s ."Rfc Cambridge Agricultural Monographs TBasic Slags m' T{pck T^hosp hates 'By G. S. 'Robertson Book JiiL CAMBRIDGE AGRICULTURAL MONOGRAPHS BASIC SLAGS AND KOCK PHOSPHATES 'b ^^ "^ u ^"'^"'^^ fN OR^.T pp,.„^_ DEDICATED TO THE MEMORY OF MY BROTHER WILLIAM SCOTT ROBERTSON LIEUT., ROYAL AIR FORCE BORN 1 JUNE, 1898, KILLED IN THE AIR OVER THE GERMAN LINES 13 JULY, 1918 A YOUNG LIFE WHICH GAVE EARLY PROMISE OF CONTRIBUTING TO THE ADVANCEMENT OF SCIENCE PREFACE By E. J. Russell, D.Sc, F.R.S. Director of the Rothamsted Experimental Station JLhe utilisation of basic slag in agriculture is an excellent example of the help that modern science affords to the working farmer. A waste product of steel-making, resulting from a modification by Thomas and Gilchrist in 1878 of the Bessemer process, it was at first considered worthless and thrown on the refuse heap. The late Prof. John Wrightson made field experiments in 1884 and 1885 at Eerryhill and at Downton, and showed that the material had noticeable fer- tihsing value: this discovery was confirmed and developed by the systematic pot experiments of Paul Wagner at Darmstadt, which began in 1885 and continued for several years afterwards. Extensive field tests were made during the 'nineties by Sir (then Professor) J. J. Dobbie and Prof. D. A. Gilchrist at Bangor, and by Prof. W. Somerville and later on by Sir T. H. Middleton at Cockle Park, with the result that a considerable body of information was accumulated as to the effectiveness of basic slag under the various conditions obtaining in practice. This has already been summarised by Prof. Somerville in the Journal of the Board of Agriculture for 1911, 1918, etc. Some ten years ago, however, it became evident that the basic open hearth process would be a serious competitor with the Bessemer process, and chemical examination showed that the slag, though correctly described as 'basic slag,' was altogether different from the material with which the agriculturist had become famihar. The upheaval caused by war and post-war conditions gave an enormous impetus to the open hearth process, and it is now extending to so many works that before long the older process will probably cease to be operated. This result is of course distinctly awkward for the agriculturist who sees a valuable f ertihser disappearing, and being replaced by one which is more costly and at first sight seems to be nothing Hke as good. Dr Scott Robertson has the great advantage of being in close touch with the steel-making industry, and at the same time of being able to carry out agricultural experiments. At the outset of his in- vestigations he made a careful selection of the types of slag likely to viii PREFACE be produced in the future and in 1915 laid out field experiments in Essex to compare these newer types with the old familiar Thomas or Bessemer slag. These experiments were continued for five years and they gave a mass of data so important in character as to deserve wide- spread circulation among farmers and agricultural experts. Separate publication was therefore advised and the Syndics agreed to its in- clusion in the Cambridge Agricultural Monographs. Fortunately Dr Scott Robertson had included also some typical mineral phosphates in the trials so that valuable information has been obtained in regard to a second problem which, while not pressing in 1915, has grown in importance since and is likely to be serious in the future. This second problem arises as a direct consequence of the circum- stance that basic slag is a by-product only, and not a primary object of manufacture. From the steel-makers' point of view it is relatively unimportant. Some 4 cwts. only are obtained for each ton of basic steel produced, and while the ton of steel has been worth anything from £27 in 1920 to £10 in 1921, the 4 cwts. of slag is worth less than 5s. to the steel-makers and only about 155. even after the slag grinder has graded, ground and bagged it. The steel-maker cannot afford to alter his processes in any way that would lengthen them or make them more costly or hazardous. The agriculturist must therefore take the slag as he finds it and cannot expect the consideration that would be shown him by the makers, say, of superphosphate, which is a primary object of manufacture and not a by-product. The practical result is that the composition of basic slag is determined by the conditions under which the steel-maker is working, and the total amount pro- ducible is regulated by the demand for steel ; neither of which factors is in any way within the control of the agriculturist or influenced to any appreciable extent by his demands. It is important that this distinction between basic slag and other fertilisers should be recognised. If the farmers of this country demanded double their present supphes of superphosphate, of nitrates, of sul- phate of ammonia or of potassic fertiHsers, the manufacturers could provide the additional material : if, however, basic slag were desired over and above the quantity determined by the demand for steel it could not be suppHed except perhaps by importation. The position thus created is being explored by the Permanent Com- mittee set up by the Ministry of Agriculture to advise on basic slag. On the agricultural side there is evidence that the farmers of the United Kingdom might with advantage to themselves and the com- PREFACE ix munity use no less than 890,000 tons per annum, equivalent to 33,820,000 units of tricalcicphosphate. On the other hand the 1920 output of British steel yielded about 560,000 tons of slag of 15| per cent, or higher content of phosphate, equivalent to 13,400,000 units of tricalcicphosphate. There is therefore a considerable gap between the farmers' potential demand and the visible supply. The difficult problems associated therewith are being fully and sympathetically studied by agriculturists and steel-making experts and no doubt various solutions will be devised. One obvious possibihty is to use ground mineral phosphates to stiffen out the supplies, and here Dr Robertson's experiments will prove helpful. Dr Robertson has not confined himself to the practical demonstra- tion of increased yields: he has gone further and endeavoured to ascertain why the increases have been obtained, thus giving the monograph a scientific as well as an empirical interest. He examines the change in herbage and he shows that the physical properties of the soil and the bacterial actions in the soil are much influenced by the phosphate in the slag, thus throwing important hght on the view now commonly held by experts that poor grassland should not be ploughed out till after it has been improved by slag. The monograph contains a store of information about the new slags and is a model of thorough and systematic investigation. I have personally inspected the plots on several occasions and have seen much of the experimental work. It deserves close study by all who are interested. E. J. R. January, 1922. AUTHOR'S PREFACE The main purpose of this book is to put on record the results of the field experiments with rock phosphates and open hearth basic slags conducted in Essex during the period 1915-20. The field trials have been confined to grass land, and the results have been measured by increases in the weight of the hay crop, and by the improvement in the quality of the crop, as determined by botanical analyses. This plan has been adopted for two reasons: first, because it is on grass that the primary and secondary actions of phosphates are most apparent, and most readily measured ; secondly, because on permanent grass, in Essex at any rate, the issue is not comphcated by previous appHcations of artificial manures, and it is therefore easier to follow out the experiment year by year than under arable conditions. The objection may be raised that increased weights of hay do not give a true test of the improvement which has taken place, and that such a test can only be obtained through the medium of the animal. While there is much to be said in favour of this contention, it may be safely assumed that, when botanical analysis shows the quahty of the herbage is similar, the increased weights of hay bear a definite relation- ship to the five-weight gains, and do afford a satisfactory method of comparing the efficiency of the various phosphates. Moreover, it must be remembered that hay is an important crop, and in Essex, as else- where, it is the prevalent custom to graze and mow the permanent grass in alternate years. The Essex results with ground rock phosphates indicate that there are soil conditions under which these types of phosphates may be expected to give as good and as quick results as the more soluble types of phosphatic fertilisers. It is equally clear, however, that under other conditions the advantage is decidedly in favour of the more soluble types. Experiments in progress in the North of Ireland on the turnip crop strikingly bear out this conclusion, and it would seem probable that an explanation of the different results secured elsewhere might be obtained by means of an examination of soil and chmatic conditions. I have to acknowledge my indebtedness to Dr E. J. Russell, F.R.S., for the great interest he has taken in the work here described. From xii AUTHOR'S PREFACE the discussions which took place during his annual visits to the various experimental centres many helpful suggestions came. For the method of estimating nitrates I am indebted to Mr D. J. Matthews, till recently of the Rothamsted Experimental Station, and, for the nitrate determinations, to my former colleague, Mr R. G. Baskett, For the bacterial counts in Tables XLV and XL VI I am indebted to Mr James Bryce, B.Sc; for the mechanical analyses in Table XXXIX to my former colleague Capt. H, H. Nicholson, M.A. (Cantab.) ; for the rainfall data to Mr Carle Salter, the Superintendent of the British Rainfall organization, and for the illustrations of Nauru and Ocean Islands to Mr A. F. Elhs, the Commissioner for New Zealand on the Board of the British Phosphate Commissioners, I should also Uke to record my thanks to Messrs B. Smith, N. F. Miles, T. Wood, C. L. Petheybridge and A. Freshwater, on whose farms the more important of the field trials were laid down, for the care they have taken of the plots and for the ready help they have given through the whole period of the trials. Finally, I have to express my keen appreciation of the kindness of the Agricultural Education Committee of the Essex County Council, who provided unique f acihties for the work and gave me a free hand in the carrying of it out. To the generosity of the members of this Committee is due, in no small measure, the opportunity of submitting this book to all who are interested in the progress of Agriculture. G. S. R. The Queen's University of Belfast, January, 1922. CONTENTS INTRODUCTION Basic Bessemer Slag Basic Open Hearth Slag . Rock or Mineral Phosphates REVIEW OP PREVIOUS EXPERIMENTS Review of Pot Experiments with Insoluble Phosphates Review of Field Experiments with Rock Phosphates THE ESSEX EXPERIMENTS . . . . . Character of the Soil Rainfall Details of Experiments Field Experiments on Boulder Clay Soils . Discussion of the Results on the Boulder Clay Soils Field Experiments on London Clay Soils Discussion of the Results on the London Clay Soils Field Experiments on Chalk Soils Conclusions drawn from the Field Experiments . The Applicability of the Results .... AN INVESTIGATION INTO THE REASON WHY BASIC PHOSPHATES HAVE CAUSED INCREASED YIELDS The Effect of Phosphates on the Botanical Composi tion of the herbage Discussion of the Results of the Botanical Analysis Effect of Phosphates on the Moisture Content and Temperature of the Soil The Effect of Phosphates on the Texture of the Soil PAGE 1 2 4 8 10 10 11 18 18 20 21 22 29 30 42 43 45 48 49 49 59 63 73 xiv CONTENTS PAGE The Effect of Phosphates on the Accumulation of Nitrogen in Grass-land 75 The Relation of Phosphates to the Accumulation of Nitrates in Grass-land . . , . . . 77 The Influence of Phosphates on Soil Bacteria . . 86 FACTORS LIMITING THE YIELD OF HAY AND THE ACTION OF PHOSPHATES ON HEAVY CLAY SOILS ... 89 The Effect of Rainfall on the Yield of Hay from the Untreated Plots 89 The Effect of Rainfall on the Yield of Hay from the Plots receiving Phosphates 92 The Second Limiting Manurial Factor .... 95 THE ACTION OF BASIC SLAG ON THE ACIDITY OF THE SOIL AS MEASURED BY THE 'LIME REQUIREMENT' AND HYDROGEN ION CONCENTRATES OF THE SOIL 99 REFERENCES 108 INDEX ... 110 PLATES I Pouring Slag from Basic Open Hearth Furnace Shipping Phosphate from Nauru Island . II Ocean Island Phosphate Workings .... III Experimental Plots at Martin's Hearne, June 1918 IV Experimental Plots at Martin's Hearne, June 1918 V Section of the Soil at Hassobury VI Experimental Plots at Horndon, July 1920 . VII Samples of Vegetation at Horndon, Aug. 1919 VIII Samples of Vegetation at Horndon, Aug. 1919 Photograph of Chalk Pit at Saffron Walden . PAGE 113 113 114 115 116 117 118 119 120 120 MAP Experimental Stations in Essex XVI INTEODUCTION Insoluble phosphates have been apphed to the land in the form of bones for a very long time, and until the beginning of the nineteenth century it was generally assumed that they owed their value to the oil which they contained. Lord Dundonald in his Treatise on the Connection of Agriculture with Chemistry, pubhshed in 1795, seems to have been one of the first investigators to reahse that the f ertihsing value of bones was due to the phosphoric acid which they contained. Kirkman writing in 1796 came to the same conclusion, and so did de Saussure in 1804. These opinions were accepted and repeated by Liebig, who was perhaps largely responsible for the widespread dis- semination of this important piece of information. Dundonald in his Treatise goes a good deal further than the other investigators, in as much as speaking of the phosphate of lime in bones he records: "It is a saline compound, very insoluble. There is reason to beheve a very considerable proportion of this nearly insoluble salt is contained in most fertile soils." It may therefore be said that Dundonald was the first investigator to estabhsh the value of insoluble phosphates. Towards the end of the eighteenth and the beginning of the nine- teenth century the use of insoluble phosphates increased with great rapidity throughout Europe, but nowhere more so than in this country. By about 1815 the home supply began to prove insufficient to meet the large demand, and resort was had to importation from Europe. The import of bones grew rapidly, and some idea of the importance then attached to the supply of insoluble phosphates may be gained from Liebig's passionate outburst : England is robbing all other coixtitries of their fertility. Already in her eager- ness for bones, she has turned up the battlefields of Leipsic and Waterloo and of the Crimea; already from the Catacombs of Sicily she has carried away the skeletons of many successive generations. Annually she removes from the shores of other countries to her own the manurial eqioivalent of three million and a half of men, whom she takes from us the means of supporting, and squanders down her sewers to the sea. Like a vampire she hangs on the neck of Europe, nay of the whole world, and sucks the heart blood from nations without a thought of justice towards them, without a shadow of lasting advantage to herself ! The discovery of large deposits of rock phosphates in Spain, in this country and in other parts of Europe, eased the situation. More- over these discoveries came close on the heels of Lawes's patent for 2 INTRODUCTION dissolving bones in sulphuric acid, and at a time when his experiments with dissolved bones, and later with dissolved rock phosphates, at Rothamsted focussed attention very effectively on the superior value of water soluble phosphates. The Rothamsted experiments seem to have very rapidly convinced those farmers who followed such develop- ments of the superior efficiency of water soluble phosphates, although attention was drawn at intervals to experiments which apparently showed insoluble phosphates to be as effective as water soluble phos- phates. BASIC BESSEMER SLAG The introduction by Thomas and Gilchrist in 1878 of their process for removing phosphorus from the molten pig-iron provided in the resulting slag a new source of phosphate for agricultural purposes. The presence of phosphorus in steel, except in very small amounts, renders the metal brittle and unfit to use for many manufacturing purposes. Most of the iron ores in this country are highly phos- phatic, and until the coming of the Thomas and Gilchrist process it was not possible to remove phosphorus from the pig-iron and so produce a good quahty of steel. The first step in the manufacture of steel is the conversion of iron- ore into pig-iron under the reducing conditions which exist in the hearth of the blast furnace. Such reducing conditions are essential for the recovery of iron from the ores, but they prevent the oxidation of phosphorus, which therefore passes into the pig-iron. The conversion of non-phosphatic pig-iron to steel is carried out in a Bessemer vessel with a sihceous hning — acid process. If phos- phorus is present in the pig-iron phosphoric acid is formed, which, being unstable in the presence of an excess of iron, reverts to phos- phide of iron, which is not removed in the slag. The Thomas and Gilchrist modification of the Bessemer process consists in hning the furnace with a basic material instead of a siliceous hning, and of adding suitable quantities of hme to the molten iron. The phosphoric acid formed combines with the hme producing a stable phosphate of calcium, which is removed in the slag which floats on top of the molten metal in the converter. The process was first tried on a large scale at Messrs Bolckow Vaughan and Co.'s Eston Works, in 1879, and a copy of the record which illustrates the manufacture of the first Basic Slag is reproduced in Table I by the courtesy of Mr Daniel SiUars, chief chemist to Messrs Bolckow Vaughan and Co. INTRODUCTION 3 Table I. Record of the Earliest Manufacture of Basic Slag (May, 1879) Metal Si Graphite Comb. Car. Phos. Time Min, Sec. Slags Fe samples SiOa CaO MgO P2O5 Pig-iron 2-89 336 •06 1-52 _ _ — — ^ — — 1 2-21 2-64 •80 1-51 3 — — ^ n — — 2 1-43 •06 2-55 1^51 6 — — tc ao — — 3 •78 Trace 2-50 — 9 — — 1" — — 4 Bad sample 5 •13 Ml •53 1-36 12 34-07 43-53 9^64 •60 5-00 6 •10 Nil 1-01 15 30-40 41-18 9^00 5-02 6-10 7 Trace •77 17 30 29-73 36-58 8^16 5-18 15-90 8 Ml •41 18 30 23^73 33-15 930 11-10 10-70 9 •12 19 30 20^93 35^62 8-50 10-94 13-50 10 •10 20 30 Bad sample Steel " •18 Mn. •IS 21 10 2110 32^84 9-95 10-78 13-60 In the first blow it will be noted the phosphorus fell from 1-52 % in the pig-iron to -18 % in the finished steel. When phosphorus has been removed to the required extent the converter is tipped forward and the slag allowed to flow over the top of the vessel into the slag pot where it is either allowed to cool or tipped molten on to the slag heap. The production of steel by this process and the consequent accumu- lation of phosphatic basic slag increased with great rapidity, and attention was turned towards the possibility of using these basic slags for fertiHsing purposes. It was at first considered that on account of the insolubihty of the phosphates in water the material would be of httle value for direct appHcation, Attempts to obtain a suitable fertihser by dissolving the slag in acid proved unsuccessful. To Wrightson and Munro we owe the discovery in 1885 that if basic slag is ground to a fine powder it has a very considerable fertihsing value. Their experiments were followed by many others including the now classic experiments at Cockle Park, which were commenced in 1896 by Professor SomerviUe and subsequently con- tinued and developed by Sir T. H. Middleton and Prof. D. A. Gilchrist. It is from the Cockle Park experiments that most of our information concerning the practical use of basic slag has been derived. These experiments continued over a period of 25 years do more than show that basic slag has a high fertihsing value. They demonstrate that under the conditions at Cockle Park basic slag per unit of phosphoric I — 2 4 INTRODUCTION acid is more effective than superphosphate, a result which was sub- sequently confirmed by the trials at Sevington, Cransley, Hatly and Yeldham(28). At Sevington where the soil is well supphed with calcium carbonate (32) the returns for the two types of phosphates are for practical purposes identical, there being only a difference of 3 lbs. hve weight gain in favour of slag over a period of nine years. The superior results from basic slag at the remaining centres is probably due to the fact that on 'sour' soils, and on soils where the calcium carbonate content is not high, as at Cockle Park (0-59 % CaC03),a certain proportion of the phosphoric acid in superphosphate is retained by the soil in the form of somewhat insoluble phosphates of iron and aluminium. With repeated dressings of superphosphate increasingly large proportions of the phosphoric acid revert to such insoluble forms. These experiments may therefore be said to have estabhshed the fact that insoluble basic phosphates have a distinct function in agriculture, and that under certain soil conditions they are to be preferred to the water soluble phosphates in superphosphate. As a consequence of the Cockle Park experiments basic slag is used for the manuring of grass-land almost to the exclusion of other types of phosphatic fertihsers. Nor has its use been confined to grass-land, where perhaps rapidity of action is not of primary im- portance. In the south of Essex, basic slag is used on the arable land almost to the exclusion of superphosphate, and many of the most progessive farmers have attributed their success to the use of basic slag instead of superphosphate on their heavy clay soils, which are either devoid of calcium carbonate or have only a very poor supply. Some idea of the extent to which basic slag has been appreciated, and the lessons which Cockle Park taught assimilated, may be ob- tained from the following figures (Table II) showing the production and consumption of basic slag during the period 1903-1920. BASIC OPEN HEARTH SLAG Unfortunately for agriculture important changes in the manu- facture of steel have been taking place during the past few years. Economic conditions and to a certain extent the working out of the higher grade ores have made the basic Bessemer process uneconomical, and it has been replaced by the basic open hearth process. In this process iron-ore and Hme are charged on to a basic hearth heated by producer gas, and the molten metal poured over the heated lime INTRODUCTION 5. Table II. Prodfctiozst and CoNSFMPTioisr of Basic Slag in the United Kingdom. In Metric Tons (thousands) Production Imports Exports Net 1903 148 9 9 ? 1907 145 ? ? ? 1910 160 ? ? ? 1913 404 51 169 286 1914 404 17 134 287 1915 400 — 117 283 1916 360 — 39 321 1917 447 — 2 445 *1917- 18 500 — — 500 *1918- 19 565 ? 2 — *1919- 20 497 ? 