-'^'•'^wimfmmm liHMKHIHB&lktt«iQlMMMaMM»*-«aMKI*M MSMi THE JOURNAL OF AGRICULTURAL SCIENCE CAMBRIDGE UNIVERSITY PRESS C. F. CLAY, Manager LONDON: FETTER LANE, E.G. 4 H. K. LEWIS & CO., LTD., 1 36, COWER STREET, LONDON, W.C. I WHELDON & WESLEY, LTD., 2—4 ARTHUR ST., NEW OXFORD ST., LONDON, W.C. 2 CHICAGO: THE UNIVERSITY OF CHICAGO PRESS (agent for the UNITED STATES AND CANADA) BOMBAY, CALCUTTA, MADRAS: MACMILLAN & CO., LTD. TOKYO: THE MARUZEN-KABUSHIKI-KAISHA All rights reserved THE JOURNAL OF AGRICULTURAL SCIENCE EDITED FOR THE PLANT BREEDING AND ANIMAL NUTRITION RESEARCH INSTITUTES AT CAMBRIDGE, AND THE ROTHAMSTED RESEARCH INSTITUTES BV Professor R. H. BIFFEN, M.A., F.R.S., Cambridge Sir a. D. HALL, K.C.B., M.A., LL.D., F.R.S., Ministry of Agriculture, London F. H. A. MARSHALL, Sc.D., F.R.S., Cambridge Sir E.J. RUSSELL, D.Sc, F.R.S., Rothamsted Experimental Station, Harpenden Professor T. B. WOOD, C.B.E., M.A., F.LC, F.R.S., Cambridge IN CONSULTATION WITH B. C. ASTON, Department of Agriculture, Wellington, New Zealand Dr C. a. barber, C.I.E., School of Agriculture, Cambridge Professor B. T. P. BARKER, M.A., National Fruit and Cider Institute, Long Ashton, Bristol W, BATESON, M.A., F.R.S., John Innes Horticultural Institute, Merton, Surrey J. R. CAMPBELL, B.Sc, Department of Agriculture, Dublin I. B. POLE EVANS, Department of Agriculture, Pretoria, South Africa F. B. GUTHRIE, Department of Agriculture, Sydney, N.S.W. Professor J. HENDRICK, B.Sc, Marischal College, Aberdeen Sir T. H. MIDDLETON, K.B.E., C.B., M.A., The Development Commission, London Dr frank T. SHUTT, F.I.C, Experimental Farms, Ottawa, Canada Professor W. SOMERVILLE, M.A., D.Sc, School of Rural Economy, Oxford Dr a. C. TRUE, Department of Agriculture, Washington, D.C., U.S.A. Sir FRANCIS W.'iTTS, K.C.M.G., Barbados Dr H. J. WHEELER, American Agricultural Chemical Co., Boston, Mass., U.S.A. VOLUME XII 1922 CAMBRIDGE AT THE UNIVERSITY PRESS 1922 PRINTED IN QREAT BRITAIN CONTENTS Part 1 (January 1922) PAGE Newton, Robert. A comparative study of winter wheat varieties with especial reference to winter-killing . . 1 GooDEY, T. On the susceptibihty of clover and some other legumes to stem-disease caused by the eelworra, Tylenchus dipsaci, syn. devastatrix, Kiihn. (With Plate I) . . 20 Salaman, R. N. and Lesley, J. W. Genetic studies in potatoes: sterihty. (With Plate II) .... 31 Crowther, Charles and Woodman, Herbert Ernest. A study of nitrogen metabohsm in the dairy cow . . 40 Armstrong, S. F. The Mendelian inheritance of susceptibihty and resistance to yeUow rust {Puccinia glumarum, Erikss. et Henn.) in wheat ........ 57 Woodman, Herbert Ernest and Hammond, John. Note on the composition of a fluid obtained from the udders of virgin heifei-s ......... 97 Part 2 (April 1922) Murray, Alan J. The chemical composition of animal bodies. (With 1 Text-figure) 103 Taylor, William and Husband, Alfred D. The effect on the percentage composition of the milk of (a) variations in the daily volume and (6) variations in the nature of the diet. (With 4 Text-figures) Ill Tocher, J. F. The citric solubihty of mineral phosphates. (With 8 Diagrams) 125 Woodman, Herbert Ernest. Comparative determmations of the digestibihty and metabolisable energy of green oats and tares, oat and tare hay and oat and tare silage 144 , Whittles, C. L. A note on the classification of soUs on the basis of mechanical analyses. (With 11 Text-figures) . 166 Salaman, Redcliffe N. The Influence of size and character of seed on the jdeld of potatoes. (With 4 Text-figures) . 182 Engledow, F. L. and Shelton, J. P. An investigation upon certain metrical attributes of wheat plants . . .197 vi Contents Pakt 3 (JuLV 1922) PAOK Deighton, Thomas. Some investigations on the electrical method of soil moisture determination. (With 6 Text- figm-es) . . . . . . 207 WooD.MAN, Herbert Ernest. The chemistry of the strength of wheat Hour . . . . . . . .231 IvANOFF, E. 1. On the use of artificial msemination for zoo- technical purposes in Russia ...... 244 Capstick, J. W. and Wood, T. B. The effect of change of temperature on the basal metabolism of swine. (W^ith 4 Text-figures) ........ 257 HoRTON, E. and Salmon, E. S. The fungicidal properties of certain spray-fluids. Ill ...... 269 CoLLms, S. H. and Thomas, B. The sugars and albuminoids of oat straw 280 Robinson, Gilbert Wooding. Note on the mechanical analysis of humus soils 287 Jones, S. G. A bacterial disease of turnip (Brassica napus). (With Plate III) 292 Robinson, Gilbert Wooding. A new metiiod for the mechanical analysis of soOs and other dispersions. (Willi 4 Text-figures) 300 Part 4 (October 1922) Amos, Arthur and Williams, Gwilym. Temperature and other factors affecting the quahty of silage . . . 323 Amos, Arthur and Woodman, Herbert Ernest. An in- vestigation into the changes which occur during the ensilage of oats and tares ...... 337 Comber, Norman M. The availability of mineral plant food. A modification of the present hypothesis . . . 363 Comber, Norman M. A modified test for sour soils . . 370 Comber, Norman M. The flocculation of soils. Ill . . 372 Hammond, John. On the relative growth and development of various breeds and crosses of pigs. (With 5 Text- figures) 387 Vol. XII. Parti January, 1922 THE JOURNAL OF AGRICULTURAL SCIENCE EDITED FOR THE PLANT BREEDING AND ANIMAL NUTRITION RESEARCH INSTITUTES AT CAMBRIDGE, AND THE ROTHAMSTED RESEARCH INSTITUTES BY Professor R. H. BIFFEN, M.A., F.R.S., Cambridge Sir a. D. HALL, K.C.B., M.A., F.R.S., Ministry of Agriculture, London F. H. A. MARSHALL, Sc.D., F.R.S., Cambridge E. J. RUSSELL, D.Sc, F.R.S., Rothamsted Experimental Station, Harpenden Professor T. B. WOOD, C.B.E., M.A., F.LC, F.R.S., Cambridge IN CONSULTATION WITH B. C. ASTON, Department of Agriculture, Wellington, New Zealand Dr C. a. barber, C.I.E., School of Agriculture, Cambridge. Professor B. T. P. BARKER, M.A., National Fruit and Cider Institute, Long Ashton, Bristoi W. BATESON, M.A., F.R.S., John Innes Horticultural Institute, Merton, Surrey J. R. CAMPBELL, B.Sc, Department of Agriculture, Dublin 1. B. POLE EVANS, Department of Agriculture, Pretoria, South Africa F. B. GUTHRIE, Department of Agriculture, Sydney, N.S.W. Professor J. HENDRICK, B.Sc, Marischal College, Aberdeen Sir T. H. MIDDLETON, K.B.E., C.B., M.A., The Development Commission, London Dr FRANK T. SHUTT, F.I.C, Experimental Farms, Ottawa, Canada Professor W. SOMERVILLE, M.A., D.Sc, School of Rural Economy, Oxford Dr a. C. true. Department of Agriculture, Washington, D.C., U.S.A. Sir FRANCIS WATTS, K.C.M.G., Barbados Dr H. J. WHEELER, American Agricultural Chemical Co., Boston, Mass., U.S.A. Cambridge University Press _ C. F. Clay, Manager London : Fetter Lane, E.C. 4 also H. K. Lewis & Co., Ltd., 136, Cower Street, London, W. C. i and Wheldon and Wesley, Ltd., 28, Essex Street, London, W.C. 2 Chicago : The University of Chicago Press (Agents for the United States and Canada) Bombay, Calcutta, Madras ; Macmillan and Co., Ltd. Tokyo: The Maruzen-Kabushiki-Kaisha 'Price Ten Shillings net [Issued 24 February, 1922.] m^ HEFFER mm " Tagore on the peasantry and Dr Leake on Agriculture supplement each other, and both are Indispensable to a true understanding of Agricultural INDIA AND ITS RVOTS {Peasants). "—The National Outlook The Bases of Agricultural Practice and Economics in the United Provinces, India. By H. Martin Leake, M.A., Director of Agriculture, United Provinces. Formerly Economic Botanist to Government United Provinces. Late Principal, Agricultural College, Cavvnpore. With a foreword by J. Macke.nna, CLE., LC.S., Agricultural Adviser to the Government of India. Demy 8vo. Cloth. Price los. net. (Postage \s.) The book brings into prominence the dual aspect of Agriculture, the necessity not only for good practice but for good organisation. The practice-and economic conditions that hold in the United Provinces, India, are throughout taken as the basis of illustration. Part I traces the history of Agriculture and shows how the basal facts of nature and chiefly the rapid multiplication of the human race have resulted in a varying standard of agricultural practice and a varying system of organisation. Parts II and III describe the fundamental facts of agricultural practice and economics respectively. Part IV traces the lines on which Agricultural practice is likely to develop in the future, and emphasises the fact that recent advances in scientific knowledge make it possible to employ directed, as opposed to undirected, or empirical, methods. Part V traces the weaknesses of the present economic system, and develops the idea of co-partnership between landlord, tenant and Government in the land. The functions of each partner are described. "Mr Martin Leake's book, which deali with fundamental* rather than details, and with a wider aspect of agricultural economics, should be read by every thoughtful student of the subject." — "Discovery CAMBRIDGE:— W. HEFFER & SONS, LTD. LONDON :-SIMPKIN, MARSHALL, HAMILTON, KENT & CO., LTD. and from all Booksellers THE UNIVERSITY OF CHICAGO PRESS Agricultural Economics By EDWIN G. NOURSE. A volume comprising several hundred carefully edited selections with valuable original introductions, which make of the material a connected whole. This book follows the plan, successful in so many recent cases, of combining in one volume the virtues of the usual text and the collateral book of readings. Particularly strong is the emphasis upon the interrelations of the technique and the economics of agriculture. Most economists have made little use of the data of agriculture in developing their exposition of . economic principles. Chapter III of this book is a notable example of the attempt to found our agricultural economics upon a sound under- standing of the technical nature of the productive processes of agriculture. Many selections also show how changing economic or commercial con- ditions are modifying the practices of agriculture, and how at the same time technical considerations, such as the exhaustion of nitrogen from the soil or the growth of knowledge about plant or animal breeding, introduce new factors into the economic problem of agriculture, xxvi + 896 pages, 8vo, cloth ; 243 net. COM P A NIO>rvOLllM E Outlines of Agricultural Economics 106 pages, paper; 4s 6d net. The Cambridge University Press Agents for the British Empire (except Canada) London, Fetter Lane, E.C. 4 Volume XII JANUARY, 1922 Part I -i^UJRAR? NEW \<)UV BUTANICAL A COMPARATIVE STUDY OF WINTER WHEAT VAIIIETIE8 WITH ESPECIAL REFERENCE TO WINTER-KILLINGS By EGBERT NEWTON, M.S., B.S.A. UniversiUj of Albert n, Edmontmi. INTRODUCTORY. Winter wheat, where it can be safely grown, usually outyields spring varieties quite markedly, but unfortunately ib is much more restricted in distribution, due to its liability to winter-kiUing. Much progress has been made with this and other crops in breeding for cold resistance by empirical methods. It would appear, however, that greater and more certain progress would be made if the nature of cold resistance in plants were well understood. This subject has long been of interest to physio- logists, and in its practical applications is of widest importance. The northern limit of profitable growth of our staple crops marks also the limit of profitable exploitation of our agricultural lands. The southern farmer has likewise to meet the problems of the frost-killing of fruit buds and flowers and the more tender winter cereals. HISTORICAL. The progress of our knowledge of the nature of cold resistance and frost effects has been reviewed quite fully at different times by Abbe(i), Blackman(5), Chandler (O) and others. Of the earlier investigations it will be sufficient to note here in their order the most significant. The theory put forward in 17.37 by Duhamel and Buffon(ii) that death from cold was due to rupturing of the cell walls by expansion on ice formation, is of historical interest. It was almost a century later that Goeppert(i3) found ice formation to occur in the intercellular spaces. Sachs (.■?8) showed this to be the usual occurrence, and developed the view, now generally considered erroneous, that disorganisation took ■ This study was carried on in the lalioratory of Dr R. A. Gortner, Chief of the Division of Agricultural Biochemistry, University of Miimesota, to whom grateful acknowledgement is made for kind help and direction. The work was aided by election to the Shevlin Fellowship. Journ. of Agric. Sci. xn 1 2 Winter Wheat Varieties and Winfr r-Killing place on thawing, and might be prevented by warming very slowly, allowing time for rcabsorpHon of the water by the cell. Later Miiller-Thiirgau(:«) proved that this ice formation in the tissues was necessary for freezing to death, and concluded that death was due to the consequent desiccation of the protoplasm. This 1i ypothesis received support from Matruchot and .Molliard (2fi), who demonstrated the identity of the modifications in cell structure produced by frost, plasmolysis and desiccation. The work of (ireeley(iT) supplied similar evidence. MezOi) opposed the theory of death by desiccation, since his investigations indicated that all solutes crystallise out at a temperature not lower than — G C. He concluded that cold desiccation must there- fore be complete at this temperature, and cannot explain injury to plants which resist much lower temperatures. He advanced instead the theory of a fatal minimum temperature for each plant. Gorke (14) showed another important effect of the withdrawal of water, namely, the precipitation of certain proteins by the increasing con- centration of the cell sap, aided by its increasing acidity on cooling. He also showed that the precipitation occurred at varying temperatures for plants of varying degrees of hardiness. This was ascribed by Scliaffnit(i(i) to the splitting in varying degrees during the hardening process of com])le.\ proteins into simjiler. less readily precipitated forms. Lidforss (2(!) found most hardy plants to have their starch reserves converted to sugar during the winter, and beheved this an adaptation for cold resistance, the sugar having a protective action in preventing the precipitation of the proteins. Schaffnit ( lo) tested the effect of adding sugars and various other substances to plant saps and to egg albumen solution, and was able to modify very greatly the precipitation by freezing. Equally striking results were secured by Ma.\imov(:<0) in increasing the hardiness of sections of red cabbage and Tradescantia by freezing them in solutions of either organic or inorganic substances, provided these were non-to.\ic and had a low eutectic point. Recent Progress. Recent inve.stigations have dealt nuiinly with the hypotheses noted above, extending them in several important respects. Attempts have also been made to determine the correlation of various physical, chemical, physiological and morphological characters with apparent frost hardiness. Several workers have drawn attention to the possible importance of the fact that plant sap is contained in cells of capillary dimensions. D' Arsonval (3) estimated the osmotic pressure in very small cells at R. Newton 3 1000 atmospheres, and notes that by the apphcation of increasinp; pressure the sohdification point of water can be lowered indefinitely. In buds with small cells, dense tissues and meagre water content. Wiegand (49) found no ice at — 18° C, though in most buds it was present in large quantity. He concluded that the degree of cold necessary to cause the separation of ice is proportional to the force which holds the water in the tissues. Lewis and Tuttle(25) reported that Hving leaves of Pi/rola wrapped around the bulb of a mercury thermometer undercooled to ~ 32-l° C. before ice formation took place. On the other liand, it has been a universal observation since the time of Goeppert(]3) that ice may form in the tissues without injury to hardy plants, so that under- cooling is not of itself a sufficient explanation of hardiness. The water content of tissues is related to structure, and it has been shown by several investigators (2, 4, 24, 37, in, 13, 14) that dry matter con- tent is directly correlated with hardiness. 8inz (44) and Beach and Allen ( 1 ) noted also the importance of structures resisting desiccation, while Pantanelli(35) found injury from frost to be always proportionate to loss of water from the tissues, even when freezing was done in a saturated atmosphere. Between the concentration of the cell sap and winter hardiness, Ohlweiler (34) and Chandler (6) found a direct relationship; Salmon and Fleming (39), working with winter cereals, found none. Pantanel]i(35) partly reconciled the conflicting evidence by reporting a relationship in some crops and none in others, including wheat. Probably the sap of all plants increases in concentration during the hardening process, but not necessarily in proportion to the degree of hardiness attained. However, the earlier evidence as to the importance of the accumulation in the sap of substances of a protective nature, especially sugars, has received further support. Gassner and Grimme(i2) and Akerman and Johansson (2) reported that hardy varieties of \vinter wheat and other grains were richer in sugar, the differences between varieties corre- sponding to differences in degree of hardiness. Pantanelli (35) found that sugar was rapidly used up during exposure to low temperatures, and that hardiness was related to the quantity of sugar retained by the plant. The association of sugar accumulation with hardening by cold has been pointed out again by Rosa (37) and Coville(7). On the other hand, Harvey (Hi) found that cabbages acquired hardi- ness on five days' exposure to + 3° C, before any great change occurred in the carbohydrate equihbrium. He beHeves the principal effect of the hardening process to be a change in the constituents of the protoplasm, 1—2 4 Winter Wheat Varieties aixl Winti r-KiHiiKj as indicated by an increase in the amino-acid content, and, on freezing the sap, by less precipitation of the proteins. He measured the increase in hvdrogen-ion concentration of the sap on cooling, and was able to produce bhe same relative precipitation of proteins by adding equivalent i|uantitie8 of acid. However, the ease of precipitation of proteins by freezing apparently cannot always be taken as an index of hardiness, since Chandler (fi) was uniible to find any difference in this respect between the sap of tender and hardy twigs of fruit trees. Again, it should perhaps 1)1' remarked that the fraction of the total proteins present in the ex- pressed sap is rather small, and of this only 31-2 per cent, at most is reported by Harvey as precipitated by freezing the juice of iinliardened cabbage. The extent to whicii this justifies a generalisation with regard to the probable behaviour of the proteins within the cell may be regarded as problematic. There is some evidence that stability of the dormant condition may be an important protective adaptation. Lidforss ('-'6) noted that a suc- cession of warm days caused regeneration of starch from sugar, with an increase in susceptibility to cold. Chandler (6) found some varieties of peaches to have a longer rest period than others and to be started into growth more slowly by warm periods in the winter. Evidently varieties which can maintain continuous dormancy during the danger period must have a distinct advantage. Our present concept of the causes of winter-killing nuiy be briefly summarised. Without doubt, the ultimate cause of death by freezing must be the disorganisation of the protoplasm. Irreversible coagulation or precipitation of the colloidal protein constituents may be caused by increase in concentration of electrolytes in the cell sap on withdrawal of water, or by increase in acidity, or by both factors acting together. The critical minimum temperature necessary to bring this about must be profoundly modified by rate of cooliug, especially if this be slow enough to give time for the hardening process, and by the presence of substances which protect the proteins from precipitation. Split ling of the proteins during hardening may be a protective adaptation. Since the funda- mental feature of the disturbance produced by freezing is withdrawal