will be retained by the mulch and never enter the soil. Mulching has the further advantage of effectually preventing the puddling of a clay soil by heavy rain. The accumulation of dead leaves, seed vessels, &c., upon the soil of a forest, is of great importance to the fertility of the land, and its maintenance is one of the cares of scientific forestry. One of the benefits derived from this layer of forest litter is a diminution in the evaporation from the surface ; it acts in fact as a natural mulch. Ebermayer, in the experiments already referred to (p. in), found that the evaporation from an artificially saturated soil within the forest when covered by litter, was in the summer months only 46 per cent, of that I 2 n6 RELATIONS OF SOIL TO WATER from a soil similarly circumstanced but without litter. Taking the evaporation from a saturated soil during the six summer months as 100 outside the forest, it amounted to 47 within the forest, and to 22 when the soil within the forest was covered by litter. The moss which covers the ground in pine forests when light gains access, acts equally as a mulch. It may, however, become injurious if it attains too great a thickness. Rain is then retained by the moss instead of entering the soil. The forest litter of dead leaves, &c., does not become injurious by accumulation, as it forms by decay a new surface soil in which the roots of trees distribute themselves. On a steep hill-side both moss and leaf-litter discharge another beneficial function, retaining the water of heavy rains which would else be lost. The effect of tillage on evaporation has been already partly discussed. Any loosening of the texture of the surface soil favours the more rapid drying of the disturbed layer, but may, as we have seen, preserve the store of water beneath. When tillage is performed with this end in view the layer of soil loosened must be shallow ; any deep tillage in summer time is out of place if the preservation of soil water is an object. The effect produced by rolling after harrowing requires some notice. Rolling consolidates the soil, and the result is that for a time the quantity of water at the surface is increased. It is obvious that the consolidation of a loose soil increases the quantity of water in a given volume. By rolling the surface soil is also brought into more intimate contact with the moist subsoil, and the transference of water becomes easier. Moreover we have just seen that a solid soil does not become dry at the surface so quickly as a loose one, owing EVAPORATION FROM BARE SOIL 117 to its less permeability to air. For a time then the rolled soil is moister at the surface than one left rough, and this fact is made use of by the farmer when sowing seeds in spring or summer. King, when sowing oats and barley broadcast, and then harrowing, found that on the portions of the land rolled a greater number of seeds germinated, and that germination was quicker than on the unrolled land. When however the same seeds were drilled, and thus buried deeper in the ground, a subsequent rolling was without advantage. Rolling is undoubtedly often of groat use in aiding the germination of turnip seed in a dry summer. A consolidated condition of the surface cannot however be usefully maintained. We have already seen that the amount of evaporation is in the long run greater in summer time from a consolidated soil, than from one covered by a thin layer of loose earth. This is due in part to the greater velocity of the wind over a smooth surface, and in part to the more continuous supply of water at the surface in a consolidated soil. There are also other distinct advantages from a loose surface. The farmer is thus adopting the best plan when he hoes the land as soon as the turnip plants are sufficiently grown. The hoeing is not effective merely because the plants are thinned and the weeds destroyed, its benefit is partly due to the renewal of the loose condition of the surface soil. The effect of saline matter in the soil, or of additions of saline manures, on the rate of evaporation, must in some cases be considerable. In the experiments of Johnson and Armsby, and of King, already quoted (pp. 97, 101), the for- mation of a slight crust of salt on the surface greatly lessened the evaporation, presumably by choking the interstices of the n8 RELATIONS OF SOIL TO WATER soil. From this point of view, salts of little solubility, as gypsum, should be those which most effectually hinder evaporation. Investigations on this subject are much needed ; but it appears quite likely that the results obtained from manuring experiments in dry seasons are a good deal com- plicated by the effect of the manures on the water contents of the soil, an effect which may be as strongly manifested by a saline manure supplying little plant food as by one rich in such constituents. The effect of farmyard manure is complicated by a variety of circumstances. Applied as a top dressing it acts excellently as a mulch, diminishing the evaporation from the soil. Ploughed in in spring or early summer, during dry weather, its effect may be to dry the surface soil considerably, its bulky nature and loose texture greatly aiding the drying effect of wind. The permanent effect of the manure is decidedly to increase the capacity of the soil for retaining water, owing to the humus produced by its decay (p. 72); but this effect is confined almost entirely to the surface soil with which the manure is mixed. Influence of a Crop on Evaporation. We have considered hitherto the case of a bare soil, and have discussed the various circumstances influencing the amount of evaporation from it. The factor, however, which more than any other determines the rate of evaporation is the presence or absence of vegetation. When the soil is covered by vegetation, a portion of the rain does not reach the ground, but remains on the leaves and is evaporated from them ; this loss probably reaches its maximum in the case of a forest. According to Weber's observations in Switzerland, Prussia, and Bavaria, the per- INFLUENCE OF A CROP ON EVAPORATION 119 centage of rain water intercepted by the foliage and branches of various kinds of forest was as follows : — Larch 15 per cent. Spruce Fir 24 per cent. Beech 19 „ „ Scotch Pine 30 „ „ The evergreen trees were thus, naturally, the most effective in retaining rain on their leaves. Some compensation for this kind of loss will occur in the case of low growing crops freely exposed to the sky ; these condense the moisture of the air as dew, and a part of this falls to the ground. The retention of rain on the leaves is, however, generally a point of minor importance ; the principal loss of water occasioned by vegetation is due to the evaporation of water from the surface of the plant, chiefly through the stomata on the under side of the leaves. This transpiration of water by the plant is a part of its life-functions, and is indeed to a certain extent proportionate to the amount of growth ; the larger is the crop, the greater being the amount of water evaporated by it (p. 52). It is often supposed that a soil covered by a crop is moister than a bare soil ; this may be true of the soil surface, which is thus shaded from the sun and protected from wind (p. 1 1 1 ), but it is not true of the soil as a whole 1. The evaporation of water from the soil particles may be diminished by covering the land with a crop, but the evaporation through the leaves of the crop which takes its place is so much greater that the total evaporation is much increased. A crop in fact dries the soil through its roots, and the greatest part of 1 The shading of forest soil is recognized as most important to its fertility, and apparently, to the storing of water in it. The influence of this continuous shading is indirect ; it leads to the accumulation of a layer of humus upon the surface of the ground, which acts as a mulch, and also greatly favours the retention of water. 120 RELATIONS OF SOIL TO WATER the water is thus removed from below the surface. A bare soil dries at its surface, and generally only to a slight depth, while a cropped soil dries from below, and often to a con- siderable depth. The distribution of the roots has a great influence in deter- mining the amount of water available to a crop, and the extent to which the soil is dried. Very deeply rooted crops, as lucerne and red clover, draw their supply from so great a depth of subsoil that they are practically independent of summer rains. Such is also the case with many forest trees. With such a crop as wheat, the extent of the development of the roots determines often the whole difference between a good and bad crop. The history of the Rothamsted wheat field shows that the best crops are obtained after a mild winter, followed by an early spring, especially when these seasons are rather dry. Under such circumstances the wheat finds itself in May provided with a maximum of root develop- ment, and it will then require little rain afterwards for its maturation. On the other hand, cold and wet weather during the early part of the plant's life prevents the development of the roots, and the crop consequently suffers in the first drought that occurs. Roots will not develop in a saturated soil : on heavy land, a dry spring goes far to ensure a good crop of corn. Examples of the quantity of water consumed by crops have been already given (p. 51), and attention has been called to the fact that the capacity of a plant to evaporate water increases with the amount supplied. A good illustration of the drying effect of a cereal crop is afforded by the deter- minations of water in land growing barley, and in adjoining fallow ground, made at Rothamsted during the drought of INFLUENCE OF A CROP ON EVAPORATION 121 1870 (Jour. Roy. Agri. Soc. 1871, 121). The samples of soil were taken on June 27 and 28. About three-quarters of an inch of rain had fallen in the ten days preceding the taking of the samples. TABLE XXI WATER PER CENT. OF SOIL AFTER BARLEY AND AFTER FALLOW AT ROTHAMSTED Barley Land. Fallow Land. Excess in Fallow. First 9 inches . . per cent. 11.9 per cent. 204 per cent. 8-5 Second 9 inches 19.3 29.5 10-2 Third 9 inches . . 22.8 34-8 12-0 Fourth 9 inches 25-1 34.3 9.2 Fifth 9 inches . . 27-0 31-3 4-3 Lawes and Gilbert calculate that the barley crop had evaporated 909 tons of water per acre from fifty-four inches of soil, an amount almost exactly equal to nine inches of rain. A comparison of the amount of evaporation from a bare soil, and from one covered with turf, is afforded by the results obtained in the different drain-gauges summarized in Table XXII. We have already referred to the amounts of evaporation from a water surface, from the surface of bare loam with flints, and from fine gravel. The results obtained by Greaves with a turfed sandy loam three feet deep, and by Evans with a turfed soil also three feet deep, show a greater evaporation in the whole year (18-1-20-0 inches) than that from the bare loam at Rothamsted (15-9-16-7 inches). This difference, however, by no means represents the full evaporating power of the grass turf. The average rainfall in the turf experiments was indeed much less than 122 RELATIONS OF SOIL TO WATER that received at Rothamsted. The turf evaporates the whole of the summer rainfall, percolation only occurring in very heavy storms. In three summers during Sir John Evans' experiments, the rainfall in six months amounted to 15-16 inches ; yet in two of these summers no percolation occurred through the turfed soil. The rainfall has thus been insufficient to show the full evaporating power of the turf. In the winter months, when a greater excess of water is available, the turf has evaporated considerably more than the water- surface. TABLE XXII EVAPORATION AND PERCOLATION UNDER VARIOUS CIRCUMSTANCES Rainfall Evaporation. Percolation. per Annum. Summer, April- Sept. Winter, Oct.- March. Whole Year. Summer, April- Sept. Winter, Oct.- March. Whole Year. in. in. in. in. in, in. in. Water Surface (Greaves)1 25-7 15-9 4-8 20.7 Loam 20 inches (Lawes and Gilbert)2 . . . 30-3 11.1 4-8 15-9 4.3 10-1 144 Loam 60 inches (Lawes and Gilbert) 2 . . . 30-3 11-5 5-2 16.7 3-9 9.7 13-6 Fine Gravel (Greaves) l . 25-7 3-0 1.2 4-2 9-1 12-4 21.5 Turfed Soil (Greaves) l . 25-7 114 6-7 18-1 0-7 6-9 7-6 Turfed Soil (Evans) 3 . 25-6 12.1 7.9 20-0 0-4 5-2 5-6 Another comparison of turf with bare soil is afforded by the results of the drain-gauges, three feet deep, established for four years, 1883-6, at the Experimental Station, Geneva, New York. The rainfall in this case was again far too small to exhibit the evaporating power of the turf. 1 Average of fourteen years. 3 Average of twenty years. 3 Average of fifteen years. INFLUENCE OF A CROP ON EVAPORATION 123 TABLE XXIII AVERAGE EE8ULTS DURING FOUR YEARS OF THE GENEVA (N.Y.) DRAIN-GAUGES Rainfall. Evaporation from Turf. Bare Soil. Cultivated Soil. inches. 25.0 inches. 21-2 inches. 174 x inches. U-01 In 1883 the evaporation from the turf reached 23-6 inches, out of a rainfall of twenty-four inches. The smallest evapora- tion in this experiment is from the cultivated soil, the surface of which was maintained in a loose condition, thus affording another illustration of the success of this plan for preserving soil water. An additional illustration of the evaporating power of turf is furnished by the series of daily determinations of water in various soils conducted by the Agricultural Department of the United States in 1895 (Soils, Bulletin 2, 8). Towards the end of June the amount of water in the surface foot of a blue grass pasture varied between 9 and 10 per cent., while in similar soil with the sod removed it amounted to 19-20 per cent. It follows from what has now been said, that when a soil is treated as a bare fallow, and left for a whole season without a crop, being submitted during that time to frequent ploughing, it is (notwithstanding its free exposure to sun and wind) in 1 The total drainage from these gauges for 1886 is apparently wrongly stated on p. 348 of the New York Fifth Report. The figures for that year which are included above have been obtained by summing the monthly items on p. 347. 