bfclav ®Il£ ^- ^. ^tU pbrarg ^ortii Carolina ^ferte Colkge S00540033 F RARIE i Results relating to the Growth of Wheat 17 Results RELATING TO THE Growth OF Barley 2ii Summary, and General OiiSERVATiONS 35 FROM THE JOUKNAL OF THF ROYAL AGRICULTURAL SOCIETY OF EXGLAND. VOI-. VIL— S. S. PART I. THE DROUGHT OF 1870, AND THE EXPERIMENTAL CROPS AT ROTHAMSTED. The rainfall of Great Britain is usually sufficient for the growth of a considerable variety of crops, in fairly abundant quantity. Indeed, so far at least as the growth of corn is concerned, our fears are of injury from an excess rather than from a deficiency of rain. It is only occasionally, and generally at long intervals, that a season of great drought occurs ; and then it is that we forcibly realise how essential for luxuriant vegetation is an abundant supply of water. Throughout the Midland, Southern, and Eastern portions of England, the year 1870, just past, has been characterised by a season of drought, commencing with the period when vegetation usually becomes active, and extending, with little intermission, to the time when its activity has upon the whole greatly diminished, and in the case of some crops entirely ceased. To find a parallel we must go back to 1844, or more than a quarter of a century. The summer of 1868 was, it is true, one of great drought ; and, being hotter than that of 1870, it is not improbable that there was at some periods of it a greater deficiency of moisture in the soil than in the latter year. In fact, those who travelled through the Southern and Midland counties of England in July, 1868, will not soon forget the almost entire absence of green in the meadows, and the intense heat of the atmosphere, resembling more what we read of in tropical countries than the usual expe- rience of our own summers. Although both the drought and heat were more extreme during the months of May, June, and July in 1868 than in 1870, the deficiency of rain commenced a month earlier and extended later last year ; and hence, not only the first crops of grass and hay, but also the second growth, suffered much more in the season just past than in 1868. It is only when crops are grown under precisely similar circumstances, as to manure and other conditions, for many years in succession, that we can obtain satisfactory data for studying the influence of variation of season on the amount and character of the produce. At Rothamsted, as is known to most of the readers of this Journal, numerous experiments on the growth of various crops, each grown year after year on the same land, with different descriptions of manure, the same description B Librai-y -lixJj^ 4 Drought of \^10 and beint; applied year after year to the same plot, have been carried on without change for many years ; in some cases reaching back as far as the drought of 1844, above referred to. Taking advantage of the results so obtained, it is proposed, in the present paper, to consider briefly : — 1. The probable amount of water exhaled during growth by some of our most important crops, 2. The source whence the required supply of -water is obtained. 3. The difference of the effects of the drought of 1870 on the different experimental crops. Amount of Water given off by Plants during Growth. A series of experiments was commenced in 1849, and was continued for ten years, to determine the amount of water given off by plants during their growth, in relation to the amount of the various constituents they assimilated. Of agricultural plants, wheat, barley, and mixed grasses, as representatives of the Graminaceous family ; beans, peas, and clover, of the Leguminous family ; and swedes, white turnips, mangolds, potatoes, and artichokes, as root-crops, were thus experimented upon. Similar experiments were also made on the exhalation by evergreen and deciduous trees, six of each being selected. The plan of experimenting was as follows : — Cylindrical vessels, first of glass and afterwards of zinc, 14 inches in depth, 9 inches in diameter, and holding about 40 lbs. of soil, were employed. Soil from the plot in the experimental wheat-field which had grown 10 successive crops without manure was selected. The general rule was to make three experiments with each descrip- tion of plant ; one with the above soil without further addition ; one with the same soil with purely mineral manure added ; and the third with the same soil and both mineral manure and ammonia- salts in addition. In the cases of wheat and barley, plants from three seeds, and of beans, peas, and clover, one plant only, were planted in each vessel. A glass plate, having a hole in the centre about three-quarters of an inch in diameter for the plants to grow through, and another smaller one, closed at pleasure by a cork, for the supply of water, were then firmly cemented upon the top of each vessel. One vessel, supplied with soil and fitted with a glass cover like the rest, was, however, always left without a plant, in order to ascertain the probable amount of evaporation from the surface of the soil itself, through the centre orifice, independently of growth ; though, in the experiments with plants, the hole was always partially closed, by laying small pieces of glass over it as far as the stems would allow. Of course in experi- menting with root-crops the holes in the glass covers were larger, but they were kept closed around the plants as far as possible, in the manner just described. Exj)erimental Crops at Rothamsted. 5 The vessel with its contents, weighing more than 40 lbs., was weighed from time 'to time, generally every ten days during active growth, by means of a delicate balance made for the purpose ; which, though carrying so heavy a weight, was capable of indicating a change of a few grains. The plants were of course supplied with water as it was needed. The earlier results, both with agricultural plants and trees, are published in the ' Journal of the Horticultural Society of London,' and to the reports there given we must refer the reader for the details of the inquiry as far as they are yet recorded.* Referring here only to the results obtained with some of the agricultural plants, it will be sufficient for our present purpose to summarise them as follows : — 1. The amount of water given off by the plants during growth was found to bear relation to the quantity of the total dry matter, or the total non-nitrogenous substance, fixed or assimilated ; and within somewhat narrow limits the same re- lation was observed in the case of both graminaceous and legu- minous corn-crops. 2. In relation to a given quantity of water exhaled, twice or three times as much nitrogenous substance is fixed by a legu- minous, as by a graminaceous corn-crop. 3. In the growth and ripening of either graminaceous or leguminous corn-crops, probably on the average from 250 to 300 parts of water are given off for 1 part of total dry substance fixed or assimilated. Before considering the application of this estimate to any special cases, it may be well to give an illustration of its bearing in general terms. Several plots in the experimental wheat-field give an average of about 3 tons of total produce (corn and straw) per acre per annum ; and if we assume one-sixth of this to be water, we have remaining 2^ tons of dry substance ripened by the end of July, or the middle of August, each year ; and if we further assume that 300 parts of water may be exhaled for 1 part of dry substance fixed, we have 300 x 2'5 = 750 tons of water evaporated per acre by the growth of such a crop. Owing to the difficulty of eliminating surface evaporation other than through the growing herbage, in experiments on the exhalation from a sod of mixed grasses, we cannot so safely adopt a figure to represent the probable average amount of water given off for 1 part of dry substance fixed in their case * ' Experimental investigation into the amount of water given off by plants during their growth, especially in relation to the fixation and source of their various constituents.' — (' Jour. Hort. Soc. Lond.,' vol. v. part i. 1850.) ' Report upon some experiments undertaken at the suggestion of Professor Lindley, to ascertain the comparative evaporating properties of Evergreen and Deciduous Trees.' — (' Jour. Hort. Soc. Lond.' vol. vi. parts iii. and iv. IS.^il.) b2 6 Drought of i^lO and as in that of their ripened allies, wheat and barley. We will assume, however, for the purpose of illustration, that in the growth of hay, as in that of the grain-crops, about 3i)0 parts of water will be exhaled for 1 part of dry substance assimilated ; and since one of the experimental plots of meadow land at Rothamsted has given an average, over fifteen years, of 3 tons of hay, or about 2^ tons of dry substance per acre per annum, its growth would again represent nn exhalation of about 750 tons of water per acre per annum — but extending in this case not later than to the middle or end of June. We will now adduce some special cases illustrating the amount of water exhaled by different crops, and their dependence on the rainfall of the period of active growth, or on the supplies of moisture previously accumulated within the soil. Eesults relating to the Gkowth op the Hay-ct.op. The following Table (I.) shows the amount of hay obtained per acre each year for fifteen years in succession (1856-1870) : — 1. Without manure. 2. With mixed mineral manure and 400 lbs. ammonia-salts per acre per annum. 3. With mixed mineral manure and 550 lbs. nitrate of soda per acre per annum (thirteen years only, 1858-1870). The Table also shows, side by side with the records of produce, the amount of rain, in inches, which fell at Rothamsted each year Table 1. Hat pee Acre. R UN AT Rothamsted. Years. Mineral Mineral Without Manure. Manure and Ammonia- Manure and Nitrate of Mean. April. May. June. Total. salts. «oda. Cwts. Cwts. Cwts. Cwts. Inches. Inches. ; Inches. Inches. 