of water from the cell, intracelhilnr adaptations to resist desiccation must be of prime importance. K. Newton 5 EXPERIMENTAL. The Problem. The present study seeks to establish a chemical or ijhvsico-chemical measurement of hardiness for winter wheat varieties. A number of varieties originated or selected by the Department of Plant Breeding of the University of Minnesota, and known to vary considerably in hardi- ness, were compared with reference to the physical constants of the cell sap, the content of amino nitrogen, water-soluble nitrogen and total nitrogen, and the content of sugars and starch. AH materia! used was grown in field plots under normal conditions. Since it was desired to compare the varieties in the hardened condition rather than to study the hardening process, collections were not made until after the advent of freezing weather. A preliminary study of physical constants was carried out with eight varieties. Subsequent stud}^ was confined to four of these, two hardy and two tender. One variety, Minhardi, was collected from two plots some distance apart, and these are reported separately as the effect of location was quite marked, Metho]is. Collection of Samples. All the samples of one series were collected from the field the same afternoon, though with the exception of the first series the leaves were frozen solid when collected, so that changes due to vital activities would be very slight. The plants were growing in rows, which were carefully gone over for the removal of dead leaves before taking the samples. For the collection of November 12, 1920, ib was necessary first to brush off a Hght covering of snow. As the leaves were cut, they were thrown on a wire screen for the removal of adhering bits of dirt and ice, then transferred at once to tight glass containers. Samples of approximately KIO grams were collected in duphcate, one lot for the study of physical constants, the other for analysis for nitrogen and carbohydrates. All samples were kept frozen until used. Physical Constants. The depression of the freezing point of the hrst collection was determined by the thermoelectric method, the accuracy of which has been shown by White (I7). The convenient arrangement of apparatus illustrated by Harvey (20, Fig. ]) was used. The leaves were packed into a section of thin-walled glass tubing 2 cm. long, in which they were held in place by a small rubber band, the thermocouple then being inserted in the centre. Undercoohng seldom amounted to more -37ti HutTuni (r4li5 (i-28i» radi.i (i-47r) u-i'sy Therefore in hiter collections preliminary freezing of the tissues was carried out, and the expressed sap used for all constants studied. The technique of (iortner and Harris (if)) was followed in the main. The rubber-stoppered bottles containing the samples were packed in a slushy mixture of ])ulverised ice and salt in an earthenware jar, which fitted snugly inside a well-insulated "tireless cooker." Jn this condition the contents remained frozen solid until required for use. never in any case for less than \'2 hours. For the freezing mixture, coinmou salt was used at first, and later calcium chloridi". To thaw th(> samples, tlie bottles were placed under running water, then rinsed with distilled water and wiped dry before ojiening. The leaves were folded in pieces of strong cotton previously boiled in three changes of distilli'd water and dried free from dust, and the sap expressed either in a hydraulic press under 400 atmospheres pressure, or in a large hand screw press with a small steel cup which permitted the application of heavy pressure. The parts of the press with which the juice came in contact w6re kept coated with a thin layer of paraffin wax. The depression of the freezing point was first determined with a Beckmann apparatus. Then the conductivity was measured with a, K. Newton 7 Wheatstoiie bridoe, using a Freas conductivity cell, and finally the hydrogen-ion concentration was determined by means of standard Leeds and Northrup potentiometric equipment. Both of the latter determina- tions were carried out at 25° C. in a constant temperature room. Usually the work on any particular sample was completed within an hour of expressing the juice. PreparuUon of Samples for Analysis. As the samples for analysis were collected in the field they were placed directly in tared one-litre erlenmeyers with rubber stoppers. They were frozen when collected, and kept in that condition overnight. Immediately after thawing in the laboratory next day, samples for the determination of dry matter were weighed out; then in addition, sufficient material was removed to leave exactly 100 grams in each fiask. To tills was added 1-5 grams of pure precipitated calcium carbonate for the neutrahsation of plant acids, and sufficient 95 per cent, alcohol to make the final concentration 80 per cent, after allowing for the dilution due to water in the leaves. The samples were then boiled half an hour under reflux condensers, and put away tightly stoppered until a convenient time for analysis. The procedure from thawing to boiling was carried out with the utmost expedition. The advantage of using calcium carbonate as noted above has been discussed by Spoehr(}.5). Davis, Daish and Sawyer (9) have pointed out the necessity for rapid destruction of enzymes. In this connection the present study afforded opportunity for sonu^ observations of interest. An additional quantity of one variety was collected for experimentation in methods. Part of this was left five days in an ice chest, and was then put through the regular preparative and analytical procedure. In Table 1 the results are compared with those for the same material disposed of promptly after collection. When it is considered that the tissues were not crushed or injured to any appreciable degree, the effect of enzyme action even in such cold storage conditions is very striking. Drij Matter. Triphcatc samples of approximately 5 grams of green material were dried to constant weight in a vacuum oven at 98° C. Total Nitrogen. The total nitrogen was determined by the Kjeldahl- Gunning- Arnold method, using the residues from the dry matter deter- minations. Extraction of Sugars and Soluble Nitrogen. Large extractors of the Soxhlet type were made by drilling a small hole close to the bottom of a 750 c.c. wide-mouthed bottle, and fitting in a glass siphon tightly by making a ground glass joint or by wedging it with a collar of rubber tubing. The bottle was closed with a large rubber stopper, in which were fi Wiiitrr W/ic(if \'(ifi. Fitr. I) under a pressure of less than 30 mm., with the dis- tilling flask in a water bath at a temperature of Uf to 50° C. When reduced to a volume of 7') to ItXlt-.c. about 200 c.c. of distilled water was added and the solution reconcent rated to get rid of the last traces of alcohol. This precaution was found necessary since the presence of alcohol affected the subsequent determination of amino nitrogen. The recoucent rated extract was then transferred to a 'jrjO cc. volumetric flask by Altering through a pad of cheesecloth in a small funnel, making it nearly to volume by several successive washings of the distilling flask with small portions of boilingwater. Apad of cheesecloth four layers thick was found the most satisfactory filter for removing the solid particles of chlorophyll which separated out. All otlier materials tried clogged at once. The extract was then cooled to room temperature and made up to volume. All volumetric flasks and jjipettes used throughout the analyses were standardised in true cubic centimetres at 20° C. Amino Nitnuferi. The amino nitrogen was determined by the usual Van Slyke apparatus, using 10 c.c. portions of the extract, filtered for removal of fine particles which had escaped the cheesecloth filter. Since preliminary trials had given a somewhat higher yield when deamination was continued for 30 minutes instead of the usual live minutes, the former period was adopted. Of this time, shaking was done during the first mintite and the last two minutes. Water-Solubh Nitrogen. The total water-soluble nitrogen in the alcoholic extract was determined by the Kjeldahl-(4unning-Arnold method, using 10 c.c. portions of the concentrated extract, filtered as for amino nitrogen. Oleariny Extract for Sugar Analysis. After the portions required for nitrogen determinations had been removed, the remainder of the extract was cleared with dry powdered lead acetate. The dry defecation method, first proposed by Horne(a), has tlie advantage not only of largely R. Newton 0 eliiiiinatiug the error clue to the voluiiu' of the precipitate, but also of obviating the necessity of mailing to vohune a second time. The leatl acetate was added in small (juantities at a time, successive small portions of the extract being filtered oft and tested for completeness of precipita- tion. The solution was then filtered through a dry filter paper and deleaded with a minimum quantity of powdered sodium oxalate, fol- lowing the same technique as for clarihcation. A second filtration through a dry filter paper completed the preparation of the extract for sugar analysis. Reducing Sugars. The reducing sugars were determined by the very excellent method recently devised by Shaffer and Hartmann (42), in which the cuprous oxide is determined directly by iodo-thiosulphate titration in the presence of an excess of potassium o.xalate. The oxalate inhibits the reaction of cupric ions with soluble iodides. The reduction of Fehling's solution is carried out under the standard conditions prescribed for Miinson and Walker's method (Official Methods, A.O.A.C.) and the sugar corresponding to the quantitv of cuprous oxide read from the tables. By connecting the gas burners witli water manometers, and carefully calibrating them in conjunction with the flasks used for the reductions, it was found possible to keep within 10 seconds of the four minutes prescribed for bringing the solution to boiling. The thiosulphate was standardised against pure copper by an ada[)ta- tion of the simple method proposed by Peters (36). Triplicate standardisa- tions by this method varied less than ()•(>] mg. copper per c.c. N/10 thio- sulphate. The accuracy of this sugar method was tested witii pure dextrose obtained from the Bureau of Standards. Duplicate portions of 100 mg. and of 150 mg. of the anhydrous sugar were weighed out. dissolved in water and carried through the analytical {)rocedure, with the following results : ixtruse jjRscuL DL'.\tru.s(' ii 100 '.l!»S 100 <1!I0 150 1411-4 150 141I-2 The nuijority of the sugar determinations made fell within the limits of these quantities. Aliquots of 10 c.c. of the cleared extract were used for the reductions. Sucrose. The sucrose was determined by the increase in reduction on inversion of the cleared extract. For inversion the citric acid method of Davis and Daish(8) was employed. Comparative tests with invertase 10 Winfer Whctd Varieties and Wintir-Killiiaj and witli the oftiiial liydrochloric acid nipthod jjavo results varyiug by less tJian U-5 per cent, of tlie ((uantitv found. It appears, therefore, that sucrose was the only disaccharide present in the leaves. Starch. A qualitative test for starch, with the usual iodine-potassium iodide solution, was made on the dried, ground residue from the alcohol extraction. Tabic 1. ChuMjt's in aminu nilroyen, lolul soluble nUrvgen and stujars of winter wheal leaves during 5 datfs' storage in ice chest. Aniinii niliuyrii VVaUii -solii bio iiitioHcii Keduciiig .siigui' as clextiosu Invert sugar as suci'o.sc ... Total sugar as dextrose . . . Fresh leaves Stored leaves perecutage percentage Percentage green weight green weight change 0-050 00(>7 + :}4-0 0157 0-2:i7 ■fSl-O a-286 4- 174 + 27-0 4-574 iWK - 52-0 7-943 (i-40-' - lU-4 Table 11. Physical constants of sap of ivinter wheat leaves. Collcctod October 29, 1920*. Variety Classification A /' pH Turkey Tender 2-08 24-;i;i ii-:j(jO Kan red Tender 1-97 2:!-(iS l>-()32 Minhar'di {ii) Very hardy 2-11 25-:j5 U-.528 Minliardi {h) V(My hardy- 2- 15 25-s:! (i-.">SO Buffuni Very hardy 2-21 2(1-55 ()-4{i5 Odessa Very hard\^ 2-09 25- 11 l>-457 Padui Wry hardy 2-05 24-(>:( (i-475 Mintnrki ... Hardy 2-(17 24-S7 ll-5:i() Itei R,.i,lv... Tender 2-02 24-2S i)-o:J2 * A delirniined ilirectly in tissue by tlu-rnioelectric method; i>H. determined on sap expressed after grinding tissue without pievioiis freezing. Tabic ill. J'hi/sical constants (f sap of irintcr irlieat leaves. Variety A 1' Collection of November 12 Turkey ... 2-26 27-15 Kanred Aliiihardi («) Minhardi (h) ButTum 2-.15 2-5(i 2-44 2-08 28-23 30-74 29-30 32-17 A' ■: lU^ 1920*. iisi; 14-31 13-25 14-87 14-83 Collection of December 9, 1920t. Turkey ... 1-99 23-92 14-94 Kanred Minhai-di (u) Minhardi \b) Buffuni 1-42 1-83 1-69 1-60 17-08 22-00 20-32 19-24 14-04 13-93 15-02 13-35 /,H .")S4U li-201 6-078 6086 5-980 5-486 5-557 5,548 5-733 5.593 * iSap e.\i)i'essed under ])ressure of 400 atmospheres after freezing tissues. -f- Sap expressed by large hand screw press afttir boiling tissues in pressure Uusks. II. Newton I I Table IV. Tlie role of suyars and eledrohjles in oamolic pressure. Variety P P, Collfftiuii of November 12, Turkey ... 27-15 11-5U Kanred Miiihardi (") MinharcU (6) Buffum 28-23 30-74 29-30 32- 17 9-51 11-54 10-51 12-32 P - l\ 1920. 15-(i."> I.S-72 19-2(1 IS- 79 19-85 10' A' . 10' 11-86 14-31 13-25 14-87 14-83 10-15 18-07 17-89 19-34 20-71 P-P> K X 10' 1-32 1-31 1-45 1-20 1-34 P-P, A-„ ■: 10' 0-97 1-04 1-07 0-97 0-90 Average ... 29-52 11-08 t'oUection of December 9, Turkey Kanred Mmhardi (u) Miivliardi (h) Buffum 23-92 17-08 22-00 20-32 19-24 7-49 5-40 7-91 0-5!( 7-27 18-44 1920. 10-43 ll-(iS 14-09 13-73 11-97 13-82 14-94 14-04 13-93 15-02 13-35 18-43 17-98 15-97 16-80 17-69 10-04 1-34 1-10 0-83 1-01 0-91 0-90 l-OO 0-91 0-73 0-84 0-78 0-75 AveraKe 20-51 6-93 13-58 14-20 16-91 0-95 0-80 Table V. Nitrogen of ivinler wheat leaves. Water- soluble Dry Amino nitrogen nitrogen Total nitrogen matter content .\ / V Ureen Dry Green Dry Green Dry Variety 0/ /o o/ o/ /O /O % % 0/ /o 0/ /o Collection of November 12, 1920. Turkey 37-54 0-054 0-14 (»120 0-34 1-394 3-71 Kanred 32-89 0-045 014 0-129 0-39 1-242 3-78 Minhardi {a) 37-99 0-051 (t-13 0-161 0-42 1-416 3-73 Minhardi {b) 33-51 0-048 0-14 0-1.53 0-4(( 1-228 3-60 Buffum 35-83 0-040 0-13 0-141 0-39 1-184 3-30 Collection of December 9, 1920. Turkey 29-65 0-049 0-17 0-1.59 0-54 1-179 3-98 Kanred 25-48 0-050 0-20 0-134 0-53 0-936 3-07 Minhardi {u) 31-74 0-058 0-18 0-192 0-60 1-276 4-02 Minhardi (b) 28-20 0065 0-23 0-175 0-(i2 1-049 3-72 Buffum 29-17 0-058 0-20 0-149 0-51 0-938 3-22 Table VI. Sugar content of ivinter wheat leaves. Keducingsugai ■ Invert sugar Total ! sugar Dry as de-xtrose as sucrose as dextrose matter content A /. . Green Dry t Green Dry Green Dry Variety O,' /o o/ o/ /o /o % % % % Collection of November 12, 1920. Turkey . 37-64 2-992 7-97 5-301 14-13 8-397 22-37 Kanred . 32-89 3-090 939 3-900 11-86 7-0.55 21-45 Minhardi (a ) 37-99 3-237 8-52 4-797 12-63 8-127 21-39 Minhardi (b\ ) 33-51 3-335 9-95 4-351 12-98 7-760 23-16 Bufium . 35-83 3-308 9-23 5-799 16-19 9-219 25-73 Collection of December 9, 1920. Turkey . 29-65 2-347 7-92 3-.598 12-14 6-004 20-25 Kanred . 25-48 1-840 7-22 2-660 10-44 4-540 17-82 Minhardi (a ) 31-74 2-579 8-13 3-356 10-57 6-991 18-88 Minhardi (b ^ 28-20 2-027 7-19 3-391 12-03 5-472 19-40 Buffum . 29-17 2-163 7-42 3-772 12-93 5-997 20-56 12 Wilder Wluat Varieties aitil Winter- Killing DISCUSSION. The depression of the freezing point A, the corresponding osmotic pressure P, and the hydrogen-ion concentration of the sap of the samples collected October 29 are given in Table II. The osmotic j)ressiires recorded are based on the freezing ])oint data, use being made of tiie ]iublished tables of Harris and (!or^ner(l^). The hydrogen-ion concentration is expressed in terms of pW value as read from the tables of Schmidt and Hoagland(4i). A classification of varieties, as determined by survival under field conditions during a nund^er of years, was supplied i)y the Department of Plant Breeding and is included in the table. It will be seen that in this collection at least there are no significant variations in the constants reported which could be correlated with the relative hardi- ness of varieties. The same absence of correlation in pliysiial constants, including specific conductivity A', holds true for the collection of Novendjcr VI. reported in Table HI. It may be noted, however, that the concentration of the sap had increased somewhat in the varieties used. In the collection of December 9, included in the same table, all varieties exhibit a faUing off in the depression of the freezing point and corresponding osmotic pressure, probably due in part to simple dilution of the sap, as the moisture content of the tissue was greater. One variety, Kanred, fell off in this respect appreciably more than the rest. It is perhaps note- worthy that this variety winter-killed somewhat more tiian tlie others during the year of this test. An unexpected difficulty was encountered in expressing the sap from the sam])les collected Decend)er 9. The samples were frozen by the method already described, using a freezing mixture of pulverised calcium chloride and snow mixed in the ju'oportions wliich should give theoreti- cally a cryohydrate mixture with a corresponding temperature of — 54-9° C. After seven iiours' freezing, the samples were thawed under running water, and refrozen for a period of 1 1 hours. They were thawed again under the tap. Even after this treatment it was found impossible to express more than 2 to 3 c.c. of juice from 100 grams of material under 400 atmospheres pressure. Further, the plants retained their bright green colour instead of a.ssuming the w'atersoaked appearance character- istic of fiost-kiiled tissuc^s. The data of Table V show that these samples contained a lower percentage of dry matter than did the earlier collection, consequently the failure to express the sap was not due to lack of moisture in the tissues but apparently to a failure to break down the colloidal R. Newton IH complex of the protoplasm and to increase the permeabihty of the cell by freezing. Since the permeability of the tissues had not been affected by freezing, it was decided to attempt to destroy the colloidal complex and increase permeability by ])lacing the material in a closed pressure flask and heating it in a boiling water bath. The juice was then expressed as readily by the hand press as by the hydraulic press, 30 to -10 c.c. (about the usual amount) being collected from each sample. These observations give rise to some very important considerations. In the first place, the wheat plants were apparently not killed by ex- posure to temperatures lower than normally obtain in many places where they suffer severe winter-killing. The specific temperature must be only one of a number of factors involved. It is also apparent that the hardening process continued long after the advent of freezing weather, as this difficulty was not met with in the collection of November 12. Further, and contrary to the findings of a number of workers, hardening was not in this case associated with an increase in the dry matter content, since as already noted the water content was greater in the collection of December 9. Nor was it associated with an increase in sugar content; this value had decreased, as will be seen later. Most significant is the evident tenacity with which the hardened tissue grips its water content. Wiegand (18) noted that as the temperature falls the quantity of water separating in the form of ice becomes con- stantly less and less. The same author develops the theory that the passage of water from the cell during freezing is due to an equalising of the force of imbibition, acting from the outer cell membrane to the centre of the system ; this follows as a consequence of the disturbance of equi- librium set up by the force with which the formation of ice crystals takes water from the surface of the cell. Wiegand supports the view that death is due to drying of the protoplasm beyond its critical water content. Evidence of other workers as tu the importance of resistance to desicca- tion has been presented in an earlier