124 RELATIONS OF SOIL TO WATER circumstances in which water will be stored up in the soil. This is indeed a fact, and the storing- of water is to be reckoned among the benefits resulting from bare fallowing, a practice which is well known to be attended with most success in dry climates. King (The Soil, 191) determined on May 13 the percentage of water in two portions of a field about to be sown with maize ; one portion had been previously a bare fallow, the other had carried clover. The results were as follows : — Water per 100 of Dry Soil in Fallow and Clover Land Depth 1-6 inches. Depth 12-18 inches. Depth 18-24 inches. Fallow ... 23-3 19-1 16-9 Clover Land 9-6 14-8 13-8 The maize sown on the fallow land would thus be far better able to withstand a summer drought, than the maize following clover. The practice of planting so-called 'catch crops' immediately after harvest, and ploughing them in as green manure in spring, is a system full of advantage so far as the conservation of soil nitrogen is concerned ; but King points out that in a dry climate such cropping and spring ploughing may be actually injurious, from the loss of soil water which they entail ; and that in such a climate pains must be taken to plough in a green crop very early, as only then will it find sufficient moisture in the soil for its decomposition, and the tillage be accomplished without robbing the soil of water. The injurious effect of weeds has a new light thrown upon it when we see that their growth dries the soil, and robs the crop of the water which it might have obtained. In the INFLUENCE OF A CROP ON EVAPORATION 125 Woburn Experimental Fruit Farm the ground surrounding certain young apple trees has been differently treated, in some cases grass has been sown round the tree, in other cases natural weeds have been allowed to accumulate, while in other cases the land has been kept clean throughout the summer by hoeing. The difference in the growth of these trees during the dry summer of 1896 was most marked, the trees standing in the bare soil worked with a hoe making an abundance of new wood, while hardly any growth took place in thejcase of the trees surrounded by vegetation. Weeds are far more injurious in a dry climate than in a wet. The amount of evaporation from a soil largely affects the amount of percolation through it. In the case of ordinary permeable soils, the amount of rainfall and of evaporation determine the quantity of water which will pass through and appear as drainage. In Table XIV we see the influence of evaporation on percolation. In the winter, 70-80 per cent, of the rain comes through the drain-gauges at Rothamsted, while in summer only 20-27 per cent, passes through the soil. A great part of this summer drainage is derived from heavy storms, a portion of the water passing through the open channels in the soil before the body of the soil is saturated. The same quantity of rain in the form of light showers would produce no percolation. In Table XIX we have seen striking illustrations of the fact that the annual evaporation from a bare soil is a fairly constant quantity, so that all rainfall over this amount appeal's as drainage. Doubling the rainfall has in fact increased the drainage five or six fold. If, however, land is covered with vegetation, the amount of evaporation ceases to be a constant quantity, every increase in rainfall up to an excessive amount 126 RELATIONS OF SOIL TO WATER determining a greater growth, and a greater evaporation from the vegetation covering the surface. Thus the growth of forests tends greatly to diminish the volume of springs and rivers, and the destruction of forests is often followed by disastrous floods. We have seen that on the bare soil at Rothamsted, active drainage commences in October, and lasts five months. In the wheat field at Rothamsted, a running of the drain-pipes is rare between March and September, owing to the drying of the land by the crop. The drain-pipes do not usually commence to run freely till November, and active drainage is limited to four months. With crops which cover the land in autumn and winter, drainage is reduced to a minimum. The farmer has thus the power, by suitable cropping, of greatly diminishing the autumn and winter drainage ; this power is of great value, as the drainage water removes from the soil considerable quantities of plant food, especially calcium salts and nitrates. Underground Water. The water which passes through the soil accumulates in the subsoil at very varying depths, forming a perfectly saturated stratum of soil or rock. The height of this saturated stratum varies with the character of the soil, and also with the rainfall and the season. When the height of this saturated stratum reaches a certain point, a discharge generally takes place in the form of springs, or as a general drainage from a wider area, into the river valleys, and finally into the sea. If no such discharge is possible, the soil becomes saturated to the surface, and a swamp is produced. The position and behaviour of the underground water may be very much complicated by the presence of beds of impervious clay in the subsoil ; the existence of such a bed determines at once the position of the underground water, which must UNDERGROUND WATER 127 necessarily accumulate on its surface. The inclination of the clay bed determines also the direction in which natural drainage will occur. When the complications of local clay beds are absent, the level of the underground water is found to follow generally the contour of the land surface, but in a less exaggerated manner, and with a decline in level in the direction of the drainage outfall. King points out that, as a result of this greater height of the water level under high ground, the land lying at the foot of hilly ground receives a continuous supply of underground water even in time of drought, amounting under favourable conditions to a veritable sub-irrigation. When the distance of the stratum of saturated soil from the surface is considerable, this store of underground water is of no direct benefit to plants. We have already mentioned, that at Wisconsin a water level 6 ft. below the surface does not prevent the effects of drought being manifested by the crops, though doubtless they derive some benefit from the water. On the other hand, King tells us that the water level should not come nearer than 4 ft. to the surface if ordinary cereal crops are to be grown, as a higher water level prevents a proper development of the roots. With crops having short roots, as grass, the water level may with advantage be consi- derably higher. A water level may generally be lowered to any desired point by the insertion of drain pipes. The water level in a soil is raised by systems of irrigation. The height of the water level in a soil varies considerably at different seasons of the year : this is best seen by measurement of the heights of water in wells. In deep wells in the chalk at Harpenden, Herts, the water is found at its lowest level in September, October or November. The rise commences at 128 RELATIONS OF SOIL TO WATER various times, according to the preceding rainfall, and may occur in October, November, December, or, more rarely, in January. The greatest height is reached occasionally in February, but more generally in March or April, after which a decline sets in. The decline is slower than the rise. We have already seen that the commencement of large percolation through the soil is generally in October or November, and continues till the end of February ; the rise and fall of the water in the wells thus follows, at a somewhat later date, the order of the percolation through the soil. In Wisconsin, with a severe winter climate and a water level only a few feet below the surface, a great rise in the well water does not commence till April, after the spring thaw. The alterations in the height of the underground water are considerable, extending to several feet ; from three to five feet is the ordinary rise in the deep wells at Harpenden. It is at first sight puzzling that a few inches of autumn drainage should produce an increase of several feet in the height of the underground water. The explanation is however simple. The water in a soil, as we have already seen, merely occupies the interspaces between the particles. Supposing then the interspaces in a perfectly dry soil to amount to 40 per cent, of its volume, it is evident that four inches of rain would saturate ten inches of soil, and raise the level of the under- ground water to that extent. The effect produced in this direction is however really far greater, for the subsoil is not dry, but already holds a considerable amount of water; a small addition of drainage water thus suffices to complete the saturation of a considerable depth of subsoil. Besides the considerable variations in the level of under- ground water determined by the season of the year, there are, UNDERGROUND WATER 129 when the water level is near the surface, a great number of lesser variations. King, who has made an exhaustive study of the movements of water in shallow wells by means of a very exact self-registering apparatus ( Wisconsin gth Rep., 129), tells us that the water level in such wells is in summer time never still, but always moving in one direction or the other. These movements in the water level are equally seen in the varying discharge of water by springs or by drain-pipes. They appear to be mainly occasioned by the expansion or contraction of the air imprisoned within the soil between the surface and the water-level. To a less extent, they are due to the alteration in the viscosity and surface tension of the water, brought about by changes in temperature. A rise in temperature starts a fresh percolation by diminishing the viscosity and surface tension of the water coating the soil particles, while a fall in temperature causes water to rise from the saturated soil into the unsaturated. The action due to the influence of temperature on the physical properties of water is thus in precisely the same direction as that due to the expansion or contraction of included air. King found the discharge from a spring to be 8 per cent, greater with a falling than with a rising barometer, and the discharge from a drain-pipe diminished 15 per cent, for a rise of 01 inch in the barometer. The height of water recorded for the wells showed a regular daily fluctuation, the fall in level during the day being more or less made up by a rise in level during the night. The drain-pipes exhibited the same changes, the discharge reaching its maximum about 7 a.m. This diurnal variation is due to alterations in temperature ; but the effect of the maximum daily temperature was not felt till early in the next morning, owing to the slow progress of K 130 RELATIONS OF SOIL TO WATER heat through the soil. A rising temperature and a falling barometer act in the same way ; the air in the soil expands, and the water filling the interstices above the water level is expelled, and causes a rise in the water level of the soil. On the air again contracting, the water is reabsorbed by the soil, and the water level again falls. King describes a blowing well in Wisconsin, in which the expansion and contraction of the underground air is manifested iri a surprising manner. The water level is in this case at a considerable depth, and the soil between the surface and the water is chiefly gravel, which from its free drainage is of course largely filled with air. A falling barometer in this case pro- duces so violent a draught of air out of the well as to blow a man's hat off; while with a rising barometer in winter time, the cold downward current is so severe as to freeze the pipes at a depth of 70 feet. Wet and Dry Soils. The characters of dry and wet soils have been already fully described. The dry soils are those composed to a large extent of coarse particles, possessing a free percolation, and little power of retaining water. The wet soils are those composed of very fine particles, having an enormous extent of internal surface, and therefore retaining much water, and offering great resistance to the passage of water through them : the colloid constituents of clay or peat help greatly to intensify these properties. The relation of soils to water is often much modified by the character of the subsoil. A sandy surface soil is agriculturally a very different thing when it has a subsoil of loam or clay, as the presence of this greatly increases the store of water at the disposal of the crop. On the other hand, a clay soil is no longer called wet when it has a subsoil of chalk to remove WET AND DRY SOILS 131 superfluous water. The character of any soil will also be much affected by its position, whether on a hill side, or in a valley receiving the drainage of higher land. The height of the water level in the subsoil is also another condition which may entirely alter the agricultural character of the land. The most suitable character for a soil must depend on the characters of the climate and situation in which it is placed, and the crops it is desired to grow ; properties of the greatest value under one set of conditions, may be those producing most evil under contrary circumstances. A coarse sand may be a very poor soil under arable culture, but it would answer admirably for sewage irrigation, and would also in time produce a good pine forest. Well supplied with manure it might make excellent early market-garden land. A clay that could only be laid down in grass with an annual rain- fall of forty inches, might be used with great advantage for arable culture where