15 — * Seasonal year June 1st to May 31st. The figxu-es i for 1913 to 1920 the Ministry of Agriculture returns. Table III. Basic Bessemer Process No. Time from beginning Metal Sl AG Si P SiOa FeO MnO MgO CaO P2O5 1 Pig-iron min. sec. 1-22 2-183 — — — — — — 2 2 46 0-72 2-148 41-15 2-40 903 4-13 41-27 0-84 3 5 21 0-15 2-224 36-30 3-97 1102 3-39 39-50 3-12 4 8 5 0-007 2157 34-41 3-60 10-72 3-35 42-80 2-99 5 10 45 0-012 2-096 31-94 4-23 9-94 4-01 43-12 4-02 6 13 28 0-005 2053 16-64 8-42 8-51 7-34 44-37 7-15 7 15 13 0-008 1-910 14-65 715 7-39 6-34 46-63 11-60 8 19 14 0-005 0-230 12-94 5-84 4-25 6-00 47-76 18-83 9 19 31 0-005 0-139 12-20 6-79 401 6-26 48-59 18-66 10 19 49 0-004 0-087 11-71 7-19 4-05 6-38 48-19 18-15 11 Rail steel 001 0145 12-77 5-94 4-80 6-75 47-87 16-92 Table IV. Z 603. Ordinary Basic Open Hearth Process Metals Slags Time p.m. No. Car. Phos. Silica Lime Total iron P2O5 Sol. P2O5 Cit. Sol. % 1 2.25 1-77 -300 20-30 33-2 8^60 17-08 15-36 89-92 2 2.40 1-68 •327 19-90 34-8 730 16-87 14-85 87-89 3 2.55 1-60 •35 18-80 35-70 8-40 17-30 15-49 89-53 4 3.10 1-57 •335 20-30 34-90 6^20 17-08 14-08 82-43 5 3.40 1-48 •321 20-20 3700 5^60 15-85 11-90 75-70 6 4.45 1-10 •19 20-50 37-70 5-50 15-66 11-78 75-22 7 6.0 •74 •083 15-50 42-50 7^20 15-47 5-38 34-77 8 7.0 •63 •07 15-10 40^50 7^00 15-75 4-99 31^68 9 8.0 •14 •026 12-60 41^80 11-50 13-65 1-79 1310 10 9.0 -09 •023 10-20 47^80 14-70 10-85 1^66 15^30 6 INTRODUCTION and ore. The oxygen necessary for the purification of the pig-iron is suppHed to the extent of about 70 % by the action of the metalloids on the oxide of iron, the balance 30 % coming from the oxidising gases of the furnace. In the Bessemer process the oxygen comes entirely from the air blast, and the combustion of the phosphorus, sihcon, and carbon generates sufficient heat to raise the temperature of the steel to the required extent. The slag formed in the basic open hearth process is much greater in volume and there is a corresponding decrease in phosphoric acid content compared with the basic Bessemer process. Tables III and IV show the changes in the composition of the slag by the two processes. Commenting on Table IV Sillars says : The decrease in P2O5 content becomes qtiite sharp after the fourth sample, and this, it will be observed, coincides with the commencement of the period at which carbon elimination becomes predominant. If high grade slag is desired, it is removed at this stage, and after charging fresh lime and oxide of iron, the carbon elimination is proceeded with. It will be noticed that the phosphorus in the first metal sample is as low as in any of the four immediately- following, and it may be asked why the slag coiold not equally well be removed at this stage instead of an hour later. The reason is that although the phos- phorus is eliminated very rapidly (sometimes it is reduced to 3 % twenty minutes after charging), yet it is necessary to delay the removal of the slag until all frothing has ceased and \xntil the whole of the hme and ore is dis- solved in the bath and the heat is sufficiently high to allow the slag formed to flow freely through the tap hole. Unless the slag is removed when it has reached the maximum concentration of phosphoric acid, the further additions of hme and ore, and the denudation of the fiirnace structure under heat, cause an increase in the slag volume which reduces the phosphoric acid content ixntil at the termination of the process it will contain from 7 to 10 % only. As the content of Hme increases, the slag thickens and reaches a viscosity which slows the progress of the 'boil.' This may be corrected by the addition of oxide of iron in the form of scale, but if siolphur has to be eliminated from the metal it is essential to keep the slag as basic as possible ; the slag is therefore thinned by the addition of fluorspar, and it is this addition more than any other con- dition which reduces the solubility of the phosphoric acid in 2 % citric acid. In Table IV 1 cwt. of fluorspar was added after the sixth sample was drawn, and the soluble phosphoric acid fell from 11-78% to 5-38% immediately afterwards. High grade slag can be obtained by pouring the slag immediately before the addition of fluorspar. In the basic open hearth process the steel and slag are tipped into a ladle — the steel ladle — which is only large enough to hold the steel. When the steel ladle is f uU the slag overflows into the slag ladle placed immediately under the spout of the steel ladle (Plate I). The significance of the change may be better appreciated by a INTRODUCTION 7 consideration of the following figures showing the output in 1920 of the various grades of basic slag. Table V. , Gbabe 1. Over 33 % CajPaOg 2. 26-33% 3. 22-26% 4. 15-22% 5. 11-15% 6. Under 11% Total all grades Production in 1920 (in tons) 46,300 121,400 90,900 302,500 118,000 22,000 701,100 The production of high grade basic slag, even if slag containing only 33 % of phosphate is so classed, had fallen by 1920 to less than one-tenth of the amount necessary to satisfy the demands of the farmer, and it is probable that a comparatively short time will see the last of this type of basic slag. Of the basic slags forming grades 3, 4, 5 and 6, a large proportion, how large it would be difficult to say, are of low citric solubility due to the use of fluorspar. It has been shown that the action of fluorspar results in the replacement of the calcium silicate in the phosphate compound of high soluble slags by calcium fluoride (19) and Bainbridge has demonstrated that the resulting slag phosphate consists largely of apatite (2). There are thus three types of slag available for agricultural purposes : 1. High grade containing 16-20 % phosphoric acid. Part of this supply consists of the rapidly diminishing remnants of the basic Bessemer slags and the other part of the slags obtained from the basic open hearth process by fractionating before the addition of fluorspar. 2. Open hearth basic slag containing 7 — 14 % phosphoric acid. 3. Open hearth fluorspar basic slag containing 6 — 12 % of phos- phoric acid. Numbers 1 and 2 have a citric solubihty of 80-95 % whilst no. 3 has a citric solubility of from 6-50 %. Open hearth fluorspar basic slag is a new material containing totally different phosphate compounds to those in nos. 1 and 2. It is not the type of basic slag which produced the remarkable results at Cockle Park and elsewhere. Its value compared with such slags is unknown, and its low solubility suggests that it will prove less effective as a fertiliser than the more soluble tjrpes. 8 INTRODUCTION The total production of basic slag, including in that term slags con- taining from 1 1 % tricalcium phosphate upwards, amounted to 680,000 tons in 1920, but, if all slags below 22 % are excluded, to only 258,600. Of the totals a very large proportion was fluorspar slag. In 1919 there was a consumption of at least 560,000 tons and the demand is steadily increasing^. Whilst therefore the steel industry may continue to be a valuable source of insoluble phosphates for agricultural pur- poses, it is becoming increasingly evident that the supply can no longer keep pace with the demand and the agriculturist must turn to other sources of supply. ROCK OR MINERAL PHOSPHATES The increasing demand for basic phosphates can most readily be met by increasing the output of the apparently inexhaustible stores of rock or mineral phosphates and utiHsing these materials for direct apphcation. Unfortunately there are no extensive deposits in Great Britain^ and there are not many sources of supply within the British Empire. (Collins in Chemical Fertilisers gives a map showing the distribution of the chief deposits of rock phosphates.) Broadly speaking the deposits may be divided into two types — the softer and woolher North African phosphate such as Gafsa, Egj^tian and Algerian phosphates and the harder North American and Island phosphates such as Florida pebble, Carolina, Nauru Island, and Ocean Island phosphates. The deposits in the majority of cases are close to the surface and can be worked, and, in the case of the Island phosphates, transported to the shore and shipped at a comparatively low cost. Plates I and II. The North African phosphates are more soluble by the Wagner test than the harder American phosphates (20). They apparently con- tain more calcium carbonate and less calcium fluoride combined in the phosphate compound than is the case with the American phos- phates (21). It may therefore be just as important to distinguish between these two types of rock phosphates as it is to distinguish between open hearth fluorspar basic slag and the open hearth basic slag produced without the use of fluorspar. Rock phosphates have the great advantage of a high phosphatic content, ranging in the case of the North African, the Island phos- ^ Middleton estimates our requirements of basic slag at 891,000 tons per annum. ^ The deposits of coprolites in Cambridge and Suffolk can no longer be worked economically. INTRODUCTION 9 phates and American phosphates from 50-88 % of tricalcium phos- phate. Moreover, it has been shown (14) that these phosphates are even more soluble in citric acid than the majority of open hearth fluorspar basic slags, and that they contain phosphate compounds which are in many respects similar to those in open hearth basic slags (20). It is thus a matter of great urgency to ascertain their precise manurial value, as it is no exaggeration to say that the future of agriculture and our national prosperity will be largely determined by the extent to which suitable phosphates can be supphed at a comparatively low cost. The problem is a big one, capable of attack from more than one point of view. Useful results are Hkely to be secured by investigating the effect of chmatic conditions, particularly rainfall, on the avail- abihty of the rock phosphates. The question of soil conditions is also of great importance in this connection. Rock phosphates for example may prove a failure compared with superphosphate on a chalky soil under dry conditions, whilst on a sour soil and under a more humid cUmate the reverse may well be the case. REVIEW OF PREVIOUS EXPERIMENTS REVIEW OF POT EXPERIMENTS WITH INSOLUBLE PHOSPHATES In order to ascertain with any degree of certainty the agricultural value of a suggested f ertihser two types of experiments are necessary — pot experiments and field trials. Russell (23) discussing the relative advantages of field and pot trials points out that as a general rule pot experiments are more accurate than field trials. The experimental conditions are more under control, and it is therefore possible to bring out small differences between materials which it might not be possible to secure under the conditions of a field trial. On the other hand, the conditions under which pot experiments are conducted are so artificial that a positive result is not always paralleled by a positive result in the field. Furthermore, though very considerable difference in the cropping power of the two materials may be shown by pot experiments, it by no means foUows that the differences will be equally marked under field conditions. Whilst therefore pot experi- ments are of great value as a prehminary method of investigation, field experiments are always essential before any deductions can be made relative to the economic importance of the factor under investi- gation. If they are to be of real value such field experiments must be carried out under varying cHmatic and soil conditions and on different types of soil, and an attempt be made to interpret the results in the light of such conditions. Dutton ( 6) during 1912 conducted a series of pot experiments designed to ascertain the fertihsing effect of that portion of the phosphoric acid in basic slag which is not soluble in citric acid, and came to the conclusion that such insoluble phosphate is active enough to feed a^ short-hved plant like mustard. Bauibridge(2), in a paper on "The Effect of Fluorspar Additions on the Phosphates in Basic Slag," describes a series of pot trials with barley, and shows that a very insoluble fluorspar slag possessing a citric solubility of only 6 %, when contrasted with a slag of 81 % solubihty, gives a yield of 61 % compared with the high soluble slag yield of 100. These two experiments, although not conclusive, clearly indicate that even short-hved crops such as mustard and barley are capable of making considerable use of phosphates which are much REVIEW OF PREVIOUS EXPERIMENTS 11 more insoluble than those phosphates which are readily dissolved in dilute solutions of citric acid. In view of these results it is reasonable to expect that promising returns would be secured from similar trials with less resistant materials hke rock phosphates. Burhson(3) made an elaborate series of pot experiments with six types of rock phosphate, the trials extending over a period of three and a half years and embracing the results from 700 pot cultures. It is difficult to interpret the exact meaning of these experiments in terms of basic slag or superphosphate as neither of these forms of phosphatic fertilisers was included in the trials. The results from this elaborate series are nevertheless of considerable interest as they show that the phosphates in rock phosphates, even of the hard re- sistant type like Canadian Apatite, can be assimilated by farm crops in sand cultures under greenhouse conditions and in the absence of decaying organic matter. Three other conclusions from his work are worth noting. Burhson found that the plants could obtain their calcium as well as their phosphorus from rock phosphates, and that the addition of calcium carbonate to the rock phosphates did not produce better results. An attempt was made to ascertain the effect of fineness of grinding on the availability of such phosphates, and the work shows that better results were secured by grinding beyond the '100' grade. Finally the author gives it as his opinion that there is no particular relation between the citric acid, soluble phos- phoric acid and the availability of rock phosphates to the plant. These pot experiments, scanty and incomplete though they may be, agree in demonstrating that, under the conditions of the experi- ments, the insoluble phosphates in fluorspar basic slag and in rock phosphates may have a very considerable agricultural value. REVIEW OF FIELD EXPERIMENTS WITH ROCK PHOSPHATES American Experiments. Although by no means exhaustive, a large number of field experiments have been carried out with rock phosphates. The subject has perhaps received more attention in the United States than elsewhere. There considerable differences of opinion exist concerning the fertilising value of raw ground rock phosphates or 'floats.' In the States the controversy centres round the relative value of ground rock phosphates (floats) and acid phos- phate (superphosphate). Most of the American experiments, a detailed account of which is given by Hopkins (ii), are confined to this aspect 12 REVIEW OF PREVIOUS EXPERIMENTS of the question. Moreover many of the American State experiments, e.g. Maine, Massachusetts, and Rhode Island, compare the two phos- phates by applying equal money values, and it is obvious that such trials have only a hmited value as far as the apphcation of the results to this country is concerned. Moreover, changing economic conditions must seriously detract from the value of their apphcation to present day American practice. The Ohio, IDinois, and certain of the Massa- chusetts experiments compare equivalent quantities of the two forms of phosphate, alone and in combination with other manures. These experiments have extended through several rotations on duphcate, and in some cases triphcate, plots. After an exhaustive review of the American experiments up to 1908 Hopkins draws the conclusion that rock phosphates are much the more economical type of phos- phate to use, and that from the point of view of the permanent f ertihty of the soil they are much to be preferred to acid phosphates. A later review of the American experiments is given by Waggaman and Wagner (31), covering the period up to 1917. These writers give a table incorporatiag the results of 232 field experiments. Only 37 of these experiments extended over a period of five years or more. Their tabulation of these experiments is given in Table VI. In explanation of this table they give the following notes: Out of the 37 tests given in Table VI, 22 were carried on with a view to comparing the relative merits of raw rock and acid phosphates. The conditions xm.der which such a comparison was attempted varied greatly, but it may be said that in a general way, 13 of these experiments, or 59-1 %, gave crop yields as favourable to raw rock as to the more soluble form of phosphoric acid. Of the 9 experiments in which raw rock did not compare favourably with acid phosphate, 2 were conducted on fields Tinresponsive to phosphate treatments and 2 gave results which could be classed as either favourable or ixnfavourable, depending on the method of interpretation employed. Of the 15 experiments in which no comparison between raw ground rock and acid phosphate was attempted, 11, or 73-3 %, gave resvilts strongly indicating beneficial effects from the application of the former material, and 2 of the remaining 4 experiments were conducted on fields showing little or no response to phosphate treatment. In 21 experiments the applications of raw rock were relatively light (250 lbs. or less per acre), yet 15 of these experiments, or 71-4%, showed distinctly favourable increases in yields on the fields treated with this material. In 16 experiments where the raw rock applications were more liberal, 13, or 81-3 %, resulted favourably to raw rock phosphate, and the remaining 3 experiments were conducted on soils showing little or no response to phos- phate treatment. Raw rock phosphate was applied in connection with organic matter in 23 experiments. Out of this number, 18, or 78-3 %, gave distinctly favourable results, and of the 5 remaining experiments 3 were conducted on fields un- responsive to other forms of phosphoric acid. > II < igqranK UMOUTian 1 1 1 1 |i-l 1— 1 1 o EH p C 11% g|S 1^ 8iqB -jtiOA-Bjan jaqratifj |||||C^-H||||||i 1 1 1 1 lOS^r— ll |l-Hl— (IC^Jl 1 I'^l 1 1 1 1 1 1 1 00 r— 1 8 ^ F^ox 1 1 1 1 |,_H(^)^| |_i,-i|tH| 1 IffOl 1 1 1 1 1 1 1 CO (7^ 1 > 03 ^ o ■S o^ tg cs 53 Sp,- 9 P. aiqu -inoABjun I9qnin^ * * CO TS WITH R. ROM 5 TO aiqB -moABj jaqinn^ 1 1 1 1 lOl-Hl— IrHlCN'— ll 1 1 1 |'~*| 1 1 1 1 1 1 1 CO % 1 o I'B^ox 1 1 1 1 |CO(N^i— l|!N'-H| 1 1 1 |'~'| 1 1 1 1 1 1 1 CO |2 P.O 9xq« -inoABjun laqran^ -1— ||||||||(M||||C^|||'-'|| * 1 1— 1 1 1 1 CO O 1 Expe RIODS aiq'E a9qran^ 1 1 1 1 \ ^ \ 1 |(N['-i-l lllll K5 i 2 "a ABY OF Results of Field State Stations over Pi l^iox 1 1 1 1 ItHI l(N TOOX 1 1 1 1 1 (TO ©q rt CO (N ir^ IM (N 1 1 1 1 > 99 99 """ 38-6 18-4 -s 22-2 9 Tunisian „ (fine) 26-0 17-9 '^ 23-2 10 „ „ (coarse) ... 23-9 19-2 i 23-8 11 Egyptian „ (fine) 37-0 23-6 ^ 23-6 12 „ „ (coarse) ... 34-7 22-5 -4^ 25-1 13 Superphosphate (200 lbs. P2O5 per acre) — 27-0 P 23-0 14 Superphosphate (50 lbs. P2O5 per TS acre) — 25-9 S 12-3 15 Superphosphate (200 lbs. P2O5 per acre) — 1 ton of groimd lime 00 per acre ... ... — 23-4 +2 27-2 16 No manure — 15-5 PM 6-4 17 Open hearth basic slag: high sol. 91-2 22-5 28-8 18 „ „ (fluorspar) 201 18-8 16-8 19 1 cwt. ferrous sulphate per acre . . . — 13-6 6-4 E Lime at rate of 1 ton per acre ... — — 5-4 F Cambridge coproKtes 25-0 — 151 G Rough slag ... — — 10-4 H Cleveland phosphate 19-5 — 190 K No manure ... — — 50 L Florida soft phosphate 27-7 — 13-0 Average gain, Plots 1 to 5 and 7 to 13 and 15, 17 and 18, over plots 6 and 16 — — — 250% Rainfall, May 1st tiU harvest (in inches) — 2-25 1-78 t 5-34 Date of cutting — July 8 — Aug. 16 * Phosphates not apphed till Feb. 27th. t Rainfall, May 1st to June 30th. 32 THE ESSEX EXPERIMENTS The summer is equally trying on this type of soil. The dry and hot weather which is usually experienced in Essex in June and the latter part of May 'caps' or bakes the soil — the soil sets hard and cracks and the crops receive a check. It is but seldom that the crop of hay exceeds 10 cwts. to the acre, and it is only too frequently left uncut altogether. The meadows which have recently been laid down contain a smaU reserve of calcium carbonate, a residuum from the heavy dressing of lump chalk (40-60 tons per acre) fairly frequently appKed up to the eighties or nineties. The soil is exceedingly poor in both 'total' and 'available' phos- phoric acid, but is well supphed with potash. Nineteen quarter-acre plots (1-19) were laid down on this field in 1918, and the manures sown on February 27th, 1918. Subsequently Plots A, B, C, D, E, F, G, H and K were added and sown on February 3rd, 1919, and finally Plot L was sown during May, 1919. The weights of hay on the various plots for the seasons 1918 and 1920 are given in Table XIV. In this experiment an attempt was made to ascertain whether better effects could be obtained from rock phosphates by finer grinding. With this object in view the Florida pebble, Algerian, Gafsa, Tunisian and Egyptian phosphates mentioned in the above tables were specially ground under the writer's supervision by Messrs Walter Packard, of Ipswich. All the phosphates were passed through a Griffin mill. For coarse grinding the miU was set to grind for the standard usually adopted when the rock phosphates are used for the manufacture of super- phosphates (90 % to pass a '60' sieve). In actual fact about 80 % of the material wiU pass the ' 100 ' sieve. For fine grinding the mill was closed down so that the output per hour was reduced by a half. A much finer product was obtained, but it has not been practical, owing to the 'wooUy' nature of the rock phosphates, to satisfactorily distinguish by means of sieves between the 'fine' and the 'coarse' grinding. During 1918 no superiority due to fine grinding was noticed. Throughout the whole season of 1920 the writer was able to visit this centre at least every week, and a close watch was kept on the progress of the various plots. The high soluble slag and the "super- phosphate and hme" plots were the first to make a start, followed by those plots receiving the finer ground rock phosphates. During the whole of May the superiority of the plots receiving the finer ground rock phosphates over those receiving the same phosphate only more coarsely ground could be distinctly seen. As the season progressed THE ESSEX EXPERIMENTS 33 the distinction became less and less visible, until at the beginning of July it was quite impossible to see any difference. The high soluble basic slag, Plots 5 and 17, and Plot 15 (super and hme) were distinctly ahead during the whole season, but the rock phosphate plots gradually lessened the difference as the season progressed, although they never actually succeeded in catching up. The weights of hay for the season of 1920, which was particularly favourable to the hay crop, give some indication of the contrast which existed between the various phosphate plots and the untreated portions. It may be of interest to mention that only the plots were cut, and no attempt was made to harvest the rest of the field, as the crop was not considered to be worth the labour involved in doing so. When the wild white clover came into flower the contrast was remarkable. Plate VI, showing a general view down Plot K (un- treated) and Plot H (Cleveland phosphate), gives some idea of the contrast which met the eye. So thick was the crop of wild white clover that the farmer decided to seed the plots. Plots 1-19 are strictly comparable, having been sown at the same time, and a useful comparison of the effectiveness of the various phosphates may be made from the respective yields of hay. There can be httle doubt that the highest soluble types of open hearth basic slag and basic superphosphate have proved the most effective phosphates at Horndon. At the same time, however, some of the rock phosphates are nearly as effective. From June onwards, for example, it was always difficult to say which of the two, Plots 3 or 5, was the better, although there was no doubt that Plot 3 was inferior to Plot 17, which is a duphcate of Plot 5. The hard American Florida pebble phosphate is inferior to the softer North African phosphates. The inferiority is not only apparent in the weights of hay, but is plainly to be seen on walking over the plots, a result which agrees with Tacke's conclusion^. No gain from fine grinding is apparent in the weights of hay, but an earlier start was undoubtedly made by the plots receiving the finer ground phosphate, and where a meadow is reserved for grazing it is possible that the extra cost of grinding would be weU repaid. The open hearth fluorspar slag, after giving promising results during the first two years, proved a poor plot in 1920 when compared with the high soluble slag, Plot 17. AU the rock phosphate plots, 1 Inter. Inst, of Agr. Bulletin, September, 1913. R.B.S. 3 34 THE ESSEX EXPERIMENTS with the exception of the two receiving Florida pebble, were much superior to the open hearth fluorspar basic slag. Plots C and E unmistakeably show that Ume without phosphate has httle or no effect in improving this type of pasture. It is difficult to interpret the results from Plots B, D, F, G, H and L in terms of the other plots. They were not sown until 1919, and the exceedingly dry season prevented a rapid response. As all the plots were grazed throughout this season, these particular plots would not receive the same benefit from the grazing as those sown the year previously, which, at the beginning of the grazing period, were already covered with a thick and close bottom of wild white clover. During the latter part of May, 1920, Plots B, D, F and H, at first backward, made rapid progress, and at harvest time there seemed to be more heads of clover on some of these plots than on the majority of the others. Plate VI illustrates the appearance of Plot H in July, 1920. It has been quite obvious during the past two years that the hght dressing of superphosphate on Plot 14 has not been effective. The improvement was much less than the weight of hay would appear to indicate, and during the seasons 1919 and 1920 Plot 14 looked very hke an untreated plot. The heavy dressing of superphosphate on Plot 13 was much more effective. It was not, however, nearly so good as the high soluble slag plots or the " superphosphate and Hme " plot. Even on a soil of this character, very deficient in phosphoric acid and with a smaU reserve of calcium carbonate, an acid manure like superphosphate is not suitable. On Plot 15 the same dressing of superphosphate as on Plot 13, namely 200 lbs. P2O5 per acre, plus one ton of lime per acre, were sown together. Under such circum- stances the reversion of the water soluble phosphate in the super- phosphate would be practically instantaneous (22) and the dressing would become a basic one comparable to the apphcation of a dressing of basic superphosphate. It is of interest to note that Plot 15 gives results practically identical with those secured on the plots receiving the most soluble type of basic slag. A close observation was kept on Plots 15 and 17 throughout the 1920 season, and the only notice- able difference was the somewhat earlier start made by Plot 15. The difference in this respect was not great, probably not more than 7 to 10 days, and had visits to the plots been less frequent, might have been entirely overlooked. During the season of 1919 the long drought lasting from the THE ESSEX EXPERIMENTS 35 Table XV. Percentage of Ground Space occupied by the Vegetation on the Plots at Horndon Analysis made : August, 1919 Plot Manube Clover Grass Weeds Bare J acre (Dressing 200 lbs. PgOg per acre) % % % space % C Lime alone 151 34-6 30-0 20-3 1 Florida pebble phosphate ... 