1856 22i 56f .. 391 2-61 4-70 i 1-91 9-22 1857 25i 57i 413 2-16 rio 1 2-21 5-47 1858 22 64 50| 45* 2-58 2-55 0-96 6-09 1859 221 5.H 541 44 2-70 2-09 2-72 7-51 1860 24f 50i 491 41f 44f 1-94 4-30 1 6-26 12-50 1861 25i 56i 52i 1-28 1-04 2-98 5-30 1862 27i 57| 51 Ab\ 2-84 2-91 3-41 9-16 1863 203 53| m 44? 0-96 1-01 ' 4-60 6-57 1864 24 50i mi 471 45 1-25 1-88 1-79 4-92 1865 Hi 34i % 0-47 3-05 0-68 4-20 1866 23| 441 58f 1-95 1-24 4-51 7-70 1867 29f 48 64| 47t 2-8.2 3-35 1-06 7-23 1868 17i 59J 69 48| 2-19 0-73 0-37 3-29 1869 38 6Sf 76| 61 2-13 3-23 1-07 6-43 1870 5| m 5ti 30* 0-46 1-35 0-98 2-79 Averag e 22| 52i 57i 431 1-89 2-30 2-37 C-56 Experimental Crops at Rothamsted. 7 during the months of April, May, and June, which may be considered as including the period of active growth of the hay- crop. Although there is much to be learnt from the results brought together in the foregoing Table, much more information than is there given would be required — as to the difference in the character of the herbage produced under the different conditions, the distribution of the rain, the degree and range of tempera- ture, and the mutual adaptations of moisture, heat, and stage of growth of the plants — to enable us to account for all the fluc- tuations in the amounts of gross produce which the records show. It is seen at a glance that the fluctuations from year to year in the amounts of produce without manure, though doubtless greatly dependent on the quantity and distribution of the rain falling during the period of active growth, by no means correspond with the fluctuations in the total amount of rain during the three months. Thus, the average fall for the three months is 6"56 inches, and the average produce of hay without manure is 22| cwts. But we have, with almost exactly the same total amount of rain during the same period in 1863 (6-57 inches), only 20^ cwts. of hay ; whereas, with even rather less (6'43 inches), in 186U, we have the heaviest produce obtained in any one of the series of 15 years, namely, 38 cwts. The fact is that, coincidently with the small produce of 1863, less than one-third of the total rainfall of the three months occurred during the first two months of the period ; whilst, coincidently with the very heavy produce in 1869, there was considerably more than the average fall of rain in both April and May, and less than half the average fall in June; the result being that more than five-sixths of the total fell during the first two of the three months, when its influence upon the growth would be the greatest. Again, the heaviest total fall within the growing period was in 1860, when there was nearly double the average amount, whilst the produce only exceeded the average by less than 2 cwts. of hay ; the facts being, that about half the total amount fell in June, that is, not until the last month of growth ; and that the temperature was very unusually low almost throughout the period of active vegetation. The lowest amounts of produce were — Vlh cwts. in 1868, 11^ cwts. in 1865, and only 5f cwts. in 1870. This last, the lowest amount in the series, is coincident with the smallest amount of total rain over the three months throughout the fifteen years, namely 279 inches. With onlv 3 '29 inches in the three months of 1868, there was a produce of 17^ cwts., but with 4 2 inches in 1865, there was only 11|^ cwts. But whilst, in the latter year, there was in April only about one-fourth the average fall, and very high 8 Drought of 1870 and temperature, there was during the same month in 1868 more than the average fall, and about the average temperature. Turning to the columns of produce obtained by the two artificial manures, it is seen that, whilst in the earlier years the mineral manure and ammonia-salts gave more hay than the mineral manure and nitrate of soda, in the later years the mineral manure and nitrate yielded considerably more than the mineral manure and ammonia-salts. It is obvious, therefore, that the fluctuations in the produce are dependent on other conditions than the variations in external or climatic circumstances alone. It will come within the special province of our subject to explain this further presently ; but, in passing, we may here remark that the character of the mixed herbage in regard to the distribution of plants, and the prevalence of individual species, was very widely different in the two cases ; and the dependence of the amount of produce on external supplies of moisture will, of course, be greatly measured by the degree of root range, and the consequent command of the moisture within the soil itself, of the particular species favoured. These few observations will be sufficient to indicate some of the points of interest which the study of the subject in detail is calculated to elucidate, and to show the complexity of the condi- tions upon which the final result — the weight of hay — depends. We will now turn to the more special object of the present communication. The following are the amounts of hay obtained per acre in 1870, on each of the three plots already referred to, and also the average amounts over 15 years without manure, and with mineral manure and ammonia-salts, and over 13 years with mineral manure and nitrate of soda. Table 11, Hay pee Acre. 1870. Average 15 (or 13) Years, 1856-70. Deficiency in 1870. Without manure Mineral manure and ammonia-salts . . Mineral manure and nitrate of soda . . Cwts. 291 561 Cwts. 523 57i Cwts. 17 22^ Thus, under the influence of the extraordinary drought of 1870, there was a variation in the amount of produce on closely adjoining plots, from only 5| cwts. of hay without manure, to Experimental Crops at Rothamsted. 9 29|- cwts. with mineral manure and ammonia- salts, and to 56|- cwts. with mineral manure and nitrate of soda. Indeed, without manure there was not only less produce than in any pre- ceding year of the fifteen, but only about one-fourth the average amount. With mineral manure and ammonia-salts there was again considerably lower produce than in any other of the fifteen years with the same manure, and a deficiency of nearly 23 cwts. compared with the average. Notwithstanding this, we have the remarkable result of 2 tons 16 cwts. of hay produced by mineral manure and nitrate of soda, or only about 1;^ cwt. less than the average amount by that manure ; about 2^ tons more than with- out manure, and \^ ton more than by the mixture of mineral manure and an amount of ammonia-salts containing about the same quantity of nitrogen as the nitrate. On the assumption that probably about 300 parts of water pass through the plants for one part of dry substance fixed, about 700 tons of water must have been exhaled by the herbage during the growth of the 56 cwts. of hay. But, reckoning an inch of rain to represent a fall of 101 tons per acre, the 2'79 inches which fell in 1870 during April, May, and June, the period of active vegetation, could only supply 282 tons of this, provided (which would not be the case) none of it was lost by drainage, and none of it passed off by evaporation otherwise than through the plants themselves. On the same assumptions, the amount which fell would be about 160 tons less than sufficient for the requirements of the crop grown by mineral manure and ammonia- salts, but more than three times as much as would be required by the growth of the unmanured produce. So striking was the difference in the effect of the drought on two plots side by side, the one manured with mineral manure and a given quantity of nitrogen in the form of ammonia-salts, and the other with the same mineral manure and the same quantity of nitrogen, but the latter in the form of nitrate of soda instead of ammonia-salts, that it was decided, on the removal of the crop, to determine the quantities of water existing in the soil of the three plots to a depth somewhat greater than the lowest to which roots could be traced ; and also to observe the difference in the development and distribution of the roots, if any, on the different plots. Accordingly, on July 25 and 26, 1870, samples of soil were taken from the three plots to the depth of 54 inchtii in each case, roots having been traced on one of them to within a few inches of that depth. The plan of collecting and preparing samples of soil for analysis will be understood from the following description of the process in the present instance : A square yard, comprising a fair proportion of the species contributing to the bulk of the herbage, 10 Druiujlit uf 1 870 and having been carefully selected on each plot, a case or frame, open at the top and bottom, made of strong sheet-Iron, 6 inches square by 9 inches deep (but which may be of any desired size), was driven into the ground in the centre of the square, level with the surface. The enclosed soil was then dug out exactly to the depth of the case. The soil around the case, to the extent of the square yard selected, was then removed to the level of the bottom of it ; it was again driven down, and its contents carefully taken out ; and so on, the process was repeated, until the desired depth was attained. The determination of the water in the samples being the special object of the experiments in question, the exact weight of the soil was taken immediately on removal, so that any loss of moisture by evaporation during preservation, or preparation for analysis, might be duly taken account of. The whole was then broken up, the stones sifted out, separating first those which did not pass a 1-inch sieve, next a i-inch, and finally a :^-inch sieve being used. The mould, or soil, passing the :^-inch sieve was weighed, a proportional part of it finely powdered for analysis and re-weiglaed. In the soils so prepared, the loss of moisture, at different temperatures, has been, and the nitrogen and some other constituents will be determined. The following Table shows the percentage of moisture, as deter- mined by the loss when dried at 212° Fahr., inclusive of that by evaporation during preparation for analysis, in the soil from each of the three plots of the experimental meadow-land, at each depth to which the samples were taken : — Table III. — Moisture in the Soil from Plots of Permanent Meadow Land differently Manured. Samples collected July 25-6, 1870. Percentages of Moisture (Soils dried at 212° Fahr.). Depth of Sample. Plot 3. ^Vithout Manure. Plot 9. Mineral Manure and Ammonia-salts. Plot 14. Mineral Manure and Nitrate of Soda. First 9 inches Second 9 inches Third 9 inches Fourth 9 inches Fifth 9 iuchts Sixth 9 inches 10-83 13-34 19-23 •22 • 7 1 24-28 25-07 13-00 10-18 16-4(5 18-96 20-54 21-34 12-16 11-80 15-65 16-30 17-18 18-06 Mean 19-24 16-7.') 