section of this paper. Whatever may be the precise mechanism by which withdrawal of water brings about disorganisation of the protoplasm, it seems clear at least that hardiness must be intimately connected with forces which oppose this desiccation. In the light of our present knowledge of the properties of colloids, it seems most probable that the principal force is imbibition. Spoehr (45) found that in cacti the pentosans increase with decreasing water supply. MacDougal(37) also points out that the conversion of the diffusible sugars to the mucilaginous pentosans is one of the alterations which may result in the cell as a consequence of partial desiccation. The 14 Wliitir Wlnnt Vftrirflcs hikI \Vi iiI< r-f\illing latter author pictures plant protoplasm as a pentosan-proteiu colloid, and considers that the character and amount of the pentosans largely determine the hydration reactions of the protoplast. Winter conditions where the soil freezes solid are really xerophytic conditions, since the plant's usual water supply is cut ofT. Having this in mind, it may be reasoned from the evidence of MacDougal and Rpoehr that under winter conditions pentosans would accumulate in the cell, contributing largely to the formation of a protoplasmic gel of high imbibitioiml powers. Investigation of this point will be reserved for a later paper. The relative importance of sugars and electrolytes in osmotic pressure is considered in Table IV. Unfortunately there is no known method whereby the relative proportions of osmotic pressure contributed by electrolytes and non-electrolytes can be calculated accuratelv from con- ductivity and freezing point data. These proportions have sometimes been estimated by taking tlie value for electrolytes as equal to that of a solution of potassium chloride having the same conductivity. But it is manifestly unjustifiable, as Miss Haynes(22) has pointed out, to assume that the cell sap electrolytes are dissociated to the same extent as potas- sium chloride, and produce ions of the same mobility. In the present instance the sugar percentages were known and these were assumed to represent the non-electrolyte materials. The theoretical osmotic pressure exerted by the sugars is calculated from the quantities of reducing sugars and sucrose determined by analysis to be present. The difTerence between this value P^ and the total osmotic pressure P (i.e. P— P,) may be attributed chiefly, though probably not entirely, to electrolytes. In the coliec'tion of November 12, the sugars present were sufficient to account for an average of about .38 per cent, of the osmotic pressure: in that of December 9, for about 34 per cent. The probable effect of non-electrolytes in depressing the conductivity cannot be determined by ordinary viscosity measurements, as these do not distinguish between substances in molecular solution and colloids, and the latter substances, as is well known, do not interfere appreciably with conductivity^. However, the correction suggested by Miss Haynes for the depression of the conductivity by the sugar content of the sap has been applied to the observed values. The corrected specific con- d)i(tivity A'„ is derived from the observed \-;iIue K . hx the formula lOOA" '""^lOO-oa;' where n = 2 (constant for sugar) and x — per cent, sugar in sap. ' Fur i'xam])l(', spp Af.rson (2H), Table V. R. Newton 15 The ratio of the value P — P, to the corrected vahie for specific con- ductivity ( X 10^) is given in the last column of the table. This should be a constant in case the cell sap electrolytes in each variety were identical in composition, and the effect of solutes other than sugars on the mobility of the ions can be considered equal for all varieties. The ratio of P — P., to the observed value for conductivity is given in the adjoining column, in order that the effect of the correction for sugar content may be seen. For the earlier collection the ratio, based on the corrected values for conductivity, varies from 0-96 to 1-07, and averages 1-00, but for the later collection it falls off somewhat, ranging from 0-73 to 0-91 with an average of 0-80. Possibly the heating of these later samples to lOO^ C. had released ions which would otherwise have remained adsorbed by cell colloids, thus increasing somewhat the value of A', though Mason's results (28, Table VI) suggest that such an increase would probatily have been very small. In any case, the diminution in the value of P — P, in the later collection is so decided as to make it unlikely that the possible disturbing factors introduced by the heating could account entirely for the failure of K to diminish correspondingly. Furthermore, the content of soluble protein (see Table V) is small and relatively constant, and the hydrogen-ion concentration is Cjuite constant, so that the possibility of any considerable variation in conductivity due to protein salts and organic acids may be excluded. The evidence suggests that substances other than electrolytes, and variable in nature, must contribute somewhat to the quantity P — P^, or in other words that sugars are probably not the only non-electrolytes which contribute to the osmotic values. The amino nitrogen, water-soluble nitrogen and total nitrogen in per- centage of green and dry weights are given in Table V. The percentage of dry matter content is also included. It cannot be said that any of the figures exhibit a marked correspondence to differences in degree of hardiness. The increase in the water content of the collection of Decem- ))pr 0 is accounted for by a mild rainy period of some days" duration which occurred during the previous week. It has been remarked already that Kanred killed somewhat more than the other varieties during the season of this experiment, and this variety was lowest in dry matter content. However, in the hght of the evidence already presented, it seems possible that our view of the nature of the correlation between dry matter content and hardiness may require modification. A smaller water content is naturally associated with the smaller cells and denser tissues characteristic of the slower growth in late autumn, so that in general hardened tissues would be expected to contain less moisture than 16 Winter Wheat Varieticfi rnid Winter- Kill Inn imhardened tissues. But iu coiiiparin;:: liardcncd tissues of different varieties, assuniinj; them to be of similar structure, the moisture content is perliaps larfjely a function of the relative imbibitional powers of the cell colloids and the defrrec of ])reviows exposure to modifyiui; environ- mental factors, for example, desiccating agents such as winds and low temperatures. It seems, therefore that both the magnitude and order of the values in any particular series of samples may be affected bv the weather conditions preceding the date of collection. The fundamental importance of resistance to withdrawal of water does not, of course, exclude the possibility of an important relationship between the character or .state of the proteins and the secondary effects following desiccation. In this instance, though the percentages do not correspond uniformly to the known hardiness, yet Minhardi, the hardiest variety used, liad somewhat the largest content of water-soluble nitrogen. All varieties show an increase in amino nitrogen and water-soluble nitrogen in tlic later collection. This is in harmony with the evidence of Harvey (i!i) that s])litting of the proteins is associated with tjic hardening process. The high content of sugars, especially sucrose, reported in Tabic \'l. is cjuite remarkaljle. It has been noted already that sucrose was ap- parently the only disaccharide {)resent. But liere again the varieties could not be classified according to hardiness on the basis of the values found. All suffered a loss between the collections of November 12 and December 9, the percentage falling lowest in Kanrcd. The greater degree of killing in this variety thus lends support to the observation of Pan- taneUi (•■«) that luirdiness was proportional to the (piantitv of sugar retained during exposure to low temperatures. A qualitative test for starch on the dried, ground residue Irom the alcohol extraction gave entirely negative results in every case. This is as expected from the observations of Miyake (32), Lidforss (26) and others. The above discussion indicates that we are still far from an exact analysis of the factors influencing winter hardiness, but certain of the observations, notably the failure of freezing the tissues to break down the protoplasmic .structure in the hardened jilants, are very suggestive. Furtlier investigation of tliis iilicnomcmm will be carried oul in the near future. R. Newton 17 SUMMARY. 1. A number of varieties of winter wheat, known to van' consider- ably in degree of winter hardiness, were compared in the hardened condition with reference to the physical constants of the cell sap, and the content of dry matter, nitrogen, sugars and starch. 2. No constant relation was found between depression of the freezing point, specific conductivity, or hydrogen-ion concentration of the cell sap and relative frost hardiness. 3. Sugars accounted for .^f to 38 per ceni'. of the total osmotic pressure of the sap. 4. The ratio of that part of the osmotic pressure not due to sugars (i.e. P— PJ to the corrected specific conductivity ( >; lO-'') is not a constant. For the samples collected November 12, this ratio varied from 0-9G to 1-07 (average 1-(K>) and for those collected December 9, from 0-73 to 0-91 (average 0-80). 5. The relation between dry matter content and hardiness was not constant, though one of the two tender varieties had the lowest per- centage. 6. All varieties increased in amino nitrogen and water-soluble nitrogen during the hardening process. The hardiest variety had the largest content of water-soluble nitrogen, but the relation was not uniform throughout the series. 7. The sugar content did not correspond uniformly with the known hardiness. The percentage decreased between November 12 and Decem- ber 9, falhng lowest in one of the two tender varieties. 8. Sucrose is an important storage^ material and is apparentlv the only disaccharide present. 9. All varieties were entirely free from starch. 10. The colloidal complex of the cell of the fullv hardened tissue could not be broken down by exposure to the teinperature of a calcium chloride-snow cryohydric mixture (theor. = — 54-9° C). 11. The hardened tissue retains its water content with great force. From tissue containing about 70 per cent, of moisture no appreciable amount of sap could be expressed by 100 atmospheres" pressure, even after severe preliminary freezing. JouTQ. of Agric. Sci. xii 18 Winter Wheat Varieties and Winter- KUlliui KKKERENCES. (1) Ahbk, C. Kxp. Sla. Record, 6 (1895), p. 777. < o (2) Akerman, a., and Johansson, H. tSuerUjen Utsmleforenim/s TiiUkriJI, 27 (I!I17). p. 77. (3) d'Arsoxvai,, M. Compt. Hend. Acad. Set. Paris, 133 (l!iol), p. 84. (4) Beach, S. A., and Allen, F. W., Jr. lounAgr. Exp. .Sin. Ue.ienrrh Hull. 21 ( 1015). (5) Blackman, F. F. New I'hyl. 8 (1909), p. 354. ((i) Chandlkr, W. H. Missouri Agr. Exp. Sla. Research Bull. 8 (1913). (7) CoviLLE, F. V. ./. Agr. Research, 20 (1920), p. 151. (8) Davis, W. A., ;ui. .■U9. (43) Shutt, F. T. Trans. Roy. Soc. Can. Ser. 2, 9 (1903). .sect. 4. p. 149. (44) SiNZ, E. ./. Landw. 62 (1914), p. 301. (45) Spoehe, H. A. Carnegie Inst. Wash. PkIk 287 (1919). (46) Van Slykb. D. D. ./. Biol. Ohem. 10 (1911), p. \r>. (47) WinTK, W. P. Plii/sirul Revieiv, 31 (1910), p. I;i.'>. (48) WiEfiANU, K. M. Plant Worhl. 9 (19(l(i). pp. 2.",. 107. (49) Bot. Gaz. 41 (1006), p. 373. (Received Aiifjusl -IGlh. 1921.) 2—2 ON THE SUSCEPTIBILITY OF CLOVKi; AM) SOMK OTHER LEGl'MES TO STEAJ- DISEASE CAUSEJ) BY THE EELWORM, TYLENCHUS DIPSACI, 8YN. DE VA ST A TRIX, K t^H N. By T. COODEV, D.Sc. DepartmenI of Helmintholof/y, London School of Tropical Medicine, hite of I he Rothmnsted Experimental Station. (With P1m(.' 1, and Tables I and II in Text.) CONTENTS. PAOE Tntioduetion . 20 Tlie pai-asite . 21 DfHoriptioii of tlif disi'asc 22 lOxppriiiH'iilal . 23 Disfussion .... . 28 RcflTl'IU'CS .... . .W i.XTHoDrc'i'rox. Tt has l)een known for a number of years that red clover and rertain other cultivated leuuniinous plants, besides many other non-lei vdiiKcd bij I Ik h'tin'onii in till' biutal t-avitv. It is usually assuiiu'd that it is by means of this organ that it pierces plant tissues and then penetrates into the plant, and that havinj^ once got inside it uses its stylet to puncture the cells amongst which it is l^-ing. and feeds on the cell sap which it sucks out from them. All this, however, needs much further investigation, I think, especially when one remembers the power pos.sessed by the infective larvae of the human ])arasite Anoi/losloma dmxJenale of boring through the skin without the aid of anv piercing mouth armature. Moreover, no one. as far as 1 know, has ever obser\'ed the stylet of Tifletichus dipsaci in use as a piercing organ. It is a very small structure, being 12-15 microns in length, and t he point is so fine that it can only be seen under the very high magnification of an oil-immersion lens. In all my numerous examinations of diseased clover i>lants. l)oth fresh and ])reserved. 1 have never found the stylet exserted from the anterior end of the body. Another point: it is a matter of observation that the parasite, in .some way or other as }et not understood, causes an increase in the size of parenchymatous cells of the host plant, and this also requires further investigation. Kitzema-Bos suggested that it was due to a secretion poured out by the worm, and considered the spatulate posterior portion of the oesophagus as probably the seat of the secreting gland; the matter is, however, still very obscure. It is hoped that later on one may be enabled to take up the elucidation of these interesting problems. DESCKll'TlON OF THE DISEASE. In regard to a suitable name for this disease, the term Sicni (/i.va/.ve sliould be used, thus keeping in line with the German expression >Slock- krankheit for the same condition. 'Die term Stem-rot is already in use as the name of tlie fungal disease of clover caused by Sclerotinia Iri- folioniin, see Anu)s (1919) and Cotton (1920). Admittedly the two names are sufficiently alike to lead to the ])ossi- biiity of confusion in nomenclature for two very different pathological conditions, but 1 would suggest that the confusion might be avoided by the use of another name for the fungal disease which is not strictly con- fined to the stem as is the eelworm disease, but attacks the whole of the foliage of the plant. The term Foliw/e-rol of clurcr. or Clorer rot would be j)referable to Stem-rot, which is not an exact descriptive designation. It is of interest to note that the Germans have avoided confusion by the use of the word Klcekrebu for the fungal tlisease. T. (;ooDKY 23 Tlio followinjj; is a brief account of tlie rliicf cliaractpri.stics of flic two diseases, the differences between wliicli are fairly well shown in tlic two pliotograpLs at the end of this paper. Stem (lisea>ie caused by Ti/lencliKs dipsaci is chiefly characterised by a stimtinfi a.nd deforming of the plant accompanied by much swclhug of the leaf stalks at their base and of the stipules. The older leaves die back and new leaves which develop, though frequently numerous, have very short stalks, and the leaf blades are twisted and much wrinkled. The leaf stalks and stipules are also very much discoloured. The plant gradually dies back and is eventually killed, though the parasite does not attack the root. The disease spreads slowly from an infective centre and persists throughout the year, being most noticeable perhaps during the winter and spring when the leafage is not profuse. Foliage rot of clover, or Clover rot, due to the fungus, differs from t he eelworni disease in the appearance of the affected plants. The foliage is first attacked and becomes covered with a white mould or niycehum which causes the leaves and stems to turn brown and rot rapidly. The spread of the disease is rapid, sweeping over a whole field in a few days, especially when favoured by mild damp weather during autunm and winter. There is no stunting or swelhng of the affected parts but just a collapse and decay of the foliage followed by attack on the roots. If not too badly attacked, a second crop of clover may spring up from the unkilled roots and a fair yield may be obtained. The disease does not apparently continue to attack the clover during spring and summer. A certain amount of confusion seems to exist among agriculturists as to the differences between these two diseases, b^it to anyone who has seen the two conditions in the held the differences are very marked and unmistakable. The confusion is partly exphcable on account of the frequent association of saprophytic eelworms along with the fungal foliage-rot disease. All these eelworms are different, however, from the disease producing Tylenchus dipsaci and are free- living soil nematodes which liave invaded the rotting clover tissue and have there found a highly favourable medium for growth and multiplication. EXPERIMENTAL. !^ome preliminary experiments were carried out in which several seedhngs of red clover, about 10 days old and showing the first true leaf, were each inoculated with a single egg containing a well developed larva. 24 SU Ill-dim (isi rtiHsa/ //// fJn hJi Inoriii 'J'lio inoiuliition was (-arried out by means of a fine capillary glass pipette. Later examination and dissection of seedlings revealed the fact that mature and sexually difTerenliated worms were d<'veloped in from 24 to 30 days from the date of inoculation. Several experiments were also made in which clovers of different kinds were grown on soil heavily infected wnth Ti/lenchu^ dipsaci. It was invariably found that the seedlings of red clover showed a high per- centage ol affected planfs which possessed, at first, swollen hj'pocotyls and later on became stunted and deformed with small wrinkled leaf- blades. Kollowini; these early experiments the main experiment which forms the basis of this paper was set up. The soil used was that from around the roots of diseased red c^lover from the Universitj' Farm at Cambridge, which had i)een kept for about two months at the bottom of a small galvanised iron bin. All the plants had died down and become withered and brown. The soil was crumbled up finely and the tops of the plants were chopjjcd up finely and thoroughly mixed into the soil, with which also a small proportion of silver sand was incorporated. This mixture, which it was considered would be highly infective for Tylenchuf! dip.mci, was put in a layer about one inch deep on the top of Hoosfield .soil plus lOper cent, sand in f O-ineh glazed pots, thus forming a shallow layer of infective soil close to the top of the pot. Each pot was divided into four quadrants for economy of space by means of glass partitions pressed down into the soil. One hundred seeds of each of the following were sown, each in its quadrant, and covered with a thin layer of .sand. Red clover — Enghsh, French, Canadian. Wild Knglish. C'oiv-grass — English . Swedish . Alsikc clover — English, Canadian. While clover — Sutton's Mammoth. Wild Cotswold. Wild Kentish, English. Dutch AVhite. Kidncif celcli, Sainfoin. Lucerne (Provence), Trejoil. All the seeds were sown on the same day, June 1st. The ])ot s were then carefully watered and put out in the wire-covered enclosure and there left for the seedlings to grow under the same conditions of temperature, light, air, moisture, etc. It was hoped by giving throughout all the pots uniform conditions of soil in which the parasites were as equally di.stributed as could be arranged, and by lea\nng them under the same climatic conditions in all cases, that anv differences shown in the inci- \ (tOODEY •in denco of attack by Ti/lenr/nis would be due to differences of suscepti- bility to the jtarasite. That this was a reasonable and sound assumption is borne out, I think, by the results obtained. On July 8th, that is, after 37 days, all the seedlings were harvested, counted, and after being separated into deformed and health}' in ap- pearance, were separately jiickled in 70 per cent, alcohol. The figures obtained in this way enabled me to arrive at Ihe per- centage of deformed seedlings in each case, as shown in the following table. Taljle [. Showing Ihe ninnherf: of heallhi/ (uul ilcfunned nvedlinys, ihe percenlage of (leforined, and Ihe sliindard error. No. of .seedling's ' , i'ciri'ntage of Stiiudiiiil iviiiil of plant Healthy Deformed Total deformed enor Hal chinr: Knglisli :i', 41 71) .