46-0 30-6 13-3 101 3 Algerian phosphate 47-4 301 7-4 15-1 5 Open hearth high sOl. basic slag 44-1 28-6 13-7 13-6 6 Untreated 4-2 14-8 310 500 8 Gafsa phosphate 41-3 32-3 17-6 8-8 9 Tunisian phosphate ... 38-5 36-9 210 3-6 12 Egyptian „ 55-5 410 0-7 2-8 13 Superphosphate (200 lbs. PgOg per acre) 23-9 57-3 0-7 18-1 14 „ (50 lbs. PgOg per acre) 18-8 25-3 18-8 37-1 15 Superphosphate (as for Plot 13) plus 1 ton of hme per acre 60-0 32-7 1-4 5-9 16 Untreated 9-4 191 26-0 45-5 17 Open hearth high sol. basic slag (same as for Plot 5) 46-2 47-2 1-4 5-2 18 Open hearth (fluorspar) basic slag (low soluble) 43-8 31-8 13-3 IM H Cleveland phosphate 431 33-3 5-6 18-0 Fig. 4. Percentage of Ground Space occupied by the Vegetation at Horndon, August, 1919. Soil London clay. 1, Florida pebble phosphate. 5, Basic slag. 6, No manure. 8, Gafsa phosphate. 12, Egyptian phosphate. 13, Superphosphate heavy dressing. 14, Superphosphate light dressing. 15, Superphosphate and hme. 16, No manure. 17, Basic slag. 18, Open hearth fluorspar basic slag. C, Lime. 3—2 36 THE ESSEX EXPERIMENTS beginning of May until the third week in June made a hay crop out of the question, and the plots were therefore grazed by cattle and sheep during the remainder of the season. The contrasts between several of the plots were, however, so great, that on the suggestion of Dr Russell a detailed examination of the ground space covered by the various types of vegetation was made on several of the plots, using the method recommended by Armstrong (i). The results are set out in Table XV, and illustrated in Fig. 4. Photographic representations of several of the plots are given in Plates VII and VIII. The poverty of the untreated plots is difficult to describe, but some idea of their unproductiveness is afforded by Table XV, by Fig. 4 and the lower figure on Plate VI. Amongst the weeds on the un- treated plots Hyjpochaeris radicata, Leontodon Mspidus, Ranunculus, Prunella vulgaris, Potentilla reptans, Bellis perennis and Plantago lanceolata are prominent, and amongst the grasses Holcus lanatus, Hordeum pratense, Agrostis vulgaris, Cynosurus cristatus, Lolium perenne are also prominent, whilst traces of Dactylis glomerata, Phleum pratense and Alopecurus pratensis can be found. The transformation which has been brought about by the various phosphates is remarkable. Weeds have been largely crowded out and the bare space reduced in some cases to vanishing point. On the untreated plots the crop was left practically untouched, whilst on the plots receiving phosphates the growth had been grazed to the ground, and even the clover runners were being eaten by the sheep. The contrast remained equally striking right through the whole winter. The untreated plot was clearly defined by its dark unhealthy appearance and the black heads of the uncropped crested dog's tail. On the plots receiving phosphates the mat of wild white clover runners remained green throughout the whole winter, and continued to afford feed for the stock wintered on the meadow. The botanical examination of the flora reveals differences between the various phosphates which do not appear so prominently in the yields of hay. Whilst the various basic phosphates show but small differences, the three plots receiving superphosphate show significant contrasts. A very decided improvement has followed the heavy dressing of superphosphate, but the small dressing of 50 lbs. of P2O5 in the form of superphosphate has had httle effect. The addition of Hme at the rate of 1 ton per acre, sown immediately the superphosphate had passed through the drill, produces an effect which affords a significant contrast with the same dressing of superphosphate THE ESSEX EXPERIMENTS 37 applied alone. Superphosphate alone has had most effect on the grasses, whilst superphosphate and Hme together — basic superphos- phate — has told mostly on the clovers. Armstrong's method of interpreting the results fails to bring out any marked distinction between the types of basic phosphates. It merely demonstrates that the improvement in quality is approxi- mately the same. When the plots are left for hay the differences between the phosphates are reflected in the hay yields, but when the plots are grazed or when the growth is short inspection of the plots gives the impression that there is little to choose between them. This difficulty in interpreting results under such conditions is also noted by 01dershaw(i6). Butterfields, Latchingdon. ThesoilatLatchingdonisnotsoheavy as that at Horndon, containing only 16-5 % of clay, against 30 % at the latter centre. The soil would be better described as a clay loam, resting on a stiff London clay subsoil. In other respects it is very similar to Horndon. It is very poor in both total and available phosphoric acid, but contains a small reserve of calcium carbonate. Eight years before the commencement of the experiments in 1915, the experimental field had received a smaU dressing of about 4-5 cwts. of basic slag per acre — a dressing which probably accounts for the comparatively high proportion of citric soluble phosphoric acid to total at this centre (Table VII). The meadow, however, was in an exceedingly poor condition when the experiments began, and it is evident from the response to the various phosphates that the effect of the small dressing applied 13 years ago was practically exhausted. The plots have been cut every year, and the hay crop weighed. The figures are set out in Table XVI and the results are shown diagrammatically in Fig. 5. The figures in Table XVE give some idea of the remarkable response to the various phosphates. The improvement in the quality of the herbage was equally marked. During the seasons 1916, 1917 and 1918 the treated plots contained a dense and vigorous growth of clover. The open hearth fluorspar slag was quite as effective in this respect, during the initial stages of the experiment, as any of the high soluble slags, and was in fact superior to either of the two open hearth high soluble basic slags on Plots 5 and 6. It has already been pointed out that it was not until the second year that the yellow suckhng clover and bird's foot trefoil, which formed the natural leguminous flora of the untreated plot, were replaced on Plot 2 (Gafsa S8 THE ESSEX EXPERIMENTS Table XVI. Weight of Hay at Bftterfields, Latchingdon Manures sown: December, 1915 Plot Manure Citric solubihty Hay (in cwts. per acre) Average J acre 200 lbs. P2O5 per acre of phos- phate (%) 1916 1917 1918 1919 1920 5 years 1 Basic Bessemer slag 92-0 44-4 24-3 22-2 35-9 20-0 29-4 2 Gafsa rock phosphate . . . 38-3 44-2 19-1 27-0 28-3 17-8* 27-3 3 No manure — 31-4 14-5 20-1 20-6 161 20-5 4 Open hearth (fluorspar) basic slag 45-0 44-7 23-5 26-2 28-7 21-4 28-9 5 Open hearth basic slag — high citric soluble (1) ... 93-4 37-6 21-9 32-4 34-5 22-6 28-9 6 Open hearth basic slag — high citric soluble (2) ... 82-2 40-9 22-7 36-3 35-7 25-7 32-3 Percent, increase of Plots 1, 2, 4, 5 and 6 over the — 34-4 53-8 43-4 62-2 39-2 — unmanured plot (Plot 3) Average rainfall May 1st to June 30th (in inches) — 3-41 2-32 2-51 1-47 2-28 — Date of cutting — July July July July July — 25 19 29 21 19 * During the late winter and early spring of 1919-20 cattle were driven without the knowledge of the farmer across a portion of this meadow on their way to a more distant pasture. Their track lay right along the length of Plot 2, which they poached badly. As a result, although this plot had the best bottom, there was not such a vigorous growth as on the other treated plots. " 40 _ 30 r— 1 1 1 1— 1 1— j a S-i ft - 20 [— 1 CO - 1.0 3 2 4 5 1 6 Fig. 5. Yield of Hay (average of 5 years) for the various Phosphate Plots at Butterfields, Latchingdon. Soil London clay. 3, Untreated. 2, Gafsa rock phosphate. 4, Open hearth (fluorspar) basic slag. 6, Open hearth (high soluble) basic slag 1. 1, Basic Bessemer slag. 6, Open hearth (high soluble) basic slag 2. THE ESSEX EXPERIMENTS 39 phosphate) by a bottom of wild white clover comparable to that on the other plots. During 1919 and 1920, and to a certain extent during 1918, it was noticeable that the open hearth fluorspar slag on Plot 4 was not doing quite so well as some of the other slags, and the weights of hay for these years confirm this opinion. The inferiority of both the Gafsa phosphate and the open hearth fluorspar slag plots during the dry season of 1919 was also obvious to the eye. The behaviour of the clover was the outstanding difference between this centre and Horndon. Whereas at Horndon the clover runners ramified over the whole of the plots receiving basic phosphates and persisted throughout the whole winter, at Latchingdon the clover seldom made its appearance before the end of May or the beginning of June, and seemed to vanish completely from the plots by the end Table XVII. Percentage of the Ground Space occupied by THE Vegetation on the Plots at Butterfields, Latchingdon Type of vegetation Plotl Basic Bessemer slag Plots No manure Plot 4 Open hearth (fluorspar) basic slag Clovers Grasses Weeds Bare space 18-1% 56-2 0-0 25-7 . 7-8% 41-5 1-6 49-1 22-2% 42-4 00 35-4 of October. During the dry season of 1919 the clover bottom on the treated plots, although vastly superior to the unmanured plot, was much inferior to what it had been during previous years, or during the following year 1920. Not till the rain came at the end of June in 1919 did the clover make any real show, and had the plots been cut early in July as was intended there would have been Uttle, if any, clover in the hay. The examination of the flora covering the ground space of Plots 1, 3 and 4 was made during the third week of September 1919, and the results are given in Table XVII. As far as the composition of the herbage is concerned, the open hearth (fluorspar) basic slag compares very favourably with the basic Bessemer slag, and both are greatly superior to the untreated plot. Butcher's Farm, Lambourne End. These trials were not com- menced until 1919, the manures being sown on January 4th, 1919. The writer had been offered a supply of a new 'ferruginous phosphate' recently discovered in Cleveland (N.R. of Yorkshire), 40 THE ESSEX EXPERIMENTS and, through the courtesy of Dr Stead, a small quantity of two open hearth slags from the same Steel Works, but of widely different solubiHties. It was therefore decided to start a new experimental centre in order that a fair comparison between the different phosphates might be secured. The size of the plots was one-quarter acre, and the usual dressing of 200 lbs. P2O5 per acre was given of the various phosphates. The plots have received no further treatment. Table XVIII. Weight of Hay at Butcher's Farm, Lambourne End Manures sown: January 4th, 1919 Plot J acre Manure 200 lbs. P2O5 per acre Citric solubiHty of the phos- phate(%) Hay (in cwts. per acre) 1919 1920 1921 Average 3 years A 1 2 3 4 5 6 7* 8* 9 Cambridge coprolites... Open hearth (fluorspar) basic slag ... Open hearth basic slag No manure Egyptian phosphate ... Florida pebble phosphate Tunisian phosphate ... Open hearth (fluorspar) basic slag (Wigan) Open hearth basic slag (Wigan) Cleveland phosphate 25 20 91 35 18 24 32 80 19 25-0 26-6 24-5 13-2 18-0 16-9 19-0 16-0 23-7 19-9 32-3 34-7 36-2 21-4 34-4 37-8 38-1 34-1 38-0 38-9 33-3 38-5 310 18-4 27-4 30-5 34-0 29-4 28-5 34-4 30-2 33-3 30-6 17-7 26-6 28-4 30-3 26-5 30-1 31-1 Rainfall, May 1st till harvest (in ins.) Date of cutting — 3-08 July 17 5-27 July 17 2-44 June 25 * Plots 7 and 8 = ttto °^ ^^ ^*^^®- The condition of the meadow was very different from that of the other centres. Instead of a bare open surface as at Horndon, the surface was covered with a thick matted turf. Down to a depth of about 12 inches the soil was of a fibrous peaty character, and although it rested on a London clay sub -soil, the first 9 or 12 inches of soil resembled a sour peat soil. Scarcely a trace of leguminous plants has been visible on the untreated plot throughout, the hay consisting largely of water grasses and the type of weeds characteristic of sour soils. The soil, as will be seen in Table VII, was very deficient in total and available phosphoric acid, it contained no calcium carbonate, THE ESSEX EXPERIMENTS 41 had the high lime requirement of -45 %, and was highly charged with organic matter. The results are given in Table XVIII. Although there are considerable differences between the effect of the various phosphates at this centre during the dry season of 1919, there are no decided indications that high citric solubihty has been of any great importance. The noticeable difference between the returns from the two types of open hearth fluorspar basic slag (Plots 1 and 7) is somewhat surprising, especially as it is the more soluble of the two slags which gives the poorer result. It has, however, been pointed out (18) that a modification of the solubihty test, so that 1 gm. instead of 5 gms. of the phosphate is used in performing the test, reverses the order of solubihty of these two slags. The slag on Plot 1 becomes 60-6 % soluble whilst that on Plot 7 is only 37-7 % soluble. Difference in the nature of the phosphates in the two slags is evidently in this case of greater importance than any question of solubihty by the Wagner citric acid test. During the dry season of 1919 clover was practically absent from these plots, and it was not until May of 1920 that it began to force its way through the matted turf on the treated plots. In June the progress made was remarkable, and by the end of the month the treated plots were covered with a thick and vigorous growth of red and white clover, which over large areas practically precluded the growth of any other type of vegetation (Table XXVIII). The clover on Plot 2 (high soluble slag) made better progress than that on Plot 1 (low soluble slag), and on the whole the high soluble slag was the better of the two. The difference was noticeable at the beginning of the season, but towards the end it became less and less visible. Throughout the whole season Plot 7 (open hearth fluorspar basic slag) was much inferior to any of the other plots, and it was rather surprising to find it weigh out so heavily. The Cleveland phosphate plot was perhaps the best plot on the field, although the superiority was not great. The Florida pebble phosphate was slow in making a start, but this plot made rapid progress and was ultimately one of the best plots on the field. In spite of the dry season the 1921 hay crop was quite a heavy one, whereas at Martin's Hearne, only a short distance away, the crop was practically a failure (Table X). The contrast is probably due to the difference in the soils. The heavy soil at Martin's Hearne 'bakes and cracks ' during a hot and dry spell of weather, and under such circumstances it is likely that the greater part of the heavy fall of 42 THE ESSEX EXPERIMENTS rain on May 26tli (1 inch) was lost by surface drainage. At Lamboume End the soil is not heavy and its peaty character no doubt enabled it to retain a much greater proportion of the rain which fell on the 26th. Over a period of three years the heaviest average crop of hay has come from the plot receiving open hearth (fluorspar) basic slag of 20 % solubihty — the slag which has given comparatively poor results at other centres (Martin's Hearne, Table X, and Homdon, Table XIV). As if to emphasise the peculiarity the next best return is given by the least soluble of the mineral phosphates used in the experiments. At Lambourne End citric solubihty is clearly of minor importance and the difference in the behaviour of the open hearth (fluorspar) basic slag (Plot 1) at this centre and at Martin's Hearne is probably accounted for by soil conditions. The Lamboume End soil has a much higher hme requirement figure (-45 % compared with -27 % at Martin's Hearne) but it is by no means certain that this difference alone suffices to explain the results. Discussion of the Results on the London Clay Soils At aU the centres on the London clay formation there has been a marked response to phosphates. The effect of the manures on the whole is even more striking than on the boulder clay soils. Differ- ences between the relative efficiency of the various phosphates have been noticeable. Where the pasture has been down for many years, and where as a consequence big stores of organic matter have accumulated and the soil has a high 'hme requirement,' and where the rainfall is adequate, then rock phosphates prove quite as efficient as the best grades of basic slag. Under these conditions open hearth (fluorspar) basic slags do not give consistent results. One type of open hearth fluorspar slag of very low solubihty proves quite as efficient as the highest citric soluble type of slag, whilst another open hearth fluorspar slag, apparently more soluble than the former one, proves decidedly inferior. Where the pasture is comparatively new (30 years or so), where the summer rainfall is low, and where a small reserve of calcium carbonate still exists in the soil, the best results are secured by the highest citric soluble types of basic slag. The improvement effected by the best grades of basic slag is, however, closely approximated to by the North African rock phosphates. In fact in some years, noticeably at Latchingdon, rock phosphates may do considerably better than the best grades of basic slag (Table XVI). THE ESSEX EXPERIMENTS 43 The open hearth fluorspar slag of 45 % solubility on the average of five years seems to be quite as effective at Latchingdon as the most soluble types of slag, although it seems to be somewhat less effective during the fourth and fifth years of the experiment. The low soluble open hearth fluorspar slag (20% soluble), although it gave promise of good results at Horndon during the first two seasons, fell far behind during the favourable season of 1920, and proved to be inferior to the more insoluble types of rock phosphate. This slag also proved less effective on the boulder clay soil at Martin's Hearne (Table X). It is not suggested that this slag is of little value. On the contrary the improvement effected is really very marked in aU cases, and only suffers by comparison with the other phosphates. FIELD EXPERIMENTS ON CHALK SOILS Wendens, Saffron Walden. The lower photograph on Plate VII illustrates the character of the soil at this centre. The chalk is covered by a thin layer of boulder clay soil, mixed with a large proportion of chalk. The first nine inches of soil at Wendens contains upwards of 36 % of chalk. It is naturally comparatively rich in phosphates, and although the available phosphoric acid figure is low (Table VII) this is no doubt due to the calcium carbonate neutrahsing the citric acid. This type of soil dries out quickly. It therefore makes an early start in the spring, and the meadow hay is always ready to cut during the second or third week of June. In this respect, therefore, the conditions are somewhat different from those at the other experi- mental centres, which are known as late meadows and which are not harvested before the second week of July. The experimental field at Wendens was allowed to faU out of cultivation about 25 years ago. No seeds of any description were sown, and the meadow is therefore a natural one. The general practice has been to cut the field every year for hay, and to fold the aftermath with sheep. The pasture is of a much superior type to that on any of the other experimental centres. The same phosphates were used at this centre as at Latchingdon and Tysea Hill Farm. The weights of hay for the five years 1916-1920 are given in Table XIX, and are represented diagrammatically in Fig. 6. The results at this centre are of considerable interest because they show that even on an early meadow the more insoluble types of phosphate, such as those represented by rock phosphates and the better types of open hearth (fluorspar) basic slag, are capable of 44 THE ESSEX EXPERIMENTS Table XIX. Weight of Hay at Wendens, Sapfron Walden Manures sown: January, 1916 Plot Manure Citric solubility Hay (in cwts. per acre J acre 200 lbs. P2O5 per acre of phos- phate (%) 1916 1917 1918 1919 1920 Average 5 years 1 Basic Bessemer slag ... 92-0 68-5 30-4 42-6 23-4 39-8 40-9 2 Gaf sa rock phosphate . . . 38-3 62-8 31-5 39-7 20-7 36-2 38-1 3 No maniire — 51-2 25-4 33-4 14-3 28-0 30-4 4 Open hearth (fluorspar) basic slag 45-0 64-9 35-1 42-3 17-7 39-7 39-9 5 Open hearth basic slag (high soluble), (1) ... 93-4 54-8 34-4 35-7 15-9 36-7 35-5 6 Open hearth basic slag (high soluble), (2) ... 82-2 60-3 40-4 42-6 17-8 34-1 39-2 Average increase of Plots 1, 2, 4, 5 and 6 over Plot 3 — 21-6 36-0 21-6 33-5 32-2 — Rainfall, May 1st till harvest (in inches) . . . — 4-00 4-00* 2-44 0-63 2-42 Plots cut June June June June June 17 16 29 19 22 * 2-3 inches feU on May 20th. 50[-| 40 30 20 10 fii Fig. 6. Yield of Hay (average of 5 years) for the various Phosphate Plots at Wendens, Saffron Walden. Soil Chalk. 3, Untreated. 2, Gafsa rock phosphate. 6, Open hearth (high soluble) basic slag. 4, Open hearth (high soluble) basic slag. 1, Basic Bessemer slag. THE ESSEX EXPERIMENTS 45 giving returns comparable to those obtained by the use of the old basic Bessemer slag. The explanation probably rests on the fact that rainfall is the most important limiting factor on this type of soil. (It wiU be observed that the hay crop varies from 14-3 cwts. per acre in the dry season of 1919 to 51'2 cwts. during the favourable season of 1916.) Shortage of phosphate is possibly the second Umiting factor, and the original dressing appUed is more than is essential. Conclusions drawn from the Field Experiments With two exceptions the field experiments show a marked response to phosphates. The failure at Hassobury is probably partly due to the fact that the soil is comparatively rich in phosphoric acid, and partly to the fact that it is very much poorer in potash than the soil at those centres where a response to phosphates was secured. The failure at Farnham is not due to the soil being well suppHed in phosphoric acid, but to a deficiency in some other factor. If the centres where a definite response has been secured are con- sidered, it is quite apparent that good results can be expected on both the London clay and boulder clay soils from the various types of rock phosphates, and that, considered over a period of four or five years, it is reasonable to expect these phosphates to give results approximately equivalent to those secured from the high citric soluble types of basic slag. Seasonal differences have, however, been apparent which suggest that the rock phosphates require a considerably higher rainfaU to produce the maximum effect than is the case with the high soluble slags. These differences are clearly apparent at Martin's Hearne, Latchingdon and Lambourne End. The seasons 1917 and 1919 were dry, or comparatively so, whilst those of 1916, 1918, and 1920 were moist. If the results for the two dry years on the high soluble slag plot and the Gafsa rock phosphate plot, and the corre- sponding results from the moist season, are compared as is done in Table XX the influence of the season on the availabihty of the rock phosphates wiU be seen to be very pronounced. During dry seasons high soluble basic slag gives considerably better results both at Martin's Hearne and at Latchingdon. Latchingdon is in the eastern and drier portion of the county, whilst Martin's Hearne is in the western and moister section, and it is of interest to note that, as might be expected, the advantage of the high soluble slag over the Gafsa rock phosphate is greater at Latchingdon than at Martin's Hearne. 