15-19 The results recorded in this Table are of great interest and significance ; and they supply important data towards the explana- tion of the extraordinary difference in the amount of produce obtained on the different plots. It should be premised, however, Experimental Crops at Rotkamsted. 11 that between the removal of the crops and the date of sampling the soils, in all nearly an inch of rain had fallen, perhaps affect- ing somewhat the actual percentages, but the relative amounts probably but little. The first point to remark is, that the first 9 inches of soil of both the heavily manured, and more or less heavily cropped, plots contained a higher percentage of moisture than that of the unmanured and lightly cropped plot. But from that point downwards to a depth of 54 inches, and doubtless further still, the manured and more heavily cropped soils contained much less moisture than the unmanured ; and the most heavily cropped soil, that of Plot 14, manured with mineral manure and nitrate of soda, contained considerably less than that of Plot 9, manured with mineral manure and ammonia-salts. And whilst at a depth of from 45 to 54 inches the unmanured soil contained 25 per cent, of moisture, that receiving mineral manure and ammonia-salts contained only 21'34 per cent, ; and that receiving mineral manure and nitrate of soda only 18 per cent,, or scarcely fths as much as the unmanured soil at the same depth. To sum up the results, there is an average amount of moisture down to the depth of 54 inches, of 19J per cent, on the plot without manure, of only 16f per cent, on the plot manured with mineral manure and ammonia-salts, and of scarcely 15j per cent, on that manured with mineral manure and nitrate of soda, or only about ^ths as much on the latter as on the unmanured plot. The subsoil of this meadow land is a reddish yellow clay, interspersed with grey veins, and the specific gravity increases by about one-half from the surface down to the greatest depth taken. For our present purpose it will be a sufficiently near approximation to the truth to assume that down to the depth of 51 inches, the soil (exclusive of stones) weighed an average of 1,000,000 lbs. per acre for every 3 inches of depth, or an aggre- gate of 18,000,000 lbs. per acre to the depth of 54 inches. Adopting this estimate, and the percentages of moisture given in Table III., it results that down to the depth of 54 inches, or 4 feet 6 inches, the unmanured soil retained 1546, the soil of Plot 9, 1346, and that of Plot 14, 1221 tons of water. That is to say, to the depth of 4 feet 6 inches, the soil of Plot 9, manured with mineral manure and ammonia-salts, contained 200 tons, and that of Plot 14, manured with mineral manure and nitrate of soda, 325 tons less water per acre than that of the unma- nured soil to the same depth ; whilst, from the great difference in the percentage at the lowest depths taken in the three cases, there can be no doubt that the difference extended considerably deeper still. Here, then, we have evidence of the source whence the ma- 12 Drought of \^10 and nured crops derived the water required for their growth, over and above that supplied by the rain actually falling during the period of active vegetation. But the questions obviously arise — if the unmanured subsoil retained so much more water, why did the crop suffer from the drought so very much more than the manured crops ? and why did the crop manured with mineral manure and ammonia-salts suffer so much more than that manured with mineral manure and nitrate of soda, and not avail itself so fully as did the latter of the stores of moisture within the soil ? To gain some information on the points here suggested, careful examination was made of the distribution of species on the square yard of the plot selected, of the section of the soil and subsoil, and of the distribution of roots within them. It should be stated that 53 species in all are found on the continuously unmanured plot; this great complexity of herbage being maintained in consequence of the little encouragement to luxuriance of any. On the other hand, by the application of mineral manure and ammonia-salts on Plot 9, and of mineral manure and nitrate of soda on Plot 14, for many years in suc- cession, and the consequent great encouragement and predomi- nance of certain individual species, the total number discernible has become reduced to 30 on each of these plots. And whils the herbage on the unmanured plot comprises 17 graminaceous, 4 leguminous, and 32 miscellaneous or weedy species, that of Plot 9 includes only 15 graminaceous, 2 leguminous, and 18 miscellaneous species, and that of Plot 14 only 14 gramina- ceous, 3 leguminous, and 13 miscellaneous species. But such, again, is the difference in the character of the two nitrogenous manures — ammonia-salts and nitrate of soda — in regard to their reactions upon the soil, and the consequent degree of rapidity and range of distribution of them or their products of decomposition within it, that they respectively encourage the development of species of widely different underground, as well as above-ground habit of growth. Thus, the dominant plants were very different on the two manured plots. Under the influence of the annual application of mineral manure and ammonia-salts, Dactylis glomerata (rough cock's-foot), Ac/rostis vulgaris (common bent-grass), Festuca ovina (sheep's-fescue), and Poa j)ratensis (common meadow-grass), among graminaceous plants, and Rumex acetosa (sorrel-dock), among the miscellaneous herbage, prevailed somewhat in the order of enumeration ; whilst under the influence of mineral manure and nitrate of soda Bromus nioUis (soft brome-grass), had become so prominent as to constitute probably about one-half the crop ; Poa trivialis (rough meadow- grass) was also very prominent, Hulcus lanatus (woolly soft-grass). Experimental Crops at Rotkamsted. 13 Festuca ovina (sheep's-fescue), Lolium perenne (rye-p^rass), Dactylis glomerata (rough cock's-foot), Avena Jlavescens (yellow oat-grass), and among weeds Anthriscns sylvestris (wild beaked-parsley), coming next in order of prevalence. And, whilst the plants most encouraged by the ammonia-salts have a tufty habit of growth above ground, and a tendency to luxuriate within a limited range beneath the surface, some of those most favoured by the nitrate of soda, and especially under its influence, are very different in character, not growing in tufts, but producing comparatively uni- formly dense herbage, with many stems, comparatively few root- leaves, and roots having a characteristically downward tendency, those of the Bromus mollis especially (which contributed such a large proportion of the whole crop) being strong and wiry, and descending far into the subsoil. The sectional examinations, indeed, showed great differences in the character of the turf, in the prevalence and character of de- velopment of the roots within and below it, and in the character of the soil and subsoil, as the following brief abstract of the observations made will show. It should be first stated, however, that whilst on the square yard selected as characteristic of the unmanured plot, there were found 9 graminaceous, 4 legu- minous, and 11 miscellaneous — in all 24 species ; on that of Plot 9, having mineral manure and ammonia-salts, there were only 6 graminaceous, no leguminous, and only 3 miscellaneous spe- cies ; and on that of Plot 14, receiving mineral manure and nitrate of soda, again only 6 graminaceous, only 1 leguminous, and 2 miscellaneous species. Owing to the great complexity of the herbage on the un- manured plot, including a comparatively large number of legu- minous, and miscellaneous or weedy species, some fleshy roots were observed at a considerable depth. The turf consisted of a complex network of fine roots and fibrils, which were much less in size and strength than in the case of either of the manured plots. These fine roots seemed to have more or less complete possession of the soil to a depth of about 6 inches, and some of them then showed a downward tendency ; becoming, however, much fewer, and even in the second and third 9 inches extremely fine ; and at a depth of about 40 inches they were as fine as a fibre of silk or a spider's web. It was concluded, though not Avith great certainty, that the roots found at the greatest depth were those of Agrostis vulgaris and Bromus mollis. The sample of the first 9 inches of the unmanured soil possessed the character of mould not much less than that of the manured plots ; the second 9 inches, too, was very much altered from the character of the clay subsoil ; but below this point very slight difference was observ- 14 Drouyht of 1870 and able ; though, of the four lower samples, the uppermost, that is, the third from the surface, perhaps showed slightly the least, and the lowest, or sixth, the brightest red tinge. The turf of Plot 9, manured with mineral manure and