the rainfall is only twenty-five inches. A marsh which would be useless in England, would in India produce luxuriant crops of rice. The kind of agriculture which can be most usefully adopted must indeed always depend on the special conditions of the locality. In the climate of England, the soils yielding the most favourable .supply of water for arable culture are loams, alluvial silts, and very fine sands containing some humus ; such soils are capable when deep of storing much water, while at the same time they allow of a free movement of water within them, and drain sufficiently speedily to favour a large development of root. Amelioration of the Physical Properties of Soil. Although the natural conditions of soil and climate have always a 132 RELATIONS OF SOIL TO WATER preponderating influence in determining the agricultural value of land, considerable improvements in the character of the soil may often be effected by artificial means, if we bear in mind the facts laid down in the preceding pages. As the same causes, and the same modes of cure, apply as a rule both to cases of deficiency and excess of water supply, and to cases of deficiency and excess in the tenacity of the soil, we shall most conveniently consider these subjects together. In a coarse sand we have a soil of minimum tenacity, and also one of minimum capacity for retaining water. The evils due to lack of tenacity are exemplified in the case of blowing sands. When these sands are in a bare dry condition, the wind separates the finer and more valuable particles, and then rolls over the coarser grains, which may thus be carried in great drifts to considerable distances, rendering much land infertile. If such land can be permanently covered with vegetation, the mischief we have described will be of course prevented. Where a coarse sand is found naturally covered by a pine forest, it will be a great mistake to cut down the forest and attempt arable culture. The establish- ment of perennial plants having widely spread roots is a first step in the reclamation of blowing sands. Shelter from wind must also be provided. Fences will check the progress of sand in the same way as the groins on the seashore hinder the movement of the shingle. King has investigated the question of the agricultural treat- ment of the blowing sands of Wisconsin (Wisconsin nth Rep., 292). He points out the great diminution in the velocity of the wind which results even from shelters of small height above the ground. Land left ploughed in ridges across the prevailing direction of the wind was little injured when a AMELIORATION OF SOIL 133 level soil had been badly disturbed by the wind. The shelter afforded by low-growing crops was considerable ; and he makes the very practical suggestion that such land should be always cultivated in strips, fifteen or twenty rods wide, grass or clover alternating with the arable culture. Both the tenacity and the water-holding power of coarse sandy soils may be greatly improved by a small admixture with clay ; this plan admits of being carried out with economical success when the clay is found in the subsoil of the field, or in the immediate neighbourhood. The applica- tion of marl (a calcareous clay) to light soils is a common practice in the eastern counties of England; from forty to sixty tons per acre are usually applied. The effect of this dressing is seen for many years. Peaty and fen land are also much benefited by the addition of clay, which increases the weight and coherence of the surface soil. On Rimpau's system for reclaiming peat land, trenches are cut at intervals across the peat down to the clay bottom which always underlies it ; the clay is then brought up and spread on the surrounding peat, the land being thus drained and clayed at the same time. A similar plan is made use of in the Lincolnshire fens ; the trenches are in this case eight to ten yards apart. Another very practical mode of improving the coherence and water-holding power of sandy soils is by increasing the proportion of humus. For garden purposes, or on a small scale, this may be done by digging or ploughing in heavy dressings of well-rotted farmyard manure. On a large scale, the same result may be accomplished by the ploughing in of green crops, which on their decay add greatly to the store of humus in the soil. Preference is usually given to legumi- 134 RELATIONS OF SOIL TO WATER nous crops for this purpose, as the soil is then enriched with nitrogen from the atmosphere. Lupins are commonly employed in Germany for the amelioration of sandy soil. The water-holding power of a sandy soil is increased by keeping it consolidated ; the treading of sheep in winter time is thus beneficial. In a light soil, crops should be planted as early as possible, so that the fullest root development may be attained before summer begins. As soon as the season for evaporation commences the surface of the soil should receive a shallow cultivation, and a layer of loose earth should be maintained on the surface throughout the summer. In the garden, mulching is a still more effective plan for preserving soil water. The temporary increase of water at the surface obtained by rolling the soil has been already mentioned. We turn now to the opposite case of a clay soil, exhibiting an excessive tenacity, hindering both root development and the production of tilth ; retaining an injurious amount of water after rain, and drying in summer time into a hard mass. We seek in this case to diminish the coherence of the soil, to make it more friable and more permeable to water. With very stiff clay soils it may be impossible to remedy the evils which they present in an economical manner, in which case the land is allowed to remain permanently in grass. When in grass, the clay soil becomes of great value ; the excessive supply of water near the surface is then an advantage, while under the protection of a sod a fairly good natural tilth is finally obtained. In the case of clay soils of a less extreme character a very considerable amount of improvement is practicable. Deep autumn ploughing, and exposure of the ridged surface to winter frosts, is a method of first-class importance for AMELIORATION OF SOIL 135 improving the condition of the surface soil, and obtaining a good tilth in spring. If a subsoiler follows the plough in the furrows, and penetrates and stirs the subsoil without inverting it, still more benefit will be obtained. By such deep tillage, the permeability of the surface soil is much increased, and rain passes through freely to the subsoil. The improvements effected by frost and tillage are not however very permanent ; every storm of rain tends to bring the clay back to its original condition. For practical purposes it is useless to attempt to improve the texture of a stiff clay by admixture with sand, the effect produced being far too small to render the measure economical. Clay-burning is a more practical scheme ; it may be seen frequently in operation in railway cuttings where it is desired to prevent the erosion of the banks by storm water. In the field, the dried clods of soil are made into heaps with hedge cuttings, or other combustible matter, and slowly burnt at a low temperature. The burnt clay is then spread over the land and turned in with the plough. Sixty to eighty cubic yards of burnt clay per acre is a usual dressing. It should be recollected that the nitrogenous matter in the soil is lost by burning ; the potash in the clay becomes, however, more soluble by this treatment, if the burning has not been con- ducted at too high a temperature, and especially if the clay contains some lime. Clay-burning is to be regarded as an extreme measure, to be adopted only when ordinary methods have not proved sufficiently efficacious. A method far more frequently adopted is the application of chalk or lime to the clay. The action of lime is peculiar. We have already seen (p. 31) that it brings about the coagulation of the colloid clay, and completely alters its physical character. 136 RELATIONS OF SOIL TO WATER The soils containing the smallest proportion of true colloid clay are those most successfully treated by lime. The action of the lime extends to a considerable depth, and endures for some time ; it is not, however, strictly permanent, as lime is continually being removed from the soil in the drainage water. It is usual to apply ten to fifteen tons of chalk, or three to nine tons of lime per acre. The land is best ploughed and then limed in autumn, and afterwards harrowed in the spring. The enrichment of a clay soil with humus very much im- proves its physical condition, and for garden purposes there is no better treatment for a clay soil than the digging in of large quantities of fresh stable manure. In the field, great advantage is experienced by growing clover and grass for several years, and then ploughing in the crop residues remaining in the surface soil. The methods we have mentioned serve chiefly to ameliorate the surface soil. If a deep clay soil is to be permanently im- proved it is generally necessary to have recourse at the same time to draining. The primitive methods of draining consisted in throwing the land by the plough into high ridges or ' lands,' or in cutting ditches ; the far more effective modern drainage systems are carried out by the insertion of drain-pipes in the subsoil, the water collected by these pipes being removed by a main drain into which they deliver. There are two very distinct cases in which draining is a remedy. If the surface soil is freely permeable to water, but a bed of clay occurs in the subsoil, the water collects upon the surface of this clay, and may rise to such a height as will injure the fertility of the land. To remove this water it is only necessary to tap the water reservoir above the clay bed at such a distance below the surface as will prevent the water AMELIORATION OF SOIL 137 standing at an injurious height. A few pipes generally suffice for this work, and the effect they produce is immediate 1. The second case is that of a deep clay. Here there is no water reservoir to be tapped ; the whole soil is equally full of water, and holds it firmly. Drain-pipes are usually laid in such a soil, 15-20 ft. apart, and 3 ft. below the surface. Little or no water is at first delivered by the pipes, and no improvement in the soil may be perceived for some time, the effect being very gradual. The amelioration consists first and chiefly in an alteration in the physical character of the clay, which commences round the pipe and gradually extends from it. The clay subsoil, being opened up by the introduction of pipes, is in fact placed under the influence of the changes of temperature and changes in the condition of dry ness, which we have already seen determine the gradual disintegration of a tenacious soil, and the formation of a looser texture. The forces thus brought to bear upon the subsoil are far weaker than those which ordinarily affect the surface ; but, on the other hand, the transformation in the subsoil is not hindered by the occurrence of tillage operations, or by the puddling effect of heavy rain. As this change in the texture of the clay proceeds, the soil is able to part more freely with its excess of water, and the drying thus effected serves to extend still further the alteration in the character of the clay. Drain- age operations will be especially successful when lime is at the same time applied to the surface. 1 If the impervious layer is quite near the surface, and is of the nature of a ' pan ' (see pp. 28, 29, 46), the most effectual remedy will be the destruction of this pan by a subsoil plough. Steam cultivation is in such cases of special value. In parts of the United States it has been found practicable to shatter the pan, and thus ensure the drainage of the surface soil, by means of dynamite cartridges. 138 RELATIONS OF SOIL TO WATER It may naturally be thought that crops on a drained elay will suffer more in a season of drought than those on an undrained clay, the quantity of water held near the surface being undoubtedly greater in the latter case. This is not so, indeed the reverse is found to be the fact. The roots of the crop are indeed far more widely distributed in the drained soil, and are thus far better able to obtain water. Moreover, the movement of water is more speedy in a drained than in an undrained soil. CHAPTER IV RELATIONS OF SOIL TO HEAT Influence of Temperature on Life— Sources of Soil Heat — Influence of Latitude and Aspect — Temperature of Surface Soil — Specific Heat of Soils — Con- ductivity of Soil Constituents — Kadiation of Heat — Influence of Water on Soil Temperature — Temperature of Subsoil — Prevention of Summer Frost. Influence of Temperature on Life. All the processes of life, whether in plants or animals, are only possible between certain limits of temperature, below or above which life cannot exist. Each kind of life has a range of temperature more or less peculiar to itself, within which its functions may possibly be performed ; it has also an optimum temperature at which the greatest amount of vigour is exhibited ; this optimum temperature may be different in different stages of growth. All living beings have, however, some power of adaptation to the circumstances in which they are placed. The life processes which occur in the soil are distinctly affected by the temperature of the soil ; the germination of seeds is one of the processes so affected. The seeds of some plants, as rye, mustard, and lucerne, may undergo a very slow germination at the freezing point, but for most seeds a higher temperature is required. F. Haberlandt (Landw. Ver sucks- Stationen, xvii. 104) experimented with a great variety of seeds at four temperatures ; the lowest temperatures at which germination took place were as follows : — 140 RELATIONS OF SOIL TO HEAT Loivest Temperatures for Germination Wheat, Barley, Oats, Rye, Buckwheat, Rye Grass, Peas, \ Beans, Vetches, Lupins, Red Clover, Lucerne, Mustard, I 32° - 40° F. Rape, Turnip, Beet, Hemp, Flax J Maize, Sorghum, Timothy Grass, Sanfoin, Carrot, Sunflower 40°-51°F. Tomato, Tobacco, Pumpkin 51°-60°F. Cucumber, Melon 60°-65°F. The time required to produce germination becomes shorter as the temperature rises, until the optimum temperature is passed. Haberlandt's results were as follows : — Days required for first appearance of Radicle at various Temperatures 40° 51° 60° 65° F. 1 ... 