46 THE ESSEX EXPERIMENTS The relative position of the two types of phosphate during the wet seasons is curious. At Martin's Hearne the rock phosphate has a decided advantage. It also does a trifle better at Latchingdon, but the advantage is so small as to be well within the limits of experi- mental error. The results at Lamboume End (Table XVIII) for the dry season of 1919 and the moist season of 1920 fully bear out the results recorded above. The soil at Martin's Hearne has considerably more organic matter in it than the soil at Latchingdon. Moreover it is a ' sour ' soil, with a Kme requirement of -27 %, whilst the soil at Latchingdon has a Table XX. Effect of Rainfall on the Availability OF Rock Phosphates Centre AvEBAGE Weight of Hay (in cwts. per acre) Gafsa phosphate High soluble basic slag in- crease due to solubility Average rainfall May 1st tiU harvest (inches) Martin's Hearne Latchingdon Martin's Hearne Latchingdon Dry seasons, 1917 and 1919. . I 24-3 I 28-7 I 4-4 I . I 23-7 I 30-1 1 6-4 I Moist seasons, 1916, 1918 and 1920. Lime require- ment of soil /o 36-9 29-7 32-6 28-9 -4-3 -0-8 4-06 3-74 8-92 609 •27 •03 •27 •03 small reserve of calcium carbonate and has only a neghgible Kme requirement. It would seem therefore that, on ' sour ' soils well sup- phed with organic matter and situated in districts with a moderately high rainfaU, rock phosphates may give even superior results to those secured from basic slag. In Table XXI the average returns from rock phosphates and basic Bessemer slag on the ' sour ' soils are contrasted with the corresponding results from those centres where the soil is 'sweet,' that is, has a reserve of calcium carbonate and no Hme requirement. The differences, although small, are probably real. The figures for Martin's Hearne, for example, are compiled from four rock phosphate plots over a period of four years, for Lambourne End from five rock phosphate plots over a period of two years, and two high soluble slags over a similar period. Those for Latchingdon and Saffron Walden represent one rock phosphate plot and one basic Bessemer slag plot THE ESSEX EXPERIMENTS 47 over a period of five years, whilst at Horndon the results are made up from ten rock phosphate plots and two open hearth high soluble basic slag plots for a period of two years. Table XXL Comparison of Results on Sour and Sweet Soils Centre Lime require- ment of soil % Ph. value of soil Eock phos- phate. Average cwts. per acre Basic Bessemer slag. Average cwts. per acre Sour soils : Tysea Hill ... Martin's Hearne Lambourne End 0-29 0-27 0-45 5-7 6-1 30-5 28-8 28-0 30-9 30-7* 30-6* Average — ... 29-1 30-7 Sweet soils : Latchingdon ... Saffron Walden Horndon 003 0-00 0-00 7-8 7-7 27-3 38-1 19-5 29-4 40-9 23-4* Average — 28-3 31-2 * Open hearth high soluble basic slags. The results shown in Table XXI, although they do not agree with those secured by Pfeiffer^, yet demonstrate that rock phosphates compare more favourably with basic slag on 'sour' soils than on chalky soils. Had the seasons of 1917 and 1919 been as favourable as those of 1916, 1918 and 1920, it is probable that the contrast would have been more marked. Open hearth fluorspar basic slags are very uncertain in their action. The improvement effected by their apphcations has in every instance been considerable, but the lower soluble types have undoubtedly proved to be less effective than either the high soluble slags or the rock phosphates. At only one experimental centre (Lambourne End) has the open hearth fluorspar slag of 20 % solubiHty given results com- parable with the high soluble slag. The open hearth slag of 45 % solubihty has proved quite as effective as the best grades of slag and rock phosphates. On the other hand, an open hearth slag of 30 % solubihty has given at Lambourne End inferior results to a similar slag of 20 % solubihty. The relation of citric solubihty to the value of the phosphate has been discussed elsewhere (18), ^ PfeifEer is quoted in the Inter. Inst, of Agric. Bulletin, September, 1913, p. 1316, as follows: "on sour soils and on peat moss soils crude earthy phosphates (Algerian, Gafsa, etc.) do better than basic slag." 48 THE ESSEX EXPERIMENTS The Applicability of the Results It must be borne in mind when considering these experiments that the rainfall conditions in Essex are not favourable to insoluble phos- phate. In view, therefore, of the fact that rock phosphates have proved, even under unfavourable conditions, to be but little inferior to the best grades of basic slag, it seems fair to conclude that the results detailed here are apphcable to the heavy clay pasture and meadow land which cover large areas in this country. The impression gained from close observation, of the various experi- ments over a period of five years leads to the conclusion that rock phosphates are slower in their action during the spring and early summer, but if the crop continues to grow until the latter end of July this disadvantage disappears. If, however, the harvest is early, the advantage is with the higher soluble phosphate. It is probable, therefore, that for root crops where the growing period continues well into the autumn, rock phosphates wiU prove almost as effective as the best grades of basic slag^, and in the northern and western parts of the country, where the corn harvest is late and the rainfall high, rock phosphates of the North African type may reasonably be expected to prove a suitable substitute for the high grade basic slags of the past. 1 Journal Department of Agric. and Tech. Institute for Ireland, Jan. 1917. AN INVESTIGATION INTO THE REASON WHY BASIC PHOSPHATES HAVE CAUSED INCREASED YIELDS THE EFFECT OF PHOSPHATES ON THE BOTANICAL COMPOSITION OF THE HERBAGE JVliDDLETON(i5) in his paper on "The Improvement of Poor Pastures " puts to himself the following question: "Why do phosphates produce so rapid an increase?" He states that a study of tables recording live weight gains or yields of hay will not supply the answer, but "that the pastures themselves when closely examined clearly explain the action of the manures." He attributes the results to: (1) the very rapid increase in the leguminous herbage which takes place; (2) a rapid improvement in the quahty of the surface soil; and (3) the accumulation of atmospheric nitrogen fixed by the nodule organisms. He gives it as his opinion, formed from his careful inspec- tion of the various experiments, that phosphates have Httle or no direct action on the grasses, and that it is the hme in the basic slag acting on the nitrogen accumulated by the nodule organisms which brings about the improvement in the grasses. This improvement does not take place until the clover has been well estabhshed. Finally Middleton concludes that it is impossible to obtain a purely legu- minous herbage, that clovers wiU partly and sometimes almost com- pletely disappear in three or four years as a consequence of the competition of the grasses encouraged by the nitrogen accumulated in the soil by the clover nodule organisms. Gilchrist comes to very much the same conclusions in his reports on the Tree Field results. In order to obtain more detailed information concerning the effect of phosphates on the composition of the hay crops in Essex, botanical analyses of the hay on the plots at several centres were made during the season of 1919, the samples being taken the same day as the hay was cut. Martin's Hearne and Tysea Hill. The results from the experi- mental centres at Martin's Hearne and Tysea Hill are set out in Tables XXII and XXIII. 50 EFFECT OF PHOSPHATES Several points of interest are brought out by these two tables. The luxurious bottom of red and white clover which covered the treated plots at Martin's Hearne in 1918 (see Plates III and IV) had aU but vanished during the 1919 season, and Leguminosae formed only a fraction of a per cent, of the hay crop at both the above centres. Nevertheless, the contrast in the botanical analysis of the treated Table XXII. Botanical Composition, by Weight, of the Hay at Martin's Hearne Farm Soil: Boulder clay. Manures sown: Feb. 28th, 1917. Sample taken: July 9th, 1919 Species Ploti Open hearth (fluorspar) basic slag % Plot 2 Open hearth high soluble basic slag % Plots No manure % Plot 4 Gafsa rock phosphate % Plots Egyptian rock phosphate % Plot 6 Algerian rock phosphate Clovers Grasses Weeds trace 85-2 14-8 trace 88-1 11-9 trace 58-5 41-5 trace 82-6 17-4 trace 96-7 3-3 trace 95-8 4-2 Composition of the grasses by weight Lolium perenne ... 9-9 22-0 6-8 26-9 19-8 17-0 Phleum pratense ... 6-0 7-7 2-8 4-5 5-7 1-9 Cynosurus cristatus 20-6 14-7 10-8 25-2 28-7 10-6 Poatrivialis 1-3 120 0-6 10-9 7-3 9-5 Avena flavescens 1-3 1-4 0-6 1-0 1-3 0-6 Festuca ovina — 0-9 — — Holcus lanatvs ... 32-5 29-7 44-3 18-0 17-0 29-0 Agrostis alba 0-7 2-6 6-8 4-5 4-8 11-2 Anthoxanthum odoratum 27-7 9-0 27-3 9-0 15-4 20-2 100-0 100-0 100-0 100-0 1000 100-0 Superior grasses 391 58-7 21-6 68-5 62-8 39-6 Inferior grasses ... 60-9 41-3 78-4 31-5 37-2 60-4 and untreated plots is very striking indeed. The hay on the untreated plots at both centres consists largely of weeds, and poor undesirable grasses such as Holcus lanatus, Agrostis alba and Anthoxanthum odoratum. The apphcation of phosphates has either directly or in- directly considerably affected the botanical composition of the grasses. The better types of grasses such as Lolium perenne, Phleum pratense, Cynosurus cristatus and Poa trivialis show a general increase on aU the treated plots. With the exception of the open hearth (fluorspar) basic slag at Martin's Hearne, all the phosphates seem to be equally effective in bringing about the change. Although the clovers have ON BOTANICAL COMPOSITION OF HERBAGE 51 Table XXIII. Botanical Composition, by Weight, op the Hay AT Tysea Hill Farm Soil: Boulder clay. Manures sown: December, 1915. Samples taken : July 9th, 1919 Species Plotl Basic Bessemer slag % Plot 2 Gafsa rock phos- phate % Plots Un- treated % Plot 4 Open hearth (fluor- spar) basic slag % Plot 5 Open hearth high sol- uble basic slag 1 % Plot 6 Open hearth high sol- uble basic slag 2 Plot? Un- treated % Clovers Grasses Weeds ... trace 89-1 10-9 trace 90-5 9-5 trace 66-7 33-3 trace 93-7 6-3 trace 91-0 9-0 trace 94-5 5-5 trace 71-8 28-2 100-0 1000 100-0 1000 1000 100-0 100-0 Composition of the grasses by weight LoUum perenne 31-1 22-5 15-3 21-2 20-6 21-5 5-4 Phleum pratense 2-0 1-3 — 50 4-1 4-6 — Cynoswrus cristatus ... 18-4 19-5 130 22-4 18-5 16-0 16-0 Avena flavescens 2-4 4-2 — 1-0 1-2 0-9 2-3 Hordeum pratense 2-4 1-3 3-9 1-2 2-5 5-0 5-9 Holcus lanatus 27-6 27-4 41-5 24-8 27-7 19-6 250 Agrostis alba 5-6 14-3 140 12-2 14-4 21-4 14-4 Anthoxanthum odoratum 10-5 9-5 12-3 12-2 110 11-0 31-0 1000 1000 1000 1000 1000 100-0 100-0 Superior grasses 53-9 47-5 28-3 49-6 44-4 43-0 23-7 Inferior grasses 46-1 52-5 71-7 50-4 55-6 570 76-3 Fig. 7. Botanical composition of the Hay, by weight, at Martin's Hearne. Season, 1919. Soil Boulder clay. 1, Open hearth (fluorspar) basic slag. 2, Open hearth (high soluble) basic slag. 3, Untreated. 4, Gafsa rock phosphate. 5, Egyptian rock phosphate. 6, Algerian rock phosphate. 4—2 52 EFFECT OF PHOSPHATES disappeared, the improvement in the grasses has succeeded in reducing the amount of weeds to about one-third of that present in the hay from the untreated plots. The main points of difference are illustrated in Figs. 7 and 8. 1-170 60 O W _ Ph Good Gross Inferior Grass Weeds '1^^^/f1 mk> 50 40 30 20 -10 1 r Fig. 8. Botanical composition of the Hay, by weight, at Tysea Hill. Season, 1919. Soil Boulder clay. 1, Basic Bessemer slag. 2, Gafsa rock phosphate. 3, Untreated. 4, Open hearth (fluorspar) basic slag. 5, Open hearth (high soluble) basic slag 1. 6. Open hearth (high soluble) basic slag 2. Table XXIV. Peecentage of Ground space occupied by the Vegetation at Tysea Hill and Martin's Hearne Tysea Hill Maktin's Heabne Plot 1 Plot 3 Basic slag Untreated Plot 2 Basic slag Plots Untreated Clovers Grasses Weeds Bare space 50-0 50-0 0-3 34-8 64-9 6-3 30-5 0-7 62-5 20 19-2 1-9 76-9 1000 1000 100-0 100-0 The aftermath on both sets of plots was grazed by cattle, and during the first week of October 1919 a determination of the ground space occupied by the various species was made, and the results are set out in Table XXIV. The clover had not reappeared at either centre in spite of the favourable chmatic conditions, and although the treated plots could be distinguished from the unmanured more than a mile away through- ON BOTANICAL COMPOSITION OF HERBAGE 53 out the whole winter and early spring, there was never any visible difference in the clover content between the treated and untreated plots. A chemical examination of the soils on the treated and untreated plots showed that at least one-half of the original dressing of phos- phoric acid was still present in an available form in the treated plots (see p, 105). The disappearance of the clover during 1919 was not therefore due to lack of phosphates. Table XXV. Botanical Composition of the Hay BY Weight at Martin's Hearne Sample taken: August 9th, 1920 Plot 2 Basic slag high soluble % Plots Untreated % Plot 4 Gafsa rock phosphate /o Clovers Grasses Weeds 27-5 630 9-5 11-2 58-5 30-3 350 54-2 10-8 Table XXVI. Botanical Composition of the Hay BY Weight at Tysea Hill Sample taken: August 23rd, 1920 Plotl Basic Bessemer slag /o Plot 2 Gafsa phosphate /o Plot 3 Untreated % Clovers Grasses Weeds 5-9 85-5 8-6 4-4 89-6 6-4 4-4 88-5 71 From March 1920 onwards the plots were inspected closely every week, and towards the end of May it became evident that the clover plant was again beginning to make headway, and by the end of July all the plots at Martin's Hearne, and particularly the rock phosphate plots, were covered with a vigorous growth of red and white clover. At Tysea Hill, less than half a mile away, there was very little clover showing, any difference there may have been between the treated and untreated plots in this respect was not discernible. Samples of hay from both centres were taken when the crops were cut, and the results of a partial botanical analysis are set out in Tables XXV and XXVI. 54 EFFECT OF PHOSPHATES A comparison of Tables XXV and XXVI brings out several points of considerable interest. In the first place the aU but complete dis- appearance of clover from the herbage at Martin's Hearne and Tysea Hill during the dry season of 1919, and the return of the clover at Martin's Hearne but not at Tysea Hill during the moist favourable season of 1920 is curious. Secondly, it will be noted that during the dry season of 1919 weeds formed about 30 % of the small crop on the imtreated plot at both centres. In 1920 weeds stiU formed about 30 % of the crop by weight at Martin's Hearne, but at Tysea HiU the crop on the untreated plot was a heavy one and the hay on this plot was as free from weeds as on any of the treated plots. Table XXVII. Botanicax, Composition of the Hay by Weight AT Lambourne End (London Clay) Sample taken: July 17th, 1919. Manures sown: Jan. 4th, 1919 Plot 2 Basic slag high soluble % Plots No manure % Clovers, etc Grasses Weeds 0-3 88-7 11-0 00 86-9 13-1 lOO-O 1000 Composition of the grasses by weight Cynosurus cristatus 7-4 6-3 Avena flavescens 4-3 11-3 Agrostis alba 320 36-9 Holcus lanatus 48-9 34-2 Anthoxanthum odoratum 7-4 11-2 Lambourne End (London clay). The botanical composition of the hay at Lambourne End is shown in Table XXVII. As will be seen from this table, the phosphates were sown six months before the plots were cut. The season (1919) was a dry one, and the phosphates were without any appreciable effect on the clovers, which could therefore not act as an intermediary in encouraging th^ growth of the grasses. Nevertheless the basic slag plots jrielded almost twice the crop secured on the unmanured plot, a result which appears to indicate that phosphates have a direct and not an indirect action on the grasses, and that it is quite possible to obtain a specific and marked response to phosphates on pastures where clover plants are absent. ON BOTANICAL COMPOSITION OF HERBAGE 55 The moist season of 1920 was more favourable to the growth of clover, and during the latter part of May and the month of June the plots were rapidly covered with a luxurious growth of red and white clover. The botanical examination of the hay crop in 1920 is set out in Table XXVIII and is illustrated in Fig. 9. Table XXVIII. Botanical Composition op the Hay BY Weight at Lambouene End, 1920 Sample taken: July 17th Plotl Open hearth (fluorspar) basic slag % Plot 2 Open hearth high soluble basic slag % Plots No manure % Plot 4 Egyptian phosphate Plot? Open hearth (fluorspar) basic slag Wigan % Plots Open hearth high soluble basic slag Wigan % Plot 9 Cleveland phosphate % Clovers Grasses Weeds 22-7 67-8 9-5 25-6 61-9 12-5 2-3* 70-3 27-4 33-5 59-3 7-2 15-2 72-5 12-3 33-7 57-9 8-4 38-6 55-7 5-7 1000 100-0 100-0 100-0 100-0 100-0 1000 * Practically all bird's foot trefoil, purple vetch and Vicia sativa. w n80 70 _ 60 . PM :rJH- FiG. 9. Botanical composition of the Hay, by weight, at Lambourne End. Season, 1920. Soil London clay. 1, Open hearth (fluorspar) basic slag. 2, Open hearth (high soluble) basic slag. 3. Untreated 4, Egyptian phosphate. 9, Cleveland phosphate. The results recorded in the above table simply afford another illustration of the effect of the various phosphates in encouraging the development of the clover plant. It would be difficult to secure poorer quahty hay than that obtained even on the slag plot in 1919. When conditions are favourable to the development of clover as was the case in 1920, phosphates, in addi- tion to an increased crop, produce a vastly better quahty of hay. 56 EFFECT OF PHOSPHATES Table XXIX. Botanical Composition of the Hay by Weight AT BUTTERFIELDS, LaTCHINGDON. SeASON, 1919 Soil: London clay. Sample taken: July 21st. Manures sown: December, 1915 Plotl Basic' Bessemer slag /o Plot 2 Gafsa rock phosphate /o Plots No manure o/ Plot 4 Open hearth (fluorspar) basic slag /o Plots Open hearth high soluble basic slag 1 /o Plots Open hearth high soluble basic slag 2 /o Clovers Grasses Weeds 20-8 74-4 4-8 15-8 81-0 3-2 6-3 87-9 6-8 17-6 80-4 20 150 82-8 2-2 18-0 80-9 11 100-0 100-0 100-0 100-0 100-0 Composition of the grasses by weight 100-0 Lolium perenne . . . 14-4 15-2 10-8 19-4 241 17-9 Phleum pratense ... 23-3 21-3 14-8 190 23-3 25-6 Cynosurus cristatus 15-7 18-3 11-6 20-4 22-3 19-3 Poa trivialis 8-5 10-8 7-6 12-4 15-7 10-8 Bromus mollis 33-9 8-7 3-6 4-2 — — Hordeum pratense 3-8 21-8 35-6 19-6 10-0 15-5 Agrostis alba — 3-0 6-8 20 2-0 2-4 Holcus lanatus ... 0-4 0-9 8-3 3-0 20 8-5 A nthoxanthum odoratum — — 0-9 — 0-6 — 100-0 100-0 100-0 100-0 100-0 100-0 Superior grasses ... 61-9 65-6 44-8 71-2 85-4 73-6 Inferior grasses . . . 381 34-4 55-2 28-8 14i6 26-4 rfloo Fig. 10. Botanical composition of the Hay, by weight, at Butterfields, Latchingdon. Season, 1919. Soil London clay. 1, Basic Bessemer slag. 2, Gafsa phosphate. 3, No manure. 4, Open hearth fluorspar basic slag. 5, Basic slag. 6, Basic slag. ON BOTANICAL COMPOSITION OF HERBAGE 57 Butterfields, Latchingdon. The effect of the various phosphates on the composition of the flora at Latchingdon is similar to that at Martin's Hearne and Tysea HiU. The botanical composition of the hay is shown in Table XXIX and in Fig. 10. The contrast between the treated and untreated plots is not so marked as at the other two centres mentioned, but the pasture on the untreated plot at Latchingdon is much superior to that at Martin's Hearne and Tysea Hill (compare Figs. 7, 8 and 10). On all the treated plots the better grasses, such as Lolium perenne, Phleum pratense, Cynosurus cris- tatus and Poa trivialis, have improved their position at the expense of the poorer quahty grasses, such as Hordeum pratense, Agrostis alba, Holcus lanatus, etc. It is worthy of note, moreover, that at Latchingdon clover forms a fair proportion of the crop by weight, whereas at Martin's Hearne and Tysea Hill it was practically absent during 1919. Table XXX. Botanical Composition op the Hay Crop BY Weight at Wendens, Saffron Walden. Soil: chalk. Sample taken: June 19th, 1919 Manures sown: January, 1916 Plotl Basic Bessemer slag /o Plot 2 Gafsa rock phosphate /o Plots No manure % Plot 4 Open hearth (fluorspar) basic slag % Plots Open hearth high soluble basic slag 1 o/ /o Plot 6 Open hearth high soluble basic slag 2 % Clovers Grasses Weeds 110 86-9 2-1 101 80-9 9-0 7-8 82-2 10-0 8-2 88-8 30 80 84-5 7-5 6-9 86-9 6-2 100-0 100-0 1000 1000 1000 1000 Composition of the grasses by weight Lolium perenne . . . 64-4 50-0 45-4 56-9 43-4 51-3 Phleum pratense ... — — 0-7 — — — Dactylis glomerata 3-7 3-5 3-8 1-2 8-0 1-6 Cynosurus cristatus 8-6 14-8 14-6 10-5 11-9 6-9 Poa trivialis 3-7 4-0 6-1 3-6 4-3 5-9 Avena flavescens ... 5-8 14-1 13-9 16-4 13-5 14-7 Festuca ovina — — 0-8 — — — Holcus lanatus 3-7 1-3 2-4 2-7 1-9 2-6 Bromus mollis 10-1 12-3 12-3 8-7 170 17-0 100-0 1000 100-0 1000 100-0 100-0 58 EFFECT OF PHOSPHATES Wendens, Saffron Walden. The botanical composition of the grass at Wendens, Saffron Walden (chalk) is shown in Table XXX. The quality of the meadow is obviously much superior to that at any of the other centres, and it is therefore not surprising to find that the various phosphates have had a comparatively small effect on the quahty of the herbage. Table XXXI. Botanical Analysis of the Hay Crop BY Weight at Horndon. Soil: London clay. Sample taken: Aug. 16th, 1920 Manures sown: Feb. 27, 1918. Plots B, C, D and H: Feb. 3rd, 1919 Plot Manure Clovers Grasses Weeds B Cambridge coprolites 67-5 28-7 3-8 C Lime 20-6 77-8 1-6 D Coarse ground open hearth (fluorspar) basic slag ... 56-9 38-1 5-0 1 Florida pebble phosphate ... 51-9 40-0 8-1 3 Algerian phosphate 51-3 45-8 2-9 5 Open hearth basic slag (high citric soluble) 45-4 53-3 1-3 6 Untreated ; 8-7 81-4 9-9 8 Gafsa phosphate 63-8 320 4-2 9 Tunisian „ 47-7 49-2 31 11 Egyptian „ 49-3 47-7 30 13 Superphosphate 340 61-4 4-6 15 Superphosphate and Ume 51-4 47-9 0-7 16 Untreated ... 7-6 85-5 6-9 17 Open hearth basic slag (high citric soluble) 441 53-4 0-5 18 Open hearth (fluorspar) basic slag (low citric soluble) 45-5 53-2 1-3 H Cleveland phosphate 58-9 38-5 2-6 K Untreated 8-0 80-1 11-9 Horndon (London clay). These plots were grazed during 1919, and samples of hay for botanical analysis were not removed until the 1920 crop was cut on August 16th. The results which are set out in Table XXXI, and iQustrated in Fig. 11, show an extraordinary contrast between the treated and untreated plots. In view of the effect of grazing during 1919 on the growth of the herbage on the plots receiving phosphates (see Table XV and Plates VI and VII), it is, however, not surprising to find that clover was the dominant constituent of the hay crop in 1920. ON BOTANICAL COMPOSITION OF HERBAGE 59 TOO B, Cambridge coprolites. C, Lime. 3, Algerian phosphate, 5. Open hearth (high soluble) basic slag. 6, Untreated. 8, Gafsa phosphate. 9, Tunisian phosphate. o a 90 80 70 60 H50 40 30 20 10 I" 13 L §15 I ^ 16 ^HL [ ^18 L H ^ 11, Egyptian phosphate. 13, Superphosphate. 16, Super phosphate and lime. 16, Un- treated. 17, Open hearth (high soluble) basic slag. 18, Open hearth (fluorspar) basic slag. H, Cleveland phosphate. K, Untreated. Fig. 11. Botanical composition of the Hay, by weight, at Great Mulgraves, Horndon-on-the-HiU. Season, 1920. Soil London clay. DISCUSSION OF THE RESULTS OF THE BOTANICAL ANALYSIS Although a rapid and large increase in the amount of clovers present in the herbage has followed the application of phosphates at the various experimental centres, and although the various phos- phates appear to be equally effective in this respect, there are clear indications that the effect of the phosphates on the herbage is not confined to the clovers alone. At Martin's Hearne, Tysea HiU and Latchingdon for example (see Figs. 7, 8 and 10), the better types of grasses are greatly encouraged as a result of the apphcation of phosphates. The meadows at Martin's Hearne and Tysea Hill have been down for a considerable time. The nitrogen content and the 60 EFFECT OF PHOSPHATES organic matter content of the soil are already high, and it is difficult to beheve that the accumulation of a comparatively smaU amount of nitrogen by the nodule organisms and its subsequent nitrification is responsible for the double crop which the treated plots bore in 1919 when no clover was present. It seems more probable that the double crop is due either to the grasses benefiting by the direct f ertihsing effect of the phosphates or to the phosphates having some action on the production of nitrates in the soil, or to the operation of both these causes. This contention is borne out by the results at Table XXXII. Botanical Composition op the Hay by Weight ON THE Plots receiving High Soluble Basic Slag. Season, 1919 BoTTLDEB Clay London Clay Martin's Heame % Tysea HiU % Famham* % Lambourne End % Latching- don % Wendens % Horndon* % Clovers Grasses ... Weeds trace 88-1 11-9 trace 89-1 10-9 50-2 33-3 13-5 0-3 . 