■)4 ±5-7 French L'(; III S7 70- 1 1 14-!) Wild English IS 47 li.") 72-3 -5-5 Canadian 10 lil StI 70-25 :14-S Cow-fjrass : Englisli .. Xi •.V.) 72 54-1 .■)-il Swedish 20 37 li3 58-7 :Jy- Al-^^-ike rloi'cr : English .. 71 12 S3 144 J:3-!l Canadian (iO 23 S3 27-7 l,4'l) White, clover: Sutton's Mammoth . 711 !l ss 10-2 ' 3'2 English r,:', 1(1 1)3 ir)S7 J:4-0 \Vild Cotswold 1- II) S2 12-2 30 \yild Kentish 1 1 III S7 ll-.-> i.3-4 Kidneif rclch .. :!l 42 7.3 00-0 _..")• S i^aiiifoin .js 13 71 I.S-3 ^4-0 Lucerne (Provence) lis 4 72 :,-i + 2-7 Trefoil 41 41 These figures of ]iercentage infection give one a rough indication of the susceptibihty of the different kinds of clovers, etc., but it was thought that if one could estimate the numbers of Tylenchus in a series of deformed seedHngs one would be able to arrive at some expression for the intensity of infection or intensity of susceptibihtv in eacli case. With this end in view I examined all the deformed seedlings and selected the ten most deformed in each case, and then proceeded to the dissection of these in order to obtain the contained Tylenchus. Each seedhng was carefully dissected by means of needles, with the aid of a binocular dissecting microscope. This, of course, was a slow 2*t SU'iH-iliseaxc (•(Dined l>ii Ihi A't/icunii proci'ss, but it, was loiind quite ]>r;ictic;ibli', :iiu| coiiiparalively fow of tiie worms were damaged or broken during these operations. Each seedling was placed in a shallow glass capsule iu distilled water, and bit by bit tlisspcted so as to free tlie tissues from the contained parasites wliich were then left in the water whilst the vegetabh; matter was gradually removed. After the dissection the nematodes were collected, concentrated and trnnsferred to a microscope slide oi- slides as the case might require, and then e.xaniiued under the microscope, i^'or the purposes of collection 1 first of all used glass capillary pipettes, bub later on abandoned these and used the centrifuge which proved very servicealile. The resulting drop for examination thus contained adults of both se.xes, larvae in many stages of development, and numerous eggs. In making my exami- nations I recorded separately the numbers of males, females and larvae : eggs were not counted. A record was also kept of whatever other kinds of nematodes might be present with the Ti/lenchus. These, however, do not concern us in the present paper. In the case of those varieties which gave a large number of deformed seedhngs, the ten most deformed were, as stated above, selected for dissection, whilst iu the other cases as for example, lucerne, sainfoin, white clover, where only a few seedhngs were deformed, the \^liole of them were dissected. In the case of lucerne only four seedhngs were defornieil. hut none of these revealed any adult Tyleiichvx on dissection. Since the same number of seedlings was not dissected tiiroughoul, it became necessary to decide arbitrarily on some number to take as an average in arriving at the index of susceptibility, and for this purpose the number of deformed lucerne seedlings was chosen, i.e. four. Since also index of susceptibihty was being interpreted as equivalent to intensity of susceptibility, the four highest totals of males plus fcnuiles were taken and an average made of these, the resultant figure being called the index of susceptibility. Whether the males and females counted are the progeny of a single fertilised female, thus giving a reproduction figure, or whether they are the result of an invasion of the tissues of the host by numerous larvae which have attained sexual differentiation and maturity witliin the seedling, the data available are not sufficient for one to say, since we do not know enough of the life- history of the parasite to say definitely which is the infective larval stage. The figure arrived at in each case gives us, I think it will be agreed, an expression of the suitability of the host plant for the needs of the parasite, and this is what the investigation was designed to reveal. 'W (JOODEY I'T Table 11. Sliowiiuj (lie four liighesl counts of male andfenude 'I'ylenchi iii seeiUiny.'i, the aceraije of llie loluh of Uiene giviiuj the index of suscepti- bility in each case. Name of plant Males Females Total Med liuLLr, Eu'Aiah Avci-ayc I!JO index of suseeptibilily Yin/ t/ui'tc, Freneli Average ... Ikil doixr. Willi En-lisli Averaii.- ... lied doixr. (.'anai-liau Avei'ane ... t'utc-tjia-'is, Enyliali Avei-aj,'C ... Co(v-'jnt-'i-:y, yu'ci-liali Average ... Alsike duccr, Eiinlisli Average ... AlsilcL clijixr, Uaiuuliaii. Average ... While ilorcr, .Sutton's Mainniotli Jiil Wliite dover, English ... Average ... ... ... 4-75 l(i:i i:tii ■2:vj !)l> lilt nil IH> II III iiiii OS !):5 KH I!JO 11(1 :2.'')l 1 :',iii) i;;i !l(i 1 .j',.i 11 HIS I5i !l i+il 14!) ■2or> ;:'' IIJ 1 17 1 <■) ■ 1 '. 1 1 -ill 141' I7J :{(t ll(i ui; Hill S7 :!17 1114 7-) ISO L'.")- (i.S L'lil :!:!!' Itii ISO 1'7(1 :5i(i 1 M 1 1 117 ■ 4 i; Ill .) !l 11 :5S-2r, v.> .Sll I2i.l .".;"> !is l.^:i :•'.) lu I2lj IIKI Hill 2(1(1 1117 III III L'd III 1 " J.") 1 ■> '!- 1 .1 1 _ U :i4 2.S-.-J IS :!ii 4S l(i 4li (12 7 17 24 n !l 14 37 ml i; 1 nil 1 o nil III (j 2 — 2 — 1 1 28 Stem-disease caused hif the Eehrorm 'I'ahlc II (continued). Name of |)laiit H'hik clover, W'ild (.'ejtswoM Avcnigf ... While cloicr, Willi KeiitiBli ■Malfs ~} 1 1 Females 7 7 1 Total S 2 2 1 3 5-5 index of Muxccptiliility 1 5 41 (17 lOH It 7.S 121' sr, 1:27 212 95 US 2i:i Ui-ilo i li • ) Id 7 ii 10 — 2 2 7 nil nil nil nil nil nil Average ... Kiilnci/ rclch Average Sainfoin ... Averajre ... Lucerne ... Trefoil BISCUSSIOX. Jn a hioloiiical iuvestitiation of this character it is practically im- possible to repeat the experiments exactly, and consequently one would not expect to get the same figures in a repeat experiment. Nevertheless, one would, I think, obtain figures having the same relative significance, and for this reason the figures put forward above as indices of susceptibility may be taken as representing approximately the relative susceptibilities of the dift'eront clovers, etc. to 7'//. dipsad attack. An examination of the figures shows that all the varieties of red clover tested are very susceptible to attack and fall into a common group (1) to which also belong cow-grass (Swedish) and kidney vetch. Arranged in order of intensity of susceptibility we have them as follows: lied clover (Canadian) ... (French) (English) Cow-grass (Swedish) Kidney vetch Red clover (Wild English) .•$ll> 'i 20.5 I 190 ,, , 163-75 I 100 T. GooDBY -29 Widely separated from this group we have another one (2) much less susceptible, comprising Cow -grass (English) ... 38-2", | AJsike clover (Canadian)... I!7 - f (roup 2. Alsike flovfi- (Englisli) ... iS-.", ) The great disparity between the figures for the Swedish and the English cow-grass is remarkable and somewhat surprising, and suggests tjie need for much further investigation along these lines. A further group (3) comprises varieties which are but very sliglitly susceptible to attack, viz. : Sainfoin ... ... ... 7 ) White clnvfM- (Wild Cot.swold) ry'y | ^, „„ , (English) ... 4-7.-. '"^""P-'- (Wild Kontish) 2 ) Lastly we have a group (4) made up of White rluver (Sutton's .Mammoth)) Lucerne [ I iroup 4. Trefoil ) which appear to be iusu.sceptible to attack. A comparison of the indices of susceptibility with the percentages of deformed seedlings shown in Table I reveals the fact that all the members of Group I have a high percentage of deformed seedhugs. Canadian red clover, having the highest index of susceptibihty, has also the highest percentage of deformed seedlings. The parallel does not hold throughout the group, however, for Wild EngHsh red clover, which has the second highest percentage of deformed seedlings, has the lowest index of susceptibihty in CTroup 1 . The low index of susceptibility shown by English cow-grass, compared with its high percentage of deformed seedlings, is an otitstanding excep- tion to the parallelism of high inde.x of susceptibility with high per- centage of deformed seedlings. In the case of Groups 2 and ?> it is not possible to establish a parallel when we have cases like English alsike clover with a lower perci'ntage of deformed seedlings I ban either English white clover or sainfoin, and a much higher index of susceptibihty than either of these. These results are in general agreement with those of Amos (1919, pp. 8 and 9), except as to one or two details. He never found sainfoin attacked, whilst I find that the seedlings are very slightly susceptible to attack. Tlie results have a practical bearing of considerable importance to the farmer whose land is infested with Tylenchun dipsaci, and whose red clover is therefore hable to attack from this parasite. '^>0 Stcni-diMase com^eff fi// the Eclworm If he wishes to avoid Stem disease lie should not sow red clover, fow-ijrass or alsike clover, but should make use of trefoil, lucerne, sainfoin or a larjje white clover, such as Sutton's Mammoth White. Recommendations similar to the above liavc been made before, notably by Amos (J919, p. 9), but they are so obviously the proper and hopeful lines to adopt that they will bear repetition and reiteration. It is no use attemptino; to fiet rid of the eelworm disease if red clover, cow- grass, alsike clover and kidney vetch are sown on infected land. RKrKRK.XCKS. (1) Amos, A. (191 it). Tlu: IJitliculties of growing lied f'lovor — Clover .Sieknes.s iiiul other Causes of Failure. Journ. Roy. Agr. Soc. Eng. 79, pp. 68-88. (2) Byars, \j. p. (1920). A Xematode Disease of Red Clover and SIrinvherrv in llie Pacific Nortli-we.it. I'liyloprithologi/, 10, No. 2. p. III. (:!) Cotton, A. I). (1020). Clover Stem Rot (Sclfrnliiiia Irifolionim ICrik.). Juxrt). Ministry of Aiji: 26. \o. 12. .March, pp. 1241-1244. (4) KiJHN, J. (18S1 ). Das Lii/.crn;ilehcii (Tylenrhii-i HiireiiMernii). Deiilsche hinthr. Prexxe, 8, p. ,S2. (5) IVIarcinowski, K. (1009). Parasitiseh und semiparasitiseh an I'llanzeii lebeudc Nematodeii. Arbe.iten mm der Kaiseriirlien liioh(jisclien Aiiiloll fiir Land- iiinl Forslimrt.schiiJ'1, 7, Heft 1. jip. 5()-()7. (()) Ormerod, E. .a. (1887-1899). lieporls on Injurious In.sect,i. (7) ]ilTZERLV-Bos, .1. (1892). L'.Vngiiillule de la Tige (Tylenchus ilenislalri.r Kidin), etc. Archives dn Musee Teyler, Haarlem, 3, p. 161 and p. .545. (8) Smith, R. H. (1919). A Preliminary Note concerning a Serious Nematode Di-sea-sc of Red Clover in tlie Xorth-western States. Journ. Econ. J'Jiit. 12, So. ti. pp. 460- 462. KA'PLANATION OF I'LATI-; I. Fitl. 1. Photograph nf pni-tinns ol jilanl iif icd clover suffernitr from Stem disease ea\iaed in' Tyhnclms dipsaci, showing; the cjiarat'teristieally swollen ir;if stalks ixntX stipules at .1 and B. Fill. 2. Photograph of a whole plant of red clover suffering from Foliage-rot caused by the fungus Hclerotiniii Irifolionim. The white mycelium oau he .seen spreading on to the surface of the .soil, whilst tlw eollap.sed nature of the leaf and leaf-slalks is well shown. [Received September Gth, 1021. JOURNAL OF AGRICULTURAL SCIENCE. Vol. XII. Part PLATE I Fig. I. Fig. 2. r4ENETIC STUDIES IN POTATOES; STERTLTTY. By R. N. SALAMAN, M.A., M.D.. Burh'n. Herts. AND J. W. LESLEY, M.A., Phi,vl Brcfdliii/ Innlihih\ Ctitiliridjir Universi///. (Witli Plato IL) Male sterility in potatoos, that is tlio ahsenco of pollon from the antliors, t" has been shown by one of us (8arlaiiuin(i)) to behave as a recessive to male fertility. The present paper is an account of some further experi- ments in the genetics of this character, particularly of crosses between two varieties in which the reciprocals give difTerent results. Both quantity and quaUty of pollen were estimated by methods similar to those used by 8alaman{i). Four empirical grades of quantity were employed, viz. "abundant." "medium," "small" and "very few grains." The quality of the pollen was determined by mounting a sample in water and examining microscopically. Under this treatment living and presumably healthy grains swell up, appear spherical and translu- cent, whilst the pores become prominent. Such grains are termed "good." Bad grains are generally smaller, irregular in outline, and often appear quite empty. Even the best samples of pollens which we have seen contain a con- siderable proportion of empty grains. The projiortion of good grains was determined approximately by counting. The Relation between Qitaltty and Qitantity of Pollen. .4 close correlation was found to exist between qiiantity and quality of pollen as indeed was observed by P]ast(2) and Salaman(i). In general the larger the quantity the higher the proportion of good grains. The constant presence of "abundant" pollen is a sure sign of a large propor- tion of good grains. Similarly the lowest grade of quantity — "very few grains" — is proof of the absence of good pollen. Of the intermediate grades "medium" quantity was generally associated with poor quality pollen. The grade which we describe as "small" rarely contained any good grains. Perhaps " medium " quantity is best counted as potentially ^ fertile, the "small" quantity as potentially sterile. 32 Genetic Studies in Potatoes: Sterilitfi lu some families the ditiiculty of classification was considi'iablc, and a satisfactory method of dealing with a quantitative and variable char- acter such as the present has yet to be discovered. In some few cases only the quantity of pollen was recorded but as a rule both quantity and quality were estimated on three or four different dates; especially was this so when flowering was continued over a long period (this varied in different seedlings from one to over fourteen weeks). Frequently the repeated observations both of quantity and quality of ))()ilen agreed closely in different flowers on the same plant but in some cases this was not so. Little is known of the factors causing such variations. The variation of quantity is chiefly to be ob.^erved in sterile or partially fertile plants, and may he, in part, a result of the irregular and delayed ripening of the anthers cliaracteristic of such ])laiits. We iiud the quality of pollen sometinu's deteriorates considerably after the flower has been open for five or si.\ days, but apart from that there are inherent differences in the degree of fertility. For instance, Edgecote Purple invariably shows abundant pollen of high quality and may be termed fully fertile. Edzell Blue on the other hand is somewhat variable in quantity, usually only "medium" and the quality is correspondingly inferior. We have not attempted to measure small (|uantitative differences but only lai'ge differences of a more obvious kind. The methods of esti- mation are, of course, appro.ximate and not suitable for accurate measure- ment, but are sufficiently effective in determining large differences. Rapidity of execution is also a ('onsideration when over three hundred microscopic examinations need to be made in the flowerinu' ])eriod. Material u.sed. Three cultivated varieties formed the basis of the breeding experi- ments. They are characf erised as follows : ( 1) E (I f/ecote Purple. The male organs are I'lillv fertile. Imvinu a.hnndant pollen of good quality of which at least 'J5 per cent, consists of ]ierfect grains. It forms berries and seed freely by luitural self-pollination. Our stock of this was brought by Professor Biffen Ironi \Viit>liii-c. where it maintains a. fairly high yield although grown year aftei' year withoni the usual renewal by means of Scotch or other seed. (2) Myatt's Ashleaf. The male organs are fully fertile like the pre- ceding. It likewise forms berries and .seed, self-setting freely. (3) Edzell Blue. The male organs are fertile but not fully so, being considerably less fertile than either of the two preceding varieties. Both R N. Salaman and J. W. Lesley 33 quantity and quality vary considerably, but as a rule it is "medium, ' containing about 5 per cent, f^ood grains. Forms berries but sparingly, such berries having fewer seeds than the preceding. Parchment bags and the usual precautions were used unless expressly noted to the contrary. The seed from each berry was kept separate, which served as a useful check when self-set or "natural" berries were used. More than 200 selfed seedhngs were raised from Edgecote Purple and many of these carried on to the second year by tubers. Twenty-five plants flowered which were recorded as follows : Abundant ... 20 Small ... ... 1 Medium ... 3 Very few grains 1 All of the nine examined qualitatively contained good pollen (see Table I). Two plants, one with "small" ([uantity, the other with "very few grains,"' were only once recorded and are of doubtful significance. Fer- tility, which has already been shown to be recessive in the potato, here probably breeds true in accordance with expectation. Five individuals formed self-set berries. Table I. Abundant Medium Small Per cent. good. < f^ Kdsecote Purple selfed ... 20 .3 Edzcll Blue selfed S (i Ednecote Purple x Myatt's AsUeaf 1!) :{ Myatt's Aslileaf x Edge- cote Purple i:{ Edzell Blue x Myatt's Asli- leaf 8 4 Edgecote Purple x Edzell Blue 4+ f 2 from Edgecote Purple X Edzell Blue 15 Edzell Blue x Edgecote Purple ... 13 10 1 n 72 ^ tC 0~ I- (i- 11 - 10- 0 I- 0- 11- 10- 0 1 1 ■A 4 2 1 2 1 3 1 .5 2 1 1 3 :! 10 I 1 1 1 34 7 1 8 18 2 5 3 4 (i Some 200 selfed seedlings from Myatt's Ashleaf were raised, mostly in 1921 which was a disastrous year for potato seedlings in this district. Of these only two flowered, both having "abundant" pollen of which some at least was certainly good. The evidence here is very meagre and merely shows that Myatt's Ashleaf gives fertile offspring when selfed. In the case of Edzell Blue only self-set or "natural" seed has been available notwithstanding that about 100 flowers were self fertilised Joum. of Agric. Sci. xii 3 34 Genetic Studies in rntiifnes: Sferilifif luulor baj;s. The seed from tlip natural berries was sowu separately ,i;ivin<^ 150 seedlings; but eight plants from one berry proved from other evidence to be the result of a cross and were therefore not included, otherwise the results are consistent and trustworthy. Twenty-one plants flowered and were classified as follows: Abundant ... 8 Small ... ... 3 Medium ... G Nil I The qualitative results are shown in Table I. It was noted that even the fertile plants, unlike those from Edgecote Purple selfed, were apt to vary considerably in different flowers and on different dates both in quantity and quality of pollen. Although itself fertile Edzell Blue selfed actually throws a large proportion of .sterile offspring, a result which contrasts strongly with that obtained from Edgecote Purple. A somewhat similar appearance of sterility is recorded by Salaman(i) in certain lines derived from a fertile Sutton's Flourball. We shall .submit an explanation after the much fuller evidence from crosses has been given. Two hundred seedlings were raised from the cross — Edgcote Purple x Myatt's Ashleaf in 1921. Of the F^ plants 22 flowered, consisting of Abundant ... 19 Medium ... 3 All were fertile and contained at least 20 per cent, of good grains. A similar number of seedlings were raised from the reciprocal cross — Myatt's Ashleaf ■: Edgcote Purple — of which 13 flowered. All were fully fertile. The results from this cross are at once intelligible and harmonise with the few selfed data available, fertility behaving as a recessive. The cross — Edzell Blue x Myatt"s Aslileaf — was raised in 1920, and grown on in 1921 when a further sowing was also made. Of the 163 plants grown 25 flowered and were recorded as: Abundant jiollen 8 Small ... ... 