am- monia-salts, consisted of a dense, almost peat-like mass, of de- composing roots, radicle leaves, and stubble, thickly penetrated with strong roots and fibrils, the whole being as much matted as on the unmanured soil, showing, however, less complexity, but greater strength of roots. The horizontal subterraneous stems of the Agrostis vulgaris greatly predominated, emitting many fibrils, and sending out many descending fibrous roots. Poa pratensis also developed a large amount of strong root, and a profusion of fibrils. Roots penetrated to about the same depth as on Plot 3, but in larger quantity, and of larger size ; being, however, in the fifth 9 inches, both very few in number and very fine. As already said, the samples of the first 9 inches of the soil of the three plots differed comparatively little from one another in the degree of their change by the action of vegeta- tion ; but, if anything, that of this Plot 9 was the darkest, indi- cating so far more of mould-like character. The second 9 inches of this plot was decidedly more changed than that of the un- manured, or of even Plot 14. The third and fourth 9 inches were, comj)ared with the unmanured, slightly darker, or less bright in colour, showing still some change. The fifth and sixth were little, if at all, distinguishable in colour from the raw, red- dish-yellow clay of the unmanured plot at corresponding depths. The turf of Plot 14, manured with mineral manure and nitrate of soda, had not the peaty appearance of that of Plot 9 ; the prevailing plant, Bromiis mollis, which made up about half the crop, possessing comparatively few radicle leaves ; whilst, especially under the influence of this manure, Poa trivialis, Holcus lanatus, and Loliiim perenne, have a tendency to assume the same character of development above ground. The Bromus mollis, too, was found in a most striking degree to send down strong wiry roots into the subsoil, leaving only its fibrils, and the roots of less prominent or smaller species, to feed near the surface. The second 3 inches of soil also held together, being full of fibre. At the extremity of the fibrils of the Broirms mollis small tubercles, much like those which occur on the roots of some legu- minous plants, were observed down to a depth of perhaps 12 or 14 inches. The roots of this grass extended, however, to a depth of nearly 4 feet, still maintaining their wiry character. The dif- ference in the character of the samples of soil, and especially of the subsoil, of this compared with those of either of the other plots, was very striking. The first 9 inches differed little from that of the unmanured plot. The second was, however, more altered Experimental Crops at Rothamsted. 15 than tliat of the unmanured plot at the correspondino: depth. The third, fourth, fifth, and sixth 9 inches were very strikingly different in appearance from the corresponding layers of either of the other two plots ; the clay, instead of being of a compara- tively uniform reddish yellow colour, was very much mottled or veined, showing a mixture of yellow, grey, red, and brown, with the yellow and grey predominating. So much was this the case that when the samples were powdered they were of a yellowish grey colour, instead of reddish yellow ; and the lighter or less yellow the greater the depth of the sample, that of the sixth 9 inches being the lightest of all. There was, perhaps, more of natural grey vein in the subsoil of this than in that of the other plots, but the difference in colour and texture was too great to be so accounted for. Upon the whole the lower layers were softer and more soapy than in the case of either Plot 3 or Plot 9 ; though, as Table III. at page 10 shows, they contained a considerably less percentage of moisture. Indeed, the subsoil of this plot had much more the appearance of disintegration from some cause than that of either of the others ; it was consequently much more easily worked, and especially more so than that of the unmanured plot, which was very tough and hard. To sum up these distinctions : it is seen that not only did different plants become dominant according to the different condition of the plot as to manure, but those which prevailed on the unmanured land, though numerous, had much finer and much less vigorous roots ; the raw clay of the subsoil was much less changed ; and it had yielded up very much less moisture to the growing crop. On the plot manured with mineral manure and ammonia-salts free-growing grasses predominated ; but chiefly those whose underground habit of growth was such as rendered them dependent for their food and moisture in great measure on that which is to be found in the upper layers of the soil. Still, owing to the increased vigour of growth under the influence of the manure, it is seen that moisture was obtained, either directly by the roots of the plants, or by capillary action induced by the pumping out of the upper layers, from the extreme depths to which the samples were taken ; and, from the great difference in the percentage of moisture at that depth compared with that of the unmanured plot, there is no doubt that the action extended deeper still. On Plot 14, on the other hand, where nitrate of soda was applied, the plant which con- tributed about half the produce had roots of a very characteristi- cally downward tendency. We find the soil, to the depths examined, pumped drier still ; and, coincidently, the drought has comparatively little affected the amount of the crop. IG Drought of 1^10 and Intimately connected with the greater change in the subsoil of the plot manured with nitrate of soda than in that manured with ammonia-salts, with the greater predominance and luxuri- ance of the deeper-feeding herbage, and with the consequent little evil effects from the drought where the nitrate was employed, is doubtless the fact that the ammonia of the ammonia-salts is much more readily absorbed and retained by the soil than is the nitric acid of the nitrate. The latter, consequently, becomes, under the influence of rain, more rapidly distributed and washed into the subsoil, whither the roots follow it. As this filtration, into and through the subsoil, of a solution of the nitrate, or of its products of decomposition within the soil, has been proceeding for thirteen years in succession, there is little cause for surprise that the subsoil should have become much more changed than where the ammonia-salts had been used. It seems intelligible, too, that those plants of the herbage, whose habit of growth is characterised by a comparatively large development of descend- ing roots, aided as they would be when once they had asserted their predominance by more and more self-sowing each succeeding year, should get such complete possession of the lower layers of the soil, with their stores of food and moisture. On this point it may be remarked, that the Bromus mollis, which so strikingly predomi- nated on the nitrated plot, and whose roots, though only a biennial, had obtained more complete possession of the subsoil than those of any other plant, is one of the earliest of the grasses, and has, in point of fact, generally seeded to a greater or less extent before the crop has been cut. It may be here mentioned in passing, that, wherever, in the course of the experiments at Rothamsted, nitrate of soda is employed year after year on the same plot of arable land, the difference in the appearance and texture of the soil is very great, and is discernible at a considerable distance. The soil appa-^ rently retains very much more moisture, becomes more aggluti- nated, and so sticky compared with that of adjoining plots under equal conditions of weather, as to be with difficulty worked at the same time, and never brought to the same tilth without the expenditure of extra labour upon it. It may be judged, indeed, that during the wet season the nitrated soil, and its more dis- integrated subsoil, would acquire more moisture, or at least more available moisture, than the soil and raw clayey subsoil of the other plots. We have, then, in the properties of the nitrate of soda and its effects upon the soil and subsoil, in the influence of these in determining the character of the prevailing herbage, and in the comparative independence of external sources of moisture which a deep root range gives to the plants encouraged, an explanation Experimental Crops at Rothamsfed. 17 of the fact that, notwithstanding the unusual drought of 1870, which almost suspended the growth of the unmanured herbage, and much diminished that manured with mineral manure and ammonia-salts, the plants which had gradually asserted posses- sion over others on the plot continuously manured with mineral manure and nitrate of soda, should have yielded, under the same circumstances of scarcity of rain, an all but average crop. Before leaving the subject of the influence of the drought of 1870 on the hay-crop, it may be added that a portion of the park adjoining the experimental plots was liberally manured with London stable-dung, but no benefit whatever was apparent, and the crop was so light as to be scarcely worth mowing. The evidence at command in regard to the effects of the drought on other of the experimental crops, is not of the same, or in some respects of so direct a kind, as that relating to the mixed herbage, and to the soils, of the experimental plots of grass land. Nevertheless, some facts of interest may be recorded illustrating the influence of the moisture stored up within the soil on the growth of both wheat and barley. Kesults relating to the Geowth of Wheat. The following Table (IV.) shows the amounts of grain, and the amounts of total produce (corn and straw together), obtained in the experimental wheat-field for 19 years in succession, 1852- 1870 inclusive : — 1. On Plot 3, continuously unmanured. 