88-7 110 20-8 74-4 4-8 110 86-9 21 441 28-6 13-6 Lime requirement RainfaUf (inches) 0-27 2-85 0-29 2-87 0-00 1-73 0-45 308 003 1-47 000 0-53 0-00 1-78 * The figures for Farnham and Horndon indicate the percentage of ground space covered by the various species and not the botanical composition of the hay. It is obvious however that had a hay crop been cut at these centres clover woidd have formed a large proportion of the crop by weight. t May 1st till Harvest. Lambourne End, where a very large increased crop of grasses foUows the apphcation of phosphates, the response of the clovers to the dressing not being manifest until the following year (see Tables XXVII and XXVIII). Table XV clearly shows the superior effect of basic phosphates in encouraging the growth of clovers, but this effect once produced is not maintained on all the soils until the dressing of basic phosphates is exhausted or nearly so. At Horndon, Latchingdon, Saffron Walden and Famham no diffi- culty has been experienced in maintaining the clover even during a dry and unfavourable season like 1919. At Martin's Hearne, Tysea HiU and Lambourne End, however, the clover completely disappeared from the treated plots during 1919. This result at Martin's Hearne was rather surprising in view of the fact that clover covered the treated plots in 1918 almost to the exclusion of the grasses (see Plate IV). ON BOTANICAL COMPOSITION OF HERBAGE 61 These differences between the various centres are brought out clearly in Table XXXII. At Latchingdon and Wendens clover has persisted on the un- treated plots as far as can be ascertained ever since the fields went down to grass. There are, moreover, no signs of the clover which was encouraged by the apphcation of slag five years ago tending to 'go off,' although, of course, it is subject to seasonal fluctuations. It wiU be noted that it is on the sour soils, and only the sour soils, that the clover failed during the season of 1919. If, however, the results for the 1920 season are tabulated, as is done in Table XXXIII, it wiU be seen that sourness is not the only factor. Table XXXIII. Botanical Composition of the Hay Crop by Weight on Plots receiving High Soluble Basic Slag. Season, 1920 Maetin's Hearne TySBA HlIiL Lam- BOITENE End (average) 2 plots Basic Slag plus slag alone lime* 0/ 0/ /o /o Basic Slag plus slag alone lime* % % HOEN- DON Clovers Grasses Weeds 27-5 630 9-5 18-7 73-3 8-0 5-9 85-5 8-6 8-5 84-9 6-6 29-7 59-9 10-4 45-4 53-3 1-3 Tiirae requirement Rainfall t (inches) 0-27 8-37 — 0-29 9-34 — 0-45 5-27 0-00 5-34 * At the rate of 35 cwts. of CaO per acre. t May 1st tiU Harvest. On the sour soils at Lambourne End and Martin's Hearne clover forms more than 25 % of the crop, and it will be noted that a dressing of hme in addition to the slag has not improved the position of the clover at Martin's Hearne. It is, of course, quite possible that the clover on this plot will benefit by the dressing of lime should another unfavourable season succeed. The application of hme at Tysea Hill has not succeeded in bringing a vigorous growth of clover even in a favourable season hke 1920, and it is quite clear that some other factor besides chmate, hme and phosphate is responsible for the failure of the clover plant. Chemically the soil at Tysea Hill differs from that at Martin's Hearne in having a lower content of available potash, and it seems 62 EFFECT OF PHOSPHATES probable that an inadequacy of potash is responsible for the failure of the clover. The effect of grazing and continuous cutting on the condition of the botanical flora is well illustrated in Table XXXIV and Fig. 12, which show the percentage of the ground space covered by the flora on the slag plots at Latchingdon and Horndon during 1919. Table XXXIV. Percentage of the Ground Space occupied by THE Vegetation on the Basic Slag Plots AT Latchingdon and Horndon Determinations made: Aug. and Sept. 1919 Latchingdon Cut 4 years in succession HOENDON Cut 1918 grazed 1919 Clovers Grasses Weeds ... Bare space 181 51-2 00 30-7 45-2 37-9 7-5 9-4 cS ft 13 o PM 60 50 40 30 20 10 i I n 1 Clover Crass Weeds Bare space x\\x\x iiiil Horndon. Cut 1918: grazed 1919. Latchingdon. Cut four years in succession. Fig. 12. Percentage of Ground Space occupied by the Vegetation on the Basic Slag Plots at Latchingdon and Horndon. Season, 1919. Soil London clay. The bottom at Latchingdon, it will be seen, is an open one, and although it shows a great improvement in this respect over the un* treated plot, it is not nearly so close as that at Horndon. At Latchingdon the clover disappears in the autumn and come again the following year towards the end of May or the beginning of June. At Horndon on the other hand the surface is covered with a network of clover runners, and there is practically no 'bare space' on the plot. The Essex farmer still holds to the practice of grazing and cutting his meadows in alternate years, and in view of these results ON MOISTURE CONTENT OF SOIL 63 and the climatic conditions of the county, there is much to be said for this practice. The botanical analyses from all the centres agree in showing that as far as the quality of the herbage is concerned there is nothing to choose between the effectiveness of rock phosphates and high citric soluble basic slags. There are, however, indications that the open hearth (fluorspar) slags of very low solubiUty are less efficient in this respect than the high soluble basic slags. EFFECT OF PHOSPHATES ON THE MOISTURE CONTENT AND TEMPERATURE OF THE SOIL It has already been stated that one of the great difficulties experi- enced on the clay soils — particularly the London clay soils of Essex — is the wet condition under which they he throughout the winter and late spring. As a rule a hot and dry spell of weather succeeds, lasting during the greater part of May and June. The soil whether under pasture or arable conditions dries up rapidly, sets as hard as a brick (caps), and cracks badly. These unfavourable conditions were very obvious at the Horndon Experimental centre during 1919. Following a wet April the remains of a heavy faU of snow were still visible on the plots on May 3rd. A spell of dry hot weather set in and lasted without any recordable rain faUing until the third week in Jime. The condition of the un- treated plot was difficult to describe. There was practically no growth and the surface was covered with innumerable cracks, some of them wide enough to aUow of the insertion of the greater part of the arm. What Httle growth there was shrivelled up by the second week in June. It was obvious, however, that the plots receiving phosphates were not suffering nearly so badly. The cracks were fewer, and required looking for, and the thick matted bottom of clover provided a con- tinuous feed for the grazing stock throughout the season (Plate VI). The marked difference between the condition of the soil on the slag plot and untreated plot suggested a better regulation of the moisture supply on the former plot. It was Tuifortunately not possible during 1919 to follow up the enquiry which these observations suggested, but during the season of 1920 a series of moisture determinations and temperature records were made on Plot 17 (basic slag) and Plot 16 (untreated) at Horndon. 64 EFFECT OF PHOSPHATES MOISTITEE Commencing in March 1920 the moisture was determined each week in a 9 inches sample of soil from each plot, and a httle later 3 inches samples were taken. At the same time temperature records were obtained at a depth of 9 inches and 3 inches. All the samples were taken by the writer and were secured by means of a 15 inches sampler of f inch diameter which could be adjusted to remove a core of 3 inches, 9 inches or 12 inches in length. The sample for each plot consisted of from 10-14 cores. Table XXXV. Moisttjre Content op the Soil at Horndon on Plots 16 and 17 at Various Depths Moisture (%) Date 0-3 inches 3-9 inches A * f • ■ -\ \ Plot 17 Plot 16 Plot 17 Plot 16 1920 Basic slag Untreated Basic slag Untreated April 26 29-1 28-6 24-0 29-6 May 5 26-1 28-6 22-8 22-6 10 26-2 29-3 22-6 25-0 17 22-8 22-6 17-8 21-6 25 201 18-4 220 21-5 31 23-7 19-0 15-4 15-4 June 8 180 15-9 12-8 151 14 19-5 16-8 15-6 170 21 21-5 19-9 18-3 181 29 18-7 17-6 13-4 16-2 July 8 26-1 25-8 20-2 23-8 12 26-3 27-0 24-3 21-8 20 18-4 18-4 17-2 19-1 26 23-2 24-7 210 22-0 Aug. 4 24-2 25-6 19-6 20-8 9 220 24-9 16-9 191 16 18-8 19-2 15-5 15-4 31 — 190 — 17-8 Sept. 7 17-9 18-7 16-9 17-2 20 26-8 26-8 23-2 23-2 Nov. 24 25-3 27-9 191 21-4 The moisture contents of the first three inches of soil and of the soil from 3 inches to 9 inches on both plots are given in Table XXXV. The results are plotted together with the rainfall record in Figs. 13 and 14. Broadly speaking the 1920 season was a moist one, and particularly favourable for the hay crop. At the beginning of May the ground, as the result of a wet April, was saturated with water. May, however, was a dry, and on the whole a hot month, and the soil rapidly began ON MOISTURE CONTENT OF SOIL 65 26 3 IT 31 14- 28 12 26 B 23 6 20 27 April May June July August Sept. Fig. 13. Moisture content of the first 3 inches of soil on the Untreated and Basic Slag Plots at Great Mulgraves, Homdon-on-the-Hill. Soil London clay. Plot 16, Untreated Plot 17, Basic Slag . Rainfall 26 JO 24- 7 21 S 13 2 16 30 13 27 April May June July August Sept. Fig. 14. Moisture content of the Soil at a depth of 3 to 9 inches on the Untreated and Basic Slag Plots at Great Mulgraves, Horndon-on-the-Hill. Soil London clay. Plot 16, Untreated . Plot 17, Basic Slag . Rainfall R.B.S. 5 66 EFFECT OF PHOSPHATES to dry up. By May 25th the untreated plot had begun to crack and by June 7th it was difficult to find a square yard of this plot that was not traversed by a big crack. On this date also the first signs of 'cracking' were observed on the slag plot, but the cracks required looking for and were of small dimensions. The dense growth on Plot 17 was obviously taking much more water from the soil than was the case on Plot 16 where the growth was neghgible and where a large proportion of the surface was bare (see Table XV). The data presented in Table XXXV and Figs. 13 and 14, however, afford some explana- tion of these differences. On May 10th, Plot 16 (untreated) contained 29 % of water, whilst Plot 17 contained approximately 26 %. By the 17th the moisture content of both plots had fallen to 23 %, Plot 16 having evidently lost its moisture at a more rapid rate than Plot 17. On the 25th the moisture content had fallen to 18-4 % on Plot 16, whilst Plot 17 in spite of the very much larger transpira- tion which was taking place from this plot contained 20-1 % of moisture, a difference of 1*7 % in favour of the slag plot. The week following the 25th was showery and hot, and on the 31st of May, when samples were again taken, there was a surprising difference between the moisture contents of the two plots to a depth of 3 inches. Plot 16 contained only 19 % of moisture, whilst Plot 17 had a moisture content of 23-7 %, a difference of 4-7 % in favour of the phosphate plot. Plot 16 received very httle benefit from the showers during the week ; much of the rain must have run down the cracks, and the bulk of the remainder, faUing on a bare surface exposed to the direct rays of the sun, was evaporated rapidly. The following eight days from May 31st to June 8th were dry and hot, no rain whatever falling. The moisture content of both plots fell rapidly, but on the 8th, Plot 17, notwithstanding the much greater demand made upon it, contained 18 % of moisture in the first three inches of soil, com- pared with 15-9 % (the lowest moisture content recorded throughout the season) on the untreated plot. The advantage of a dense crop on this type of soil is fairly obvious, and it is in fact the only practical method of conserving the soil moisture. It is difficult to emphasise the importance of this indirect action of basic phosphates on this type of soil during a dry season when the absence of rainfall in May and June or a small precipitation makes the growth of a hay crop impossible. On the 10th of June the fine weather broke and during the sub- sequent fortnight unsettled conditions prevailed. Even at the end of a fortnight Plot 17, in spite of the rapid growth which was taking ON MOISTURE CONTENT OF SOIL 67 place, still contained a higher moisture content in the surface 3 inches of soil than on Plot 16, where as far as the eye could judge the growth was all but neghgible. Apparently a good deal of the rain had drained down the cracks on Plot 16, which were stiU as prevalent as during the preceding fortnight. On July 1st a speU of wet weather set in, there being only two dry days during the first twelve days of the month. For this period a total of 1-97 inches of rain was recorded. On July 8th and 12th, when the samples were taken, both plots appeared to be equally wet and the cracks had all disappeared. The analytical results showed that on both these dates the surface three inches of soil on both plots had approximately the same moisture content. The third week of July was dry and growth on Plot 17 was rapid. At the end of the week both plots had the same moisture content, namely 18-4 %. The remaining week and the first week in August were wet and the moisture content of the surface three inches varied from 23-26 % ; the treated plot during the period 26th July to 9th August being distinctly drier. From the 9th of August to the 19th no rain fell, the untreated plot dried more rapidly and on the 16th the moisture content of both plots was approximately the same. On this date the plots were cut dead ripe; they were weighed on the 21st and carted to the stack and subsequently threshed foi wild white clover seed. The weights of hay on the two plots were, Plot 17, 28-8 cwts. and Plot 16, 6-4 cwts. per acre. It will be noted that from the 16th of August onwards (Plot 17 no longer being covered by a dense crop) the moisture on both plots remained practically the same. By determining the moisture content of the soil to a depth of 9 inches and 3 inches on both plots it was possible to calculate the moisture content of the layer of soil 3 inches to 9 inches on both plots. This was done, and the figures are given in Table XXXV, and are shown graphically in Fig. 14. The calculations were made with a view to ascertaining whether the crop on the basic slag plot was able to draw more water from the lower depth than was the case on the untreated plot. It is natural to expect this to be so under dry climatic conditions for either or both of two reasons. Firstly because of the increased root action which follows the application of basic phosphates to clay pastures, and secondly because it seemed possible that the remarkable root development which took place on Plot 17 would affect the texture of the soil to some extent and thereby facihtate the up- ward passage of capiUary water. An inspection of Fig. 14 shows that though the first three inches of soil on the basic slag plot remained 68 EFFECT OF PHOSPHATES moister than on the untreated plot, throughout the dry periods, the opposite was the case as far as the moisture content of the 3rd to 9th inch was concerned. Whenever a period of wet weather succeeded a dry spell, as for example during June 14th to 21st, the moisture content on both plots at a depth of 3-9 inches rose to approximately the same level. A dry spell invariably resulted in the moisture content of the 3rd to 9th inch on Plot 17 faUing more rapidly than on the un- treated plot. In view therefore of the high moisture contents which have prevailed at various periods throughout the season, it seems evident that the crop on Plot 17 has been able to utiHse the moisture at this depth, particularly during dry spells, to a much greater extent than was the case on the untreated plot. Mechanical analysis (admittedly imperfect for this purpose) fails to detect any difference in the mechanical structure of the soil on these two plots, and it is therefore probable that during the first years which follow the appH- cation of basic phosphate to this type of meadow land, the crop on the slag plot is able to draw upon the moisture content of the soil at lower depths than is the case on the untreated plot largely because of the increased root development. The behaviour of the two plots from August 16th, when the plots were cut, until September 20th^ tends to confirm this view. It will be seen that the moisture content of the section of the soU from 3-9 inches remained practically the same on both plots during this period and as far as the eye could judge no growth took place. Temperature The temperature records were taken by means of a special thermo- meter recording between 0° and 30° C, and graduated to one- tenth of a degree. During the latter stages of the work this thermometer was replaced by one registering only to one-fifth of a degree. It was, however, a simple matter to get results to one-tenth of a degree by interpolation. The temperature records of the soil on Plots 16 and 17 are given in Table XXXVI and are recorded graphically in Figs. 15 and 16. Examined in conjunction with the figures in Table XXXV (repre- sented in Figs. 13 and 14) the results have an important bearing upon the action of slag under dry cKmatic conditions on such heavy London clay soils. During the whole of the period from May 17th till the crop was carted off the plots on August 21st, the surface three 1 Unfortunately the sample drawn from Plot 17 on August 31st met with an accident. ON THE TEMPERATURE OF THE SOIL 69 Table XXXVI. Temperature op the Soil at Horndon ON Plots 16 and 17, at Depths op 3 inches and 9 inches Temperature (degrees Centigrade) Date At a depth of 3 inches Plot 17 Plot 16 At a depth of 9 inches Plot 17 Plot 16 1920 Basic slag Untreated Basic slag Untreated May 5 — — 9-8 10-3 10 — — — — 17 140 160 12-5 13-5 25 17-2 20-3 14-7 17-7 31 15-4 21-4 14-8 16-2 June 8 14-4 18-1 13-5 15-4 14 170 200 150 18-0 21 16-3 20-0 15-8 17-5 29 17-5 20-5 16-4 19-2 July 8 15-3 16-8 14-8 16-1 12 17-5 19-8 15-4 17-9 20 16-5 200 15-6 17-6 26 15-3 16-3 14-9 16-2 Aug. 4 14-6 17-6 14-6 16-2 9 15-2 17-4 14-9 16-6 16 15-5 19-2 15-0 16-8 21 130 13-8 — — 31 15-3 14-7 14-2 13-9 Sept. 7 15-8 15-5 15-3 14-8 20 12-2 12-5 12-6 12-7 Oct. 4 13-4 13-5 13-6 13-4 Nov. 24 4-5 4-2 5-2 50 o 23r n 24 7 21 S t9 Z 16 30 /4- Z8 5 May June July August Sept. Oct. Fig. 15. Temperature of the Soil at a Depth of 3 inches on the Untreated and Basic Slag Plots at Great Mulgraves, Horndon-on-the-Hill. Soil London clay. Plot 16, Untreated Plot 17, Basic 70 EFFECT OF PHOSPHATES zor 7 21 June S /9 July /6 30 /3 27 4 August Sept. Oct. Fig. 16. Temperature of the Soil at a Depth of 9 inches on the Untreated and Basic Slag Plots at Great Mulgraves, Horndon-on-the-Hill. Soil London clay. Plot 16, Untreated Plot 17, Basic Slag . inches of soil on the slag plot remained considerably cooler than the surface soil on the untreated plot. The importance of securing an efficient covering of the surface soil, so as to protect it from the direct rays of the sun, is well brought out in Fig. 15. The thick bottom of clover has not only succeeded in retaining the moisture on Plot 17, but it has very effectively kept the plot cool during the hot spell of weather in May and June. A comparison of Figs. 13 and 15 shows quite clearly moreover that the lower temperature of the surface soil on Plot 17 is not due to the higher moisture which it contains, but is almost entirely due to the superior covering effect of the crop on this plot. During the hot period May 1 7th-25th the temperature of the surface soil on both plots rose considerably, and on the 25th there was a difference of 3-1° C. between them. The subsequent week was showery, a total of -24 inch falling on four of the seven days. Of this amount •14 inch feU on the 29th. On the 31st the temperature of Plot 17 had fallen from 17-2° C. the previous week to 15-4°. Plot 16 on the other hand had risen from 20-3° on the 25th to 21-4° on the 31st there being now a difference of 6-0° C. between the two plots. In degrees Fahrenheit the temperatures at a depth of 3 inches on the two plots were— on Plot 16, 70-5° and on Plot 17, 59-7°. During the whole of May and June, Plot 17 (slag) at a depth of 3 inches was never less than 3° C. cooler than the untreated plot. During the wet month of July the temperature on the two plots more closely approximated, but whenever a warm and dry speU of ON THE TEMPERATURE OF THE SOIL 71 weather ensues the difference between the two plots becomes greatly- accentuated. On August 16th the plots were cut shortly after the temperature readings were taken, and the hay lay on the swathe until the 21st, when it was raked up, weighed and carted. The reason for the difference in temperature which had existed throughout the season soon became apparent. With the removal of the dense covering from Plot 17 the temperature records corresponded very closely with those of Plot 16, and in fact on several occasions were a trifle higher. The temperature records taken at a depth of 9 inches on each plot serve to emphasise the importance of an adequate covering of the soil. The beneficial effect of the covering action of the clover on the slag plot is very marked even at a depth of 9 inches, and during the warmest periods there is often a difference of 3° C. between the two plots. Plot 17 being invariably the cooler, despite the fact that the moisture content at this depth was generally lower than on Plot 16, the untreated plot. With the removal of the crop on August 21st this difference in the temperature records of the two plots at a depth of 9 inches becomes neghgible and the two curves follow each other very closely indeed. It is of course difficult to say how far this 'secondary effect' of the action of slag has contributed to the difference in the cropping power of the two soils (28-8 cwts. on Plot 17 and 6-4 cwts. on Plot 16), but there can be Httle doubt that it is of very considerable importance, and that during a dry season it may weU be the most important action. To retain moisture and keep the soil cool is the great difficulty experienced on this type of soil during the late spring and early summer. It is obvious therefore that whatever is possible should be done to maintain a thick close bottom, and for this purpose it is very desirable that the meadows should not be cut every year, but should be cut one year and grazed the next or two following years. To reserve a meadow for hay for several years is far from desirable under the climatic conditions prevaiUng in the east of the county, unless a hay crop varjdng from 1-7 cwts. of hay per acre should prove sufficiently profitable. Martin's Hearne. Itisof interest to compare the results at Horndon with those at Martin's Hearne. At Martin's Hearne the plots were cut for four years in succession and are known to have been cut for hay during the two years prior to the commencement of the experi- ments. There is no close bottom of clover at this centre such as that at Horndon (compare Tables XV and XXIV). At Martin's Hearne the clover dies down at the end of the season and in the following 72 EFFECT OF PHOSPHATES year makes its appearance towards the end of May if the season is a suitable one. During a dry unfavourable season Uke that of 1919 clover was almost entirely absent (Table XXII). Moisture and tem- perature determinations were made on Plot 2 basic slag, and Plot 3 untreated, throughout the season of 1920, and the records are shown in Table XXXVII. Table XXXVII. Effect of Basic Slag on the Moisture Content AND TeMPERATUEE OF THE SOIL AT MaRTIN's HbARNE Temperature (degrees Centigrade) Moisture Date Plot 2 Plot 3 Basic slag Basic slag Basic slag Untreated Plot 2 Plot 3 1920 3 ins. 9 ins. 3 ins. 1 9 ins. % /o Apr. 19 1 32-2 35-1 26 — 10-5 — — 341 33-2 May 5 — 9-9 — 9-7 29-9 31-6 10 — — — 36-7 34-6 17 — 13-5 — 13-5 25-3 25-5 25 18-5 15-9 — — 19-9 21-3 31 16-9 16-8 16-8 15-8 27-9 28-5 June 8 — — — — — — 14 17-6 15-6 18-1 160 23-3 25-2 21 171 16-8 17-4 16-8 24-8 23-4 29 17-8 16-7 17-8 17-5 16-7 17-5 July 8 — — — — — — 13 170 16-2 17-3 16-4 25-6 24-3 19 18-2 15-8 17-8 16-4 170 18-3 26 14-7 14-8 14-6 14-7 25-1 23-5 Aug. 4 14-9 14-6 14-9 150 26-8 26-8 9 15-2 14-6 15-5 14-8 26-0 23-6 Plots cut Aug. 9th Aug. 16 17-8 16-4 17-8 16-4 211 20-2 24 15-7 15-2 15-6 14-7 22-8 24-3 31 14-6 140 14-7 14-3 20-2 18-6 Sept. 7 151 151 15-4 15-3 19-5 200 20 14-8 14-6 14-8 14-5 25-3 25-2 Oct. 4 140 13-8 14-1 140 — — At Martin's Hearne there is no appreciable difference in the tempera- ture records of the two plots either at a depth of 3 inches or at 9 inches. There is a big difference in the yield of hay on the two plots, namely, 31-9 cwts. per acre on Plot 2, and 22-0 cwts. per acre on Plot 3. The bottom in both plots however is an open one, due to the recent practice of cutting every year, and the better growth on the slag plot does not act as a cover to any appreciable extent. ON TEXTURE OF THE SOIL 73 The moisture figures also present no important point of difference. During the dry spells the slag plot, as might be expected considering the heavier crop which it carries, loses moisture at a more rapid rate than the untreated. The rainfall at Martin's Hearne and in the west of the county generally is considerably heavier than at Horndon-on-the-Hill, which is situated in the eastern part of the county. As a general rule it would be quite a safe practice in the west of the county to cut the meadows for hay every year, and Httle would be gained by alternating with grazing, except perhaps during a par- ticularly dry season hke 1919. THE EFFECT OF PHOSPHATES ON THE TEXTURE OF THE SOIL Collins (4), discussing the effect of phosphates on grass-land, states that on the untreated plot at Cockle Park yellow clay still remains close to the surface, yet on the plot which has been manured with basic slag for over twenty years a very useful loam extends to 10 or 12 inches below the surface. The steady downward trend of the roots on the slag plot opens up the clay soil, and by admitting air and supplying organic matter gradually transforms the soil into a kindly loam. Such a transformation must of necessity have an important influence upon the movement of water in the soil. The experiments described here have only been running for a comparatively short time, and such an effect as Collins indicates would only be starting, and would not be readily noticed. The records of the moisture content at various depths on the two plots at Horndon (16 and 17), and the extraordinary contrast afforded by the root development, suggested that an appreciable alteration in the mechanical condition of the soil might have been brought about on the slag plot. In order to ascertain whether this difference is measurable by the ordinary methods of mechanical analysis, samples to a depth of 9 inches were removed from the slag and untreated plots at several of the centres during the autumn of 1919, and subjected to mechanical analysis. It cannot be said that the results, which are given in Table XXXVIII, afford much positive information. At Tysea Hill, Martin's Hearne and Butterfields, the clay fraction is appreciably less on the slag plots. At Farnham there is no difference, whilst at Horndon, where the experiment had only been iri progress for a year, the results are contradictory. In any case, the positive differences are so small as to be in each case within the limits of experimental error. 