1 2 ... 1| 2| ... 2 3| ... 3 11 ... 11 2 ... 2 21 ... 2 11 ... 1 4f ... 41 1 ... 1 1 ... 1 2 ... 1| 31 ... 3| 2 ... 2 31 ... 3 31 ... 3 41 ... 31 3£ ... 3 3 ... 2 6 ... 31 9 ... 61 101 ... 4 The temperature at which germination was most speedy, and the largest proportion of seeds germinated (optimum Rye 4 2£ ... Wheat and Barley 6 3 Oats 7 3f Rye Grass 10 5i ... Peas 5 3 Vetches... 6 5 Lucerne 6 ... 3| ... Red Clover 7£ 3 Beans 7 ... 6* ... Mustard 2 H ... Rape 6 2 Turnip 8 4 Sugar-beet 22 9 Flax 8 4£ Timothy Grass 6^ ... Maize ... Ill -. CaiTot 64 Sanfoin ... 71 ... Sunflower 25 Tomato Tobacco ... Pumpkin INFLUENCE OF TEMPERATURE ON LIFE 141 temperature), and the highest temperature at which ger- mination was possible, were approximately ascertained by Haberlandt by experiments at 61°, 77°, 88°, 100°, 111°, 132° F. Optimum and Maximum Temperatures for Germination Optimum Maximum Temperature. Temperature. Barley, Vetches 61°-77D ... 88° -100° Mustard ... 61°-88° ... 88° - 100D Rye, Wheat, Oats, Timothy Grass, Beans, \ ^0 ^0 ^^ Carrot, Flax, Tobacco j Turnips 77D-88° ... 88°-100° Buckwheat 77° - 88° ... 100° -111° Red Clover, Lucerne 77°-100:> ... 100°- 111° Sorghum 77°- 100° ... Ill0- 122° Rye Grass 88° ... 88° - 10(P Lupins, Sunflower 88° ... 100°-111° Cucumber 88° ... Ill0- 122° Maize, Melon 88°-100° ... 111°-122° Rape 100° ... 100°- 111° Pumpkin 100° ... 111D-122° These results afford ample illustrations of the dependence of germination on the temperature of the soil. The warmer the soil is, the quicker will the seed germinate, and the earlier will be the ensuing crop. A warm soil in spring time is thus of immense advantage to the agriculturist. The temperature of the soil has an equally great influence on the subsequent development of the plant. Bialoblocki (Landw. Versuchs-Stationen, xiii. 424) grew barley in pots of sand maintained at various temperatures. The sand was watered with a solution of nutritive salts, the influence of heat on the chemical changes proceeding in a natural soil was thus eliminated. In the first series of experiments, the pots were maintained at their respective temperatures from the time when the seeds of barley were sown. In the second 142 RELATIONS OF SOIL TO HEAT series the barley plants were allowed to develop for two months at the temperature of the air before the temperatures of the soils were altered ; in consequence of this change in the method, the ill effect of the higher temperatures is less apparent in this series than in the first. TABLE XXIV INFLUENCE OF SOIL TEMPERATURE ON THE YIELD OF BARLEY Temperature of Soil. Dry Matter in Final Produce. Series I. Series II. 10° C. or 50° F. grams. 7-64 grams. 7-33 20° „ „ 68° „ 30° „ „ 86° „ 40° „ „ 104° „ 8-22 3-85 0-93 9-15 5-33 3.47 A soil temperature of 50° F. was apparently sufficient for the normal development of the barley plant, but the ripening of the crop was slow, and none of the corn was more than milk-ripe when the experiment concluded. The soil tem- perature of 68° gave a larger produce and better ripened ears. With much higher soil temperatures the amount of produce was much reduced. In an experiment lasting only twenty days, it appeared that the optimum soil temperature for wheat was somewhat higher than that for barley, and this again somewhat higher than that required by rye. The great influence of soil temperature on plant growth is well known to gardeners, who frequently employ hot-beds, and make use of a bottom heat for striking cuttings and INFLUENCE OF TEMPERATURE ON LIFE 143 other purposes. In extreme northern latitudes the low temperature of the soil limits both the kind of crops which can be grown, and the amount of produce obtained. The much greater length of day in summer time does not com- pensate for the deficiency of soil temperature. Besides the direct influence of the temperature of the soil on the growth of crops, it has a very considerable influence on the activity of the lower forms of life with which a fertile soil abounds. The dead organic matter of the soil, the remains of vegetable and animal tissue, is made again avail- able as plant food by the successive action of various kinds of fungi and bacteria. The production of nitrates in the soil, a process having a most intimate connexion with its fertility, is, for instance, brought about by the successive action of several species of bacteria. The activity of these living agents is entirely dependent on the temperature of the soil in which they live. At the freezing point their action is practically nil ; it increases as the temperature rises till the optimum temperature is passed, when a rapid decline sets in. Vital action seldom continues beyond 50° C. The soil tem- perature most favourable to the chemical activity of bacteria varies with different species ; it generally lies between 30° and 40° C. The temperature of the soil also affects the physical pro- cesses occurring in it. We have already noticed (pp. 87, 129) the great influence of temperature on the movements of water and air within the soil. Temperature also affects the move- ments of salts (p. 193). Indeed, there is probably no physical process within the soil which is not affected by temperature. The chemical processes within the soil are equally influenced by temperature. The chemical changes in dead matter which 144 RELATIONS OF SOIL TO HEAT occur without the intervention of living organisms are in fact confined to certain limits of temperature, and are pro- moted or retarded as the temperature approaches or recedes from the optimum point, in a similar manner, though within far wider limits, as we have already seen happens in the case of the chemical changes produced by the agency of life. Thus, speaking generally, the whole of the processes within the soil become more active as the temperature rises. Winter is a time of sleep, summer is a time of activity. The productive power of a soil depends largely on its temperature. Sources of Soil Heat. The surface of a globe moving alone in space must very shortly reach a condition of intense cold. The present temperature of the surface of our earth is almost entirely maintained by the radiation received from the sun. To a very small extent the temperature of the soil will be due to the heat evolved in chemical and physical actions ; this heat is, however, in most cases merely a reappearance of solar energy previously consumed in the production of chemical or physical work, which afterwards, by a reverse action, is resolved into its original elements. To a small extent, the temperature of the surface of the earth is also raised by the gradual outward passage of internal heat. The great internal heat of the earth is a familiar fact, shown by the rise of temperature in the rocks when mines are sunk into the earth, and also by volcanic phenomena. The rate of increase in temperature on sinking below the surface is not uniform in every place. If we assume as an average a rise of i° F. for each 50 or 60 feet of descent, we shall have a temperature equal to that of boiling water at about a mile and a half below the surface. It is impossible to say SOURCES OF SOIL HEAT 145 to what extent the temperature of the surface soil is affected by the internal heat of the earth. The average temperature of the surface soil is in England about 1° F. higher than the temperature of the air above the surface, and the average temperature of the subsoil becomes slowly higher as we descend. Since, however, the air at the surface of the earth derives almost the whole of its heat from contact with the earth, and is itself cooled by mixture with the air of the upper regions of the atmosphere, the surface of the earth should be a little warmer than «the air even if the whole of its heat was derived from solar radiation. The heat derived from the interior of the earth will be equally dis- tributed through all the seasons of the year ; it will probably differ in amount in different places, owing to the greater or less thickness and conductivity of the earth' s«crust. We have already seen, p. 62, that a thoroughly dried soil rises somewhat in temperature when moistened, especially if it is rich in colloid constituents. A still greater rise in temperature is observed when a soil condenses water from the air. Under special circumstances, considerable quantities of heat may be generated by chemical action. It has been well said that when a log of wood decays in the forest it produces as much heat as when burnt in a furnace. If the products — water and carbonic acid — are the same under both conditions, the heat produced must also be the same. The heat evolved is, however, in one case spread over many years, and in the other case is probably concentrated into one hour. The rise of temperature is thus in the first instance imperceptible, and in the latter very great ; the total quantity of heat being the same in each case. L 146 RELATIONS OF SOIL TO HEAT If in place of a log of wood we deal with some vegetable matter which more easily undergoes chemical change, the rise of temperature during fermentation and decay may become very perceptible ; we have an excellent example of this in the case of the heating of a damp hayrick or of a silo. For the evolution of heat it is by no means essential that oxidation to water and carbonic acid should take place. An extreme oxidation, having these final products, will indeed produce the whole of the heat which the fuel substance is capable of yielding ; but heat is also produced by fermenta- tive changes in which oxidation plays an insignificant part. In the alcoholic fermentation of sugar, by far the greatest part of the sugar is simply split up into two bodies, alcohol and carbonic acid, no oxidation of the sugar occurring, and the action taking place in the absence of air, yet the pro- duction of heat is well marked throughout the whole opera- tion. The evolution of heat in the stack, the silo, and the hot-hed, is generally the result of fermentive changes. Only a portion of the stable manure or hay is capable of rapid fermentation ; the active production of heat thus soon ceases, and the subsequent changes become very slow, if a secondary action does not occur leading to ignition. Nearly the whole of the processes bringing about the natural destruction of organic matter, within the soil or out of it, and whether by means of fermentation or oxidation, are the work of living organisms —animals, fungi, yeasts, bacteria; that this is the case is proved by these actions ceasing when the organisms present have been destroyed. As these agents can only work within certain limits of temperature, a rise of temperature beyond a certain point destroys the agent, and brings the work to an end. Thus SOURCES OF SOIL HEAT 147 it is well known that sweet silage is produced by allowing the mass of green matter to heat in the first place to a tempera- ture at which the bacteria producing lactic acid are destroyed, and then checking the action. Both in the silo and in the hot-bed it is however quite easy to obtain temperatures considerably exceeding those at which most organisms perish. It is now known that a special class of bacteria exists in the surface soil, and is probably widely distributed, which is capable of living and performing energetic work, at temperatures distinctly exceeding 70° C. (158° F.1). These bacteria doubtless take an active share in the chemical changes producing heat both in the hot-bed and in the silo. Dybowski (Annales agronomiques, 1887, 268) has made careful experiments on the course of the development of heat in hot-beds containing a 2 ft. layer of various materials. The highest temperature, 75° C., was obtained from horse manure ; the lowest, 37°C., from a mass of mixed dead leaves. For the maximum heat to be obtained the material must be in a fresh state. Cases sometimes arise in which the actual ignition of a haystack or manure heap occurs. The rise of tempera- ture above 75° is clearly the result of chemical oxidation, carried on without the aid of living organisms. We must assume that fermentation has produced substances which at the temperature of the rick are rapidly oxidized if air can obtain access, and that the production of heat is thus carried to the point of actual combustion. 