2. On Plot 2, receiving ,14 tons farmyard manure per acre per annum. 3. On Plot 7, receiving, annually, mixed mineral manure, and 400 lbs. ammonia-salts per acre. 4. On Plot 9a, receiving, annually, the same mixed mineral manure as plot 7, and 550 lbs. nitrate of soda per acre. The Table also shows, side by side with the amounts of produce, the fall of rain each year during the months of April, May, June, and July, which may be said to include the period of active vegetation and accumulation of substance. It should be further explained, that, in order that the different amounts of grain from year to year may be more strictly comparable one with another, and to avoid the necessity of recording and con- sidering the weight per bushel in each case, the total weight of dressed corn has been divided by 61, and the Table shows, there- fore, not the actual number of measured bushels in each case, but the number of bushels of an assumed uniform weight of 61 lbs. IS Droiu/ht of 1870 and i 1 P3 1 ^ 1 Inches. 9^34 12-69 6-50 11-35 10-70 7-08 9-28 10-53 14-49 8-49 10-96 7-27 5-81 7-13 10-71 11-33 3-66 7-40 3-91 00 CO 00 1> i |S?SS;^S2§g2§ggSSS?;jS2 CO CO i-1 cc C(N.*0«D — — coco — C0lr-(MO5t-(M03Tj'CDr-«5>fiOC0Oa5 ^ j^TfcOO— -^C'lOCMCOtNCO't^C'i'rHO — C T3 STtfCOOOiMOO-^OJOTf — — OOinrflfiCOCOin ^oor-cocor ^lnscooo50ooo in I- m ?i =2 CO C.2 .^oo-j'-tys — lnt-cn'MoocJ>(^^-tc■. 00 — — tC l^ 'a < 5 O CO O O « OJ C) C-l ^ — Ol O — O — (.N ^ -* O !M in 05 Jr^ -* -^ if ' t^ t^ 00 05 (N T') (N CO CO 2 CO ;::; ,i s 00 -f Tf< CO CO on 05 o o nn CO o CO ; i CO -f 00 t^ in CO n sS (^ ,^ in c o 00 (^ oo (^ t^ (35 ,^ o n c 111 CO ^ in ^ ^ on rv) m ■* ^ t^ -# ^ 1 in 't o ■^ 00 t>. t^ 05 (35 I 6^ (N o in o CO rr CO -t" o Cf) o in o if5 ^5 -* '~' '^ "" ^ '^ '^ ■"■ _ o ^ -t o in CO ^ CO in CO CO c. s 3 w CO in -* — ' ou 00 in o o C5 ,^ -f CO c» CI ^ m in CO CO _). in !>) (M (M (M (M (M (M (M (>J 8 ^Pi CO CO O) ^- H. C5 (35 on ^ (35 CO o o Hi!' in Ol t^ 05 i CO (M o -f -f -H CO 00 •^ on 00 f^ in . i (M t^ CO C5 CO CO g •^ CO '- en :n CO CO :i CO CO in 2 -M ^^ ,s s; 1 sl-i . O I# 00 CO 1-5 CO 4i< o CO ■■£> 00 Ci CO in 00 f- 00 CO 00 (io (35 (N I ^ ^ ^ ^ S 111 7 o t- 00 f w C) a-i CO CO ■* i O ►J — -^ ■—1 — ' '-' — -- d " cS ■rj o m o CO 05 'f CO -+ -* Tf O (M a> (N (M o 00 <~i ■<* o ,:= 5 ■* ,^ on o ^ CO 'f in CO m ■* in (M c. ^ c oS-|. c o> 00 ^ lO CO t^ an 7g 48J lbs. 2333 S8.'i3 2090 lbs. 28 C 8 4265 2050 lbs. 4311 6701 4287 lbs. 5454 7194 4621 Mean \A\ 5(:1 2759 3061 5100 5756 Here, then, we have again a similar result. There is, too, proportionately a greater increase with the nitrate, especially of corn, in the two drier and hotter seasons of 1868 and 1870 — vears, in fact, of summer drought. The following Table shows the produce of barley without manure, with farmyard manure, and with mixed mineral manure and 200 lbs. ammonia-salts per acre, in 1868, and in 1870, the two recent years of summer drought; and also, under the same condi- tions as to manure, the average produce over the nineteen years of the experiment. As before, the number of bushels of dressed corn, reckoned at an uniform weight of 52 lbs. per bushel, is given. And, side by side with these records of produce, is given the Experimental Crops at Rothamsted, 31 amounts of rain at Rothamsted, in April, May, June, and Jtdy, each year, those being the months of active growth of the barley crop. Table YIII. Dressed Corn. Total Proddce. Rainfall at Rothamsted. (In bushels of 52 lbs.) (Corn and atraw). t g 1 1. u So II Mean. i 1 ■s il 11 Mean. April. May. June. July. Total. ^ fo Si is ^ % Bush. Bush. 1 Bush. Bush. lbs. lbs. lbs. lbs. Ins. Ins. Ins. Ins. Ins. 186:! .. .. 1 Ui 471 374 32 1902 5281 4311 3831 2-19 0-73 0-37 0-37 3-66 1870 .. .. 1 131 52i 41| 351 1489 4949 4287 3575 0-46 1-35 0-98 1-12 3-91 Average, 19 Years, > 1862-1870 . . ( ! 20 50i 48i 39^ 2433 5856 5786 4698 1-72 2-36 2-43 2-37 8-88 As there has been a decline in the produce without manure during the second as compared with the first half of the period over which the experiments have extended, the difference indi- cated between the unmanured produce in the years of drought and that over the nineteen years will exaggerate the deficiency due to the deficient rainfall alone during the four growing months of the two years in question. On the other hand, the produce by farmyard manure has considerably increased during the latter half of the period, and hence the deficiency in the years of drought which the figures show for that manure is less than is due to the characters of the seasons alone. With the artificial manure the produce was, however, very much more nearly equal during the first and second halves of the total period, and the indicated deficiency in the years of drought probably more nearly represents that really due to the characters of the seasons in its case. With this manure there was a deficiency compared with the average, of 11 bushels of corn in 1868, and of 6f bushels in 1870 ; or, of total produce, of 1475 lbs. in 1868, and of 1499 lbs. in 1870. There was not far from an equal total amount of rain during the four months in the two seasons ; but whilst there was more than an average fall in April, 1868, and only about one-fourth the average fall in April, 1870, there was a greater deficiency in May, June, and July, 1868, than in the same months in 1870. The result was a greater deficiency of corn, but a less deficiency of straw, in 1868 than in 1870. We are enabled to adduce more direct experimental evidence 32 Dromjht of 1870 and showing' the extent to which the barley-plant can avail itself of the stores of moisture within the soil, than that which was at command relating- to wheat. Before considering the results themselves, to which reference is here made, it will be well to describe briefly the circumstances under which they were obtained. With a view to the determi- nation of what proportion of the rainfall passes to given depths in the subsoil, under different conditions of season, manuring, and cropping, a series of experiments has been commenced, for the cutting off, and the collection, of the drainage-water from the land at different depths — an essential condition being that neither soil nor subsoil should be disturbed. Leaving out of view for the present the questions of the influence of different manures, or of the growth of different crops, early in 1870 three plots of un- cropped land, each of one-thousandth of an acre area, were selected, with a view of determining the amount of water passing below the depths of 20, 40, and 60 inches, respectively. The plan of operating was, to cut a sufficiently wide trench for men to work in, down one side of the plot, to a considerably greater depth than that at which the drainage was to be cut off. The plot was then carefully undermined and shored up at the depth decided upon, until a cast-iron plate, rather more than the length of the plot, 8 inches wide, and having small holes for the water to drain through, could be got in and fixed underneath. The plot was then further undermined, until another plate could be put in ; and so on, until the whole w^as supported at the proper depth, without disturbance, by a perforated iron flooring, which finally was itself supported on three sides by brickwork, and on the fourth and across the middle by iron girders. The three as yet undisturbed sides of the plot were then trenched round ; a 4^-inch brick and cement wall was built round the plot, resting on the projecting rim of the iron flooring below, and finished level with the surface above. The trench outside the wall was then filled in again. Thus, the exact area required was cut off from the surrounding soil by brickwork at the sides, and below, at the depth required, by a perforated iron flooring. The field in which these drain-gavges were made, had grown wheat in 1869, and was sown with barley in March, 1870, and the drill by mistake was allowed to sow two rows of seed on the plots along one side of them. As the excavations proceeded, barley-roots were observed to have extended to a depth of between 4 and 5 feet, and the clayey subsoil appeared to be much more disintegrated, and much drier, where the roots had penetrated than where they had not. Accordingly, it was decided to make careful notes on the sections under the two conditions, and also to take samples of soil and subsoil to a depth below that at which roots Experimental Crops at Rot/tamstcd. 