74 EFFECT OF PHOSPHATES EH O ft <1 EH a fi X X ^^ W O00OCf cd •53 ^^ 0000«>CDOt-C5T}* C;-0(N05l>'7iC>(N (N i-H ^ (N I— I w HPh Cv) r-H r— ( 1— I I— I (M 1— I 1— I 1— I I— I w pq <»p:; Kq c3 o P (D r-H hPH OOii-HOO<:DQOt> ph :::::: o.o '.JS .-+2 § . g .;S • "" S tH • cs • :d • el S ao • M 'to '.so ■S is .S iS -S 4S o o piHOiPqaifJ-fOi-li-^ ON ACCUMULATION OF NITROGEN 75 It seemed probable in view of these results that three inch samples might show more clearly any alteration in the mechanical composi- tion which were in progress on the slag plots. Three inch samples were accordingly carefully removed from Plots 16, 17 and C at Horndon during August 1920 (two and a half years after the manures were sown), and submitted to mechanical analysis. The results are given in Table XXXIX. The differences between the figures for Plots 16 and 17 are smaller than the experimental error. It would seem therefore that the effects of basic slag on the moisture content of the soil which are described in the preceding section are not due to any improvement in the mechanical structiu-e of the soil that can be disclosed by the ordinary methods of mechanical analysis. This result must not of course be taken as indicating that the apphca- tion of basic phosphates to heavy clay pastures, and the consequent development of the clover, does not affect the mechanical condition of the soil, but simply that under the conditions of this particular experiment the differences in the behaviour of Plots 16 and 17 at Horndon with regard to moisture cannot be attributed to changes in the mechanical condition of the soil on the respective plots. Table XXXIX. Mechanical Analysis of the Soil at Horndon TO A DEPTH OF 3 INCHES Plot 17 Plot 16 PlotC Fraction Basic slag % Untreated % Lime % ■ Fine gravel) Coarse sand j 203 1-82 2-27 Fine sand ... 10-30 10-98 13-33 Coarse silt ... 19-63 18-61 22-16 Fine silt 18-73 18-62 17-38 Clay 27-55 27-74 25-25 Loss in solution ,, on ignition J *"' Undetermined Undetermined Undetermined THE EFFECT OF PHOSPHATES ON THE ACCUMU- LATION OF NITROGEN IN GRASS-LAND Field experiments at Rothamsted, and later at numerous other places, have conclusively demonstrated that leguminous plants leave the soil richer in nitrogen in spite of the fact that they are highly nitrogenous themselves. Collins (4) has shown that the appHcation of phosphates, whether in the form of superphosphate or basic slag, on 76 EFFECT OF PHOSPHATES Tree Field, Cockle Park, has resulted in a considerable increase in the nitrogen content of the soil. In 1908 the percentage of nitrogen in the soil receiving phosphates was -236 %, whereas the untreated soil contained only -185 %. From analyses of the Tree Field soils in 1919 the results shown in Table XL were obtained. Table XL. Percentage op Nitrogen in the Soil OF Tree Field, Cockle Park Plot Nitrogen % 4 6 8 Basic slag Untreated Super and lime till 1905. Basic slag and lime 1905-1919 •249 •172 •228 Analyses of the soil at the various Essex experimental centres after intervals of four, three and two years show that the gain in nitrogen on the plots receiving phosphates is considerable. The results are given in Table XLI. Table XLI. Percentage of Nitrogen in First 9 inches of Soil BoTTLDER Clay London Clay Chalk Martin's Hearne 3 years Farnham 3 years Butterfields, Latchingdon 4 years Homdon 2 years Wendens 4 years Phosphate plots Untreated plot •338 •299 •231 •208 •260 •248 •234* •210t •244 •219 * Average of 8 plots. ■ f Average of 2 plots. At Horndon samples were withdrawn from several of the plots in order to ascertain whether there was any difference in the influence of the various phosphates on the collection of nitrogen by the nodule organisms. The figures are set out in Table XLII. The samphng errors are probably considerable, and it would be unfair to argue too much from the comparison of one plot with another. If, however, the various rock phosphate plots are grouped together, the two basic slag plots, and the two untreated plots, more reUable data are obtained, as appears in the lower part of Table XLII. ON ACCUMULATION OF NITRATES 77 The figures suffice to demonstrate that there is no difference between rock phosphates and basic slag so far as their effect on the collection of nitrogen is concerned, a conclusion which is borne out by the botanical analysis of the 1920 hay crop (Table XXXI, Fig. 11) and also by the botanical examination of the pasture in 1919 (Table XV, Plate VI). Table XLII. Percentage of Nitrogen IN Various Plots at Horndon Plot % nitrogen 1 Florida pebble phosphate ! -240 3 Algerian phosphate •239 5 Basic slag •244 6 Untreated •212 7^ 8) Gafsa rock phosphate .. •227 11 Egyptian phosphate •247 13 Superphosphate •226 15 Superphosphate and lim( i ^222 16 Untreated •208 17 Basic slag •226 erage of Rock phosphate plots .. •238 „ two basic slag „ •235 „ untreated „ •210 THE RELATION OF PHOSPHATES TO THE ACCUMU- LATION OF NITRATES IN GRASS-LAND It has been suggested by Middleton(i5) that the secondary action of basic slag on pastures is due to the nitrification of the nitrogen accumulated by the nodule organisms, and that this resulting nitrate nitrogen is responsible for the vigorous growth of grasses which follows after the clover has been stimulated. In order to ascertain what effect the application of phosphates had on the production of nitrates in grass-land it was decided to make as far as possible weekly determinations of the nitrate nitrogen con- tent of the soil on the basic slag and untreated plots at three of the experimental centres. The centres selected were Martin's Hearne, where the soil has a hme requirement of -27 %, Lambourne End, a more sour type of soil with a hme requirement of -45 %, and Horndon- on-the-HiU, a 'sweet' soil containing a small reserve of calcium carbonate. 78 EFFECT OF PHOSPHATES It was realised, however, that the experiment was complicated by the fact that the growing crops would probably remove nitrate as rapidly as it was formed, and that the much heavier crops on the slag plots would make a bigger demand on the nitrate supplies than would the crop on the untreated plots. To overcome this difficulty a supply of soil, taken to a depth of 9 inches, from the slag and untreated plots at Martin's Hearne and Horndon-on-the-Hill was removed, transferred to Chelmsford, broken up, all the green growth removed, and then firmly packed into 10-inch glazed pots provided with a suitable drainage outlet. Particular care was taken to consohdate the soil in the pots by placing weights on the surface for some time and ultimately by tramping. The pots were placed in an open space under atmospheric conditions, and the nitrate nitrogen in the soil determined every fortnight. Samples for analysis were removed both from the field and from the pots by means of a small soil sampler similar to a cheese sampler. Estimation of the Nitrate Nitrogen. As soon as possible after the samples were taken — ^generally 4-5 hours in the case of the field soils and 30 minutes in the case of the pots — they were dried at a temperature of about 50° C. When dried the samples were finely ground and bottled. The nitrate nitrogen was estimated in the dry ground sample within four days of the sample being removed from the field. Method. From 50-100 gns. of soil were placed on a Biichner funnel and washed with several portions of distilled water until about 400- 500 c.c. of filtrate were collected. The filtrate was transferred to a conical flask and then rapidly concentrated to a very small bulk — about 50-60 c.c. with 10 c.c. of normal caustic soda. The concentrated liquid so obtained was diluted with distUled water and boiled again for 10 minutes. The flask was cooled and 1 gram of finely powdered Devarda's alloy added. The contents were distilled into 10 c.c. of N/50 sulphuric acid, a specially prepared trap being used to prevent spitting. The burner was adjusted so that distillation proceeded slowly for about 5-10 minutes, at the end of which period distilla- tion was quickened so that about half the liquid passed over in 30 minutes. Methyl red (-05 % solution in alcohol) was used as an indicator. The accumulation of Nitrate in the Pots. The soils representing Plots 2 and 3 at Martin's Hearne were potted on March 29th, and 16 days later the first nitrate determinations were made. The soils from Plot 16 (untreated) and Plot 17 (basic slag) at Horndon were ON ACCUMULATION OF NITRATES 79 not removed and potted till April 19th, the first nitrate determination being made sixteen days afterwards on May 5th. The nitrate contents of the potted soils as determined at various dates throughout the season are given in Table XLIII. Table XLIII. Accumulation op Nitrate in the Potted Soils PROM Martin's Hearne and Horndon-on-the-Hill Date 1920 Martin's Hearne Boulder clay Basic Un- treated HORDON-ON-THE- HlLL London clay , Basic Un- treated Rain- fall in inches Mean max. temp. °C. Remarks Nitrate nitrogen parts per million of dry soil March 29 0-96 1-12 — — — — April 14 6-16 3-92 — — 1-89 55-3 Drains had run 19 — — 2-24 1-12 — — 21 5-60 3-50 — 1-00 54-4 Drains had run May 5 616 3-04 504 1-68 0-68 66-0 '9 99 19 9-52 6-16 5-60 1-68 0-24 61-6 June 5 12-88 5-04 9-52 5-04 2-10 69-4 Drains had run 17 17-04 5-60 12-88 6-72 0-22 64-3 — July 2 22-96 7-28 22-40 7-00 1-04 68-8 — 14 25-76 7-84 17-36 6-84 1-04 63-3 — 29 22-40 7-54 20-14 504 1-55 69-6 Drains had run Aug. 11 14-56 8-40 16-24 5-60 1-37 67-7 99 99 25 20-16 18-48 25-20 4-48 0-98 71-6 Sept. 9 19-36 20-72 28-00 4-48 0-21 65-6 , 15 23-52 25-56 — — 0-13 69-6 27 26-32 10-64 14-56 7-84 3-19 68-8 Drains had run The general trend of the figures for both centres is very similar. There is a rapid accumulation of nitrate on the slag plots during the period May 19th to July 2nd. The results are illustrated in Pigs. 17 and 18. The curves indicate a much greater and a much more rapid accumulation of nitrate in the soil from the slag plots than in the soil from the corresponding untreated plots. This can only be due to one or both of two causes: 1. The nitrification of the nitrogenous matter accumulated in the slag plots by the nodule organisms. 2. The direct effect of the slag on the soil organisms which bring about nitrification. It is difficult to understand how the addition of a comparatively small amount of nitrogenous organic matter to the already big accu- mulation in the experimental soils can be responsible for such a difference. Particularly is this the case at Martin's Hearne where 80 EFFECT OF PHOSPHATES the untreated soil contain -299 % of nitrogen and 11-80 % of organic matter (loss on ignition). It seems far more probable, therefore, that the result is mainly due to the direct effect of the phosphates on the soil organisms bringing about nitrification. 29 March J2 2e IQ S4- 7 21 S /9 Z 16 30 /3 27 April May June July August Sept. Fig. 17. Nitrate content of the Potted Soil from the Untreated Plot and the Basic Slag Plot, at Martki's Hearne. 8oil Boulder clay. Season 1920. Soil from Untreated Plot (3) Soil from Basic Slag Plot (2) . RainfaU za 1 cS u 1 y 24 - .«^ f \ s /\ / \ 20 as 60 / ^ / \ \ \ /€ 2 - +3 / V \ ^ 1 •s J 12 • a -1 - « -^x^_.^^^^ ^^ 4 •^ t y^ J — 1 — 1 I — 1 , 1 „ , 1 1 1 1 -.1. .._ r 1 19 April n May /4- June 12 Z6 July 9 23 August 20 27 Sept. Pro. 18. Nitrate content of the Potted Soil from the Untreated and the Basic Slag Plots at Great Mulgraves, Horndon-on-the-HiU. Soil London clay. Season 1920. Soil from Untreated Plot Soil from Basic Slag Plot . The difference between the slag and untreated plots is striking, and seems to indicate, in view of the fact that the Martin's Hearne soil is sour, that on both these tjrpes of soil a deficiency in phosphates is a more important factor in limiting nitrification than a deficiency in lime. ON ACCUMULATION OF NITRATES 81 There is a slightly greater accumulation of nitrate in the pot representing the untreated soil at Martin's Hearne than in the corre- sponding pot for Horndon-on-the-Hill, which may be due to the fact that the former is a more open soil. On the other hand, it may be due to the fact that, although the soil at Martin's Hearne is sour, it has a considerably higher content of total and available phosphoric acid than the soil at Horndon. The figures are as follows : Martin's Heabnb Horndon-on-the-Hill 0/ 0/ Total P2O5 -089 -078 AvailaMe P2O5 -0046 -0030 Lime requirement ... 0-27 « 0-00 At no period throughout the season does the nitrate content of the untreated soil from Horndon ever approach that of the soil receiving basic slag. The two pots representing the treated and un- treated soils from Martin's Hearne behave somewhat differently. Until August 11th the figures are comparable with those representing the Horndon pots, but during the hot spell which succeeded, there is a rapid accumulation of nitrate in both the Martin's Hearne pots, and when sampled on August 25th and September 9th and 15th, the nitrate content of the slag and untreated pots was approxi- mately the same. The temperature during the period August 11th to September 15th was higher than at any other period during the season, and although the pot drains did not run, there was a sufficient precipitation to keep the soil moist. On August 18th-19th -88 inch of rain fell, and there was a fall of -10 inch on two consecutive days out of the remaining 13 days in August. Four out of the first five days in September were showery with a total precipitation of -21 inch. There was no further rain until September 14th, when -13 inch fell. After the 15th September (which was the last date on which the nitrate content of the two pots was similar) until the 22nd the weather was wet, 3-19 inches of rain falling between the 15th and 22nd inclusive. The drains from both pots ran freely, but unfortunately the drainage water was not collected. From the 22nd to the 27th the weather was dry and hot, no rain falling, and on the 27th when the pots were sampled the slag pot contained 26' 32 parts per miUion of nitrate and the untreated pot 10-64 parts. The relative nitrate content of the two pots was therefore similar to what it had been up to August 11th. It is difficult to account for the curious results obtained during the period August 11th to September 15th. It may be that under R.B.S. 6 82 EFFECT OF PHOSPHATES the cKmatic conditions then prevailing the untreated soil at Martin's Hearne is capable of yielding sufficient phosphate to enable nitrifica- tion to take place at a much more rapid rate than at any other time during the season. 8r3z Fig. 19. Nitrate content of the soil on the Untreated and the Basic Slag Plots at Great Mulgraves, Horndon. Soil London clay. Season 1920. Untreated Plot Basic Slag Plot . Moisture Content 2-8 ?•«•- 2-4- Z-2 - IS 1-6 I '4 /■2 1-0 •8 •6 •4 ■2 s. ft A ; \ >-*"t •y».«' I ♦«♦ ■ ^•>^ ■ -—I" fS S 19 3 17 3i /A za /£ Z6 3 23 March April May June July August 6 20 Sept Fig. 20. Rainfall at Great Mulgraves, Horndon. Season 1920. If, as is postulated here, it is correct, in view of these results, to assume that the main eJffect of phosphates on the production of nitrate in soils well stored with nitrogenous organic matter is due to their action on the nitrifying organisms, it is possible to explain the large increase in the hay crop obtained on the treated plots at ON ACCUMULATION OF NITRATES 83 Lambourne End in 1919. The various phosphates were not sown until January, 1919, and although clover was absent from aU the plots throughout the season, the treated plots gave almost twice the yield of the untreated. The result was not due to any stimulation _g 14 a /2 Z & 4-, 2Z S 19 S 17 31 J-^ 28 12 Ze S 23 ZO March April May June July Attguit Sept. Fig. 21. Nitrate content of the soil on the Untreated and Basic Slag Plots at Martin's Hearne. Soil Boulder clay. Season 1920. Untreated Plot . Basic Slag Plot . Moisture Content 2 2 8-; -36 • .•• . .• .* • -32 . ,.** *V *• • * .• *• : 8 •* -28 u •• • «•• < • or A ..••"-. -2* S *'. ; ••..••. •• • • «4-l A •••**.• '• •• •* o -20 a A ' \/ A V ''-'' \J 1 \ • / \ * o '"I •c A /a\ ^\ /A II 1 1 1 1 1 1 1 1 f /4 r32 J^"'»- *• ." c V \ /2 -28 8 / • o. • ; '• J! -,••*• /O -24^ *. r \ : \ ?.. \ : '% : *. .* '•• •V **.. 8 -20^ i • : / '••••• 6 4- 2 ■q = A. t t 1 A 'A 1 i 1 J 1 1 1 1 1 1 1 22 Match S 13 Z 17 2i iA 28 /2 26 9 Z3 6 20 April May June July Augasb Sept. Fig. 22. Nitrate content of the soil on the Untreated and Basic Slag Plots at Butcher's Farm, Lambourne End. Season 1920. Soil London Clay. Untreated Plot . Basic Slag . Moisture Content of the clover, and as the crop was composed entirely of grass and weeds, it seems probable that the much heavier crop on the treated plots was mainly due to the direct effect of the phosphates in stimu- lating the production of nitrates. 6—^ 84 EFFECT OF PHOSPHATES Field Results. The nitrate figures for the samples of soil taken weekly from the slag and untreated plots at Horndon, Martin's Hearne, and Lambourne End are given in Table XLIVand the results are shown graphically in Figs. 19, 20, 21 and 22. Table XLIV. Nitrate Content of the Soils on the Basic Slag AND Untreated Plots during 1920 Date Horndon Plot 16 Untreated Plot 17 Basic slag Lambourne End Plots Untreated Plot 2 Basic slag Martin's Hearne Plots Untreated Plot 2 Basic slag Parts per million of nitrogen as nitrate Mar. 15 — — 5-04 2-8 2-24 2-80 22 2-8 112 2-24 1-68 1-12 1-68 29 •56 2-80 2-24 2-8 1-12 •96 Apr. 12 — — 2-24 1-12 1-68 •56 19 112 2-24 5-6 1-68 1-68 1-68 26 2-24 1-68 1-68 1-68 2-24 1-68 May 5 4-4 5-6 2-8 3-36 2-80 8-40 10 S-36 1-68 SS6 4-48 5-04 9-52 17 1-12 2-24 — — 1-68 2-24 25 2-24 S-36 112 112 112 2-80 SI 2-24 2-8 112 2-24 112 1-12 June 8 4-48 12S2 3-92 616 — — 14 2-24 5-60 2-24 2-24 2-80 2-80 21 6-7 S-64 2-8 3-36 4-48 5-04 29 SS6 112 1-68 2-8 1-68 2-80 Jtdy 5 — — 2-8 4-48 — — 8 2-24 2-24 — — — — 12 1-68 3-36 — — — — IS — 1-68 6-72 5-60 8-40 19 — — 3-36 1-68 1-68 5-04 20 SS6 S-92 — — — — 26 S-92 2-24 1-68 2-24 3-92 1-68 Aug. 4 2-24 5-60 2-80 504 2-24 4-48 9 2-8 7-28 1-68 1-68 S-S6 5-04 16 2-8 5-6 — — 2-8 2-24 21 5-6 S-36 — — — — 2S — — S-92 5-92 — — 24 — — — — 3-36 1-68 SI S-92 10-64 S-92 6-16 5-6 5^04 Sept. 7 7-28 112 1-12 2-80 6-72 4-48 20 2-80 1-68 1-12 1-68 0-56 112 The curves at all three centres have a general similarity of appear- ance, and they demonstrate that at certain periods during the season there is a much greater accumulation of nitrate in the slag plots than in the untreated. Even on the very sour soil at Lambourne End nitrification seems ON ACCUMULATION OF NITRATES 85 to be much more active in the slag plots than in the untreated. There is a distinctly greater accumulation of nitrate during the periods May 5th to 10th, May 31st to June 8th, July 5th to 17th, August 4th and August 23rd to September 7th. Reference to Fig. 19 shows that these periods roughly correspond to those periods at Horndon during which there is a much greater accumulation of nitrate in the slag plots, and with two exceptions, viz., May 31st to June 8th, and August 23rd to September 7th, these dates hold good for Martin's Hearne. On June 8th samples were not taken from Martin's Hearne, so this exception is easily accounted for. The figures for this centre for August 23rd to September 7th are curious. There is an accumula- tion on both the treated and untreated plots, but it is greater on the untreated than on the slag plot. Reference to Fig. 17 shows that the same result was obtained from the pots, and that at this period, and only at this period, did the nitrate in the untreated pot accumu- late to an extent at all comparable with the slag pot. This result in the field would seem to lend some weight to the suggestion that at this period of the season the soil on the untreated plot has been able to furnish sufiicient phosphoric acid to meet the requirements of those organisms engaged in the production of nitrates. These results are not in accordance with the conclusions come to by RusseU(25). As the result of his work on the "Nitrate Content of Arable Soils," he says: "that only in one year (1911) was there any evidence of the organisms responsible for nitrification being retarded by a deficiency of phosphates and potash." It must be noted, however, that the soil even on the untreated plot of Broadbaulk Field contains considerably more phosphoric acid (-114 %) than the soils at Horndon, Martin's Hearne, or Lambourne End. It is of interest to note that at Martin's Hearne and Lambourne End periods of high nitrate accumulation coincide as a rule with high moisture content of the soil, whilst at Horndon they coincide with periods of low moisture content. At Martin's Hearne and at Lambourne End the periods of high nitrate accumulation on the slag and untreated plots occur as a rule at the same time (Figs. 21 and 22). At Horndon, on the other hand, periods of high nitrate accumula- tion on the untreated plot follow, about a week later, similar periods on the slag plot. This difference might possibly be due to some influence the crop may have on nitrate production, but RusseU(25), when investigating this subject, was unable to secure any definite data supporting such a contention. 86 INFLUENCE OF PHOSPHATES It may be that an inadequate supply of available phosphoric acid on the untreated plot prevented the crop from utilising the accumu- lated nitrate when suitable conditions occur, and that the subsequent depressions in the nitrate content are due to rain washing the nitrate down to below the 9 inch level. THE INFLUENCE OF PHOSPHATES ON SOIL BACTERIA At the suggestion of Dr RusseU an attempt was made to ascertain what effect, if any, the apphcation of phosphates has had on the soil bacteria. PreUminary counts were made at Rothamsted during the autumn of 1919, but the results were contradictory. Phosphates are essential for the development of all types of bacteria. Fred and Hart (9) in an investigation on the comparative effect of phosphates and sulphates on soil bacteria show that phos- phates increase the number of soil bacteria, and they suggest that increased crop production of a soil resulting from the application of soluble phosphates is in part due to the promotion of bacterial activity. The work of Hoffman and Hammer (lO) demonstrates that phosphates greatly increase the amount of nitrogen fixed by Azoto- bacter and they came to the conclusion that for this purpose di- and tri-calcium phosphates are more effective than mono-calcium phos- phates. The soils at Martin's Hearne and Horndon are very deficient in phosphoric acid and if any positive results were to be obtained it seemed probable that it would be at these two centres. Samples for bacteriological examination were taken every month from March to August. The samples were secured by means of a small soil sampler which removed a 9 inch core of about | inch diameter. The sample from each plot consisted of four cores. Before use, the soil sampler was sterilised by means of a methylated spirit lamp, and the samples when drawn were placed in previously sterihsed bottles. The total counts for the treated and untreated plots at Martin's Hearne and Horndon are given in Table XLV. There is very Kttle difference between the bacterial counts repre- senting the two plots at Martin's Hearne, and apparently phosphate has had little effect in this direction. The two plots at Horndon show decided differences. During the months of May, June and July there are twice as many bacteria in ON SOIL BACTERIA 87 the slag plot as on the untreated, but in April and August the position of the two plots in this respect is reversed. Whenever possible during the season counts of the Azotobacter and nitrate organisms were made. The results are set out in Table XL VI. Table XLV. Bacterial Counts (in thousands) at Martin's Hearne and Horndon. (Agar-Albumose media) Mabtin's Heaene Horndon Date Plots Untreated Plot 2 Basic slag Plot 16 Untreated Plot 17 Basic slag 1920 March 6 29 April 26 May 25 June 25 July 20 Aug. 17 2935 575 1540 6869 6206 5347 7070 4632 1234 1638 3810 6039 4500 7211 5430 8235 5349 4120 8609 2783 19550 9303 8568 4000 Table XLVI. Counts op Azotobacter and Nitrate Organisms IN the Soil at Martin's Hearne and Horndon. (Thousands per grm.) Azotobacter Nitrate Organisms Date Martin's Hearne Horndon Martin's Hearne Horndon Plots Untreated Plot 2 Basic slag Plot 16 Untreated Plot 17 Basic slag Plots Untreated Plot 2 Basic slag Plot 16 Untreated Plot 17 Basic slag 1920 May 26 June 21 July 19 Aug. 17 4722 2154 1330 3302 4837 1735 1771 4685 1800 1082 1031 7828 3171 4983 149 298 483 1600 2680 2903 6401 8411 There are no important differences between the numbers of Azoto- bacter present on the treated and untreated plots at Martin's Hearne, but at Horndon these soil organisms have been considerably developed by the apphcation of phosphates. The numbers of nitrate producing organisms have been greatly increased on the basic slag plots at Horndon, a result which is in keeping with the much greater amounts of nitrate found in this soil throughout the season. (Tables XLIII and XLIV and Figs. 18 and 19.) 