1 These high-temperature bacteria are readily separated from the other organisms in soil by inoculating with soil tubes of broth maintained at 65° C. Under these circumstances only the high-temperature organisms will develop. L 2 148 RELATIONS OF SOIL TO HEAT Although farmyard manure furnishes a considerable source of heat when employed in a hot-bed, its application has little influence on the temperature of farm soils. This is chiefly owing to the small proportion of farmyard manure which is mixed with the soil. The farmer also seldom applies the manure in a fresh state ; a portion of its heat-producing power is thus lost before it reaches the land. A few experiments were made by Georgeson at the Imperial College at Tokio (Agricultural Science, i. 251) on the altera- tion in temperature which followed the application of various quantities of farmyard manure to the soil. The soil used was peculiarly light and porous, being in fact a volcanic ash. The manure was partly decayed, but still rather long. The soil and manure were well mixed, and wooden frames were filled with the mixture to a depth of i ft. ; one frame received no manure. The frames were sunk in the open ground, the top of each frame being level with the surface. The experimental soils were thus under perfectly natural conditions in respect to rainfall, drainage, &c. The average temperatures of the soils, in successive five-day periods, are shown in Table XXV. The temperatures are given in Fahrenheit's degrees. The increase in temperature was greatest during the first five days, and then rapidly diminished. After the first fifteen days the increase became almost imperceptible where only ten tons of manure per acre had been applied, but con-' tinued to be distinct after twenty-five clays when forty and eighty tons of manure had been employed. As an ordinary dressing of farmyard manure does not exceed ten tons per acre, and a very liberal dressing seldom exceeds twenty tons per acre, it would appear that the rise of temperature pro- SOURCES OF SOIL HEAT 149 duced in the soil is not great, and soon ceases ; it may, however, have a distinct effect in hastening the germination of seeds, and must aid in protecting a spring-sown crop from the effects of frost. The deficiency in temperature observed in the last five days of the experiment in the case of the soils receiving the smallest dressings of manure is attributed by Georgeson to the cooling effect of the larger amount of water held by the manured soils (see p. 72). As the result of adding farmyard manure to a soil is in nearly every case to increase the amount of water retained near the surface, it would appear that the after effect of the manure is rather to cool than to warm the land. TABLE XXV INFLUENCE OF FAKMYAKD MANURE ON TEMPERATURE OF SOIL Farmyard Manure per acre. None. 10 tons. 20 tons. 40 tons. 80 tons. Temperature, Oct. 27-31 . Excess over unmanured . 60°-5 62°.5 2°.0 63°.8 3°-3 63°.l 2°.6 65°. 1 4°-6 Temperature, Nov. 1-5 . . Excess over unmanured . 58°-5 59°-5 i°.o 60°.2 l°-7 61°.3 2°.8 62°.2 3°-7 Temperature, Nov. 6-10 . . Excess over unmanured . 57° 2 57°.8 0°.6 58C4 1°.2 59°-3 2°.l 60°-4 3°-2 Temperature, Nov. 11-15 . Excess over unmanured . 54° 7 54°.8 0°.l 55°.3 0°.6 56°.2 1°.5 56°.8 2°.l Temperature, Nov. 16-22 . Difference from manuring 50° 8 49°.8 -1°.0 50°-1 -0°-7 51°.6 0°-8 52°.5 l°-7 Average excess with man- ure in 1st twenty days. 0°-93 1°.70 2°.25 3°.40 150 RELATIONS OF SOIL TO HEAT Experiments by F. Wagner on the same subject (Wollny, Forsck. der Agrikulturphysik, v. 373) showed that with the heaviest dressing of farmyard manure, about twenty tons per acre, the temperature of the soil remained above that of the unmanured land for four to twelve weeks. The greatest excess of temperature observed was 5°F., but the highest average excess of temperature was i° F. Generally the average excess of temperature by manuring was from 0°'2-0°-7. The ploughing in of green crops has a similar effect on the temperature of the soil as the application of farmyard manure. The after result of such organic manuring is to somewhat lower the temperature of the soil, probably for the reason already stated. For the greatest rise in temperature to occur, a soil must be porous, and moist, and of a temperature not below 50° F. ; the rise is in fact greatest in summer time, all processes of fermentation and oxidation being then most vigorous. As tillage greatly promotes oxidation, the cultivation of a soil in spring, especially if it contains vegetable residues, (as for instance a clover ley), will tend to increase its temperature. We have seen that neither the internal heat of the earth, nor the chemical and physical changes which occur in the soil, have any considerable effect on the temperature of the earth's surface; this depends almost entirely on the amount of radiant energy received from the sun. The amount of heat received from the sun depends of course primarily on the activity of solar processes, which may apparently vary from time to time. The great altera- tions in climate in the past history of our globe are difficult SOURCES OF SOIL HEAT 151 to explain without assuming a difference in the amount of heat received from the sun. It is now generally assumed that the solar energy is greatest during the period of most numerous sun spots, which returns about every eleven years. The heat received from the sun by any portion of the earth's surface depends greatly on the transparency of the atmosphere ; our cold and hot summers are chiefly determined by the presence or absence of cloud. For the greatest amount of heat to reach the earth the air must not only be clear, but also dry. Dry air has but little power of absorbing the heat rays from the sun, but water-vapour is an active ab- sorbent. It is thus in elevated regions, having a clear dry atmosphere, that the heating power of the sun is most strongly felt. The highest temperature shown each month during the winter of 1870-1 by a blackened thermometer in vacuo at Greenwich, and at Davos in Switzerland, 5,400 ft. above the sea, was as follows :— Greenwich. Davos. November 95°-2 ... 115°.3 December 78°-8 ... 115°.0 January 79°-9 ... 117°.l February 101°.8 ... 126°.0 The sun's rays were thus far hotter at Davos than at Greenwich, although the ground at Davos was continuously covered by snow. Although cloud, and to a less extent a moist atmosphere, greatly hinder the sun's heat from reaching the earth, they do in an equal degree prevent a loss of heat from the earth. The condition of clear sky and dry air which admits a free passage to the sun's radiation during the day, allows an equally free radiation from the earth during the night; a climate of great extremes of temperature, of hot days and 152 RELATIONS OF SOIL TO HEAT cold nights is thus produced. With the cloudy sky, and moist air, which characterize the climate of an island, the extremes just mentioned are never found, but the climate is remarkably uniform in temperature. The same conditions of sky which give us a cool summer produce also a mild winter. Influence of Latitude and Aspect. The angle at which the sun's rays strike the earth has a very important influence upon their heating effect at the earth's surface. Every flat portion of the earth's surface receives the same number of FIGURE 4. hours of sunshine in the course of the year, and at the spring and autumnal equinox the day and night are of equal length over the whole globe ; yet how extremely different is the temperature in the tropics and in the arctic regions! The enormous influence of latitude on climate is simply due to the varying angle at which the sun's rays reach the earth The simplest illustration is however furnished by an ordinary summer's day. How different is the intensity of the sun's rays at sunrise or at sunset, when the sun is near the horizon, to what it is at noon, when the [sun has risen to a great height! This difference in the intensity of the sun's heat INFLUENCE OF LATITUDE AND ASPECT 153 is simply due to the different angle at which the rays fall on the earth at different times of the day. The reason why so much depends on the angle at which the sun's rays strike the earth will appear by an inspection of Figure 4. Here three sunbeams of equal dimensions are represented as falling on a flat surface, one vertically, another at an angle of 30°, and the third at an angle of 10°. It will be seen that the beam at 30° spreads itself over twice as FIGURE 5. much earth surface as the vertical beam, and consequently only supplies half the heat per unit of area ; while the beam at 10° spreads itself over more than five times the surface, and thus supplies less than one-fifth of the heat per unit of area which the same beam would furnish if falling vertically. Another cause, but one far less important, of the diminishing heating power of the sun's rays as it approaches the horizon, is the greater atmospheric absorption of heat rays which then takes place. In Figure 5 we have a representation of an 154 RELATIONS OF SOIL TO HEAT atmosphere surrounding a globe. It is at once evident that the vertical rays falling on the globe pass through a thinner stratum of air than the rays falling on the same point from a smaller angle, the maximum amount of atmospheric inter- ference occurring plainly at sunrise and sunset. The facts we have just brought forward not only elucidate the influence of latitude on climate, they also help to explain the well-known effect of aspect on the fertility of land. In BAG FIGURE 6. i our northern hemisphere a field or garden facing or sloping towards the south will for most purposes be greatly preferred to one sloping towards the north, and will yield much earlier crops. The inclination towards the south does indeed^ in part counteract the prejudicial effect of north latitude, and causes the sun's rays to fall at a higher angle upon the surface. In Figure 6 we have a flat surface BC, divided into two equal portions AB and AC, and receiving equal solar radia- INFLUENCE OF LATITUDE AND ASPECT 155 tion represented by the beams KL and LN. If now we replace this flat surface by a hill having a slope of 20°, facing south and north, it is at once evident that the equality of the radiation on the two portions is destroyed, and that, with the sun at the angle assumed, the southern slope will now receive twice as much heat as the northern. The greater heat obtained from the sun on a southern slope is, of course, received only in the daytime, and when the sky is clear. The increase of heat is confined to the surface of the ground, which becomes sensibly warmer ; the vegetation upon it does not obtain any increased radiation from the sun, (though it will from the ground), the growth of plants being always perpendicular, whatever the slope of the soil. The general temperature of the air is also unaffected by the inclination of the ground, except near the surface during the hours of sunshine. In addition to the increased intensity of solar radiation, the southern slope has also the advantage over the northern of more hours of sunshine, and of protection from cold winds. As the southern slope gains heat only during sunshine, and falls at other times to the general tem- perature, it follows that the extreme range of temperature is greatest with a southern aspect. Wollny (Forsch. der Agrikulturphysik, i. 263) determined the temperature of the soil at 6 inches below the surface on various sides of artificial hills of sandy soil containing humus. The experiment was made at Munich, and extended throughout a whole year. When the slope of the hill was 15°, the average temperature of the south side was i°-5 F. more than that of the north. When the slope was 30°, the average excess of temperature on the south side was 3°-! F. From May to August the south-east side was the warmest; in September 156 RELATIONS OF SOIL TO HEAT and October the south side; from November to April the south-west. A south-east aspect is the one generally pre- ferred by gardeners, as the sunshine in this case begins at an earlier hour. Such an aspect may, however, be attended with disadvantage in the case of spring frosts, owing to the too rapid thawing of the vegetation. King (The Soil, 228) determined the temperature of a red clay soil on the south shore of Lake Superior on Ju]y 31, both where the slope was 1 8° and on the level. The results were as follows: — Firstfoot. Second foot. Third foot. South Slope 70°.3 ... 68°.l ... 66°4 F. Level 67°-2 ... 65°4 ... 63°.6 „ Excess on South 3°-l ... 2°. 7 ... 2°.8 Thus, in the middle of summer, the greater heat on the south side had penetrated to a considerable depth. When planting in rows or ridges the exposure to the sun should always be considered. Rows running north and south will be equally exposed to the sun on both sides, while those running east and west will, in the case of tall plants, receive but little sunshine on their northern side. Temperature of Surface Soil. In the case of a bare dry soil, freely exposed to the sky, the range of temperature at the surface is very great, far greater than that of the air above it. The maximum temperature reached by such a soil is chiefly determined by the intensity of the solar radiation which it receives. Sir J. F. W. Herschel sunk a thermometer 4 inches deep in the sand in South Africa, and observed the temperature rise to 159° F. Schiibler (Jour. Roy. Agri. Soc. 1 840, i. 206) determined during two years the temperature of the garden mould on the south side of his TEMPERATURE OF SURFACE SOIL house at Tubingen in Germany, between noon and one o'clock, on every day when the weather was perfectly fine ; the bulb of the thermometer was placed one-twelfth of an inch beneath the surface. The highest temperature thus observed was I53°'5 F.J the temperature of the air at the same time being 78°. The average of the temperatures recorded for each month of the year, with the average temperatures of the air in the shade taken at the same time, will be found in Table XXVI. In the same table will be found the average midday temperatures observed by Schlibler at Geneva during a single year ; in this case, however, the temperatures were observed every day, and not only in fine weather. TABLE XXVI MEAN TEMPEKATURES OF SOIL AND AIR AT MIDDAY (SCHUBLER) In Fine Weather, Tubingen. In Variable Weather, Geneva. Surface Soil. Air in Shade. Soil in excess of Air. Surface Soil. Three inches deep. Air in Shade. January . . . February . . March . . . 54°. 1 86°-2 99°-5 24°-6 43C-0 46°-6 29°.5 43°-2 52°-9 43°.0 45°.7 53°.2 38°-5 39°.8 43°.2 38°-2 36°-8 38°. 1 April .... May .... June .... 121°.6 131°.2 139°.8 61°.7 67°.3 75°-2 59°-9 63°-9 64°-6 78°.9 80°. 1 89°.l 60°-7 64°4 73°-6 50°. 1 55°-9 60°-9 July .... August . . . September . . October . . . U6°-3 130°-1 119°-8 80°.8 81°-3 68°-9 68°-0 42°.8 65°-0 61°.2 51°-8 38°-0 93°4 96°.0 82°-8 59°-8 73°-3 76°-9 70°.2 54°4 63°-2 65°-8 62°-4 51C.8 November . . 72°.7 40°. 1 32°.6 47°-3 43°. 7 41°-6 December . . 