33 were traced, with a view to the determination of the amounts of moisture at the different depths in the two cases. Portions of the barley-ground and the fallow-ground, closely adjoining the drain-gauge plots, but undisturbed by the excavations in connec- tion with them, were selected, and from each six samples, 6x6 inches superficies by 9 inches deep, that is, in all to a depth of 54 inches, were taken. The following Table shows the percentages of moisture in the different samples, including that lost during their preparation, as well as that aiterwards expelled at a temperature of 212° Fahr. : — Table IX. — Percentages of Moisture in Uncropped and in Cropped Land, at different depths. Samples collected June 27th and 28th, 1870. Depth of Sample. Fallow Land. Barley Land. Difforencp. First 9 inches Second 9 , , Third 9 ,, Fourth 9 , , Fifth 9 ,, . Sixth 9 ,, 20 -.S6 29 - 53 34-84 34-32 31-31 33-55 11-91 19-32 22-83 25-09 26-98 2G-38 8-45 10-21 12-01 9-23 4-33 7-17 Mean 30-65 22-09 8-r,a Before commenting on these results, it should be stated that, ten days previous to the collection of the samples, about two- thirds of an inch of rain had fallen, and only three days before the collection about one-tenth of an inch ; and hence, perhaps, may in part be accounted for the somewliat high percentage of moisture in both soils near the surface at that period of a season which was upon the whole one of unusual drought. Further, for a iew days during the interval since the heavier rainfall, some soil, thrown out from the excavations near, had laid upon the spot whence the samples from the uncropped land were taken, and hence, again, may be accounted for part of the excess near the surface in the uncropped as compared with the cropped land. The difference between the amounts of water retained at the depths examined by the uncropped and the cropped ground, at points only a few feet apart, is very striking ; and that it should be greater in the upper portions of the subsoil, which had pro- bably contributed more to the exigencies of the growing crop than the lower layers, is what would be expected. The percentage of water in the subsoil even of the cropped land was very high — indeed nearly as high at corresponding depths as in that in the 34 Drought of 1870 and experimental wheat-field in January, 1869, when it was supposed to be in a state of saturation; whilst the amount in the subsoil of the uncropped land was not only considerably higher than in that of the cropped land, but considerably higher also than in that of the saturated wheat soil. We shall recur presently to the difference in the percentage of moisture in the soils and subsoils of the different fields which have been referred to, but must first direct attention to the more special application of the results now under consideration. The following Table shows the number of tons of water per acre retained to the total depth of 54 inches, or 4|^ feet, by the uncropped and the cropped land, and the difference between the two. The upper line gives the amounts calculated according to the actual weights of the measured samples of soil (exclusive of stones), and the lower line the amounts, assuming that (exclu- sive of stones), the dry or barley soil would weigh 18, and the wet, uncropped or fallow soil 19| million lbs., to the depth of 54 inches : — Table X. — Tons of Water per Acre to the depth of 54 inches, in Fallow Land, and in Land Cropped with Barley. Samples collected June 27th and 28th, 1870. w ATER PER A-IKE. Fallow Land. Barley Land. Difference. According to experimentally determined! weights of soil / According to assumed average weights! of soil / Tons. 287.5 2Gfi8 Tons. 1951 177.^ Tons. 924 893 Mean 2772 1863 909 On whichever basis the calculation is made, the indication is that there were about 900 tons less water per acre in the soil and subsoil, to the depth of 4 feet 6 inches, where the barley had grown than where the land was fallow. It may be that jiart of the excess in the uncropped land was due to the shelter from surface evaporation since the last preceding heavy rain, by the laying of soil upon it for a {ew days, as above referred to. But even supposing a liberal deduction on this account, the evidence would still point to the conclusion that there had been a higher rate of exhalation by the growing crop than oOO parts of water for every 1 part of dry substance fixed ; for it may Experimental Crops at Rothamsted. 35 safely be assumed that the dry matter of the crop at the time of the experiment would be under rather than over 2 tons per acre, which, at the rate of 300 parts to 1, would only account for an exhalation of 600 tons of water per acre. Further, since there was such a great difference in the per- centage of moisture in the two cases at the lowest depth taken, it is only reasonable to conclude that the difference extended lower still. To conclude, in reference to these particular experiments, it is clear that we have in the facts adduced sufficient evidence, and a striking illustration, of the enormous extent to which, in a time of drought, our crops may rely upon the supplies of moisture previously stored up within the soil. At the same time it cannot fail to be recognised how dependent must be the result upon the character of the soil and the subsoil with which the farmer may have to deal. Summary, and General Observations. Leaving detail, it will be of interest to summarise the results illustrating the difference of effect of the drought of the past year on the different crops, and also to bring together those relating to the amount of water retained by the soils and subsoils of the different fields, under the various conditions as to season, manuring, and cropping. It has been already said that although the summers of both 1868 and 1870 were seasons of drought, yet, chiefly owing to the facts that the deficiency of rain commenced later, and the tem- peratures ruled higher in 1868, there was in reality considerable difference in the characters of the periods of growth of the two seasons, and in their consequent effects upon the different crops. To save space, however, we will confine attention here to the effects on the different crops of the more continued drought of 1870. Table XI. shows the average annual produce obtained, under selected conditions as to manure, of hay, of wheat, and of barley ; also the produce of each in 1870, and the deficiency compared with the average. In the case of the hay, the average is taken over 15 years, and in that of wheat and barley over 19 years. For simplicity of comparison, the produce is, for all three crops, given in lbs. ; and the figures relating to wheat and barley repre- sent the total produce, corn and straw together — which, of course, more clearly indicates the total amount of vegetable growth, com- pared with that of the hay, than the records of corn and straw separately would do. D 36 Drought of 1870 and Table XI. — Produce of Hay, Wheat, and Barley in 1870 compared with the average. Hay; 15 Years. Total Produce, Corn and Straw. Wheat ; 19 Years. Barley ; 19 Years. Without Manure. Average produce per acre per annum . . Produce in 1870 lbs. 2391 644 lbs. 2398 2002 lbs. 2453 1489 Deficiency in 1870 1747 396 964 With Farmyard Manure. Average produce per acre per annum .. Produce in 1870 4604* 1556 6016 5092 5856 4949 Deficiency in 1870 3048 924 907 With Mixed Mineral Manure and Ammo nia-salts. Average produce per acre per annum . . Produce in 1870 5794 3306 6267 5836 5786 4287 Deficiency in 1870 2488 431 1499 It is remarkable that, notwithstanding the great fluctuation in the amounts of produce of each of the three crops from year to year according to season, and also the difference in the degree in which each will vary from the average in one and the same season, still, when the average is taken over a considerable number of years, hay, wheat, and barley, are seen to yield tcithout manure almost identically the same average weight of produce per acre per annum. On this point it should be mentioned that the second crop of grass is never removed from the land, being either consumed on it by sheep having no other food, or mown and left to rot as manure. The deficiency without manure, due to the drought of 1870, is seen to be 1747 lbs. of hay, 964 lbs. of barley (corn and straw), and only 396 lbs. of total produce of wheat. Thus, the deficiency Avas much the greatest in the hay ; there being a reduction in its case by nearly three-fourths, in that * For the hay crop, farmyard manure was only applied in the first 8 years ; but the average produce is taken over the 1 5 years. Experimental Crops at Rothamsted. 37 of the barley by scarcely two-fifths, and in that of the wheat by only about one-sixth, compared with average amounts. For the hay-crop, farmyard manure was only applied during the first 8 years of the 15 ; but as the average produce was as great over the succeeding 6 years without the manure, as over the first 8 years with it, and as there was a heavier crop in 18G9 than in any of the preceding 13 years, the deficiency in 1870 compared with the average, may be taken as at any rate mainly due to the drought, and but little to the cessation of the manuring. The figures as they stand show, as without manure, again, a much greater deficiency than in either wheat or barley ; the crop amounting in fact to only one-third the average. Of total produce of wheat and barley, there is, with farmyard manure, again nearly the same average amount over 19 years in the two cases. The deficiency in 1870 compared with the average is also very nearly the same with the autumn-sown wheat and the spring-sown barley ; amounting in each case to scarcely one-sixth. In the wheat the reduction is actually much greater, but in proportion to the average, only about the same as without manure; but in the barley it is actually less, though in proportion to the average very much less, than without manure. The greater power of retention of water which a dunged soil has been shown to possess in its upper layers, has doubtless much to do with the result. With the artificial mixture, in the case of the hay and the wheat supplying 400 lbs., but in that of the barley only 200 lbs. of ammonia-salts per acre per annum, there is not the same uniformity in the average annual produce of the three crops ; the wheat giving nearly 500 lbs. more gross produce than the hay with the same amount of ammonia applied, and the barley about the same as the hay, with only half the supply of ammonia- salts. The deficiency in 1870 amounts, in the hay to more than two-fifths, in the barley to rather more than one-fourth, and in the wheat to little more than one-fifteenth, compared with the average. Thus, then, with a drought extending over the months of April, May, June, and July, the mixed herbage of permanent meadow land suffered, under the different conditions of manure in question, very much more than either wheat or barley ; and the spring-sown barley suffered, both without manure and with the artificial manure, very much more than the autumn-sown wheat. With the farmyard manure, however, the barley would appear to have been as little adversely affected by the deficiency of rain during the period of actual growth as