3 Medium jjoileu 4 Very few grains 10 As in the case of Edzell Blue selfed the quantity varied considerably in different records from the .same individual. Unfortunately the quality of pollen here was only examined in three cases (see Table I). It is clear, however, that a large proportion of sterile plants result from this cross between two fertiles. In 1919 the crosses — Edzell Bhie x Edgecote Purple and Edgecote Purple ;< Edzell Blue — were made. In the families raised in 1920 there R. N. Salaman and J. W. Lesley 35 was no significant difference eitlior in tlie number of seeds per berry or in the germination of the seed or in the mortahty among seedhngs. Neither flower nor tuber colour showed any significant difference. Two types of inflorescence occurred in these famihes. In the " simple '' type (see plate) the primary axis divides into several (usually two) secondary axes, each of which forms a scorpioid cyme. In the " compound "' type (see plate), however, the two or more secondary axes divide again, forming tertiary axes and some of these divide so that axes of the fourth or fifth order are formed each of the ultimate branches forming a scorpioid cyme as before. Similar types of inflorescence have been studied by Crane (8) in tomatoes. The proportions of these two types in the present case showed no significant difference in the reciprocal crosses. The proportion of plants which flowered in Edgecote Purple >: Edzell Blue, however, was nearly double that in the reciprocal family. (In 1920, 23 per cent, as against 13 per cent. ; in 1921, 72 per cent, as against 37 per cent.) Both families were equally vigorous. In the late summer when many of the plants had ceased to flower several seedlings in Edge- cote Purple :■' Edzell Blue were seen to bear berries, often a dozen or more on a plant. In the reciprocal cross, on the contrary, extremely few plants had berries and these in but small numbers^. It occurred to us that this difference might be due to a difference in pollens, as seed and berry production in potatoes is frequently determined by pollen (of. Stuart(3)). As a first step, all the available pollens were examined. In Edgecote Purple x Edzell Blue all the 18 plants tested had "abundant" pollen. This was in accord with the frequent presence of self-set berries in this family. In the reciprocal cross 1.5 plants were recorded where pollen was described as follows: Abundant ... 1 Small ... ... 8 Medium ... 1 Very few grains 5 Only one plant was "abundant" as against the entire 18 in the reciprocal family. A remarkable difference in the male organs of the reciprocal cross thus came to fight, a difference which is paralleled by previously observed difference in fruit and seed jiroduction. Both famihes were grown on by tubers in 1921 and a further sowing made of 1919 seed of the cross Edzell Blue x Edgecote Purple. ' The berries both from Edgecote Purple •: Edzell Blue and Edzell Blue y Edgecote Purple contained seed. The average germination capacity was 90 % and the lowest 38 %. We are indebted to Mr Saunders of the National Institute of Agricultural Botany for testing the seeds. 3—2 36 Genetic Studies in Potatoes; Steri/iti/ From the cross Edgecote Purple x Edzell Blue 41 plants flowered. Both quantity and quality of pollen were repeatedly tested; all were fully fertile. The constant production of "abundant" pollen of high quaUty was a characteristic of every plant. The results from separate berries were consistent and the observations of 1920 were therefore fully confirmed (see Table I). An F2 family from a fully fertile F^ plant from this cross was raised in 1921 and 15 plants flowered. All had "abundant" pollen of high (jiiality (sec Table I). Thus there is no evidence here of sterility arising either in ^1 or -Fj- In Edzell Blue < Edgecote Purple grown on 19 plants flowered; both quantity and quahty were again repeatedly observed with the following results : Abundant ... 13 Small 8 Medium ... 10 Very few grains 1 8 In the qualitative test (Table 1) only four of the plants showed any con- siderable proportion of good grains; the majority of jilants tested had no good grains. Again the results from separate berries were consistent. The difference in berry production which first attracted our attention was as remarkable in 1921 as in the previous year. In Edgecote Purple X Edzell Blue over 50 per cent, of the individuals bore self-set berries, often in the greatest profusion ; of those grown under more favourable con- ditions at Oruiskirk, actually 20 out of 28 (over 90 per cent.) bore berries. In the family Edzell Blue x Edgecote Purple only si.x plants bore self- set berries and here it is interesting to note the correlation between good quality pollen and the setting of natural berries. All of the four plants having good pollen had self-set berries, but only one of the sixteen plants having no good pollen had self-set berries and even these were probably due to insect-borne pollen. These figures give no idea of the contrast in quantity of berries per plant which was even more remarkable (see Plate). That there is no evidence of female sterility here is indicated by the presence of natural berries on the six plants and on two others which set seed when artificially crossed with fertile plants. It has therefore been shown that whilst Edgecote Purple x Edzell Blue gives a wholly fertile Fj^ the reciprocal cross gives some sterile and some fertile plants. R. N. Salaman and J. W. Lesley 37 Discussion. The two main facts to which we wish to draw attention are: (1) The appearance of sterile (or female) plants from the fertile variety, Edzell Blue, when selfed and from crosses between it and other fertiles, Myatt's Ashleaf and Edgecote Purple. (2) The different results obtained in F^ from a reciprocal cross. The evidence certainly points to a particular one of the three varieties as the principal agent introducing sterility. Both Edgecote Purple and Myatt's Ashleaf are themselves remarkably fertile. They have been crossed reciprocally and gave a wholly fertile Fj^; there is some evidence that when selfed they give fertile offspring. With Edzell Blue it is other- wise. It is itself only moderately fertile. Both selfed and crossed with Myatt's Ashleaf it gives a majority of sterile offspring. It is noteworthy that sterihty has arisen in both cases where Edzell Blue is the mother parent. It behaves as a sterile when used as female and as a fertile when used as a male parent. In plants male sterihty arising from a cross between fertile parents is, of course, a well-known occurrence, but in the present case sterility arises from a cross in one direction but not in the other. This difference in the reciprocal cross between Edzell Blue and Edge- cote Purple points to some difference in the eggs and pollen of one or both of the parents. The evidence seems to point strongly to such a difference in the sexual cells of Edzell Blue. It is not impossible that in Edzell Blue the sterihty is attached to the cytoplasm of the egg, but that the generative nucleus of the pollen grain carries the basis of fer- tihty. As sterihty normally behaves as a dominant we should then ex- pect all the F^ to be sterile where Edzell Blue is the mother. This, how- ever, was not so. A few were fertile, but we know of no evidence sug- gesting that the cytoplasm of eggs may differ in constitution. Indeed the heterogeneity of the eggs suggests segregation. It is more likely that a factor or factors are at work which are localised in the nucleus. As_sierility is dominant it would seem that the eggs of Edzell Blue are of two kinds, some — possibly half — carrying male sterihty, the re- mainder carrying male fertihty. The data at present available only give an approximate idea of the proportions. The pollen on the other hand all carries fertility. Following the same hypothesis both eggs and pollen of Edgecote Purple and Myatt's Ashleaf would appear to carry fertihty. In order to account for the difference between the eggs and pollen of Edzell Blue we suggest that at some stage of development a process 38 Genetic Studies hi Potatoes; Steriliti/ of segregation has occurred at which the basis of sterihty has dropped out of the hneage of the tuale germ cells. On the other hand the germ lineage of the eggs is unaffected so that male sterility is present in some of the egg cells and is absent from others ; this may well be merely the result of normal segregation. Possibly in the developing bud at the time when the male organs of the flower are being differentiated a differentia- tion of genetic characters takes place. This hypothesis affords an ex- planation of the rather paradoxical male fertility of Edzell Blue in spite of its containing the dominant element of male sterility. For as a result of the premature segregation the primordia of the antliers would only contain the basis of fertility and are therefore able to form good pollen. It would account for the fertility of t\ and F2 from the cross Edgecote Purple X Edzell Blue and for the fertihty of the reciprocal crosses be- tween Edgecote Purple x Myatt's Ashleaf. The F-^ result from the cross Edzell Blue x Myatt's Ashleaf which was similar to that of Edzell Blue X Edgecote Purple is also intelhgible on the above hypothesis. A diffi- culty is met however in the result from Edzell IMue selfed. According to the theory about half the eggs carry sterility and the other half fertility while all the pollen carries fertihty. Consequently selfing should give half sterile and half fertile offspring, i.e. the same result as Edzell Blue < Edgecote Purple. But the data (Table I) indicate an excess of fertiles. The numbers are small but the discrepancy demands further investiga- tion. Again, no explanation is afforded of the rather imperfect pollen production of Edzell I51ue. To sum up it appears that in male sterile varieties the eggs either all carry male sterility or some carry male sterility and the remainder male fertility. The eggs of male fertile varieties (except Edzell Blue) and the pollen of all male fertile varieties tested carry male fertility. As sterility behaves as a dominant the F^ from a cross, sterile x fertile, consists of all steriles or partly of .sterile and partly of fertile plants. Miss Saunders (4) in her classical experiments with Matthiula has shown that the pollen of certain "singles" is all of one kind, but that the eggs are of two different kinds with respect to the "doubling" factors. Other experiments (5) suggest that in Petunia the pollen is heterogeneous and the ovules homogeneous for a "singleness" factor. The present data point to an explanation of the former kind in potatoes. If we regard a male-sterile plant as a functional female it follows that the gametes of Edzell Blue are differentiated in regard to sex, the eggs carrying either Hermaphroditism (cj Fertility) or Femaleness (o Sterility) and the sperms all Hermaphroditism (cJ Fertility). JOURNAL OF AGRICULTURAL SCIENCE. Vol. XII. Part I. PLATE II Fig. 1 Fig. 2 Fig. 3 R. N. Salaman and J. W. Lesley 39 Tliat a process of differentiation jjrior to normal segregation may occur in plants is shown by the work of Bateson(6) and (7) and others. In the present case there is reason to suspect a segregation by which sterility is dropped out of the germ lineage of the sperms but not of the In concluding it is a pleasure to acknowledge the help and encourage- ment given us by Prof. R. H. Biffen and the willing assistance of Miss E. Hagger at Barley and Mr Hamilton at Cambridge. REFERENCES. (1) Salaman, R. N. Limi. 8oc. Jourii. Botany, 39, Oct. 1910. (2) East, E. M. Rep. of the Agronomist, 1907, Conn. Agr. Expt. Sta. (3) Stuart, W. U.S. Dept. of Agr. Bull. 195, \^15. (4) Saunders, E. R. Journ. of Genetics, 1, No. 4, 1911. (5) Journ. of Genetics, 1. No. 1, 1910. (6) Bateson. Journ. of Genetics, 5, No. 1, 1915. (7) Journ. of Genetics, 6, No. 2, 1916. (8) Crane, M. B. Journ. of Genetics, 5, No. 1, 1915. EXPLANATION OF PLATE II. Fig. 1. "Simple" cymose system of male-steiilo F^ jjlant from Edzell Blue x Edgeoote Purple. Fig. 2. " Compound" cymose system of male-sterile F^ plant from Edzell Blue x Edgeeote Purple showing variation from the less "compound" type at the apex to the more "compound" type at the base of the shoot. Fig. 3. Bunch of self -set berries developed from a compound mflcjrescence of male-fertile Fj^ plant from Edgeeote Purple ;■: Edzell Blue. (Received November 3rd, 1921.) A STUDY OF NITROGEN METABOLISM IN THE DAIRY COW. By CHARLES CliOWTHEH. .\1.A., rii.JJ. AND HERBERT ERNEST WOODMAN, D.Sc, Pii.l). The work outlined in the present communication icas carried out during the years 1916-19 in the Institute for Research in Animal Nutrition of the University of Leeds. In an earlier conunuiiicatioii' dealing with the results of a series of digestibility determinations carried out by us with two sheep, attention was directed to the fact that, w^hen the averages for consumption and excretion of nitrogen in the different periods were arranged in the order of increasing nitrogen consumption, the retention of nitrogen in the body of the sheep rose only up to a certain point and then fell. The essential data are reproduced below: Period Nature of ration 1 Hay I palm kernel oake I IV Hay alone II Hay + palm kernel cakes I and 1 1 V Hay + extracted palm k. meal ... Ill Hay + undec. cotton cake VI Hay + dried yeast It will be noted that each sheep showed the maximuni retention of nitrogen in Period II, in which about 12 gni. digestible nitrogen, or roughly 17 gm. total nitrogen (= 106 gm. total crude protein) per head per day (= 2-4 kg. total crude protein per 1000 kg. live-weight) were consumed. In the periods following this in the table, as the nitrogen- consumption increases the nitrogen-retention falls steadily. These results suggest that, judged by nitrogen-retention, there is an optimum point of protein supply, above or below which nitrogen-reten- tion is reduced. It was realised, however, that this series of observations could not be regarded as in any way conclusive on this point, in view of ' Journ. ofAgric. Sci. 8 (1917), 447. Average per day ' Nitroger L retained Nitrogen digested { -f ) or lost ( - ) by sheep 8heep 1 Sheep 2 Sheep 1 Sheep 5 gm. U-88 gm. 7-38 gm. -0-37 gm. 1-94 9-0!) 9-61 + 1-93 214 11-4(1 l-'-48 -H3-80 4-7G Vl-li 1312 + .3-33 2-1.') i2;n 12-9() ■f3-25 3-7(i 1810 18-28 + 2-94 l-S.'i C. Crowtfier and H. E. Woodman 41 the varying sources of protein-supply in the different periods, and the fact that the order in which tlie diverse rations were fed was not that of increasing nitrogen-content. It woukl be ei(ually legitimate, for example, to account for the reduced nitrogen-retention in Periods III and VI, by assuming the proteins of cottonseed cake and yeast respectively to be of lower metabolic efRciency than those of palm kernel cake. Both alternatives engaged our interest, as we were at the time con- templating a comprehensive study of the nutritional requirements of the dairy cow, in which, as is well-known, protein-supply plays a part of outstanding importance. It has been established that up to a certain point increased protein consumption leads to increased secretion of milk. Does this point coincide with that of maximum retention of nitro- gen in the body? Does the amount of food-protein recj[uired to produce the latter vary with the qualitative character of the ration? Is the maximum of nitrogen-retention constant for any one individual or variable according to the feeding? Before attempting to secure answers to these questions we decided first to test whether, by feeding cows on rations of the same qualitative composition, but of successively increasing protein-content, and determining tlie nitrogen-retention, we could detect an optimum of protein-supply, such as had been suggested by the records of the sheep experiments. Further, in order to avoid the complications that pregnancy and lactation would introduce, we selected as test animals two fully grown Shorthorn cows, not in calf, and not producing milk. FIRST EXPERIMENT. (196 days, November, 1916— June, 1917.) For reasons stated later this experiment cannot be regarded as en- tirely satisfactory and consequently need only be dealt with in outhne. The two cows were first brought approximately into nitrogenous equihbrium on a daily ration of 141b. "seeds" hay, supplying about 122 gm. nitrogen. After determining over a given period the daily balance between the amounts of nitrogen consumed and excreted, the ration was increased by adding 2 lb. maize meal per day (containing about 13 gm. nitrogen) and the effect on the nitrogen balance was then investigated. This was followed by several further experimental periods, in each of which the ration was increased by 2 lb. maize meal. The general arrangements for the experiment and the methods of analysis, etc., followed were similar to those described in detail later in this paper. 42 J Sfiidji of Nitroijen Metabolism, in tin Dairy Con- 111 all changes of feeding a transitional period of at least nine days was allowed, whilst the experimental period consisted of at least twelve days, giving a minimum period of three weeks between each change. The essential data obtained are summarised in Table T. Table I. Cow A. (Initial weight. W?,^ II).) Gain ( -^ ) or Daily ration Loss ( - ) of A Total N consumed per day N by cow per day ?criod Duration JI^ Maize days lb. lb. «m. gm. 1 24 14 0 12in - 8-2 11 32 14 2 1341 + 7-6 III 21 14 4 1.51-8 + 181 IV 21 14 6 1G8-4 + 20-6 \' 28 14 8 l()l-0 + 10-4 VI 21 16-5 ,S 172-4 + 11 \11 28 16-5 10 1 S.