88 PHOSPHATES AND BACTERIA It is difficult to explain why at one centre the numbers of bacteria should show marked increases as the result of the application of basic slag, whilst at another centre, where the effect of the basic slag on the crop is quite as marked, the bacterial content does not appear to have been appreciably affected by the apphcation of basic slag. It may be that the lack of effect at Martin's Heame is due to the sourness of the soil, but counts made on the portion of the basic slag plot which had received a dressing of ground hme in April 1920 showed no appreciable advantage in this respect over that portion of the plot which had not been so dressed. FACTOHS LIMITING THE YIELD OF HAY AND THE ACTION OF PHOSPHATES ON HEAVY CLAY SOILS THE EFFECT OF RAINFALL ON THE YIELD OF HAY FROM THE UNTREATED PLOTS I HE weight of the hay crop on the untreated plot at each of the experimental centres varies within very wide limits from year to year. When a dry season is experienced the crop is often a failure, whilst the same plot given a moister and more favourable season may reach the comparatively high production level of two tons to the acre. In Table XL VII the yield of hay on the untreated plots for the years 1916-20 is compared with the rainfall for the period May 1st till harvest at the corresponding rainfall stations. The results for four of the centres are shown graphically in Figs. 23 and 24. Table XL VII. Comparison of the Weights op Hay on the Untreated Plots and the Rainfall from May 1st till Harvest at the various Experimental Centres EXPEKIMENTAIi Centre 1916 1917 1918 1919 1920 Tysea Hill Hay cwts. per acre Rainfall in inches 31-6 5-94 20-4 5-36* 17-7 4-47 11-6 2-87 38-3 9-34 Martin's Hearne Hay cwts. per acre Rainfall in inches — 14-3 6-27 23-4 11-51 10-4 2-85 22-0 8-37 Lambourne End Hay cwts. per acre Rainfall in inches — — 13-2 308 21-4 6-27 Hassobury Hay cwts. per acre Rainfall in inches 11-1 4-82 23-4 7-73 10-9 0-58 Wendens, Saffron Walden Hay cwts. per acre Rainfall in inches 51-2 4-00 25-4 4-00 1 33-4 2-44 14-3 0-53 250 2-42$ Butterfields, Latchingdon Hay cwts. per acre Rainfall in inches § 31-4 3-41 14-5 2-32 20-1 2-51 20-6 1-47 161 2-28 * -71 inch of rain fell three days before plots were cut. t 2-32 inches of rain fell on May 20th, 1917. j -57 inch of rain fell two days previous to cutting. § Rainfall figures for period May 1st till June 30th. 90 EFFECT OF RAINFALL The figures for the centre at Wendens are perhaps the most striking, the hay crop on the untreated plot varying from 51 cwts. to 14 cwts. per acre. Similar, although not quite so marked, fluctuations occur at all the other centres. The yearly rainfall figures afford no adequate explanation of these differences. At Wendens, for example, the rain- fall for the year was 27-33 inches in 1917, and in 1918 when a bigger yield of hay was obtained, 25-68 inches. The distribution of the rainfall, however, seems to be of great importance, particularly during 2 4 6 8 Inches of Rainfall, May.l till Harvest Fig. 23. Influence of Rainfall on the Yield of Hay on the Untreated Plots at Tysea Hill # and Martin's Hearne O- the months of May and June. If the rainfall figures for the period May 1st till harvest are tabulated with the yield of hay on the un- treated plot, as is the case in Table XL VII, it wiU be seen that, with one or two exceptions which can be readily accounted for, there is a very close connection between the two sets of figiu-es. The rainfall figures have been taken from the records of the British Rainfall Organisation. Their station at Havering-atte-Bower is within two miles of each of the first three experimental centres. At Hasso- bury there is a rainfall station on the farm, whilst at Saffron Walden ON THE YIELD OF HAY 91 and Latchingdon the respective rainfall stations are within from two to four miles of the experimental centres. The curve representing the correlation of the hay yield with the rainfall at Saffron Walden is a very steep one, and shows quite distinctly that rainfall is the most important factor. At Tysea Hill and Martin's Hearne even a high rainfall from the 1st of May till harvest of 9-34 inches and 11*51 inches respectively 10 12 4 6 8 Inches of Bainfall, May 1 till Harvest Fig. 24. Influence of Rainfall on the Yield of Hay on the Untreated Plots at " ~ -©, and Cockle Park ■ Wendens Butterfields produce crops of only 38-3 cwts. and 23-4 cwts. The manurial factor is clearly of greater importance at these two centres and particularly at Martin's Hearne. The curves are not nearly so steep as at Saffron Walden, and there is a much greater response to manuring. At Latchingdon the rainfall has more influence on the crop than at Martin's Hearne or Tysea Hill and less than at Saffron Walden. It was difficult to get any correlation between the rainfall and the 92 EFFECT OF RAINFALL yield of hay on the untreated plot at this centre. Owing to a wet July in 1918 and 1920 the crops had to remain uncut during a spell of wet weather, which had little or no influence on the growth of the crop. In order to overcome this difficulty the rainfall figures from 1st May to June 30th have been taken for each year. One point in the curve falls far out of fine, namely that representing the rainfall and the hay yield for 1919. The reason for this divergence, however, is clear. After an exceptionally dry May and June, the first week of July was wet, -47 inch of rain falUng on the 2nd of the month and '29 inch on the 3rd. The crop was not cut till three weeks later, and during that time considerable growth was made, particu- larly by the clover plants. If the rain faUing on the first four days of July is taken into consideration the divergence of this particular point is rectified. In Fig. 24 the effect of rainfall on the yield of hay from the un- treated plot at Tree Field, Cockle Park, is shown, and the curve affords an interesting contrast to those representing the Essex centres. At Cockle Park rainfaU does not influence the yield of hay on the untreated plot, whilst in Essex rainfall at certain centres is the most important hmiting factor, and at all centres it has a great influence on the yield of hay. THE EFFECT OF RAINFALL ON THE YIELD OF HAY FROM THE PLOTS RECEIVING PHOSPHATES The field experiments recorded show that at all these centres the appHcation of phosphates results in a considerable increase in crop. The increase is least at Saffron Walden and greatesti at Horndon. The results at Tysea HiU indicate, however, that even poor as this soil is in phosphoric acid, the heavy dressing of 200 lbs. of phosphoric acid per acre is more than the soil requires over a period of five years, as equally good results accrue from the fighter dressing of 100 lbs. In Figs. 25 and 26 the increase resulting from the appHcation of phos- phates at each centre is correlated with the rainfall. It wiU be seen that at Latchingdon and Saffron Walden the increase in the hay crop on the phosphate plots steadily progresses with the rainfall, clearly demonstrating that rainfall is the controUing factor, and that with the limited rainfall at these centres little or no increase may be expected from other than phosphatic manures. At Tysea HiU and Martin's Hearne on the other hand the increase in the yield of hay ON THE YIELD OF HAY 93 due to phosphate varies within extremely narrow limits, and is not dependent upon the rainfall. The curve for Tysea Hill is a perfect limiting factor curve and indicates that some factor other than the rainfall and phosphates is limiting the yield of hay. The absence of any increase in yield due to phosphates in 1920 is curious. The season was a particularly favourable one, and owing to the rainy weather in July and the beginning of August the crop was not cut until August 23rd. As will be shown later, at least haK of the original dressing of phosphoric acid apphed in 1915 was stiU present in the 16 14 12 10 8 6 12 3 4 5 Inches of Rainfall, May 1 till Harvest Fig. 25. Influence of Rainfall on the Increase due to Phosphates at Wendens- # and Butterfields O. soil in an available form in October, 1919, so that the neghgible increase of the treated plots over the untreated in 1920 caimot be due to a deficiency in phosphates. It can only be concluded, therefore, that with a high rainfall (9-34 inches from May 1st till harvest) and a long growing period, no increase will be obtained from phosphates, unless the second hmiting manurial factor is first satisfied. The curve for Martin's Hearne closely resembles that at Tysea HiU, and as the two fields are on the same soil formation, have practically identical chemical and mechanical compositions, and are only a short distance apart ; the result at Martin's Hearne satisfactorily confirms the conclusion that a second hmiting factor comes into operation as s y 9 -1 o J^ o 1 mt^ ^ -s ^ ^ • 1 ^ / O c / / / / -/ / / / ^ 1 1 1 1 J 94 EFFECT OF RAINFALL soon as the need for phosphates is satisfied. The soil at Martin's Hearne is appreciably poorer in total and available phosphoric acid than at Tysea Hill, and even with a rainf aU of 11 inches from May 1st i__J L_J L CD S I ' "d (u ^■. ^ eg (B s w o m :^ O -TS 1—1 02 in CM o CO CO till hay harvest and a long growing period (cut August 10th), the soil cannot provide an adequate supply of phosphates. It is too soon to draw definite conclusions from the results at Lamboume End and Horndon-on-the-Hill. There are indications OTHER MANURIAL FACTORS 95 that once the need for phosphates has been satisfied at Lambourne End the hay crop could be increased by the addition of some other essential plant food. It seems very probable, however, that at Horndon-on-the-Hill, after the need for phosphates has been satisfied, rainfall is the Umiting factor as far as the hay crop is concerned. THE SECOND LIMITING MANURIAL FACTOR It has been previously stated that the experiments were started with the object of ascertaining the relative manurial value of various types of insoluble phosphates. No attempt was therefore made to include dressings of potassic and nitrogenous manures. Out of the eight experimental centres dealt with here there are two — Hassobury and Farnham — at which the response to phosphates, measured by the hay crop, is negHgible. A very marked improvement in the quality of the meadow has resulted at Farnham, as has already been pointed out (Table XIII), but even in a favourable season the increased weight of hay resulting from the apphcation of the various phosphates has been very small indeed. Moreover, the productive level of this type of soil is exceedingly low, and the same remark apphes to Hassobury, where also the crop seldom passes the 10 cwts. per acre level. An examination of the analytical data presented in Table VII shows that the Hassobury soil is reasonably well supphed with phos- phoric acid, and has, in fact, practically twice as much available phosphoric acid as any of the other experimental soils. The percentage of available phosphoric acid is well above Dyer's limit of 0-01 %(7). This is not the case at Farnham, where the soil is markedly deficient in available phosphoric acid, and it seems reasonable to conclude that a deficiency of another essential plant food is the cause of the low productivity at Hassobury, and is preventing a response to the dressings of phosphates applied at Farnham. Both soils are weU suppHed with nitrogen, and as readily available nitrogen has been slowly accumulating at Farnham without having any appreciable effect on the hay yield, it does not seem that a lack of nitrogen is responsible for the poor crop returns. The Hassobury soil, though devoid of calcium carbonate and possessing a Hutchinson and MacLennan hme requirement of -13 %, gives no response to the heavy dressing of lime appHed to Plot 15. Moreover, the soil at Farnham has an adequate supply of calcium carbonate. The low 96 OTHER MANURIAL FACTORS level of production at these two centres is clearly not due to soil sourness or lack of lime. Table XL VIII. Chemical Analysis of the Soils AT Hassobury and Farnham Hassobury Farnham Nitrogen Loss on ignition Calcium carbonate ... Total phosphoric acid Available „ „ Total potash Available potash ... Lime requirement . . . 0/ /o 0-180 % 0-208 7-23 8-23 0-00 0-45 0-190 0-118 0-0123 0-0056 0-435 0-644 00194 00165 0-13 0-00 The Hassobury soil has a lower content of 'total potash' than the other centres, but the Farnham soil is better off in this respect than several of the other centres. At both stations, however, the 'available potash' is markedly lower than in any of the other clay soils, and although in both cases the figure is distinctly above Dyer's{7) limit, it seems difficult to come to any other conclusion than that a soil deficiency in available potash is responsible for the poor yields of hay at these two centres. At two of the remaining centres, namely Tysea Hill and Martin's Hearne, the curves in Fig. 26 show the operation of a second hmiting factor which comes into play after the need for phosphates has been satisfied. The soil at both these centres is very similar in composition. It is wen suppMed with organic matter and nitrogen, and as this store has been considerably added to by the accumulated residues from clover plants, it does not seem probable that there is any deficiency in nitrogen. Both soils are sour. They contain no calcium carbonate and have a high hme requirement. Nevertheless the production of nitrates in this soil compares very favourably with that at other centres better suppHed with calcium carbonate and where the soil is sweet. (Compare Figs. 21 and 19.) At Martin's Hearne and Tysea Hill the plots were cross dressed with hme at the rate of 35 cwts. per acre during the early part of 1920, and at Martin's Hearne another plot was marked off and received a dressing of approximately 10 tons of farmyard manure to the acre. The results are recorded in Table XLIX. RESPONSE TO FARMYARD MANURE 97 There is a small gain due to lime at Martin's Hearne, but considering that the 1920 season was a particularly favourable one for the hay crop the result suggests that very httle can be expected from Hme imtil some other fertihsing constituent is supphed*. The same con- clusion holds true for Tysea Hill, where the increase due to Hme is insignificant. In taking the average it is probably not fair to include Table XLIX. Effect of Cross Dressing with Lime at Martin's Hearne and Tysea Hill Martin's Heaene Tysea Hill Hay cwts. Hay cwts. Plot per acre, 1920 Plot per acre, A 1920 UnUmed Limed Unlimed Limed 1. Open hearth 1. High grade basic (fluorspar) slag 28-4 28-5 slag 40-2 40-1 2, Ditto, high soluble 31-9 36-3 2. Gafsa rock phos- phate 41-2 43.6 3. Untreated 230 25-5 3. Untreated 38-3 36-0 4. Gafsa phosphate 35-2 39-6 4. Open hearth (fluorspar) slag 46-4 49^9 5. Egyptian phosphate 29-0 31-6 5. Ditto, high soluble 45-2 47-2 6. Algerian phosphate 34-6 34-7 6. Ditto 42-1 41^6 A. Farmyard manure 7. Untreated 45-6 42^7 (applied Autumn 40-3 38-7 Half dressing of of 1919) phosphate 8. Same as 2. 48-3 48^4 9. Same as 6 44-8 430 10. Same as 4 44-8 46^3 Average 31-8 33-5 Average 43-7 43^9 Inches of rain. May 1st till harvest 8-37 9-34 Lime requirement •27% •29% the results from the untreated plots, because it may with reason be argued that no result from the apphcation of Kme could be expected until the need for phosphates was first met. If the figures for the untreated plots at Tysea Hill are excluded, the average yield becomes 44-1 cwts. per acre on the unhmed plots and 45-0 cwts. per acre on the hmed plots, giving an average increase of 0-9 cwt. in favour of Hme. The response to farmyard manure on Plot A at Martin's Hearne is significant. Assuming that 10 loads of farmyard manure are equivalent to 8 tons, and that the farmyard manure contained -4 % of phosphoric acid and -4 % of potash, this plot received in addition * The yields of hay for 1921 show equally poor returns from the use of lime. 98 RESPONSE TO FARMYARD MANURE to other materials a dressing of phosphoric acid and potash equivalent to about 72 lbs. of each per acre. This amount of phosphoric acid was presumably sufficient to meet the requirements of the 1920 season, and it is therefore interesting to note what effect the dressing of potash had^. There was no doubt throughout the whole season that the farm- yard manure plot was the best on the field. For the first time that portion of the field not within the experimental area, and which had received a similar dressing of dung, looked better and bore a better crop than the experimental plots. The average weight of hay on the phosphate plots was 31-8 cwts. per acre, and on the plot receiving a small dressing of phosphates and potash in the form of dung 40*3 cwts., leaving a gain of 8*5 cwts. per acre which can only be attributed to potash. During the whole season the plots were inspected once a week, and it was early evident that the clover was making a more vigorous growth on the farmyard manure plot. Owing to an oversight a sample of hay was not removed from this plot for botanical analysis. The aftermath was allowed to grow until the beginning of October, and not only was there a more vigorous growth, but the bottom of clover on the farmyard manure plot was closer and more regular than on any of the other plots. In view of the evidence there can be Httle doubt that on this type of soil, after the need for phosphate has been met, potash is the second hmiting manurial factor. Moreover it is very probable that in all but the exceptionally dry years a profitable return from the apphcation of potash will be secured. It should be possible by judicious apphcation of phosphates and potash to raise the produc- tion of meadow hay to the 2 tons an acre level in all but exceptionally dry years. Such results serve to confirm the conclusion that potash is the second hmiting manurial factor at Hassobury and Farnham, and they incidentally suggest that on grass-land in Essex profitable results from the apphcation of potash are hkely to accrue when the soil contains less than -03 % available potash — a figure considerably above Dyer's hmit. ^ It is very improbable in view of the particularly moist season that the organic matter or the nitrogen in the farmyard manure plot had any effect on the yield of hay. The meadow has been down to grass for at least 80 years, and a large store of organic matter and nitrogen has been accumulated. The action of Hme would pre- sumably be to release these materials for the plant, and the lack of response to the apphcation of lime suggests that the soil can normally provide all the nitrogen the crop requires in a suitable form. THE ACTION OF BASIC SLAG ON THE ACIDITY OF THE SOIL AS MEASURED BY THE ' LIME REQUIREMENT ' AND HYDROGEN ION CONCENTRATES OF THE SOIL It is a matter of common knowledge that clovers, particularly wild white clover, are very sensitive to what is vaguely called soil acidity. The extent to which the clover plant is able to persist on a 'sour' soil probably depends, however, not only upon the degree of acidity as measured by the soil lime requirements, but upon such other factors as climatic conditions and the water holding capacity of the soil. On the Harpenden Common, Hutchinson and MacLennan^ found that wild white clover persisted where the soil had a lime requirement of -22 %, and they illustrate this by the following table. Table L. Relatioist of Lime Requirements of the Soil TO THE Vegetation on Harpenden Common Average lime require- ment of soil Dominant flora Approx. 0-22 % CaCOa Wild white clover „ 0-26 Fescues „ 0-31 Mixed; yarrow, woodrush and moss „ 0-39 Gorse „ 0-43 Yorkshire fog „ 0-53 Sorrel The botanical analyses of the hay at the various Essex experi- mental centres shows that on poor heavy clay soils basic slag is able to induce a vigorous growth of clover even if the soil has as high a lime requirement as -45 %. The abiUty of the clovers to persist on such sour soils is, however, in Essex at any rate to a great extent dependent upon the distribution of the rainfaU. A comparison of Tables XXXII and XXXIII shows that during a moist growing season clovers form a large proportion of the hay crop on the basic slag plots, even when the soil has as high a lime requirement as -45 %. On a dry season, however, clovers are absent ^ Journal of Agric. Science, vn. p. 102. 7—2 100 LIME REQUIREMENT OF THE SOIL from the hay crops on all soils with a lime requirement between -13 and -45 %, but are present on meadows which contain a small reserve of calcium carbonate, and whose hme requirement is neghgible. The absence of clover (which 'fills up the bottom') during a dry season adversely affects the yield of hay, and any factors which tend to produce conditions unfavourable to the clover plant obviously limit the yield. The acidity of the soil as measured by its 'hme requirement' does therefore to some extent hmit the action of basic slag, and it becomes of importance to ascertain to what extent the apphcation of basic slag affects favourably or unfavourably the acidity of the soil. Table LI. Lime Requirement and Ph. Value of the Soils IN THE Basic Slag and Untreated Soils at the Various Experimental Centres 9 inches samples Ph. value 3 inches samples Centre liime requirement , * ^ Basic slag Untreated % % Lime requirement Basic slag Untreated , * , Basic slag Untreated % % TyseaHiU Martin's Hearne Farnham Latchingdon •30 •29 •01 •04 •29 •27 •00 •03 6^2 6-3 7-4 7-5 6-2 1-5 7-6 •35 •03 •31 •04 •13 Samples of soil to a depth of nine inches and three inches were removed from the basic slag and untreated plots at several of the experimental centres during October 1919. The hme requirements of all the soils were ascertained and in some cases the Ph. value also. The results are set out in Table LI. In every case the hme requirement figures are higher for the soil on the basic slag plots than on the untreated, and although the differences are not great, they suggest that the apphcation of even a heavy dressing of basic slag is not sufficient to counteract the acidity which develops from the decaying organic matter which accumulates on such plots. The Ph. values also show but smaU differences, and with one exception, namely, the soils from Martin's Hearne, they confirm the hme requirement figures and indicate a tendency towards greater acidity on the basic slag plots. As it seemed probable that the continued use of basic slag over a long period of years would accentuate this tendency, samples of soil were secured from Plots 4, 6 and 8 at Tree Field, Cockle Park, LIME REQUIREMENT OF THE SOIL 101 through the courtesy of Professor Gilchrist, and the lime require- ment of the soil determined with the results given in Table LII. The acidity of the soil on Plot 4 is quite appreciably greater than on the untreated plot in spite of the fact that it has received, during the twenty-four years the experiment has been in progress, a total dressing of about two tons of basic slag to the acre. The botanical analyses of the hay from Plot 4 show moreover that it is not possible to maintain a permanent bottom of clover on this plot, and were it not for the comparison with Plot 8, it might reasonably be assumed that the sourness of the soil was responsible for the partial failure of the clover plant. Plot 8, however, has received a dressing of hme Table LII. Lime Requirement op Soil Samples for Plots 4, 6 AND 8 AT Tree Field, Cockle Park Soil Samples taken 1919 Plot Tkeatmbnt (Dressing of phosphate equivalent to 100 lbs. PgOg per acre) Lime requirement CaCO, /o CaCOg content of soil % 4 6 8 5 cwts. of basic slag every three years (1897-1919). Last dressing 1918 Untreated Superphosphate + 10 cwts. ground lime every three years (1897-1905). 