the wheat. We need not here again refer to the special conditions already ex- plained, under which the hay crop was as little, or less, affected by the drought than the other crops. D 2 38 Drought of ISIO and The difference between the conditions of growth of the chiefly perennial (or biennial) plants composing the complex mixed herbage of permanent meadow land, and those of an annual, like wheat or barley, sown at a stated period of the year in arable land, and having a fixed, and in the case of barley only a limited time for distributing its underground feeders, and so availing itself of the resources of nutriment and moisture within the soil, are obviously very great. The perennial, or biennial, character of most of the plants composing the mixed herbage, would seem at first sight to give the grass a great advantage over the corn crops. But observa- tion shows, that although the immediately superficial layers of the soil may be more thoroughly penetrated by the roots of the perennial grasses than by those of either wheat or barley, yet it is only a very few of the former, encouraged to great pre- dominance only under special conditions, that seem to get any- thing like the same possession of the lower layers of the soil as the two corn crops. Careful examination has also shown, and it is probably generally assumed, that the winter-sown wheat secures possession by its underground feeders of a more extended range c>nd greater bulk of soil, and consequently is better able to avail itself of the supplies of food and moisture existing below a certain limited depth from the surface, than the spring-sown barley. The wheat-plant, indeed, has the advantage of making root, more or less according to season and manure, throughout the winter months, during periods of which, at any rate, the soil Avill be saturated with moisture ; and in the case of moderately retentive and well drained soils, it will be able to establish its independence of rain falling during the period of active above- ground growth, very much more than will a spring-sown crop like barley. But there are other points of distinction between the growth of the corn and the hay crops. Thus, most of the grasses, which comprise the greater proportion of the latter, flower earlier than the wheat or the barley ; and the mixed herbage is cut by, or before, the end of June, when very little, if any of it, has arrived at the degree of ripeness in which the corn crops are cut. These, on the other hand, are not only allowed fully to ripen, but direct experiments made at Rothamsted upon wheat have shown that a very large proportion, probably about half, of the total dry vegetable substance, or of the total carbon of the crop, is fixed in it under the influence of the greater power of the sun's rays after the time at which the hay crop is usually cut. These facts are obviously an element in the explanation of another fact, to a certain extent commonly recognised, and which a careful comparison of the results of the field experiments at Experimental Crops at Rolhamsted. 39 Rothamsted, with the records of the conditions of heat and moisture under which the crops have been g'rown, brings clearly to view — namely, that, as compared with the hay crop, the corn crops are not only less dependent on the amounts of rain falling during the period of active vegetation, but more on a relatively high degree of temperature during that period. This is more strikingly the case when wheat is grown by means of readily soluble mineral and nitrogenous manures, than when it is grown without manure, or with farmyard manure. Without manure the produce is comparatively more dependent on the amount of certain constituents brought down by the rain, or rendered available by its means from the stores of the soil itself; and it would seem that where farmyard manure is employed, a con- siderable amount of rain is required during the early growing period to aid its decomposition, and so to set free, distribute, and render available, its fertilising constituents. In the case of the artificial manures, on the other hand, some of the most active fertilising constituents are supplied in a much more soluble form, and require a less amount and continuity of rain for their solu- tion and distribution throughout the pores of the soil within a given range. It is seen, then, that several reasons concur to render corn crops less dependent on the fluctuations in the amount of rain falling during the period of active vegetation and accumulation of substance than is the hay crop growing under otherwise parallel conditions as to soil and manure. It is quite intelligible, too, that the autumn-sown wheat, with its much longer time for the formation and distribution of root, and its tendency to develop proportionally more in the lower and proportionally less in the upper layers of the soil, than the spring-sown barley, should be less adversely affected than the latter by a deficiency of rain during the period of active above-ground growth. Table XII. brings together at one view the percentage amounts of water retained by the soils and subsoils of the different fields, under the various conditions as to season, cropping, &c. The results so summarised relate to samples collected as under : — 1. From the experimental wheat field, just before harvest, 1868; mean of three plots differently manured. 2. From the experimental wheat field, in January, 1869, when the land was supposed to be saturated ; mean of the same three plots differently manured. 3. From uncropped land, near the end of June, 1870. 4. From land cropped with barley, closely adjoining the un- cropped land ; samples collected at the same date, end of June, 1870. 40 Drought of 1870 and 5. From permanent meadow land, in July, 1870, after the removal of the crop ; mean of three plots differently manured. Table XII. — Summary of Percentages of Moisture in Soils and Subsoils from different Fields, and under difierent conditions as to Season, Cropping, &c. Depths of Samples. Experimental Wheat Field. Barn Field. Samples collected, June 27th and 28th, 1870. Perm A KENT Meadow Land. Samples collected, Samples collected, July, 186S ; Jan. 6tb and 7 th, Mean of 1S69 ; Mean of Plots 3, 2, and 8a. ' Plots 3, 2, and 8a. Samples collected, July 25th and 26th, 1870 ; Mean of Plots 3, 9, and 14. Un^ropped Land Growing Barley. First 9 ins. Second 9 ,, Third 9 ,, Fourth 9 , , 6-23 27-17 11-19 22-70 15-02 1 25-27 16-13 ; 25-65 20-36 29-53 34'84 34-32 11-91 19-32 22-83 25-09 11-99 11-77 17-11 19-32 Mean 36 , , Fifth 9 ,, Sixth 9 ,, 12-14 2519 29-76 31-31 33-55 19-79 26-98 26-38 15-05 20-67 21-49 Mean 54 , , | 1 30-65 22-09 17 06 The special application of the detailed results having been already fully considered, attention must be confined here to the more general indications only of the foregoing summary. In the first place, it should be observed that all three fields have a subsoil of reddish yellow clay, resting upon chalk, at a varying depth, but of not many feet from the surface. All, therefore, have good natural drainage ; and it is very seldom that any water collects in the furrows, and then only for a very few hours. The experimental wheat field is, however, pipe-drained at a depth of about 30 inches, and at a distance of about 25 feet from drain to drain. It is of interest to observe that there is no wide difference in the amount of water retained at corresponding depths in the experimental wheat-field in July 1868, when the crop was nearly at maturity, and in the permanent meadow land in July 1870, after the removal of the hay crop. The percentages are, however, rather lower in the drained land ; which, at the time, had probably supported a higher average amount of produce also. Towards the end of June 1870, the undrained arable land, which then carried a crop of growing barley, representing per- haps from 1^ to 2 tons of dry substance fixed, retained only about the same amount of water near the surface as the meadow land in July 1868 ; but, lower down, it held considerably more than either the drained wheat land in July 1868, or the undrained meadow land in July 1870. Experimental Crops at Rothamsted. 41 It is remarkable that the uncropped and undrained land, though retaining much less water within 9 inches from the sur- face, from that point downwards retained, in June 1870, con- siderably more at every stage than the drained wheat soil in January 1869, when the drains were running, and the land was supposed to be saturated. From this comparison, it is obvious that no safe conclusion can be drawn from the percentage of water in the subsoil of the uncropped but undrained land, as to the probable amount retained by the subsoil of the drained land at the commencement of active vegetation in the spring. The amount retained in the subsoil of the uncropped and undrained land is indeed enormous ; but the comparison of it with that in the adjoining cropped land shows clearly enough that it was readily available for the purposes of vegetation. In reference to this latter point, the fact of the good natural drainage by the chalk must not be overlooked. There is, upon the whole, general consistency in the results brought together in Table XII. It may, perhaps, safely be concluded that, notwithstanding the natural drainage by the chalk, the pipe-drains had contributed to reduce the percentage of moisture retained by the subsoil of the experimental wheat field, to the depth examined ; but that they had, at the same time, rendered the clay more permeable by roots, and the water that was retained more readily available. The evidence is, at any rate, very striking as to the degree in which, in a time of drought, our crops are enabled to rely upon the water previously accumulated within the subsoil — provided the latter be of sufficient depth, of sufficient retentive