-)-7 + 60 VIII 21 16-5 12 201-9 + 17-8 IX .31 16-.5 1(1* l!)4-3 + 15-7 Cow B. (Ir litial weight, 959 lb.) 1 32 14 0 124-3 + 50 11 32 14 2 133-3 + 8-4 III 21 14 4 148-9 + 18-9 IV 21 14 6 1G4-3 + 20-4 V VI vu 28 14 8 161-4 + 18-1 28 14 10 170-3 + 21-0 VIII — — — — — * The concentrated food diirini; this period consisted of 91b. maize + 1 lb. linseed cake. The investigation proceeded quite smoothly up to Period V, when it became necessary to draw a fresh consignment of hay. This proved un- fortunately to be decidedly poorer in nitrogen than the previous con- signment, and conse(|uently, despite the added 2 lb. of maize, the nitro- gen consumption for Period V worked out rather less than for Period IV. In order to overcome this difficulty the basal ration of hay was increased for Period VI to 16-5 lb. per day, the maize meal being retained as in Period V at 81b. Unfortunately, although Cow A consumed the increased ration of hay satisfactorily, Cow B could not be induced to do so, and in consequence this cow had to be kept for the remainder of the experiment on the lower ration (14 lb.) of hay, so that from this point the records of the two cows were not strictly comparable. Further difficulty was experienced towards the end of the experiment in inducing the cows to consume the rations completely and it was ob- vious that we were approaching the limits of their ajipetitcs. This was the C. Crowthbr and H. E. Woodman 43 more marked with Cow B, and for this reason we were obliged to reject the records of this cow for Periods VI and VIII. If the data for nitrogen-retention be examined, it will be seen that with each cow up to Period IV there were distinct indications that the daily retention of nitrogen was approaching a maximum value and, but for the regrettable difficulty with the hay, it seems probable that this limit would have been reached in Period V, since in the later periods, despite increased intake of nitrogen, no higher retention than that of Period IV was recorded. A slight increase is indicated with Cow B in Period VII but it will be noted that the nitrogen intake in this case was practically the same as for Cow A in Period IV, so that the maximum retention of nitrogen under the conditions of experiment would seem to be associated with a nitrogen intake of roughly 170 gm. per day {= 1060 gm. crude protein). Taking the average weight of the cows for Period V this represents a daily intake of crude protein of 2-34 kg. per 1000 kg. live-weight, or almost exactly the same figure as was dedxiced for the sheep from the data obtained in the digestibility trials. This confirmation is particularly interesting in that the diets in the different periods of the sheep tests varied considerably in qualitative character, and the con- clusion suggests itself that, when fed along with hay, the proteins of maize, cottonseed, palm kernels and yeast are of equal value so far as capacity to effect protein storage in the body is concerned. It would appear therefore from these experiments that for the last- named purpose the optimum protein-supply in the food is in the neigh- bourhood of 2-4 A-(/. crude protein, or say hi kg. digestible crude protein per 1000 kg. live-weight. This, of course, appUes only to animals such as those under test which were not subject to any losses of nitrogen other than those voided in faeces and urine, and were free from any exceptional internal needs for protein such as arise during pregnancy. It will be noted from Table I that the duration of the different periods varied from 21 to 32 days. It was anticipated that a period of 21 days would afford ample time for the re-establishment of nitrogen equilibrium after the disturbance due to change of ration. It soon became evident, however, that this was not the case, and the records of Periods IV to VI bring this out clearly. For reasons already given the nitrogen-consump- tion remained relatively constant throughout these three periods, which may therefore be grouped together as one long period. Viewed in this light it is seen that the establishment of nitrogen equilibrium is a much slower process than we had assumed, the nitrogen-retention persisting, though at a steadily diminishing rate, throughout the whole interval of 44 A SftKfi/ of Xifrof/en Metabolmn in the JJnirij Cow 70 days. In tliis particular, therefore, our scheme of experiiuent would appear to have been defective in that the change of ration from period to period was evidently made before the influence of the earlier ration on nitrogen-retention had been exhausted. Accordingly, before proceeding further with the comprehensive programme of experimental work on these lines which lay before us, we thought it desirable to make a more prolonged study of nitrogen-retention under conditions of roughly con- .staiit nitrogen-consumption, and a further experiment was undertaken in November, 1917, and developed later into a study of nitrogen-retention during pregnancy and lactation, the whole experiment covering a period of two years. SECOND EXPERLAIEXT. (November, 1917 — November, 1919.) (icneral Plan of Experiment. Two Siiorthorn cows, C and D, similar to those used for the first experiment, both "dry" and not in calf, were weighed and placed on a daily ration consisting of 20 lb. "seeds" hay. The nitrogen-balance in both cows was followed continuously, except for the one week monthly during which the cows were weighed and an interval of 10 weeks at the end of the first year. After a period of 302 days (including 255 days for which the nitrogen-balance records were obtained) during which both cows received identical treatment. Cow D was put to the bull, whilst Cow C remained unserved as control. The nitrogen- balance measurements were continued for both cows tjiioughout the period of jjregnancy of Cow D and subsequent parturition, and were further extended well into the period of lactation. For Cow D the experi- ment extended over periods of 302 days "dry"' and not in calf, 2Sf davs "dry" but in calf, and 136 days in milk, a total of 722 day.s throughout the whole of which period Cow C was maintained dry and not in calf. Samples for nitrogen-balance determinations were taken on 54() days during this period. Both this and the foregoing experiment were carried out in a building specially designed for work of this character. The floor was of cement, graded and grooved to secure ra]>id drainage, each cow-stall being pro- vided with a separate drain, so arranged that the drainage could be easily collected for purposes of weighing and sampling. Preliminary tests showed that any liquid falling on the floors of the stalls could be recovered within 2 or 3 per cent., whilst by spraying the floors occasionally with C. Crowthbr and H. E. Woodman 45 water a still higher degree of completeness could, be obtained in the recovery of urinary nitrogen. E.rpni'mcntal detail.';. Collection, Sampling and Analysis of Faeces. The faeces were collected into weighed covered buckets. The 2-i-hours' collection was weighed, thoroughly mixed, and by quartering, reduced to a small sample. One- tenth of the daily faeces was preserved in a closed vessel with the addition of a little toluene. In this manner, composites were made up twice weekly, representing the collection over three and four days respec- tively. Preliminary tests had shown that the loss of nitrogen was inap- preciable when the faeces were stored in this manner over a period of several days. The composite samples were well mixed and small samples were drawn from the bulk for analysis. The nitrogen-content was deter- mined in triplicate by means of the Kjeldahl method. Collection, Sampling and Analysis of Urine. The urine was collected in weighed buckets, the drainage being assisted by frecjuent spraying of the floors with water. No litter was used in the stalls. The daily output of urine was weighed and, after thorough stirring, one-tenth of the total weight was ladled out, acidified with 20 c.c. H0SO4 (1:1) and preserved in a Winchester bottle. Composites were made up in this way twice weekly and the nitrogen-content of the samples was determined by the Kjeldahl method. Under the conditions of collection the faeces were always con- taminated with urine, and the urine was diluted to some extent by the water used in spraying. This was immaterial, however, as it was only required to measure the total voided nitrogen. Length of periods and 'weighing of cows. The jseriods were mainly of about 18 days' duration. At the end of each period the cows were weighed on three successive days under comparable conditions. Feeding of cows and analysis of foodstuffs. The hay was fed in four equal allowances during the day, care being taken that the cows had consumed the hay completely before being left for the night. Water was given ad lib. Composite samples of the feeds were made up by reserving weighed representative samples daily. Two independent hay composite samples were made up in each period, and the dry matter and nitrogen-contents of these were determined, the mean results for the two samples being used in the calculation of the nitrogen-balances. From time to time representative samples were made up from the period composites and complete analyses were carried out. Moisture deter- 46 A Study of yitriHjeii Metabolisni in the Dairy Cow minations were made daily on independent hay samples, since this factor was found to be sul)ject to consitleiable variation. Analysis of milk. Cow J) was milked twice daily during its period of lactation and half-weekly com])osites were made up. The sam])les were preserved by the addition of one or two drops of formalin. The deter- mination of the nitrogen in the milk samples was carried out in duplicate by the Kjeldahl method. The experiments were carried out at the Manor Farm, (iarfortii, (Experimental Farm of the University of Leeds and the Yorkshire Council for Agricultural Education). The care of the animals was in the experienced hands of Mr H. J. Hargraves, N.D.A., to whom we are greatly indebted for the skill and accuracy with which this important side of the work was carried out. Table II. Analysis of seeds hay and linseed cake composile. (Calculated on dry matter.) Seeds hay Days i-epresented by sample ... 1-08 69-194 195-295 296^99 500-660 661-722 Linseed cake Crude protein ... 12-46% 12-41% 11-08% 12-86% 10-27% 10-29% 3318% True protein ... 10-25 1103 9-75 11-45 8-73 9-17 23-86 Ether extract ... 1-52 1-95 1-79 1-65 1-65 1-55 10-15 Nitrogen-free extractives ... 50-23 45-68 51-50 48-50 49-23 49-92 38-95 Crude til)re 28-90 33-42 29-64 30-94 31-82 31-22 10-71 Ash 0-89 0-54 5-99 6-05 7-03 7-02 7-01 It will he noted that the composition of the hay, and in particular, its content of protein, were subject to considerable variation during the course of tlie trial. This rendered neces.sary, in the later stages of the experiment, a slight increase in the amount of hay fed per day, in order to maintain the amount of protein consumed per day at its initial level. Another factor disturbing to some extent and difficult to regulate was the great variation to which the moLsture-content of the hay was subject from day to day. For these reasons it was impossible to avoid an ap- preciable variation in the nitrogen-.supply without recourse to methods which the resources at our disposal did not permit. It will be seen from the second column of Table III what the extent of this variation was. Comments on Tabic III (First Main Period of Experiment). We cannot attempt to tabulate here in detail the whole of the data accumulated day by day during the trial. The figures given in Table 1 1 1 therefore represent daily averages for the difFerent periods, the nitrogen- balance columns giving the average daily gain or loss of nitrogen by the C. Crowther and H. E. Woodman 47 cows for each respective jjeriod. The weight columns give the gain or loss of weight of the cows over the whole period. Table III. Sunuiiurij of nitroyen-halances for Cow C and Can' D up to da// 295. Daily ration: 20 llj. seeds hay (average dry matter per day = 7600 gm. ). Cow C Cow D N con- Total N Mean ^ Total N Mean -^ sumed voided nitrogen- voided nitrogen - Change Period (average (average balance Change in (average balance in in days per day) per day) per day weight per day) per day weight gm. gm. g m. lb. gm. gm. lb. 1- 19 1360 130-1* + 5-9 -13| 132-4* + 3-0 -43 23- 47 1511 143-7 -1- 7-4 - 5* 139-7 -I-11-4 -1-15 52- 08 144-5 139-2 + 5-3 + llf 136-2 + 8-3 + lli 72- 8!) 152-9 1.50-3 j- 2-(i - ^ 147-7 -1- 5-2 - 1 93-110 151-8 1.50-2 4- 1-6 + 2 148-2 -f 3-0 + 121 1 14-131 148-0 151-2 _ 2-6 + 30^ 148-9 - 0-3 + e| 135-152 152-3 149- 1 + 3-2 .> 140-8 -f ,5-5 + 15 150-173 148-9 1.50-9 - 20 Tioi 145-4 -i- 3-5 -f253 177-194 147-3 147-S - 0-5 - 3| 142-5 + 4-8 -14j 198-215 105-3 1.55-6 + 9-7 + lo!i 151-8 ^-13-5 -1-19 219-230 141-8 1430 - 1-2 + 22| 141-0 + 0-8 + 33 J 240-257 136-6 135-0 -f 1-0 + 7 130-2 + 6-4 -flO 201-270 146-5 135-9 + 10-0 + 15§ 135-7 -flO-8 + 4 282-295 138-7 134- 1 + 4-6 + n 1310 + 7-7 + 24f Analytical Total N days consumed 2.55 37,016 gm. Condensed suni.tnanj. Total N retained Nett change of weight Cow D Cow C Cow D 571 gm. -1- 109 lb. + 118^ , lb. 1 milk (c-( JW.-5 not quite d] ■y at beginning of ( Cow C. 1: .'29 lb. 1 Cow D. 1: 260 lb. Cow C 800 gm. * Includes small amount of nitrogen tnn trial). Initial weights of eow^ According to current views on protein metabolism we might expect to find, on introducing a given ration supplying protein in excess of the requirements of the basal metabolism, that nitrogen storage took place at the outset, but that the rate of storage steadily fell until nitrogen- equilibrium was re-established. This process is clearly evident in the records of our two cows. Taking Cow C for example, and overlooking the first 20 days as preliminary, it will be seen that the nitrogen-retention steadily fell and nitrogenous equilibrium was ultimately roughly estab- lished, though not until the lapse of about 90 days. Subsequently, between days 93 and 194 the cow was in almost perfect nitrogen-equi- librium, but this was then disturbed by a rise in nitrogen-consumption due to an abrupt rise in the protein-content of the hay. The effect of 48 A Study o/yitrugen Metaholisin in t/ic Ddirij Cow this seems to have passed away ((uickly and between days 219 and 257 iiitiogeii-eqiiilibriuin again prevailed. Thus, but for the abnormal period of days 198-215, e(juilibrium was maintained over a period of 164 days. It was all the more surprising therefore in the subsequent periods from day 261 to day 295 to tind that nitrogen storage was again taking place, a phenomenon of which the explanation is not very ob^aous. It certainly cannot be attributed to more than a small extent to the comparatively small rise in nitrogen-consumption in the ])eriod of days 261-278. The record of Cow D is very similar for the first part of the period, the nitrogen-retention falling steadily until the 100th day or thereabouts when nitrogen-equilibrium was established. The jjeriod from the 114th to the 131st day shows an almost perfect nitrogen-balance, but subse- quently a curiously persistent small retention of nitrogen was recorded throughout the remaining 164 days, increasing appreciably, as in the case of Cow C, in the last stages of the period. Over the whole period, out of 37-6 kg. nitrogen consumed by each cow, Cow C retained 8(i() gm. and Cow D 1571 gm., whilst the gains in live-weight were 109 lb. and 118 lb. respectively. No clear correlation is evident between the nitrogen-balances and the weight changes. For instance, between days 114 and 131, Cow C showed a daily negalire nitrogen-balance of 2-6 gm. and an Increase in weight of 30j lb. for the period. Between days 72 and 89, however, where the cow f/ained 2-6 gm. of nitrogen daily, the weight of the cow suffered a loss of 3?5 lb. for tiie period. The extremely irregidar variation of the figures in the nitrogen- balance and weight columns, and the disconcerting rise in the nitrogen- retention after 240 days of relatively uniform feeding, show clearly the danger of placing too much reliance on conclusions drawn from the results of short-perjod experiments with cattle. A three-weeks' period in work of this character is obviously far too short. Comments on Table IV (Second Main Period of Experiment). Cow D was put to the bull on September 6, 1918 (day 303), and the measurements of the nitrogen-balances were suspended until November 8, 1918 (day 366), This break in the measurements was unfortunate, since, on resumption of the trials, distinct indications were obtained that the protein metabolism of Cow D had undergone a marked dis- turbance in the early stages of pregnancy. Up to day 295, Cow D had shown throughout a uniforndy higher positive nitrogen-balance than Cow C. On resumption, however, Cow D was found to be distinctly in C. Crowther and H. E. Woodman 49 nitrogen deficit, whereas Cow C still retained a positive nitrogen-balance. If the nitrogen-balances throughout w'ere plotted graphically against the days of duration of the trial, it would be found that the curves for the Table IV. Summarji of nilro(/en-balances during period of pregnancy of Cow D. (Days 36G-575.) Daily ration: Gorv C (to day 558): 20 lb. seeds hay* (average dry matter per day = 7340 gm.). Cow D (to day 407): 20 lb. seeds hay (average dry matter per day = 7250 gm.). (Day 408-562); 21 lb. seeds haj' (average dry matter per day = 7740 gin.). Cow C Cow D Period in days ,366-383 387-407 41(5-432 436-153 457-474 478^95 499-523 534-558§ 562-575 N con- sumed (average per day) gm. 143-7 144-6 1.36-8 138-9 1.50-2 152-6 149-1 145-2 146-6 Total N voided (average per day) gm. 132-8 131-8 129-9 1441 160-0 154-8 161-6 148-4 141-6 Mean daily nitrogen- balance gm. -1-10-9 4-12-8 -I- 6-9 - 5-2 - 9-8 - 2-2 -12-5 - 3-2 + 5-0 Change in weight fur period-l- ib. 4-32 - % + H -m + 5J -m - 3s -1-20 N con- Total N Mean sumed voided daily (average (average nitrogen - per day) per day) balance gm. 143-7 144-6 143-6 14.5-8 1.57-7 160-2 156-6 145-6 149-4 gm. 148-6 145-7 144-5 143-8 15.5-9 1.5.5-1 154-7 140-8 131-4 :m. 4-9 1-1 0-9 20 1-8 5-1 1-9 4-8 -1-18-0 Change in weight for periodj lb. -13| -Hl2f + 25i -F21i -1-19 ■t-22 + 26i 4-104 •fl5S Condensed summary. CowC CowD Number of analytical days 174 428 Total N consumed gm. 25,309 62,925 Total N retained or lost gm. - 50 -t-816 Nett change in live- weight lb. -18J 4-9111 Total N consumed gm. 24,019 Total N retained or lost gm. 4-386 Nett change in live- weight lb. 4-139-4 * From day 562 started feeding new consignment of hay. This proved poorer in nitro- gen and ration for days 562 to 575 was consequently altered, and average consumption as follows: Cow C: 21-4 lb. hay 4- -43 lb. linseed cake ( =87.50 gm. dry matter per day). Cow D: 21-8 lb. hay 4- -43 lb. linseed cake ( =8920 gm. dry matter per day). t Weight of Cow C at beginning of period, 1378 lb. X Weight of Cow D at beginning of period, 1407 lb. § The records for Cow D in this period relate only to days 534-544. This cow was withdrawn from trial between days 544 and 558 on a"count of slightly swollen hocks. II Does not include weight changes from August 29 to October 31, 1919, during which period no determinations of nitrogen-balance were made. two cows intersect shortly after service of Cow D ; previous to this, the curves run approximately parallel. Thus, on a diet more than sufficient for its protein requirements in the ante-pregnant period, Cow D, as a Journ. of Agric. Sci. xii 4 50 A Shiflfi of Xifroficn Metaho/iniii in tli< Dah-ij Coir result of changos in the early stages of pregnancy, was suffering a loss of protein. This behaviour is in accord with the findings in other investigations carried out on different species. Murlin^ investigated the weekly nitrogen- balances in a ]jregnant bitch and showed that there was a large loss of maternal protein, commencing immediately after conception and continuing for six weeks. Only during the last two weeks before parturi- tion was there a marked conservation of protein, as manifested in the pronounced nitrogen-retention. Murlin attributed the destruction of maternal protoplasm which accompanies the development of the foetus to the necessity for providing "hereditary building stones for the laying down of the youthful protoplasm in accordance with the type character- istic of the species." The diet of Cow D was increased by 1 lb. seeds hay per day on day 408 and later again on day 502 (see Table IV). This slight increase enabled the cow to begin to retain nitrogen. Reference to Tables \\ and V shows that the rate of storage, however, was not very considerable until within three or four weeks of partTirition. During this time, al- though the amount of nitrogen excreted in tlic faeces remained roughly at its former level, yet tiie nitrogen ajjpearing in the urine underwent a marked diminution. This was a consequence of protein-retention. Ex- amining the period as a whole, it would appear that the demands made on the food protein for the single purpose of foetal development were relatively small, since the average rate of nitrogen-retention was only about 2-4 gm. per day. The behaviour of Cow C, dry and not in calf, during this period con- trasted curiously with its behaviour in the preceding period, when it was able to make a distinct gain of protein and body-weight. In this period, though receiving approximately the same ration, Cow C suffered a slight loss of protein and its weight dropped about 18 lb. These results serve to illustrate further the uncertainty of short period work with animals. Comments on Table V (Third Main Period of the Experiment). In Table V are given the full experimental details for the period im- mediately preceding and following the calving of Cow D. Particulars of mUk-yield, etc., for this and the following period will be found in Table VIII. For some days after parturition, each day's collections of urine, faeces and milk from Cow D were analysed separately. ' A'meriran Journal of Phii.noloqii, 1910, xxvii 177. C. Crowther and H. E. Woodman 51 Table V. Nitrogen-balances for Cow D in period of parlurition. Daily ration: 21-5 lb. of seeds hay + linseed cake (for amounts on difi'crcnt days see Table). Consumed Nitrogen voided .V . Daily nitrogen- Days of Lb. of Total dry % In In In period cake matter Total N faeces urine milk Total balance Sia. gm. gm. gm. gm. gm. gm. 580-581) 1* 9,104 141S 76-1 .50-3 126-4 + 15-4 587 1 9,174 142-5 — . 1.58-6t 51-11 481-5 ( - 99-5 588 1 8,973 1400 143-7 40-4 87-7) 1 - 99-5 (2 davs) 589 1 9.033 140-8 90-0 61-1 100 0 251-1 -110-3 590 8 11,826 280-3 98-9 88-2 105-7 292-8 - 12-5 591 8 11,832 280-4 94-6 88-8 76-2 259-6 + 20-8 692 6 10,960 239-5 114-5 111-1 79-5 30.5-1 - 65-6 593 6 11,024 240-4 129-3 99-3 81-7 310-3 - 69-9 594 6 11,010 248-1 91-4 125-5 78-2 29.5-1 - 47-0 595 8 11,718 286-4 100-8 108-2 78-4 287-4 - 1-0 596 8 11,730 286-6 101-9 102-1 76-5 280-5 + 6-1 597 8 11,830 288-1 103-4 lOS-0 74-8 286-2 + 1-9 598-603 8 11,715 286-5 111-6 111-2 69-2 292-0 - 5-5 Average from part urition 252-5 102-2 103-5 76-8 282-5 - 30-0 Average including N of calf and afterbirth ... 91-8 ... 