59°-2 35°-6 23°-6 35°-3 33°.3 32°. 1 Means . . 103°4 54°.6 48°.8 67°.l 56°.0 49°.7 158 RELATIONS OF SOIL TO HEAT The results at Tubingen show that in perfectly fine weather the surface of the soil reached a midday temperature of 120°, or more, from April to September. The highest soil temperature was generally reached in July, the average for that month being 146°«3. Taking the mean for the whole year, the midday temperature at the surface of the soil in fine weather is seen to be 48°- 8 above the temperature of the air. The daily determinations at Geneva, made in all kinds of weather, show that the average midday temperature of the surface soil is about 30° above that of the air in June, July and August, and that at 3 inches below the surface the midday temperature is in the same months io°-i3° above that of the air. For the whole year, the average midday temperature of the surface is i7°-4, and at 3 inches below the surface 6°-3 above that of the air. The minimum temperatures reached by the surface of a bare, dry soil, freely exposed to the sky, are, on the other hand, considerably below the minimum temperatures of the air. These low temperatures of the soil are the result of the loss of heat by radiation, and occur in the night or early morning when the air is clear, and the sky free from cloud. A ther- mometer lying on the ground, in an open place, will usually show a lower temperature each 24 hours than a thermometer a few feet above the surface. When a soil is shaded, or protected by any covering, its range of temperature is much diminished ; it receives less heat from the sun during the day, and it loses less by radiation during the night. A mulching of straw or manure will have the greatest influence in this direction. A layer of snow is also very effective in preventing the extreme TEMPERATURE OF SURFACE SOIL 159 cooling of the soil in winter. Boussingault placed a ther- mometer upon the soil beneath a layer of snow 4 inches thick ; another thermometer lay on the top of the snow freely exposed to radiation ; a third thermometer was suspended in the air, 39 feet above the ground. The readings of these thermometers during three days of clear weather were as follows. The degrees are centigrade. TABLE XXVII TEMPERATURES BENEATH AND ON THE SURFACE OF SNOW (BOXJSSINGAULT) In the Air. Upon the Snow. Under the Snow. February 11, 5 p.m. + 2°-5 -1°.5 o°.o „ 12, 7 a.m. . -3°-0 -12°.0 -3°-5 „ „ 5.30 p.m. + 3°.0 -1°.4 o°.o „ 13, 7 a.m. . -3°.8 -8°-2 -2°.0 „ 5.30p.m. + 4°-5 -1°.0 o°.o Thus the temperatures of the surface of the ground under the snow were on February 12 and 13, 6°-2-8°«5 C. (u°-i5° F.) higher after the night's radiation than those shown by the thermometer freely exposed to the sky. The shading of a soil by vegetation has a considerable influence in diminishing the extremes of temperature. Crops covering the ground with an abundant foliage will have a distinct effect in this direction. The soil of a forest, shaded by trees, and further protected by a thick layer of forest litter, affords an extreme example of the exclusion of solar radiation. According to Ebermayer (Lehre der Waldstreu, 188), the mean temperature of the soil l6o RELATIONS OF SOIL TO HEAT of Bavarian forests to a depth of 4 feet is, in June, July, and August, nearly 7° F. less than that of similar soil covered by turf and freely exposed to the sky. In winter, the mean temperature of the two soils was nearly the same. The forest soil was thus on the whole distinctly cooler than the grass land. This coolness of a forest soil doubtless favours that accumulation of humus which is so characteristic of soils of this description. Had the forest soil been compared with arable, instead of with grass land, the differences in the range of temperature observed would have been much more con- siderable ; the turf was, indeed, itself an efficient protection to the underlying soil. The range of temperature of soil permanently covered with grass is much less than that of soil not so protected ; it is, indeed, a common observation, that winter frosts do not penetrate to such a depth under turf as they do in bare soil. The temperature of the soil may be considerably affected by other circumstances besides the amount of solar radiation. Every circumstance affecting the general temperature of a locality, as altitude, and prevalence of hot or cold winds, will clearly affect the temperature of the soil. Prominent among these circumstances is the neighbourhood of large masses of water. As the temperature of such masses of water is far more constant than the temperature of either the atmo- sphere or the soil, the neighbourhood of such masses will generally considerably diminish the extremes of heat and cold. The moist and cloudy state of the atmosphere arising from the presence of much water will also act in the same direction. The range of temperature on an island is thus distinctly less than on a continent. INFLUENCE OF COLOUR 161 The neighbourhood of large masses of water, by reducing the severity of winter, tends to bring about an early spring. Such situations are thus often extremely favourable for the production of early crops, and are of great value to market gardeners. Most striking are the differences of temperature determined by the neighbourhood of cold or warm ocean currents. The west coast of Scotland, and the coast of Labrador on the opposite side of the Atlantic Ocean, are in the same latitude, and receive the same solar radiation ; but the climate of the first is so warmed by the gulf stream, and the prevalent westerly winds, that fuchsias will live through the winter in the open air ; while the temperature of Labrador is so reduced by a cold arctic current that the sea freezes in October, and remains in this condition until April. Influence of Colour. The amount of heat absorbed by a soil when exposed to the sun's rays depends partly on the colour of the soil. Schubler made the difference of colour as great as possible by lightly sprinkling different portions of the same soil with lamp-black, and with magnesia ; the soils thus treated were then exposed to the sun towards the end of August, and the maximum temperature reached one-eighth of an inch below the surface was then observed. The black soils under these circumstances became 13°-! 5° F. warmer than the white soils. The original character of the soil had little or no influence on the temperature attained in this experiment, provided all the soils were dry ; the colour had clearly the preponderating effect. The excess of temperature shown by a darker soil depends of course on the intensity of the sun's rays. Humboldt, when in the Canary Islands, observed a difference of 25° F. M 162 RELATIONS OF SOIL TO HEAT between the temperature of a black and white sand. Working with natural soils, in a European climate, Schiibler, Oemler, and Wollny always obtained the highest temperatures with the darkest soils, humus always heading the list. The extreme differences observed among natural soils did not however exceed 6°-8°. No difference was observed on cloudy days. During the night the dark and light soils cooled to the same point. The facts just mentioned are easily explained. The white or light coloured soils reflect a portion of the radiation from the sun, while the black soils absorb the whole of it, the total energy of the sun's rays appearing in this case as heat. It is sometimes erroneously supposed that the darkest soil must cool most, and that the gain of temperature due to a dark colour is followed at night by a loss of temperature due to the same cause. This is not necessarily the case. The absorption in sunlight is influenced by colour, the cooling by radiation at night is not influenced by colour. The absorption in day by black and white surfaces is different, because these surfaces behave differently to rays of high refrangibility ; at night both behave alike, because both are then emitting similar rays of very low refrangibility. Melloni found that when lamp-black and white-lead were both exposed to the rays from a lamp flame their relative absorption of heat was 100 : 53 ; but when the same surfaces were exposed to the radiation from hot water, their capacity for absorbing heat was identical. The emission of heat by lamp-black and white-lead at the temperature of boiling water would necessarily be as equal as their rate of absorption. The extra heat obtained from the sun by a dark soil is thus a substantial gain. Ahr, in his investigation on the radiation INFLUENCE OF COLOUR 163 of heat by soil (Forsck. der Agrikulturphysik, xvii. 397), found that while different soil constituents radiated heat with different facility, these differences were quite independent of their colour. The fact that the rays of heat emitted by the warm earth are different in character from the rays emitted by the sun is in several ways very important. The solar rays are but little absorbed by the atmosphere ; the radiation from the earth is completely absorbed by moist air. The heat received from the sun is thus retained around the earth when the atmo- sphere is moist, and night frosts only occur with a clear sky and dry air. The vicinity of water is a great advantage where early crops are to be grown, or where there is a danger of frost in spring or early summer ; a small island or peninsula is thus especially well suited for early market gardening. Land irri- gated in the spring is also largely free from the effects of frost. The same facts serve to explain the high temperatures easily obtained in glass-houses exposed to sunshine. The sun's rays pass freely through the glass and are absorbed by the wood, brick, and soil within ; these radiate back heat rays of a different kind, which are absorbed by the moist atmosphere in the house, and also by the glass roof. The energy furnished by the sun is thus caught in a trap, it enters the greenhouse easily, and leaves it with difficulty ; the tem- perature within the house thus quickly rises. Frankland (Pro. Eoy. Soc., xxii. 3 19) placed a thermometer in a box lined with black cloth, with a sheet of plate glass as the lid. Exposed to bright sunshine at Davos on December 22, the thermometer rose to 221° F., or 9° above the ordinary tem- perature of boiling water. Herschel, at the Cape of Good Hope, cooked a beef-steak, and boiled eggs hard, by simple M 2 164 RELATIONS OF SOIL TO HEAT exposure to the sun in a box covered with a pane of window glass, and placed in another box so covered. The influence of the colour of the soil on its temperature has been recognized by practical men. In some of the Rhine vineyards it is usual to scatter fragments of black basalt on the surface of the ground, with the object of gaining more heat to ripen the grapes. In parts of Spain, on the other hand, having a much hotter climate, it is only on white soils that certain grapes can be successfully cultivated. Colour ceases to have a considerable influence on the tem- perature of the soil when soils are very wet, the extra heat received by dark soils being consumed in the evaporation of water. The power of humus to warm a soil is thus seriously diminished, humus always favouring the retention of water at the surface. Specific Heat of Soils. Different substances require different quantities of heat to bring them to the same temperature ; or, in other words, different substances receiving the same quan- tity of heat will rise to different temperatures. Thus if a pound of iron and a pound of tin are both placed in boiling water till they have gained that temperature, and are then each of them placed in a similar bulk of cold water, it will be found that the cold water is raised to a higher temperature by the iron than by the tin. By proceeding in this way the relative amounts of heat contained by substances at the same temperature may be measured. As water is of all ordinary substances the one requiring most heat to raise a given weight to a given temperature it is taken as the standard. The numbers representing the specific heat of other substances express the fraction of a unit of heat which is required to raise them to the same temperature that would be reached if SPECIFIC HEAT OF SOILS 165 one unit of heat were imparted to an equal weight of water. The calculation may of course be varied so as to show the specific heat of equal volumes. The following determinations of the specific heat of various constituents of soil are quoted from C. Lang. TABLE XXVHI SPECIFIC HEAT OF SOIL CONSTITUENTS Kelative Specific Heat of Equal Weights. Equal Volumes. Water 1-000 0477 0.260 0.20 - 0.28 0-233 0.206 0-189 0.163 1-000 0-587 0-754 0.54 - 0-84 ? 