power, and at the same time sufficiently permeable. Before concluding, it will be well to call attention to a very important bearing of some of the results adduced. Assuming, as we may be allowed to do for the sake of illustration, that a good crop of hay, wheat, or barley, will probably exhale not less, and perhaps more, than 700 tons of water per acre during growth, we still have only about 7 inches of rain, out of an average annual fall of say 25 inches, thus directly disposed of by the growing crop ; and, taking the amount retained by the soil itself as practically a constant quantity from year to year, there re- mains to be disposed of by evaporation from the surface, and by passage into the drains or otherwise beyond the reach of the roots of the crop, an average of about 1 8 inches of rain annually, equivalent to more than 1800 tons of water per acre. How much of this large quantity of water passes off by evaporation from the surface of the soil itself, inducing by capillary action the withdrawal of water, carrying with it, it may be, essential plant-food, from tlie lower to the upper layers of the soil ? — or, how 42 Drought of 1870 and much passes downwards, carrying in solution any manurial matters in excess of the quantity which can be absorbed and retained within the pores of the s(nl and the upper layers of the subsoil ? These questions cannot be so satisfactorily answered in regard even to any particular soil, or season, as is desirable ; and could they be so, the answers would vary greatly with variations of soil and season. As already stated, direct experiments are now in progress at Rothamsted with the view of acquiring useful data on this subject. With regard to the results hitherto obtained, it may be remarked that, from September 1st to December 31st, 1870, that is, commencing after the unusual drought of the preceding summer, it Avas found that, out of a rainfall of about lOS inches within the same period, about 50 per cent, had passed below a depth of 20 inches, about 40 per cent, below 40 inches, and about 20 per cent, below 60 inches from the surface. Calcula- tion further showed that, even supposing there were some accu- mulation during August, still, a very large proportion of that which did not so pass, would be required to bring the previously very dry soil to the point of saturation — judging this require- ment from the results which have been already given bearing upon the point. That is to say, as would be expected, a com- paratively small proportion of the rainfall was evaporated at that season of the year. Much more would, of course, so disappear taking the whole year round ; the quantity varying considerably with the characters of the soil and the season. Towards the end of the last century, Dr. Dalton * devised an apparatus for the determination of the proportion of the rainfall which passed off from the soil by drainage, and by evaporation, respectively. It consisted of a cylinder, 10 inches in diameter, 3 feet deep, open at the top, and closed at the bottom ; but having one small exit tube near the top, and another near the bottom, for the escape of water into bottles placed to receive it. The vessel was filled with earth, and sunk into the ground level with the surface, one side being left exposed for access to the bottles. He continued the experiment for three years, 1796-7-8, and found the drainage to average, over that period, 25 per cent., and the evaporation to be, therefore, equal to 75 per cent, of the rainfall. This was exclusive of any evaporation of dew, but in- clusive of that resulting from vegetation, as the surface of the soil became, after the first year, covered with grass ; a circum- stance which, however. Dr. Dalton considered immaterial. For eight years, 1836-1843, Mr. Dickinson, of Abbott's Hill, King's Langley, Herts,t experimented with a modification of * Mem. Lit. Phil. Soc. of Manchester, vol. v., part 2. t ' Journal of the Royal Agricultural Society,' vol. v. Experimental Crops at Rothamsted. 43 Dalton's apparatus. The cylinder he employed was 12 inches in diameter, and 3 feet deep, but provided at that depth with a perforated bottom, and a receptacle beneath for the collection of the water ; and there was an arrangement of tubes for the escape, and measurement, of the drainage water. Grass was grown on the surface of the soil in the cylinder. The drainage would doubtless be more free in the experiments of Mr. Dickinson than in those of Dr. Dalton ; and the results, over 8 years, showed, with a less rainfall, a larger actual amount of drainage ; the latter representing 42 J per cent., and the evaporation, therefore, only 57^ per cent, of the rainfall. This amount included, of course, the exhalation due to vegetable growth. From results obtained by gauging the flow of water from pipe- drains, it has been concluded that a still larger proportion of the rainfall passes off by evaporation than that indicated by the experiments of either Mr. Dickinson or Dr. Dalton. But results obtained by deducting the amount passing through drains from the total rainfall may be judged to be quite untrustworthy, from the fact that, before the pipe-drains in the experimental wheat field had passed any water at all in the autumn of last year, the dj-ain- gauges already referred to had indicated that, of the rain which had then fallen since the 1st of September, nearly 25 per cent, had passed below 20 inches, nearly 10 per cent, below 40 inches, and nearly 4 per cent, below 60 inches from the surface. It is clear, therefore, that the amount of water passing through arti- ficial drains may be no measure whatever of the total quantity passing below the reach of the roots of growing crops.* In the admitted defect of satisfactory evidence from which may be deduced the probable average amount of evaporation from the surface of the soil independently of vegetation, we will assume, by way of illustration, that, taking the average of many * It was not until after this Paper went to press for the Journal, that we recol- lected the experiments of Maurice, Gasparin, and Risler, relating to this subject, and therefore make brief reference to them here. M. Maurice, experimenting at Geneva, over two years, 1796 and 1797, sought to measure the evaporation from the soil by means of a cylindrical iron vessel filled with earth, the changes in the weight of which he determined daily during that period. His results indicated an amount of evaporation corresponding to about 61 per cent, of the rainfall, which latter averaged over the two years about 26 inches per annum. (' Bibl. Britan. de Geneve : Sciences et Arts,' t. i.) M. Gasparin, experimenting in a somewhat similar way at Orange, in the south of France, found the evaporation so deter- mined to amount, in the two years 1821 and 1822, to about 80 per cent, of a rain- fall of about 28 inches per annum. (' Cours d'Agriculture,' t. ii. p. 116.) M. Risler, again, at Caleves, near Nyon, Switzerland, by gauging drains r2 metre (about 4 feetj deep, in a very compact and impervious subsoil, estimated the evaporation over the two years, 1857 and 1868, to amount to about 70 per cent, of a rainfall which averaged over the period about 41 inches per annum. The land was cropped, as usual, during the period of the experiment; so that the amouut of evaporation indicated includes that due to vegetable growth. (' Archives des Sciences de la Bibliotheque Universelle,' Sept., 1869.) E 44 Drouf/ht of 1810, .S'c. soils and seasons, three-fourths of a total rainfall of 25 inches will pass off by the combined action of evaporation from the sur- face of the soil itself, and of the exhalation due to the growth of a good crop of hay or corn. On this supposition there would still remain more than 6 inches of rain, equivalent to more than 600 tons of water per acre, annually passing downwards, and carrying with it more or less of fertilising matters. Fortunately, some of the most important mineral constituents of soils and manures are, in the case of the heavier soils at any rate, almost wholly retained by them within the range of the roots of our crops. Nitrogen, whether supplied in the form of ammonia- salts or nitrates is, however, much less completely so retained ; being, in whichever state supplied, carried off in greater or less quantity in the drainage water, chiefly in the form of nitrates. According to results obtained independently by Professor Frank- land and Professor Voelcker, on the analysis of drainage water frorn the experimental wheat-field at Rothamsted, that col- lected during the winter, from land manured in the autumn by an amount of ammonia-salts supplying 82 lbs. of nitrogen per acre, may contain from 2"5 to 3 parts, or even more, of nitrogen, as nitrates and nitrites, per 100,000 parts of water. Assuming that only 2'5 parts of nitrogen were so carried beyond the reach of roots for every 100,000 parts of water passing downwards, there would still be, for every inch of rain so passing, a loss per acre of between 5 and 6 lbs. of nitrogen, supplied in manure at a cost of not much less than Is. per lb. The above estimate of quantity must be understood to be adopted only provisionally, and by way of illustration. It is, however, a sufficiently near approximation to what must happen in the case of many soils and seasons at any rate, to show the very great importance of further investigating the reactions of various descriptions of nitrogenous manure on different descrip- tions of soil, and of determining the best modes, and the best periods of the year, for the application of such manurial matters, so as to reduce the loss by drainage to a minimum. This subject is now receiving- attention at Rothamsted. Ruthamsted, January, 1871. Library N. C. State Collog:e London Printed by William Clowes and Sons, Stamford Street, and Charing Cross.