252-5 _ _ _ 344.3 * 1 lb. linseed cake contained 399-0 gm. dry matter and 19-93 gm. N. I This figure gives N of urine + fluids + washings. Loss in weight of Cow D durmg this period was 161 lb. Cow D calved on the morning of June 17th, 1919 (day 587), the period of gestation being 284 days. The weight of the calf was 84 lb. Estimating the protein-content of the calf at 16 per cent.^, its nitrogen- content would be 975 gm. The placental fluid, urine and washings were collected together and analysed. The placenta weighed 11-5 lb. and had a nitrogen-content of 74-6 gm. {= 1-43 per cent. N in placenta). Taking the figure 2-4gm. as representing the rate of storage of nitrogen per day during the period of i^regnancy, it follows that the total storage of nitrogen by Cow D during the period of gestation was apjjroximately : Up to day 579 ... ... 665 gm. From day 580 to parturition 108 ,, Total 773 „ The losses of nitrogen during parturition were approximately: In calf ... ... ... 975 gm. In placenta, etc. Total 75 1050 Vide Armsby, The Nutrition of Farm Animals, p. 62. 4—2 52 A Studji of Nitrogen Metabolism in tin' Ihih-ij Cmr This figure represents a iiiiniimiiu estimate, since it does not include the nitrogen of the fluids, wliich were collected with the urine and not analysed separately. Thus, whereas 1050 gni. nitrogen had been stored as tissue in the form of calf and placenta. Cow D had been able, on a ration only very slightly heavier than that which served for its protein requirements when not in calf, to obtain 773 gm. from its food for this purpose. The deficit of 277 gm. must have been supplied from the maternal protein at the average rate of about 1 gm. per day. A cow in calf is thus able to maintain a positive nitrogen-balance on a ration differing only slightly from that requisite for nitrogenous equi- librium in the ante-pregnant period, though under such circumstances, it may have to supply from its own protein a fraction of the nitrogen necessary for foetal growth, and thus be unable to build up the desirable reserves during pregnancy to enable it to come into lactation with a good milk flow ca])able of being sustained over a long period. More generous feeding than was given in this investigation is obviously necessary in actual farm practice, especially in view of the heavy negative nitrogen-balances shown by Cow D in the days following parturition. For the first three days after calving it was thought inadvisable to increase the ration, but on the fourth day (day 590) the allowance of linseed cake was raised to 8 lb. This immediately checked the loss of nitrogen from the body, but as the cow at this stage could barely con- sume the ration, the allowance of linseed cake was reduced two days later (day 592) to 6 lb., whereupon a marked nitrogen deficit was again established. On day 595 the allowance of linseed cake was again raised to 8 lb. and nitrogen-equilibrium was practically restored, this changing to nitrogen-retention when on day G08 a further 1 lb. of cake was given {vide Table VI). For the period covered by Table V Cow C, dry and not in calf, was again in nitrogen deficit (see below) following a short period of nitrogen storage from day 502 to 575. N consumed N voided Mean daily Change in (average (average nitrogpn- weight for Daily ration (days 580-603) per day) per day) balance period gm. gm. gm. lb. 215 lb. hay + 1 lb. linseed cake 143-8 ISoO - U-8 +2i Comments on Table VI (Fourth Main Period of E.xperiment). It will be noted from Tables V and VI that a positive nitrogen- balance was not estabUshed in Cow D until three weeks after parturition. It was then consuming daily aiiout .3.30 gm. nitrogen and was yielding about 27 lb. of milk, containing roughly GO gm. N per day. That the C. f*ROWTHER AND H. E. WoODMAN 53 ration was then ample for the extra requirements of milk production was evidenced by the consistent positive nitrogen- balances recorded. Towards the end of the trial, when the daily yield of milk had fallen below 20 lb. (containing about 48 gm. N) the amount of cake fed daily was reduced to 5 lb. This resulted in the estabUshment of a decided negative nitrogen-balance. Summary of nitrogen-balaii Daily ration; 21-5 lb. seeds Table VI. ces for Cow D during period of lactation. hay + linseed cake* (for amounts see Table). Consumed N voided (average per day) (average per day) Mean Change Days daily \ in of Lb. of Total dry Total In In In nitrogen - ■weight period cake matter N faeces urine milk Total balance period gm. gm. gm. gm. gm. gm. gm. lb. 608-624 9 12,013 329-9 107-2 153-8 .59-9 320-9 -1- 9-0 - f 629-645 9 12,0.39 3499 109-1 169-2 58-9 337-2 -I-12-7 •f 15J 650-6.59 10 12,440 340-5 112-8 161-4 57-4 331-6 •f 8-9 - If 66.5-680 ot 11,952 323-5 111-9 143-4 .53-6 308-9 + 14-6 -33f 685-701 / 11,1.53 287-7 106-2 124-6 54-7 285-5 + 2-2 -1- 5 706-722 .J 10,370 241-3 97-4 106-5 48-7 252-6 -11-3 - 2J ( Condensed Summary f or Cow D. Analytical days Total N consumed Total N retained Nett change i in weight gm gm. lb. 608-624 29,131 -1- 536-8 -18 546 96,051 -1-2091 + 78H 546 96,051 -f 10411) * 1 lb. cake contained 400 gm. dry matter and 21 gra. nitrogen (approx. average for whole lactation period). t Amount of cake reduced to 9 lb. on day 068. t See note ([!) Table IV. § Including N of calf and placenta (estimated at minimum of 1050 gm.). During this period of about 90 analytical days in the lactation period, Cow D actually retained about 537 gm. of nitrogen and only lost 18 lb. body-weight. For the whole trial of 546 experimental days, it will be seen that this cow consumed about 96,000 gm. nitrogen and of this retained 1040 gm. {i.e. slightly over 1 per cent.), this being roughly the amount of nitrogen absorbed in foetal development and lost from the body at parturition. It was originally intended to continue the experiment through the whole period of lactation, but circumstances rendered this impracticable. It is of interest to compare in the case of Cow D for the peiiod sub- sequent to calving (Tables V, VI) the consumption of nitrogen over and above "equihbrium requirements" with the amount of nitrogen secreted in the milk. For this purpose we may take the "equilibrium require- 54 A Study of Nitrogen Mctaholis, in ill tin Dalrif Cow ment," judged 1 )\ the previous records of the cows, at roundly 145 gm. per day. Surplus X consumed (total Nitrogen- N-145gm.) N in milk Ratio of balance Day (average (average surplus N (average number per day*) per day*) to milk N per day) gm. gm. gm. 590-591 135 91 1-5: 1 + 4-1 592-594 98 SO 1-2: 1 -60-9 595-603 142 li 20: 1 - 2-9 608-645 195 .-)!) 3-3: 1 + 10-8 650-659 195 57 3-4: 1 + 8-9 665-680 178 .'")4 3-3: 1 + 14-6 685-701 143 55 26: 1 + 2-2 706-722 96 # 49 To nearest gm. 20:1 -11-3 It is clearly evident why during days 592-594 a heavy nitrogen deficit w'as recorded, since the "surplus" nitrogen, even without allow- ance for digestibility, was barely greater than the nitrogen removed from the body in the milk. Even in the period, days 595-603, when the "surplus" nitrogen was roughly twice that secreted in the milk, equi- librium was barely established, and again later, in the period, daj's 706-722, when a similar proportion prevailed, nitrogen deficit again set in. In the intervening jjeriods when the proportion was well over 3:1a substantial nitrogen-retention was effected. It would appear therefore that in order to prevent loss of nitrogen from the body of the lactating cow the "surplus" nitrogen must amount to well over twice the amount of nitrogen secreted in the milk. This is a condition which may well be difficult to satisfy in practice. Cow D was only a moderate milker (about 28 lb. milk daily at maximum), and yet it required the heavy and highly nitrogenous rations indicated to prevent loss of nitrogen from the body. With the more liberal milk-flow so often achieved with the modern dairy cow it would be even more difficult to secure nitrogen-equihbrium, if not indeed beyond the food-consuming capacity of the cow. This is fully in accord with practical experience of the difficulty of maintaining the "condition" of the milch cow in the early stages of lactation. The full details for Cow C (dry and not in calf) are not given for this period, as they present no new features of interest. The establishment of a po.sitive nitrogen balance is noteworthy, since on the same ration in the preceding period, a decided negative balance was recorded. An examination of the figures for the complete trial shows, however, that Cow C had been able to store about 1000 gm. of nitrogen from the 81,000 gm. consumed. C C*ROWTHKR AND H. E. WoODMAN 55 Table VII. Condensed summary for Cow C between the days 608 and 722. Daily ration: 21-5 lb. seeds hay + 1 lb. linseed cake. Analytical Total N Total N Nett change days consumed retained of weight lb. + 13f + 107 94 546 Total N consumed gm. 14,039 81,015 gm. + 472 + 1005 Days from calving 1 2 3 4 5 6 7 • 8 9 10 11 12-14* 15-17 22-24 25-28 29-31 32-35 36-38 43-45 46-49 50-52 Daily yield lb. 3-62 8-75 20-09 28-87 25-56 27-91 28-56 28-93 29-71 29-19 29-87 29-56 28-25 26-79 26-85 27-36 27-21 27-35 26-72 26-75 26-14 Table VIII. Mtll- records of Cow D. N /o 3-110 2-210 1-097 0-807 0-657 0-628 0-631 0-596 0-582 0-578 0-552 0-525 0-531 0-507 0-494 0-472 0-477 0-490 0-479 0-498 0-498 N in milk gm. .51-07 87-73 KMXKI 105-711 76-17 79-52 81-74 78-22 78-44 76-.52 74-80 70-40 68-07 61-63 60-20 58-57 58-88 60-80 5813 00-45 59-07 Days from calving 53- 56 57- 59 64- 66 67- 70 71- 73 79- 81 82- 84 85- 87 88- 91 92- 94 99-101 102-105 106-108 1U9-112 113-115 120-122 123-126 127-129 130-133 1.34-136 Daily yield lb. 26-34 25-75 24-80 24-95 25-77 23-25 22-70 22-49 22-27 22-32 21-62 22-50 20-92 22-44 21-66 19-37 19-50 18-,59 19-28 18-. 54 N % 0-499 0-497 0-485 0-511 0-510 0-509 0-511 0-534 0-523 0-529 0-555 0-547 0-5.56 0-545 0-552 0-570 0-558 0-569 0-549 0-575 N in milk gin. 59-58 56-75 54-57 57-83 59-63 53-70 52-40 54-50 53-87 53-57 54-40 .55-85 .52-73 55-50 ,54-17 .50-07 49-35 48-00 48-03 48-37 From analysis of composite milk samples. SUMMARY. This communication deals with the results of two experiments in which the "nitrogen-balance," or difference betweezi nitrogen-consump- tion and nitrogen-excretion by the cow, has been studied over prolonged periods, starting with the "dry" cow, not in calf, proceeding through the whole period of pregnancy and well into the period of active lactation. Experiment 1 was performed throughout with the two cows, " dry "' and not in calf, receiving a basal ration of hay to which was added in- creasing amounts of maize meal with a view to securing a progressively increasing consumption of nitrogen. This experiment lasted 196 days, during which period determinations of the nitrogen- balance were made on 90 days. Experiment 2 was carried out similarly with two cows, and covered a period of 722 days, including 546 days on which determinations of the nitrogen-balance were made. Throughout the whole of this period one 56 A Study of Nitrogen Metabolism i)i tlie Dalrij Cov cow (Cow C) was maintained "dry" and not in calf, as control cow, whilst the other cow (Cow D) after 302 days became pregnant and its record was followed throughout the stages of pregnancy and parturition and for the first 13C days of active lactation. The outstanding features of the results are as follows: (1) With the progressive increase of nitrogen-consumption beyond the fundamental requirements of the dry cow the rate of nitrogen-reten- tion steadily increases to a maximum and then falls. The maximum appears to be attained under the con()0 347 107 391 294 200 155 The final condition of the F.^ generation is shown by Tal)]!' II. Of 829 plants in the autumn-sown portion, 202 remained rust-free through- out the season, while the remaining (i27 plants showed evidence of more or less susceptibility. This is a close approximation to the 3 : 1 Mendelian ratio. In the spring-sown portion, it will be obsen^ed, the proportion of completely rust-free individuals was considerably less tiian one-fourth of the total number, but, on the other hand, the proportion of plants bearing only traces of rust was much higher than in the autumn-sown crop. The whole F^ generation contained 1213 rusted and 347 rust-free in- dividuals. If the badly rusted plants (giades 3 and 4) be separated from those that wei(( lijss severely attacked, we obtain the following totals: Attack, bad to very severe 361 plants Attack, moderate or less 852 plants Entirely free from rust 347 plants While these figures do not show a very rlose approximation to the simple Mendelian ratio, they are certaiidy very suggestive of it. Fuither consideration of these figures is deferred until the later results have been given. S. F. Armstrong 67 (3) The F, cultures. 1919. Since it was impossible to deal with sufficiently large F^ cultures raised from all the Fn plants, a number of the latter were chosen so as to include individuals showing every grade of attack, as well as a number of those that had remained rust-free. Altogether 198 F.^ plants were taken, and the grain sown in tiie autumn of 1918. (A) Climatic rondi/io)}f:, nidiuiring, avd qrnuih nf Ihr niltiirex in 1919. Monthly records of the weather are given in Table T. The season proved a most trying one for the plants, there being three prolonged periods of extreme chmatic conditions. The fir.st of these was from January IStJi to February 13th, during which the ground was frozen and snow-covered. So unfavourable was the spring weather that as late as the beginning of May the cultures were still in a very backward condition. Before they had sufficient time to recover they were faced by a period of drought which lasted from May 13th until June 19th. Throughout this period of 38 days the average daily sunshine was 10-1 hours, but the total rainfall only 6-2 mm. Indeed the total fall of rain from May 1st until June 19th amounted to onlv 8-9 mm., i.e. 0-35 inch during .50 days. On May 27th a top-dressing of nitrate of soda was apphed to all the cultures, at the rate of 4 cwt. per acre. It was hoped that this heavy apphcation would counteract the drought effects to some extent as soon as rain fell. However, rain did not fall in sufficient cjuantitv to carrv tlie nitrate into the soil until 23 days later. By June i7th it appeared prob- able that most of the cultures would be ruined, as many of the plants were in a critical condition and the basal foHage was dying off. Three days later (June 20th) a heavy fall of rain carried the nitrate to the roots. The change was sudden and most marked, for by June 23rd most of the plants showed a notable recovery, and soon afterwards they began to send up numerous side tillers. The period of drought referred to was immediately followed by a lengthy period of the very oppo.site conditions. From June 20th until the end of July the weather was cool, dull, and moist. During this time (42 days) 4-4 inches of rain fell, but the average daily sunshine was only 3-6 hours. One of the general effects of this weather — combined with the action of the nitrate — was to delay greatly the ripening of the plants. 68 Mendellaii Inheritance and Yellov Rust in Win ni (15) Spread of ru.sl during 1!U9. Yellow Rust W.1S observed for the first time on May 8tli, a luontli later than in tlie previous year. During May, and up to about the 20th of June, its s])rea(l was very slow, except in the case of very susceptible varieties and cultures. This difference in regard to the "|ieriod of in- fection" was very remarkable. For example, a variety of wheat which may be referred to as "B"" was grown alongside tlie f\ cultures, and proved to be extremely susceptible to Yellow Rust. Se\'eral of the plants were rusted by May 16th, i.e. fairly early in the period of drought. By May 24th almost every plant bore numerous pustules, and on the 31st of May, out of 136 plants in the plot, 135 were excessively rusted. A large plot of another variety (Sudanese wheat) was grown in tlie neighbourhood, and this was also attacked very early by Yellow Rust. This attack was so severe that, by the middle of .Tune, all the plants were prematurely destroyed, and no grain was formed. These cases are here referi'ed to because the attacks developed with great intensity durivg n period of drought, and afforded a striking contrast to the scarcely perceptible attacks made at the same time on moderately susceptible varieties. They demonstrate that the more susceptible varieties are liable to an earlier successful attack than less susceptible kinds; and also that, on very susceptible wheats at least, Yellow Hu.st is quite capable of making rapid progress in the tissues of tlie liost during the hottest weather we are likely to experience in this country. Similar differences as regai-ds the period of infection were observed among the Fg cultures. By June 14th rust was spreading rapidly on all those cultures which finally proved to be pure susceptibles; on the others, there was either no rust at all, or its progress was very slow (see Table XI). In order that all the plants might be exposed to an equal chance of infection, on the evening of June 13th all the cultures were sprayed with water teeming with fresh uredospores of Yellow Rust obtained from a neighbouring crop. This procedure was {>robal)ly quite unnecessary, since the pure susceptible cultures (already containing many attacked plants) were well scattered over the entire experimental area. Almo.st immediately after the fall of rain on June 2()th, a distinct increase in the rate of rust-spread was noticed, and during the cool, moist weather which prevailed in July tiie attack developed into an epidemic of llie greatest severitv the writei' has ever seen. 8. F. Armstron« 6i> ((J) Re.'iidls obtained from (he t\ cidtures (191'J). (a) General. Tlie spring drought proved a very severe test, for all the cultures, and some ol them, which were on rather poorer ground, had to be abandoned. Ample material was provided by the remaining 170 cultures which contained some 8500 plants. These cultures were examined in regular order, and the first appearance of rust on each noted. Immediately an attack was observ^ed, all the plants were numbered, and the rusted ones indicated in a book. The cultures were repeatedly ex- amined, from the first week in May until the end of .luly or later, to see to what extent the attack had spread. At. the last examination the intensity of attack on each plant was noted as in 1918. The conditions under which the plants were grown proved so favour- able to rust attack tliat. almost without exception, the cultui'es were more severely rusted than their F^ parents Jiad been in the jjrevious year. This is indicated to some extent by the results given in Tables III to VII. Cultures raised from badly rusted F., plants were still more severely attacked in 1919. It wiU be seen also that, in the 69 cultures which gave evidence of segregation, very few plants actually remained rust-free. Table III. Results of analt/sis of the F^ cultures (1919) grown f 7-0 in rust-free F.2 plants. Extent of the rust attack at final e.vaniination Plants (28th July to 11th August) Culture No. in culture None Trace flight M oderate Bad Very severe 66/9 49 49 — — — — 11/21 28 28 — — — — — 2/10 45 44 1 — — — — 19/12 46 41) 6 — — — — 2/18 82/14 8/17 68/19 29/1.5 41 25 60 42 43 27 1 25 13 111 (1 14 (1 (i II 12 10 20 18 11 .1 15 4 14 1 — 57 /S 62 1) 11 41 10 — — Totals 441 243 38 112 45 3 — 26/22 28/24 3/1 15/5 7 42 51 34 0 0 7 0 1 7 II 1 4 ■*."