0-568 0-561 0-499 0-831 Humus (Peat) Magnesium Carbonate .... Lava and Basalt Clay , Calcium Carbonate . ... Quartz, Orthoclase, Granite . . Looking first at the calculations by weight, it appears that the same quantity of heat which would be required to raise i Ib. of water i° F., would suffice to raise about 5 Ib. of dry chalk or quartz sand, and 2 Ib. of peat, to the same temper- ature. Or, looking only at the solid soil constituents, the same amount of heat from the sun would equally warm about 8 Ib. of quartz sand and 3 Ib. of perfectly dry peat. It follows, as a matter of course, that when cooling, 3 Ib. of peat will evolve as much heat as 8 Ib. of quartz sand ; and 5 Ib. of chalk or quartz as much heat as I Ib. of water. The specific heat of a perfectly dry arable soil, reckoned by weight, will generally be -20- -23. It follows that 4-5 Ib. of dry soil, and i Ib. of water, will be raised to the same tem- perature by the same supply of heat. 166 RELATIONS OF SOIL TO HEAT When we compare the specific heats of equal volumes, we find that water is still far ahead of all the other soil constituents, but the numbers for humus, clay, calcium carbonate, and quartz have become nearly equal. The same quantity of heat would raise to the same temperature 2 cubic feet of quartz, and i cubic foot of water. If soils were of the nature of a solid rock, without interstices, and without moisture, they would very uniformly have a specific heat of .50— «55 when compared with their own volume of water, unless indeed much iron or magnesium were present. We have already seen that the water-holding power of a soil, viewed in its relation to fertility, is far more accurately stated per unit of volume than per unit of weight (p. 69) ; the same may be said of the relations of soil to heat. It is the bulk, or depth of the soil that is warmed by the sun, and not the weight, which is important to the plant. To become acquainted with the relations of natural soils to heat on this basis of volume or depth, we must clearly take into account the varying proportions of the bulk which consist of the spaces between the particles; we must also take into account the proportion of water which will be normally present in the soil. The information already given as to the weight of different soils per cubic foot, and as to the proportion of water contained by various soils, both when in an air-dry and in a drained condition, will enable us to calculate the specific heat of soils per unit of volume in various natural conditions. If, in the first place, we regard soils in the air-dry state, that is containing only hygroscopic water, we find that there is little difference in their specific heat per unit of volume, the coarsest sand, the purest clay, and an air-dry peat, all having specific heats varying from about -30 to -42, if the specific SPECIFIC HEAT OF SOILS 167 heat of the same volume of water is reckoned as i-o. This uniformity is largely due to the presence of two factors in varying proportions. The coarse sand contains the smallest proportion of hygroscopic water, but it also possesses the greatest weight per cubic foot. The peat has by much the smallest weight per cubic foot, but it also contains the largest amount of hygroscopic water. In a naturally dry condition a soil will thus have about one-third the specific heat of water ; or three cubic feet of soil will be warmed by the sun to the same degree as one cubic foot of water. Of soils in this air- dry condition, peat will have the lowest specific heat, and clay the highest. The commonest condition, however, in which soils are met with is the condition which is found after rain has ceased, and all excess of water has been removed by prolonged percolation. If we calculate the specific heats of different soils per unit of volume when in this moist, but fully drained condition, we no longer find the uniformity belonging to the air-dry state, the result is now chiefly determined by the amount of water which the soil has retained. A coarse sand, retaining but little water when drained, is now the soil most easily warmed ; the fine sands, loams, silts, and clays show a higher specific heat, the figure rising as the proportion of water retained increases. A wet peat is now at the head of the list, and its specific heat differs not greatly from that of its own bulk of water. Thus, under natural conditions, the driest soil is the one having the lowest specific heat, and therefore (other conditions being equal) the one reaching the highest temperature when exposed to solar radiation. We have already mentioned that it is the soils with coarse particles, retaining little water when 168 RELATIONS OF SOIL TO HEAT drained, which are always chosen as specially suitable for early garden crops (pp. 20, 57) ; the reason of this suitability is now apparent — they become warm in spring time earlier than any others. The influence which oxide of iron has on the specific heat of soils demands further study. The oxide referred to in Table XXVIII was the anhydrous mineral, while the hydrated oxide is the form usually present in soils. Tillage, by loosening the soil, and thus diminishing the weight in a given volume ; and also by diminishing the pro- portion of water retained at the surface, tends to decrease the specific heat per unit of volume, and so far facilitates the warming of the soil. We shall see however presently that tillage diminishes the power of the soil to conduct heat. Conductivity of Soil Constituents. The temperature of both the surface soil and subsoil depends in part on the power of conducting heat which the soil possesses. A soil having little power of conducting heat l will become very hot on the surface when exposed to the sun during the day, and the surface will become very cold during a clear night when heat is lost by radiation. A soil having a greater power of conducting heat will be warmed to a greater depth by the sun during the day, and the surface will be longer in cooling during the night, as heat will then travel to the surface from the interior. Good conduction thus tends to equalize the tem- perature of the surface soil, while it occasions a greater range 1 The ' conduction of heat ' is throughout this section used in its popular sense, which would perhaps be more accurately expressed as the ' propagation of temperature.' The quantity of heat transmitted, and its facility of move- ment, cannot be measured by the rise of a thermometer at a distance from the source of heat. A smaller rise of the distant thermometer may be attended with a greater transmission of heat if the specific heat of the intervening matter is higher than in the comparative instance. CONDUCTIVITY OF SOIL CONSTITUENTS 169 of temperature in the subsoil. Looking at the soil and subsoil together, a greater conduction of heat (other conditions being equal) will be attended with a higher summer temperature, and a lower winter temperature. E. Pott (Landw. Versiichs-Stationen, xx. 288) has compared the power of conducting heat possessed by the principal con- stituents of soil, and has studied the influence of various conditions on their conductivity. The soil to be studied was placed in a cylindrical vessel, about one foot in length, lying on its side, the bulbs of six thermometers being sunk at equal distances along its axis. The cylinder was surrounded by a non-conducting material. One end of the cylinder consisted of a copper plate ; this during the experiment was in contact with the flat surface of a vessel containing hot water. The temperature of the water at starting was in every case 35° C. above that of the soil. The temperature of the water was maintained without alteration during the whole experiment, which lasted in every case twelve hours. The mean rise in temperature of the five thermometers furthest from the source of heat was taken as the figure for comparison. The sub- stances experimented on were quartz powder, and quartz sand of different degrees of fineness; kaolin and potter's clay ; levigated chalk (whiting) ; peat powder. The average results of Pott's experiments are shown in the following state- ment : — Relative Conductivity for Heat i. In air-dry condition Lightly shaken together. Quartz Powder ......... 100-0 ...... 106-7 Peat „ ......... 90-7 ... 98-1 Kaolin „ ... ...... 90-7 ... 96-4 Chalk 85-2 92-6 I7o RELATIONS OF SOIL TO HEAT Quartz Sand, Fine ... 100.0 „ „ Medium ... 103-6 „ „ Coarse ... 105-3 Dry Quartz Powder = 100. Dry Clay = 100. Clay Powder 94-1 100-0 „ with Limestone Stones ... 112-1 118-8 „ „ Quartz Stones ... 115-6 ... 122-5 2- In wet condition, not compressed Dry Quarts Powder = 100. Wet Quarts Poicder - 100. Quartz Powder 201-7 100-0 Kaolin „ 155-6 77-1 Chalk „ • 153-2 75-9 Peat „ 94-3 ... 46-8 Quartz Sand, Dry ... 100 „ Moist ... 174 „ „ Wet ... 189 Of all the soil constituents experimented with, quartz showed, under every circumstance, the highest power of con- ducting heat ; in this respect indeed it exceeds at least one of the metals, namely bismuth. The conductivity of every substance is greatly lessened when it is in the form of powder ; the finer are the particles of the powder the less is the power of conducting heat. The comparative conductivity of different substances cannot therefore be ascertained from experiments on their powders, unless the powders are in every case com- posed of particles of the same size ; the results obtained by Pott are thus only strictly true of the particular powders which he examined. The diminution of conductivity observed in a powder is due to the small extent of contact among the particles ; in the case of a dry powder each particle is surrounded by air, in a wet powder each particle is enclosed by water. The far greater conductivity of a solid rock than CONDUCTIVITY OF SOIL CONSTITUENTS 171 of a powder of similar composition is well shown by the ex- periments made by Forbes and Thomson on the conductivity of the sand of the Experimental Garden at Edinburgh, and of the sandstone of the Craigleith Quarry ; the latter was found to conduct heat about four times better than the former. Illustrations of what has just been said will be found in the statement of Pott's results already given. It will be seen that the compression of a dry powder increased its power of conducting heat. The coarse sand conducted better than the fine sand. A mixture of equal bulks of stones and clay conducted heat much better than clay alone. The circumstance having the greatest effect on the con- ductivity of sand, clay, and chalk was, however, the presence of water. Water is not itself a good conductor of heat ; it is indeed in this respect inferior to the solid constituents of soil, but it is a much better conductor of heat than air; the displacement of air by water thus serves greatly to facilitate the transmission of heat. The dry quartz powder had its conductivity doubled when thoroughly wetted. The presence of a little water is sufficient to produce a large effect in this direction. The moist sand contained only 9-9 per cent, of its volume of water; its conductivity was, however, 74 per cent, greater than that of the dry sand. The presence of much water is probably not favourable to the propagation of temperature, owing to the high specific heat which water possesses. It was perhaps for this reason that Pott found that wet peat was apparently a much worse conductor of heat than wet chalk, clay, or sand. It appears from the facts now mentioned, that a fine, dry, loose soil is the one which conducts heat worst ; such a soil will have its surface much heated by the sun's rays, but the heat 172 RELATIONS OF SOIL TO HEAT will penetrate to a comparatively small depth. On the other hand, a consolidated, stony soil, especially when moist, will prove the best conductor of heat. In such a soil the heat gained from the sun will be most evenly distributed, and will penetrate to the greatest depth. Practical experience has shown that a gravelly soil, having a good aspect, is one especially suited for the production of early garden crops ; this is probably due to the rapidity with which such a soil is warmed in spring time. The favourable properties of soils of this character would probably be improved by a top dressing of soot. Such soils provide also some protection from early morning frosts, their good conduction of heat serving then to warm the surface, as during the day it had proved effective in warming the subsoil. The operations of tillage have a very sensible effect on the propagation of heat through the soil. If the surface soil is brought into a loose pulverulent condition, the passage of heat downwards is retarded. If, on the other hand, the surface is rolled after cultivation, the temperature of the soil beneath the surface is distinctly increased. These results might have been predicted from the facts shown in the ex- periments described above ; they have, however, been confirmed by actual observations in the field made by King in Wis- consin. In experiments made in six counties, and on soils of various character, he found (Wisconsin Jth Report, 120) that the temperature of rolled land at a depth of i i inch was i°-9° F., and at a depth of 3 inches i°-6°F., higher than that of similar unrolled soil. The average excess of temperature by rolling was 3° F. The temperatures were all taken between i and 4 p.m. RADIATION OF HEAT 173 We have already seen (p. 113) that keeping a few inches of the surface soil in a loose pulverulent condition by frequent cultivation is an effective method of conserving the moisture of the soil beneath during the heat of summer. One result of this cultivation is to diminish somewhat the temperature of the underlying soil, and this diminution of temperature is in itself favourable to the preservation of water. King (Wisconsin loth Report, 190) records observations on the temperature of the soil made in a maize field during the month of July, in ground cultivated ij and 3 inches deep. In both the third and fourth foot beneath the surface, the average temperature observed in the day time was i°-i F. less where the deepest cultivation was maintained. In later experiments (Wisconsin nth Report, 283) he shows that the daily variation of temperature at a depth of i ft. was slightly less where the deeper cultivation was adopted. Radiation of Heat. Several facts relating to the radiation of heat by soil have been already noticed (p. 162), and we have seen that the radiation is independent of the colour of the soil. Ahr (Forsch. der Agrikulturphysik, xvii. 397) has investigated the relative radiating power of various soil constituents. The substance was placed in a thin layer, on a cube of warm water, maintained at a constant temperature, and the heat radiated by the substance measured by means of a thermo- pile and galvanometer. When in a perfectly dry condition, the various mineral constituents of soil proved to be better radiators of heat than the organic constituents. The mineral constituents showed no great difference among themselves ; quartz sand was the best radiator. Water has a greater capacity for radiating heat than any other soil constituent, it even somewhat exceeds