5cwts. basic slag+ 1 ton ground hme every three years (1905-1919). Last dressing 1918 0-23 0-20 007 0-00 0-028 0-29 every three years since 1897, and since 1905 each application has been at the rate of 1 ton per acre, the plot receiving in the form of basic slag the same amount of phosphate as has been appHed to Plot 4. In all five tons of hme have been apphed to the plot, more than four times the amoimt required to satisfy the Hme requirement of Plot 4. There is now a small reserve of calcium carbonate in the soil on Plot 8, but in spite of this fact the soil has still a small 'hme requirement' and neither the crop nor the herbage are any better than on Plot 4. It seems fair to conclude from this evidence that the continued apphcation of heavy dressings of basic slag over intervals of three years does not suffice to supply the hme requirement of heavy clay soils imder grass. On the contrary the results indicate that such soils are hable to become even more sour than similar soil left untreated. 102 LIME REQUIREMENT OF THE SOIL Although on sour clay soils basic slag fails to maintain a permanent plant of clover, yet the addition of heavy dressings of Hme fails to improve matters in this respect. At Cockle Park, a more vigorous growth of clover follows each successive dressing of slag, whilst in Essex on soils well supphed with calcium carbonate there is no difficulty in maintaining a permanent bottom of clover by the applica- tion of phosphates. (See results from Saffron Walden.) Why the Clover Plant Fails at Cockle Park The most important conditions necessary for the proper growth and development of the clover plant in conjunction with the various grasses are: 1. A suitable supply of phosphate, 2. A suitable supply of potash. 3. The presence of calcium carbonate in the soil. 4. Constant grazing to prevent the grasses shutting out the light and air, and thereby choking out the clover plant. At Cockle Park the plots have been grazed by sheep annually, so that the conditions in this respect are the most favourable possible for the permanent estabhshment of a bottom of clover. Potash in addition to basic slag on Plot 7 has not materially increased the returns, nor has it benefited the clover plant, and as has been indicated previously, no better results have attended the addition on Plot 8 of ground Ume to the standard dressing of basic slag. A comparatively heavy dressing of phosphates (equivalent to 100 lbs. P2O5 per acre) has been given to Plot 4 every three years and it would seem scarcely probable that a lack of phosphate was the cause of the wild white clover plant being unable permanently to estabHsh itself. Nevertheless, if the botanical composition of the herbage is examined over a period of years, it will be noted that following every dressing of basic slag there is a marked response by the clover plant. The results on Plot 8 apparently preclude any possibihty of the Ume in the basic slag being responsible for the improvement. By a process of ehmination one is forced to conclude that the various dressings of basic slag have never sufficed to meet the need for phosphates, and that at Cockle Park the level of produc- tion could be still further raised by increasing the dressing of phos- phoric acid or by repeating the present standard dressing at more frequent intervals. With the object of obtaining more precise information on this LIME REQUIREMENT OF THE SOIL 103 point, the total phosphoric acid in the soils from Plots 4, 6 and 8 was determined with the following results : Plot 4 1919 •088 % P2O5 •052 •076 Plot 6 untreated at the beginning of the experiment contained •071 % of phosphoric acid. Plots 4 and 8 have each received 800 lbs. of phosphoric acid during the period of the experiment, sufficient, were there no losses, to raise the soil content of phosphoric acid to •107 %. Although the soil samples were removed less than two years after the previous dressing of basic slag had been suppUed, it will 200 180 /60 100 80 60 40 20 O .g — K 4- 5 Year Tho. 27. Live Weight Gains on Basic Slag and Untreated Plots at Cockle Park. First Period, 1897-1905. Basic Slag Plot (3) . Untreated Plot (6) . be noted that the content of phosphoric acid in the soil on Plot 8 is httle better than at the beginning of the experiment, and that the reserve of phosphoric acid in the soil from Plot 4 is much less than might have been anticipated. If the suggestion that phosphoric acid is still the Hmiting factor is correct, it would be natural to expect Plot 4 to give superior results to Plot 8. This is in fact the case(i3), and the inferiority of this latter plot over Plot 5 is not due to the depressing effect of hme on the hve weight gain, but to the fact that the soil on this particular plot contains a smaller supply of phosphoric acid than on Plot 4. If the increase in hve weight gain from Plot 3* over Plot 6. at Cockle Park during the period of the experiment is plotted out as is done in * Receives 200 lbs. of phosphoric acid as Basic Slag ev^ery six years. 104 LIME REQUIREMENT OF THE SOIL Eigs. 27 and 28, it will be seen that each successive application of basic slag results in a big increase in live weight gain during the two seasons following its apphcation. Thereafter the hve weight increases rapidly dechne until a fresh dressing is appHed, clearly indicating that during the third, fourth, fifth and sixth seasons following the apphcation of the heavier dressings of basic slag bigger returns could be obtained by a further dressing of phosphates. Gilchrist (i4) and SomerviUe(29) have pointed out that far from there being a falling off in the response to basic slag at Cockle Park, the hve weight gains are gradually increasing over each six year period. The improvement is slow, but it is due to the very slow building up of the phosphoric /80r ^ f ^ • / ^v 100 - o 80- 20 7 7. 3 „ * sr 6 Year Fig. 28. Live Weight Gains on Basic Slag and Untreated Plots at Cockle Park. Second Period, 1906-1911. Untreated Plot Basic Slag Plot . acid content of the soil. Such a result serves to confirm the conclusion that phosphoric acid is still the hmiting factor at Cockle Park, and that until the demand for phosphates is satisfied it will not be possible to estabhsh a permanent plant of clover and no improvement in the condition of the clover plant or in the live weight gains can be antici- pated by either the addition of Hme or of potash. Why the Clover Fails on some Pastures in Essex DURING THE DrY SeASON If the failure to secure a permanent bottom of clover on Tree Field at Cockle Park is due to an inadequate supply of phosphates in the soil, such is not the case at the experimental centres in Essex where this difficulty has been experienced. An inspection of Fig. 26 shows quite convincingly that at Martin's Hearne and Tysea Hill phosphoric acid is no longer a hmiting factor LIME REQUIREMENT OF THE SOIL 105 on the treated plots. A chemical analysis of the treated soils, more- over, reveals the fact that even after four years one-half of the original dressing of 200 lbs. of phosphoric acid is stiU to be found in the first nine inches of soil. Table LIII gives the total and available phosphoric acid found in the soil from the basic slag and untreated plots during the autumn of 1919. By assuming that one acre of soil to the depth of 9 inches weighs 1000 tons, the actual quantity of available phosphoric acid in the two plots has been calculated, and the excess in the basic slag plot taken to represent the amount of the original dressing still left in the soil. Table LIII. Total and Available Phosphoric Acid in the Soil FROM Basic Slag and Untreated Plots, in the Autumn oe 1919 Samples taken Autumn of 1919 BUTTERBIELDS, Latchingdon. Manures sovm winter of 1915-16 Martin's Heabne. Manures sown winter of 1916-17 Tysea Hill. Manures sown winter of 1915-16 HORNDON. Manures sown Feb. 1918 Basic slag Un- treated Basic slag Un- treated Basic slag Un- treated Super, and lime, 15 Un- treated, 16 Total P2O5 ... Available PgOg % -088 -0134 0/ /o •077 •0066 % •101 -0108 0/ /o -089 •0046 /o •109 -0102 % -101 •0051 % -082 -0106 /o •078 •0030 Amoimt of PjOs added ... Amoimt found in citric acid solution lbs. 200 300-2 lbs. 147-8 lbs. 200 241-9 lbs. 1030 lbs. 200 228-5 lbs. 1142 lbs. 200 237-4 lbs. 67^2 Excess of avail- able phosphoric acid 152-4 138-9 114-3 170-2 The above table shows that from a half to three-quarters of the original dressing of 200 lbs. of phosphoric acid still remains in the soil in an available form, and such results but confirm the conclusion that lack of phosphates can not be the cause of the clover faihng at Martin's Hearne and Tysea Hill during the dry season of 1919. The appHcation of hme at Tysea HiU and Martin's Hearne at the rate of 35 cwts. per acre of ground hme is more than sufficient to satisfy the hme requirements of these soils, and it would be reasonable to expect that if soil sourness is the only limiting factor to the 106 LIME REQUIREMENT OF THE SOIL growth of clover, a pronounced improvement in this respect will follow the appKcation of such a dressing. Throughout the whole of the 1920 season the plots were examined carefully every week. At Tysea Hill clovers were practically absent from the limed and unhmed portions of the plots, and it was quite obvious that some other factor than Hme and phosphates was pre- venting the development of clovers. At Martin's Hearne there was a good bottom of clover on all the plots although it was not so good as in 1918 (see Plate IV), and no improvement in this respect was evident on those portions receiving a dressing of Ume. Table LIV. Botanical Analysis of the Hay on Limed and Unlimed Plots at Tysea Hill and Martin's Hearne Tysea Hill Per cent, of Clovers in the Hay by weight Plotl Basic slag Plot 2 Gafsa rock phosphate Plots Untreated Plot 7 Untreated Plot 10 Open hearth slag Ught dressing Unlimed portions of plot Limed portions 5-9 8-5 4-4 6-4 4-4 71 0-8 2-6 3-8 Martin's Hearne Per cent, of Clovers in the Hay by weight Plot 2 Basic slag Plots Untreated Plot 4 Gafsa phosphate Unlimed portions . . . Limed portions 27-5 18-7 11-2* 7-2* S50 20-0 * Mostly purple vetch and bird's-foot trefoil. Less than 3 % clovers. These observations were fully borne out by the botanical analysis of the hay at both centres on the unhmed and hmed portions of the various plots. The figures are given in Table LIV. Whether the appUcation of lime will enable the clovers to maintain their position at Martin's Hearne remains to be seen. If they fail in a dry season as was the case in 1919, then clearly some other essential, probably potash, is the factor limiting their growth. There can be little doubt that at Tysea HiU no further improve- LIME REQUIREMENT OF THE SOIL 107 ment in the yield or quality of the hay can be secured without the application of potash, and that neither Ume nor phosphates nor a combination of the two will suffice to maintain a permanent bottom of clover. One other result calls for explanation. The superphosphate plot at Horndon, in spite of the fact that the soil contains a small reserve of calcium carbonate, has never held the same bottom of clover as any of the basic phosphate plots (see Plate VII and Fig. 11). Samples of soil were drawn in the autumn of 1919 from this plot, and from Plot 15, which received the same dressing of superphosphate (200 lbs. P2O5 per acre), and in addition 1 ton of lime per acre. On both samples the amount of citric soluble phosphoric acid and the Ume require- ment were determined, the results being as foUows: Plot 15 Plot 13 Superphosphate Superphosphate and lime Total phosphoric acid •084 % •082 Available phosphoric acid •0046 •0106 Lime requirement •10 •05 Calcium carbonate •00 •13 The figures indicate that the inability of the clover to grow so vigorously on Plot 13 as on Plot 15 is not caused by sourness alone, but is mainly due to the phosphoric acid having been retained by the soil in a more unavailable form than is the case on Plot 15. REFERENCES (1) ARMSTRONG, S. F. The Botanical and Chemical Composition of the Herbage of Pastures and Meadows. J. Agric. Science, vol. ii, p. 283. (2) Baestbridge, R. The Effect of Fluorspar Additions on the Phosphates in Basic Slag. Iron and Steel Instit. Carnegie Schol. Memoirs. 1919. (3) BxjRiiisoN, W. L. The Availability of Mineral Phosphates for Plant Nutrition. J. Agric. Research, vol. vi. (4) Collins, S. H. Chemical Fertilisers. BalUere, Tindall, and Co. (5) Daubeny. On the Use of Spanish Phosphorite as a Manure. J. Boyal Agric. Soc. 1845. (6) DuTTON, F. V. Report on the Result of Field Experiments, 1912-14. Devon Cotmty Agric. Committee. (7) Dyer, B. Chemical Study of Phosphoric Acid and Potash in Wheat Soils of Broadbaulk Field, Rothamsted. Philosophical Trans. 1901, p. 235. Report on French Experiments. J. Royal Agric. Soc. 1896. (9) Fred, E. B. and Hart, E. B. The Comparative Effect of Phosphates and Sulphates on Soil Bacteria. Research Bulletin 35, Wisconsin. 10) Hoffman, C. and HL^mmer, B. W. Some Factors Concerned in the Fixation of Nitrogen by Azotobacter. Research Bulletin 12, Wisconsin. LI) Hopkins, C. G. Soil Fertility and Permanent Agriculture. Ginn and Co. L2) Gilchrist, D. A. Guide to Cockle Park, 1915, pp. 39-43. [3) Ibid. 1914, pp. 11-15. L4) Best Methods of Laying Down and Improving Grass-land. J. Farmer's Club. 1920. MiDDLETON, T. H. The Improvement of Poor Pastiires. J. Agric. Science, vol. I, p. 134. Oldershaw, a. W. Yield of Grass from Variovis Basic Slags. J. Board of Agric. vol. xxiv, p. 819. Paterson, J. W. Utilisation of Phosphate Deposits in AustraUa. Bull. 7, Advisory Council of Science and Industry (Australia). (18) Robertson, G. S. A Comparison of the Effect of Various Types of Basic Slag on Grass -land. Trans. Faraday Soc. vol. xvi. (19) Influence of Fluorspar on the Solubility of Basic Slag in Citric Acid. J. Soc. Chem. Ind. vol. xxxv. (20) SolubiHty of Mineral Phosphates in Citric Acid. J. Soc. Chem. Ind. vols. XXXIII and xxxv. (21) Notes on the Nature of the Phosphates Contained in Mineral Phosphates. J. Agric. Science, vol. vni. (22) Reversion of Mixtures of Superphosphate. J. Soc. Chem. Ind. vol. XXXVI. REFERENCES 109 (23) Russell, E. J. The Utilisation of Basic Slags. Trans. Faraday Soc. vol. XVI. (24) Notes on Manures, Jan. 1920. J. Ministry of Agric. vol. xxvi. (25) Nitrate Content of Arable Soils. J. Agric. Science, vol. vi. (26) Decomposition of Organic Matter in Soils. J. Agric. Science, vol. VIII. (27) SiLLARS, D. Formation of Basic Slag in the Manufacture of Steel. Trans. Faraday Soc. vol. xvi. (28) SoMEEViLLE, W. Manuring of Pastures for Meat and Milk. Ministry of Agric. Miscell. Pub. No. 30. (29) Grass. Presid. Address, Sec. M, Brit. Assoc. 1919. (30) Veeney, H. On the Spanish Phosphorite and other Manures. J. Roy. Agric. Soc. 1845. (31) Waggaman, W. H. and Wagneb, R. Agrictdtixral Availability of Raw Ground Phosphates. J. Ind. and Expt. Chem. vol. x. (32) Wood, T. B. and Berry, R. A. Soil Analysis as a Guide to Manuring. J. Agric. Science, vol. i. INDEX Agrostis alba, 32, 35, 36, 50, 51, 54, 56 Algerian phosphate, 8, 21 effect of, on accumulation of nitrogen in soil, 77; on the herbage, 35, 50, 58 ; fine grinding on availability of, 32 field trials with, 25 et seq. Alopecurus pratensis, 36 Anthoxanthum odoratum, 50, 51, 54, 56 Apatite, 11 Armstrong, S. F., 36, 37 Avena flavescens, 50, 51, 54, 57 Azotobacter, 86 Bacteria, influence of phosphates on soil, 86 Bainbridge, F., 10 Basic slag comparison of, with superphosphate, 4 effect of, on accumulation of nitrate in soU, 77 et seq. ; on accumulation of nitrogen in soil, 75; on bacterial content of soil, 86 et seq. ; on soil acidity, 99 et seq. ; on soil moisture content, 63 et seq.; on soil tem- perature, 68 et seq. ; on soil texture, 73 et seq. grades available, 7 high grade, 7 ; manufacture of, by open hearth process, 6 production and consumption of, 5 use on arable land, 4 Basic Bessemer slag early experiments with, 3 effect on botanical composition of herbage, 51, 52, 53, 57 field experiments, 23, 27, 38, 44 manufacture of, 2 Basic open hearth slag comparison with Bessemer process, 5 manufacture of, 4 Basic open hearth slag (high sol.), 21 effect on botanical composition of herbage, 28, 35, 39, 50 et seq. field experiments with, 23 et seq. Basic open hearth slag (fluorspar), 21 effect on botanical composition of herbage, 28, 35, 39, 50 et seq. field experiments with, 23 et seq. manufacture of, 6 pot experiments with, 10 Belgian phosphate, 15 Bellis perennis, 36 Bones, 1 dissolved, 2 Boulder clay, soils, trials with various phosphates on, 22 et seq. Burhson, W. L., 11 Butterfields. See Latchingdon Cambridge coprohtes, 21 effect on botanical composition of herbage, 58, 59 field experiments with, 31, 40 Carolina phosphate, 8 Centaurea nigra, 24 Chalk soils, field experiments with various phosphates on, 43 Cleveland phosphate, 21 effect of, on botanical composition of herbage, 55, 58, 59 field experiments with, 31, 40 Clover, faUure of, 102, 104 Cockle Park, 3, 4, 17, 24, 76, 100, 102 CoUins, S. H., 8, 75, 78 Coprohtes, 8, 15 Cynosurus cristatus, 36, 50, 51, 54, 56, 57 Dactylis glomerata, 36, 37 Daubeny, Prof., 14 Dundonald, Lord, 1 Button, F. v., 10 Dyer, B., 14, 96, 98 Egyptian phosphate, 8, 21 effect on accumulation of nitrogen in soil, 75 ; on botanical composition of herbage, 35, 50, 55, 58, 59; of fine grinding on availabflity of, 32 field trials with, 25 et seq. Farnham Hall accumulation of nitrogen in soil at, 76 effect of phosphates on the herbage at, 29; on the soil texture at, 76 field trials with various phosphates at, 27 limiting manurial factors at, 95 Ferrous sulphate, influence of, on hay- crop, 31 Festuca ovina, 50 Field experiments, with various phos- phates, 18 et seq.; conclusions drawn from, 45 ; apphcabihty of the results of, 48 INDEX 111 Florida pebble phosphate, 8, 16, 21 effect of, on. accumulation of nitrogen in the soil, 77; on herbage, 35; fine grinding on availability of, 33 field experiments with, 27, 30 et seq. Florida soft phosphate, 21 field experiments with, 31 Fluorspar, 6 Fred, E. B., 86 Gafsa phosphate, 8, 16, 21 effect of, on accumulation of nitrogen in soil, 77 ; on botanical composition of herbage, 28, 35, 50 et seq.; rain- fall on availability of, 44; soil acidity on availabihty of, 47 field experiments with, 32 et seq. Gilchrist, D. A., 3, 15, 49, 101, 104 Grazing and cutting, effect of, on com- position of herbage, 62 Great Mulgraves. See Horndon Hammer, B. W., 86 Hart, E. B., 86 Hassobury field experiments with various phos- phates at, 26 limiting manurial factors at, 95 Hay effect of rainfall on yield of, 89 et seq. factors hmiting yield of, 89 et seq. yield of, at experimental centres, 18 et seq. Hoffman, C, 86 Holcus lanatus, 50, 51, 54, 66, 57 Hopkins, C. G., 11, 12 Hordeum pratense, 36 Horndon, 39, 40 accumulation of nitrate in soil at, 77 et seq. ; of nitrogen in soil at, 76 effect of grazing on herbage at, 62; of phosphates on bacterial content of soil at, 86; on herbage at, 35, 58; on soil moisture at, 63 ; on soil tem- perature at, 68; on soil texture, 73 field experiments with various phos- phates at, 30 et seq. influence of rainfall on yield of hay at, 89 et seq. Hutchinson, H. B., 99 Hydrogen ion concentration, 26, 99 et seq. Hypochaeris radicata, 36 Iron sulphate. See Ferrous sulphate Jamieson, 15 Kirkman, I Lambourne End accumulation of nitrate in soil at, 77 effect of phosphates on herbage at, 54 field experiments with various phos- phates at, 39 Latchingdon accumulation of nitrogen in soil at, 76 effect of continuous cutting on herbage at, 62; of phosphates on herbage at, 39, 59; on soil texture at, 73; of rain- fall on yield of hay at, 89 et seq. field experiments with various phos- phates at, 37 Lawes, Sir John, 1 Leontodon hispidus, 36 Liebig, 1 Lime effect on clover plant, 61, 105; on herbage, 58, 59, 61; on yield of hay, 31,96 Lime requirement, effect of basic slag on, of soU, 99 et seq. Lolium perenne, 36, 50, 51, 56, 57 London clay, field experiments with various phosphates on, soils, 30 et seq. MacLennan, K., 99 Martin's Hearne, 45 accumulation of nitrates in the soil at, 77 ; of nitrogen in the soil at, 76 clover failure at, 96 effect of lime on yield of hay at, 96; of phosphates on herbage at, 49, 59; on soil bacteria at, 86 ; on soil mois- ture at, 71; on soil temperature at, 71; on soil texture at, 73; of rainfall on yield of hay at, 89 et seq. field experiments with various phos- phates at, 24 Hmiting manurial factors at, 96 Middleton, Sir T. H., 3, 49, 77 Mineral phosphates. See Rock phos- phates Moisture, effect of phosphates on soil, 63 Nitrates, estimation of, 78 effect of phosphates on accumulation of, in soils, 77 et seq. Nitrogen effect of various phosphates on accu- mulation of, in soils, 75 fixation by nodule organism, 49 Ocean Island phosphate, 8 Oldershaw, A. W., 15 112 INDEX Paterson, J. W., 17 PfeifEer, 47 Phleum pratense, 50, 51, 56, 57 Phosphates action of, on heavy clay soils, 89 et seq. effect of, on accumulation of nitrates in soil, 77 et seq. ; on accumulation of nitrogen in soil, 75; on herbage, 49 et seq.; on soil bacteria, 86; on soil moisture, 63 et seq. ; on soil tem- perature, 68 et seq. ; on soil texture, 73 et seq. ; on yield of hay, 22 et seq. ; of fine grinding on availability of, 32; of rainfall on availability of, 46; of sour soils on availability of, 47 Poa trivialis, 50, 56, 57 Potash, influence of, on yield of hay, 97 Potentilla reptans, 36 Prunella vulgaris, 36 Rainfall, effect of, on availability of rock phosphates, 46; on perma- nency of clover, 60; on yield of hay, 89 et seq. Ranunculus, 24, 36 Rock phosphates Algerian, which see American, 8, 16 composition of, 9, 21 deposits of, 8 Egyptian, which see field experiments with, in America, 11 ; in France, 14; in England, 14; in Scotland, 16; in Wales, 16; in Essex, 18 et seq. Florida pebble, which see; soft, which see Gaf sa, which see Nauru, 8 North African, 8, 33, 42, 48 Ocean Island, 8 Tunisian, which see Rothamsted, 75 Mumex acetosa, 24 RusseU, E. J., 16, 36, 85, 86 Saussure, de, 1 Sillars, D., 2, 6 Soil analyses of, at experimental centres, 19 bacteria, effect of phosphates on, 86 moisture effect of phosphates on, 63 et seq. som"ness, effect of phosphates on, 99 sourness, effect of, on clover plant, 60 61 temperatm-e, effect of phosphates on, 68 et seq. texture, effect of phosphates on, 73 Somerville, W., 3, 104 Spanish phosphorite, 1 Stead, J., 40 Stellaria media, 24 Temperature, effect of phosphates on soil, 68 Tree Field, 49 Tunisian phosphates, 15, 21, 27 effect of, on herbage, 35, 58, 59; of fine grinding on availability, of, 32 field experiments with, 31, 40 Tysea HUl Farm accumulation of nitrogen in soil at, 76 effect of Ume on yield of hay at, 96; of phosphates on herbage at, 49 et seq.; on soil texture at, 72 et seq.; of rainfall on yield of hay at, 89 failure of clover at, 104 field experiments with various phos- phates at, 23 limiting manurial factors at, 89 et seq. Verney, Sir H., Bart., 14 Waggaman, 12 Wendens accumulation of nitrogen in soil at, 76 effect of phosphates on herbage at, 57, 58; of rainfall on yield of hay at, 89 field experiments with various phos- phates at, 43 FEINTED IN ENGLAND BY J. B. PEACE, M.A. AT THE CAMBRIDGE TJNIVEESITY PEESS PLATE I 113 Poniing Slag and Metal from Basic Open Hearth Furnace. The molten slag is seen overflowing from the steel ladle into the slag ladle. Nauru Island. Shipping phosphate in bulk from Nauru Island. The phosphate has to be lightered off in surf-boats — over 1000 tons can be shipped in 9 hours. The steamer is lying in 150 fathoms of water. 114 PLATE II Ocean IslcDid Phospliate Workings. Coral pinnacles after most of the phosphate has been removed. A few feet more phosphate available below rail level. Ocean Island Phosphate Workings. Foreground: Most of the phosphate has been removed exposing the coral limestone pinnacles. Background : Phosphate deposit is intact and exists naostly in the form of gravel with occasional large boulders of phosphate rock. Coco-nut and other vegetation all growing in phosphate. PLATE III 115 Plot 1. Open Hearth Fluorspar Basic Slag. Martin's Hearne. June 3rd, 1918. Plot 2. Open Hearth High Citric Soluble Basic Slag. Martin's Hearne. June 3rd, 1918. 116 PLATE IV Plot 3. Untreated. Martin's Hearne. June 3rd, 1918. Plot 4. Gafsa Rock Phosphate. Martin's Hearne. June 3rd, 1918. PLATE V 117 Section of the Soil at Hassobury showing the presence^of chalk about 3 feet below the surface. (Photograph taken from the ditch at the bottom of the experimental field.) lis PLATE VI View looking down Untreated Plot K. Horndon. July 1920. View looking down Cleveland Phosphate Plot H. Horndon. July 1920. PLATE VII ]19 Basic Slag plot at Horndon. August 1919. Untreated plot at Horndon. August 1919. 120 PLATE VIII Gafsa Kock Phosphate plot at Horndon. August 1919. Photograph of chalk pit at Saffron Walden illustrating character of soil at Wendens' Experimental Centre.