soot in this 174 RELATIONS OF SOIL TO HEAT respect ; a moist soil thus radiates heat much better than a dry one. In an air-dry condition, only hygroscopic water being present, the differences between the radiating power of the different solid constituents are only slightly perceptible. In a thoroughly moistened state the distinctions previously observed disappear. All thoroughly moistened soils radiate to nearly the same extent, and the rate of radiation is not increased if the soils are saturated with water. The radiation of heat by a moist soil is nearly equal to that shown by soot. Ahr points out that the radiating power of a soil by no means determines its rate of cooling, this is largely deter- mined by its specific heat, and its conductivity. A high specific heat will tend to slow cooling. High conductivity will favour a more rapid emission of heat, but a relatively slow cooling of the surface, owing to the passage of heat to the surface from the interior. Influence of Water on Soil Temperature. We have in the three preceding sections found abundant evidence of the leading part played by water in determining the relations of soil to heat. The action of water in the soil is to diminish its summer temperature. From its very high specific heat, a wet soil shows little rise in temperature when exposed to sunshine. The better conduction of heat in a wet soil (wet peat excepted) tends also to equalize the temperature of the surface and subsoil, and cause it more generally to approach the mean annual temperature. The active radia- tion of heat from a moist soil leads to loss of heat. The evaporation of water from a wet surface has a still greater effect in cooling the soil, owing to the large amount of heat consumed in converting water into vapour (p. 107). King has calculated that the evaporation of one pound of water INFLUENCE OF WATER ON SOIL TEMPERATURE 175 from a cubic foot of wet clay would lower the temperature of the clay io°-3 F. The loss of heat is nearly the same at whatever temperature the evaporation takes place. The cooling influence of. water is so great as to override the influence of other conditions tending to warmth ; thus Schiibler found that dry soils, coloured white, exposed to the sun, became about 6° warmer than a wet, dark coloured, humus soil, equally exposed to sunshine. The coldness of a soil in summer time is generally in proportion to- the amount of water which it permanently retains ; it is only during the severer months of winter that a wet soil is superior in temperature to a dry one. With these facts before us it is easy to understand why a stiff clay or an undrained meadow are said to be cold soils ; while a soil of open texture, and well-drained, is said to be warm. Both soils may receive the same heat from the sun ; but, in consequence of the properties of water just mentioned, the temperature of the wet soil is always, save in the depth of winter, below that of the dry one. The coldest soil will be one having a permanently saturated sub- soil, from which the surface soil is continually replenished with water. The first step towards the amelioration of cold, wet land, will be the removal of the excess of water by draining. Land in which there is a free percolation of water enjoys generally a special rise in temperature from spring rains. In spring time the rain is frequently warmer than the soil, and if the rain can penetrate the soil it becomes an effective agent for warming it. One pound of rain water at 60° F. would be able, from its high specific heat, to raise the tempera- 'ture of 10 Ib. of dry sand from 45° to 50°. King's observations in Wisconsin supply some illustrations I76 RELATIONS OF SOIL TO HEAT of the influence of the water contents of a soil on its tem- perature. Thus on April 24, between 3.30 and 4 p.m., the temperature of the air being 6o°-5 F., the temperature of the surface inch of a wet and dry soil was as follows : — Well-drained Sandy Loam . . . 66°-5 Undrained Black Marsh ... 54°-0 On July 31, the temperature of two soils, each more than half saturated with water, was as follows :— First Foot. Second Foot. Third Foot. Alluvial Sand ... 71°-2 ... 70°.l ... 67°-6 Red Clay 67°-2 ... 65°4 ... 63°-6 On August 6, two soils, less than half saturated with water, were found to have the following temperatures : — First Foot. Second Foot. Third Foot. Sandy Loanx ... 76°-5 ... 74°-7 ... 72°.l Clay Loam ... 69°-5 ... 69^3 ... 67°.0 Thus throughout the summer the wetter soil is in every case distinctly the colder. One of the most instructive illustrations of the influence of water on the temperature of the soil is furnished by the observations made long ago by Parkes on the temperature of a peat bog, known as the Bed Moss, in Lancashire (Jour. Roy. Agri. Soc. 1845, 140). This bog was 30 ft. in depth, and of the wettest description. Below the depth of i ft. the natural bog was found to have a constant temperature of 46° F. during three years of observation. A portion of the bog had been drained by open drains 3 ft. in depth, and the drained portion had been dug to the same depth. Ther- mometers were sunk to various depths both in the drained and undrained portion. The observations of temperature INFLUENCE OF WATER ON SOIL TEMPERATURE 177 given in Table XXIX were made two years after the draining operations just mentioned. The land was without any crop. TABLE XXIX TEMPERATURE OF DRAINED AND UNDRAINED BOG (PARKES) Depth 7 inches. Depth 13 inches. Depth 19 inches. Depth 25 inches. Depth 31 inches. Drained Bog. June 15, 9 a.m. „ 16, 9 a.m. „ „ 3p.m. „ „ 3.30 p.m „ 17, 9 a.m. 57°-6 60°.0 62°-5 66°.0 58°-0 53°.0 54°.2 54°-0 57°-0 55°.6 50°.8 51°4 51°.8 52°-0 52°-8 48°-6 49°.0 49°-6 49°.8 50°.0 47°-3 47°.6 47°-8 47°.8 48°-0 Undrained Bog. „ 15-17. . . 47°-0 46°-0 46°-0 46°-0 46°.0 June 15 was a hot cloudless day. June 16 was hot. A heavy thunderstorm took place at 3 p.m. ; the temperature of the rain was 78°. The temperature of the drained bog at 7 inches below the surface is seen to be io°-i9° above that of the undrained bog, and the temperature of the drained portion remains superior even at a depth of 31 inches. The effect of the warm rain on the temperature of the drained soil is very distinct ; the temperature at 7 inches beneath the surface rises 3°«5 in half an hour, and the temperature of the subsoil is found next morning to be perceptibly increased. The warm rain has no perceptible effect on the undrained bog. A saturated peat bog is thus the coldest of all soils in summer time, and the one showing the smallest variation in temperature throughout the year. A dry peat — a heath soil for instance — is, on the other hand, the soil exhibiting the N 178 RELATIONS OF SOIL TO HEAT greatest variation in temperature. According to Mayer, the soils of highest average temperature are those which are dry, dark, and specifically heavy, as those composed of black basalt or dolerite, clay slate, and some sands. Temperature of Subsoil. The temperature of the subsoil may be affected to a considerable depth by the conditions of temperature prevailing at the surface. At the Observatory at Bombay the soil has an average temperature of about 83° at a distance of n ft. beneath the surface. At Jakutsk, in Eastern Siberia, an attempt was made to obtain water by sinking a well, but at a distance of 382 ft. beneath the surface the soil was still frozen. The mean temperature of the air in this locality was about 15° F. In many parts of North America and Siberia crops are annually raised in summer time upon the thawed surface soil, the subsoil remaining permanently frozen. In the tropics, where day and night are of nearly equal length throughout the year, and summer and winter are unknown, the variations in the temperature of the subsoil are but small ; but in higher latitudes, where the seasons are very different in character, the range of temperature in the subsoil becomes very considerable. The range of variation is in all cases greatest near the surface, and becomes smaller and smaller at increasing depths. At a certain depth in every soil a point is reached at which the variations of the temperature at the surface cease to be felt. The daily variations of temperature caused by the alternations of day and night are seldom perceived below 3 ft. from the surface. The effects of summer and winter may be ielt at a much greater depth, but a point is at last reached at which the thermometer remains unchanged throughout the TEMPERATURE OF SUBSOIL 179 year. The position and temperature of this point vary much in different climates, and in different soils. It is to be regarded simply as a point of equilibrium. In hot climates there is a warmer soil above this point, and a cooler soil below it. In cold climates the soil is cooler above and warmer below. The depth at which a constant temperature is reached is greater the greater are the variations in annual temperature at the surface. In the tropics a constant temperature is usually found a few feet beneath the surface; in higher latitudes the depth is generally very considerable. The dis- tance from the surface at which variations of temperature can be felt is however considerably affected by the conductivity and specific heat of the soil or rock of the locality ; observa- tions at different places thus frequently differ even when these places are in nearly the same latitude. We have already seen that in the case of an undrained Lancashire bog, the tem- perature remains practically constant throughout the year at a depth little exceeding one foot. The average annual range of temperature observed at various depths of soil at the Observatories of Greenwich, Brussels, and Edinburgh is shown below ; the degrees are Fahrenheit. The soil at Greenwich is gravel. At Edinburgh the temperatures are taken in a drill-hole bored into a porphyry trap- tuff rock. Average annual Range of Temperature at various depths Greenwich. Brussels. Edinburgh. 1 inch ... 7^ inches 1-5 feet ... 25°.4 23°.9 22°.3 2.5 , 20°.5 N 2 i8o RELATIONS OF SOIL TO HEAT Greenwich, Brussels. Edinburgh. 3-2 feet 21°-4 ... 19°-3 ... 16°-8 6-4 „ 15°.l 10°.2 12-8 „ 9°.3 ... 8°.l ... 5°.0 25.6 „ 3°-4 ... 2°.0 ... 1°.3 In none of these instances is the point reached at which no variation of temperature is observed. H. Fritsche (Reper- torium fur Meteorologie, 1872) has calculated from the ob- servations made at Upsala, Edinburgh, Paris, Strassburg, and Zurich, the average depths at which the annual variation will amount to i°-o, o°-i, and o°«oi C. ; his results are as follows : — Annual Temperature Variation. Average Depth of Soil. Centigrade. Fahrenheit. Metres. Feet. 1°.00 ... 1°-80 ... 8.55 ... 28.04 0°-10 ... 0°.18 ... 15.35 ... 50-34 0°.01 ... 0°.02 ... 22-37 ... 73-36 The variations of temperature immediately below the sur- face depend very much on the exposure at the surface. At observatories a hut is usually placed over the soil ther- mometers for the protection of the stems which rise above the surface. The soil in which the shallow thermometers are placed is thus neither heated by direct sunshine, nor exposed to night radiation. The extent of variation at a short distance below the surface is thus much smaller than would occur in an open field. Schubler's determinations of midday tem- peratures at Geneva (Table XXVI) show a variation between the monthly means of 43°-6 at 3 inches, and 3O°«5 at 4 feet. The mean monthly temperatures in Pennsylvania (Table XXX) also show a variation of 4o°-8 at a depth of 6 inches. Soil being a bad conductor of heat the alterations of tem- perature at the surface pass but slowly downwards, and affect the temperature of the subsoil long after the effect on the TEMPERATURE OF SUBSOIL 181 surface has ceased. Thus the date of maximum temperature in the subsoil occurs later and later in the year with increas- ing depths, till at a certain point the maximum temperature in the subsoil takes place at the same time as the minimum temperature at the surface, and the seasons are reversed. At still greater depths the maximum may occur yet later, till its influence disappears altogether. The average dates of maximum and minimum temperature at various depths of subsoil at Greenwich and Edinburgh are as follows :— Greenwich. Edinburgh. Depth. Maximum. Minimum. Maximum. Minimum. 3.2 feet ... Aug. 9 ... Feb. 8 Aug. 16 ... Feb. 21 6-4 „ ... „ 25 .- ,, 24 ... Sept. 2 March 19 12-8 „ ... Sept. 25 ... March 27 ... Oct. 15 April 22 25.6 „ ... Nov. 30 ... June 1 Jan. 6 JulyS The passage of heat through the Edinburgh soil is evidently slower than through that at Greenwich. In temperate climates, in the northern hemisphere, the sub- soil at moderate depths will be cooler than the surface soil from April to August, and warmer than the surface from September to February or March. The precise date at which the relation changes depends on the depth in question, and on other local conditions. Towards the end of September, and in March, a fairly uniform temperature will be found to a considerable depth in the subsoil. The general relations of the temperature of the subsoil at moderate depths to those observed at the surface, and in the air, will be gathered from the following table containing the observations made at the Pennsylvania State College in 1893. The soil is described as a compact loam for the first seven inches, the subsoil being a stiff clay. The neighbouring soil is covered with turf, but 182 RELATIONS OF SOIL TO HEAT the soil immediately over the thermometers is kept free from vegetation. TABLE XXX MEAN MONTHLY TEMPERATURES OF AIR AND SOIL, PENNSYLVANIA, 1893 AIT- Soil. 1 inch. 6 inches. 12 inches. 24 inches. January . . . 18°.0 26°.3 29°-0 31°.4 33°-8 February . . 26°.0 28°-3 29°-5 31°-2 32°.6 March . . . 33°-5 31°.5 31°.3 31°.9 32°.8 April .... 46°.3 42°.9 40°-9 40°.8 39°. 7 May .... 57°.l 54°.5 53°-7 52°.9 50°.6 June .... 68°-7 66°.3 65°.6 64°.7 61°-8 July .... 7