> 19 18 o 8 25 8 2 7 — 24/3 19 1 I.I 4 11 3 — 25/9 29 11 111 o 22/27 30 — — — 11 14 ,5 {b) Cultures raised from rust-free F^ plants. Seventeen of these cultures were grown, but in only two of theni did every plant remain absolutely free from attack. Of these seventeen cultures, the first ten in Table III 70 Meiideliaii Inheritance and Yellow Rust in Wheat suffered least from drought, and their growth was almost normal; if wc limit our attention for the moment to these we note the following facts: (1) they all remained perfectly free from rust up to about July 8th (see Table XI); (2) cultures 66/9 and 11/21 were entirely rust-free until maturity; and (3) the other eight cultures each contained some plants that were attacked to various extents. In most cases these attacks were of a shght character only, though three plants out of a total of 364 had actually bad attacks. Careful observation showed that the parasite did not thrive on the plants in tliese cultures, except in the rare cases just noted. The pustules, though often numerous, were abnormally small, and a large proportion of them failed to burst the epidermis, or only did so as the foliage was beginning to shrivel up. Owing to the difficulty in making due allowance tor tiie milder nature of the attack, as distinct from the relative number of pustules borne, it is probable that the sus- ceptibihty of some of these plants has been overstated by the grades accorded to them. Moreover, a comparison of these cultures with the homozygous susceptible ones growing alongside showed that the rekUive difference in the extent of rust attack was as great as had existed between their respective F., parents in the previous season. Cultures 26/22, 28/24, 3/1, 15/5, and 24/3 are grouped by themselves in Table III because they were much affected by drought, and their growth was far from normal. Although linally rusted to a greater extent than the first ten cultures, hke them they remained uninfected till as late as July 8th, and the evidence strongly suggests that they were also derived from "genetically immune" F.,'s. It was obvious that the extent to which the power of resistance was disturbed varied largely from culture to culture, and also among plants in the same culture. Further, the greatest "disturbance" occurred in those cultures that were most affected by drought, and possibly this may also partly explain why the attack varied so widely on ilift'erent individuals. Although it is impossible to state the exact extent of this ■' disturbance, '" from the evidence obtained it seems safe to conclude that it was sufficient to allow "genetically iniinuiie" plants to become either slightly or in some cases moderately attacked. This must be borne in mind when considering the remaining Fj results. Biffen(3, p. 123) found in his experiments that the immune F., jilants bred true to that character; and in the present case, when all the experi- mental evidence was taken into consideration, the only conclusion one could arrive at was that these fifteen cultures were "immune" in the genetic sense, although in uiauy instances the plant's power to resist 8. F. Ahmstkong 71 attack had been nioditied or partially broken down. This greater or le.s.s degree of predisposition to attack was probably due to the interaction of other causes or "factors '" such as are discussed later under section III. The last two cultures in Table III (25/9 and 22/27) were attacked considerably beyond the average, thougli not nearly so severely as the pure susceptible cultures were. They were probably the progeny of heterozygous susceptible plants that escaped infection in 1918. (c) Twelve cultures (Table IV) were raised from F.^^ plants which showed only traces of attack. Three of these proved highly resistant and compared closely in this respect with the iirst fifteen cultures given in Table III. One of these (82/9) was further tested on a large scale; forty-seven F^ cultures were grown containing some 850 plants, all of which proved highly resistant in 1920. It is therefore reasonable to conclude that the two cultures, 31/23 and 70/13, were also the offspring of homozygous "immune" plants, although these plants showed traces of attack in 1918. Table IV. Results of amdysis of F-^ riiUares (1919) grown from F^ plants whicJi had onlij traces of rust attack in the previous season. Extent of the nist attaek at final examination Plants (28th July to 11th August) Culture in No. cultuie None Trace .Slight A Moderate Bad Very severe 12/20 48 — — 1 9 17 21 14/18 54 — — 4 29 14 7 9/0 53 — — 6 32 9 6 16/15 67 — — •> 51 12 ■t 10/10 2B — — 3 14 9 — 72/7 53 1 — 21 21 9 1 32/1 40 9 (1 11 14 5 1 66/1 63 20 .s 13 14 7 1 41/17 (44) Totals: 404 30 8 61 184 82 39 [8 cultures] ^ ^ 1 99 305 31/23 48 4 — 24 20 70/13 44 2 7 30 5 — 82/9 61 — 10 49 2 ^ The other nine cultures gave evidence of segregation into shghtly, moderately, and badly rusted types, and were clearly the offspring of heterozygotes for rust resistance. In eight of these the plants were fully graded and included 99 individuals with a shght attack, or none, and 305 with a moderate or bad attack. (d) Of the cultures raised from only shghtly rusted -^2'^' ^^ gave distinct evidence of segregation (Table V). Twenty-two of these plants I'l Mendd'uiH I nlierltancc and Ycllutn Ruxf In Wheat were graded at the final examination, and these contained "291 individuals with only a slight attack or none, and 893 on which the attack was either moderate, bad, or very severe. These numbers are very close to the 3 : 1 Mendelian ratio. Of the other cultures in this group, seven showed much less evidence of segregation, but, as the majority of the plants in each case were much less severely attacked than those of the homozygous susceptible cultures growing alongside, it is probable that these were the offspring of heterozygotes. Two other cultures, 1/3 and 38/22, proved to be as resistant as the first 15 given in Table 111. Table ^ . Analyaii: of F-j cullitres (191'.)) (jivwujrom F<, plants which had a slight attack of rust in 1918. Extent of the rust attack at filial cxamiiiation Plants (28th July to 1 1th August) OiJture in * , No. culture None Trace Slight Moderate Bad Very severe 2/8 52 12 — 9 20 8 3 7/15 48 II — 9 5 12 1 9/3 7(t 15 — 22 28 4 1 11/3 51 11 — 7 26 5 2 07/10 54 12 — 6 10 11 15 67/11 73 n 3 12 24 8 15 13/13 lit 21 0 9 9 8 2 39/7 87 — — 15 48 14 10 41/16 48 — — 8 28 6 6 52/13 48 — — 8 24 10 fi 52/23 67 _ _ 13 31 17 6 56/14 65 — — 20 21 l> IS 72/2 44 — — 5 13 7 19 72/12 50 — — 3 15 12 20 9/18 29 — — 7 12 9 1 10/9 49 _ _ 5 23 15 6 17/13 66 — — 20 42 4 0 17/24 62 — — 10 43 7 2 22/1 48 — — 2 30 16 0 23/17 50 — — 3 21 25 1 24/15 31 — — 1 10 IS 2 8/5 43 — — 1 2S 10 4 30/11 (63) 32/16 (34) 33/23 (44) 38/6 (32) Totals: 1184 93 3 195 511 238 144 [22 cultures] ' ^^ — — ' "^ v ' 291 893 (e) Si.xty-threc cultures were raised from F^ plants which had been moderately attacked. Fourteen of these proved to be homozygous susceptibles, for every plant in these I 1 cultures (791 plants in all) was either badly or very severely iiisted. S. F. AUMSTRONG 73 Thirty-one cultures gave clear evidence of .segregation, thougli only nine of these actually contained any rust-free plants at maturity. Details of rust attack on each plant in '2i of these cultures are given in Table VI. It will be noted that the extent of "'disturbance of resistance" varied considerably fi'oni culture to culture. In the first four given in this table, it was practically non-existent; in the next four it was ap2)reci- able, while in most of the others it was very great. Nevertheless, the very wide difference between the nature of the attack in the extreme grades, and the proportion of cases of intermediate attack, indicated that JVIeudehan segregation had occurred in all these cultures. Table VI. Results of analysis of the F.^ ciiliures (1919) raised from F., plants which had a moderate rust attack in 1918. //; this tabic are grouped 24 cultures which gave clear evidence 20 9 3 8/15 71 — — ■ 13 40 1.-. 3 11/10 34 — — 1 18 10 5 12/3 52 — — 2 23 20 7 18/2 48 — — 3 34 s 3 18/4 55 — — 1 17 33 4 28/13 40 — — 8 9 15 8 40/5 40 — — 4 24 14 4 40/17 44 — — 16 25 2 1 54/4 78 — — 5 55 10 o 24/18 39 — — . 2 2 15 20 73/8 76 — — 4 23 20 29 77/10 57 — — 15 14 10 12 78/2 55 — — 9 21 13 12 Totals: 1359 55 4 191 549 338 222 250 1109 (339-7) (1019-3) In addition to tlie above, seven otlier cultures gave clear evidence of segregation, but full records of rust on each plant were not made. Most of the remaining 18 cultures had suffered very severely from drought. In these segregation was much less evident, the attack varying from a moderate to a veiy severe one. They were, however, distinctly 74 MendeliiiH Inkerilance and I'lilui'' Hnxt In Whral intermediate between the homozygous susceptible and "imuiune" cultures in regard to their rusted condition, and, taking their abnormal growth into consideration, one is probably correct in concluding that they were the offspring of F., heterozygotes. (/) Nineteen cultures were grown from i\ phmts which had been badly rusted in 1918. Fifteen of these proved to be homozygous sus- ceptibles, every plant being either badly or very severely attacked. One culture (14/2) was seriously affected by di'ought, but was 2)robabiy also of the same genetic constitution for susceptibility. The three other cultures, 74/12, 6/4, and 31/1, gave clear evidence of segregation. {(j) The 24 cultures raised from very badly rusted i\ plants all proved to be homozygous snsceptibles. In each of these all the plants were distinctly worse attacked than their parents had been in the pre- vious season (Tabic VII). Table VJl. Results from 21 7'^^ vultures (1919) raised from F^ plaids wkicJi were very badly rusted in 1918. Every plant in this grouj) of cultures was very severely rusted in 1919. Percentage of plants rusted C'ultino Plants in Kust HrsL .May June July July 28tli No. culture seen 16tli to 3Ist 14th to ISth 4tli to 10th to Aug. 9th 9/li) 33 May 28 6 33 91 UMI 14/21 34 „ 16 6 82 100 100 27/21 54 ,. 17 4 88 100 101) 28/ U 33 „ 17 3 81 100 lOO 29/(i 21 ., 17 5 72 10(( 100 31/7 44 ,. 17 2 20 75 100 .•i3/2 23 „ 29 12 62 83 100 34/ 1;! 24 ., 28 12 63 100 100 42/5 33 „ 31 3 51 93 100 .54/l(> 20 ., 29 10 45 85 100 55/17 47 ., 29 13 49 98 100 57/3 62 ., 19 1 50 97 lOd 1)4/20 24 „ 27 8 50 100 loo 1)8/ 12 44 „ 29 4 57 98 10 sprinji-sown ]K)rtion of the Fo generation: 189-7 DD : ;378-2 DR : Hi.'M RR. Table X. Approximate genetic constitution of the F^ generation {spring- sown portion) as regards susceptibility to Yellow Rust. The.'^e figures are arrived at bi/ the direct application of the F^ results. Xiiiiilx'r iif A'j plan |)l!i<^«'(l in each gra( for rust attack* A ts Ir RcsnUs ill /■'., ) Approxiiiiatt' <;f'nPtir ronstitntion of tlif I''., prop as inflioak'd liy the ^'3 results A 4 3 2 1 1 1 T ( DD DR nil 4- 100",, proviMl /)?) Sfi — — i: 184 .. DD |l(i .. /)R 74-7 — 14-:! — !2 22 ., DD 7S .. D/{ 20 — — m:! — u U liU .. Dl! I C. .. Jill — 17n II 1 9S 175 ., DIt |25 ., UK — 73-5 24-5 1- .-. (12 .. Dl! /SS .. HI! — 17-4 127-r. Totals: 189-7 :578-2 1631 riK 1:2:1 ratio would be (183) : (:?r.,->) : (183) * For rust grades see p. Gl. The total number of plants was 1560, and the numbers expected according to the 1:2:1 ratio are (;}9()) : (780) : (:59()). Applying the F3 results to the wliole F^ crop, we (Ind tlie following cotn])osition indicated : 391-7 homozygous susceptible individuals. 797-2 heterozygous susceptible individuals. .'>71-1 homozygous imnmne individuals. 6. Results obtained from F2, F^ and ^^4 ctiltures, etc., in 1920. Tli(^ foregoing results were supplemented by further data obtained (luring 1920. All the rust cultures of that year were grown on the Uni- S. F. Armstrong 79 versity Seed Farm near Cambridge. The ground, wliich had received a dressing of 10-12 loads of farmyard manure and had been cropped with potatoes in the previous year, was in good condition for wheat growing. Cultures of tlie Fo, F^ and F^ generations of the cross Wilhelmina X American Club were raised, in addition to others. During the winter 1919-1920 abnormally mild weather was almost continuous. One interesting feature associated with this mild weather was that freshly formed uredospores of Yellow Rust were found from the first week in October, 1919, onward to the following summer on certain susceptible wheats which had been sown at the end of August, 1919. During the spring and summer of 1920 the weather was favourable to the normal growth of the cultures (see Table I). Yellow Rust attack was general amongst the field cultures at a very early date, and by the beginning of May in many of the susceptible cultures all the plants were rusted. By the first week in June the attack was at its Jieight. (n) Further Fn statiMici nhUiined in 1920. In raising the i^., generation (cross No. 120) in 1918, no attempt was made to keep separate the grain of the several F^ plants. In a later cross, however (Brooker's x American Club, cross No. 154), this was done, and several distinct F^ famihes were raised in 1920. Two of these families were examined in the same manner as the F^ of cross No. 1 20 had been in 1918 with the following results: A population of 198 7^., individuals raised fi'oni a moderately rusted f J plant contained 42 badly rusted plants and fO whicli remained abso- lutely rust-free. On the remaining plants, the rust varied from the merest trace up to a moderate attack. The second i^, family consisted of 2-58 individuals and was raised from an F-^ plant which had a shght attack only. It contained 62 badly rusted and 61 perfectly rust-free plants, the remaining 135 indixaduals being attacked to not more than a moderate extent in any case. These figures indicate that similar results are obtained whether the analysis be made upon the progeny of separate i^j plants or upon the offspring of several ^''j^'s combined. The small F^ crop of the original Wilhelmina x American Club cross grown in 1920 contained 114 plants, and by the end of July it was found that 86 of these were rusted and 28 were rust-free. Of the rusted plants six had a much severer attack than Wilhelmina grown under similar conditions. 80 Mendeliau Inheritance and Yellow Btisf in Wheal {b) Direct aiial/jsis of un F., Ihroiujh the F^ results, 1920. A small Fn crop of the same parentage (Wilhelmina x American Club) consistiiij; of 86 plants was raised in 1919. No attempt was made to grade these for rust attack as it was decided to grow an F.^ culture from each plant and to use the F^ data for placing each F.^ plant in its proper category for susceptibihty or resistance. These 86 F^>^ were raised in 1920. In 22 of these cultures all the plants were readily su.s- ceptible; 20 others contained only rust-free or highly resistant plants, while the remaining 44 cultures consisted of readily susceptible and resistant plants. The infeience is that the A', grown in 1919 consisted of 20 rust- resistant plants, 44 impure susceptible and 22 pure susceptible individuals. In four of the pure susceptible F-^ cultures all the plants were excessively attacked. (c) Results from F^ cidliires. (Cross No. 120. Wilhelmina x American Club.) A considerable number of F^ cultures were grown in 1920. Twenty- seven of these were raised from plants taken at random out of the wholly susceptible F^ cultures of the previous year. Without exception these F^ plants (433 altogether) were badly rusted. Thirty-.six F^ cultures were grown from plants which had been extracted from obviously segregating F^ cultures. Space will not allow a full iircount of the results, but they showed definitely that the three genetic types — pure resistant, pure susceptible and impure susceptible —were present in the F^ cultures. At the same time they indicated again that some allowance must be made for fluctuations in suscepti- bihty due to season and other external conditions. The 19 F^'f. raised from rust-free F^ plants consisted entirely of rust-resistant individuals. Similarly all the extracted badly rusted F^ plants produced badly rusted F^ cultures, while slightly or moderately rusted F.-^ j)lants gencniJlN- produced cultures in which rusted and rust-free plants occurred. Two hundred and eighty-eight F/^ cultures were raised from F.^ plants which were the offspring of rust-free F.;, individuals. Some of these were picked from cultures like 66/9, 11/21, 2/10 and 19/12 (Table III) in which all the plants were completely or almost completely rust-free in 1919. The F^ plants nimTbered several thousands and it was impossible to examine each individual very closely, but a general inspection showed that all the plants were either rust-free or possessed a very high degree of resistance. S. F. Armstrong 81 Many plants were purposely taken from cultures 15/5, 24/3, 29/15, 57/8, etc. (Table III), in which rust-resistance had apparently been dis- turbed or "broken down" under the adverse conditions prevailing in 1919. A few of these extracted jPg's had, indeed, been badly rusted, e.g. 15/5/15, 24/3/18, etc. Nevertheless in 1920 all the F^ plants proved to be either completely rust-free or highly resistant. Table XI. Progress uf infedion bi/ Yellow Rust on the various F-^ cidlmes in 1919. (Cross No. 120.) 169 cultures*. Percentage of cultures on wliich infection had occurred for the weekly periods ending t\ Number May June July Final examination cultures of ^''3 , ^' , '• , , * , July 28th to 1919 cultures 17th 24th 31st 7th Uth 21st 2Sth 5th 12th August 1 1th DD\ 5.3* 13-2 340 773 94-3 100 — — — — — DK'a 'Mi 3-1 15-6 22-4 49-0 61-4 IJ4-5 80-2 81-2 93-7 100 lilCs 20 0 0 0 0 0 0 0 0 Of 90 * One culture omitted because no record of date of first attack was made. t Only one small pustule seen on one plant by this date out of a total of 817 [ilants. Section- 111. .SOME OF THK FACTORS WHICH JIAV IXCREASIC Oil DIMINISH SUSCEPTIBILITY TO RUST ATTACK. Introduction. In the previous section of this paper it was seen that a sharply defined 1:2:1 ratio was not directly found in the F.^ generation in 1918, although the Fg results showed that it undoubtedly existed. The varia- tions observed in the F^, and from season to season, appear to indicate that, in addition to the inherited factors which primarily determine resistance, etc., there exist other factors which may tend either to increase or reduce a plant's predispo.sition to attack. The now generally accepted view is, that immunity is due to the production of specific toxins or anti-toxins which have the power to neutraUze the action of the attacking fungus. If this is correct, an im- mune plant is one whose normal metabolism provides such a substance, whilst a susceptible plant is one which is either unable to form such a substance at all, or can only do so to an ineffective extent. But even in the case of a plant which is able to produce such protective substances, it is obvious that such production may be subject to modification as regards quantity, rate of formation, etc. There is, indeed, evidence to show that a state of complete immunity depends not only upon the inherited factor for resistance, but also upon a properly balanced con- Joura. of Agric, Soi. xu 6 H'2 Meiuh'luni I iiln rlttnicr ((ml Yellow Rf(st in Wlmd ditioii of the plant's normal nu'taholisni. For example, American Club, thoufjh normally immuiio, has been found to develop pustules of Yellow ]?ust under abnormal conditions of gro\vth(iO) or nutrition (12). It is therefore probable that any factor or factors which bring about a condition of growth unfavourable to the production of those substances upon which immunity depends, may lead to a greater or less degree of susceptibility. Ward(20) especially emphasized the fact that "the physiological condition of the host is always a factor of prime import- ance" in considering the beliaviour of the host towards the parasite. Since all physiological processes are dependent upon both internal (inherited) and external factors, in matters of this kind we must be careful to distinguish between the inherited and non-inherited factors concerned in bringing about the net result. Consequently, in trying to discover the causes of the "disturbance in rust resistance" as observed for example in 1 919, we nuist recognize the possible effects of: (1) External environmental factors. (2) Inheritable factors (other than those prinuirily concerned in causing resistance, etc.) which lead to fresh combinations of parental features in the hybrid and its descendants. These will now be con- sidered. (a) Effect ok climatic conditions. The greater severity of attack in 1919 is partly explained by the different weather in the two seasons. That of 1918 favoured a regular and normal development of the plants. On the other hand the cultures of 1919, while still in a backward condition, were subjected first to a 7 weeks" drought, and afterwards to G weeks of almost continuous dull, cool, wet weather. During both years records were taken of the extent of the Yellow Rust attack on 13 distinct varieties of wheat grown in the cages. In eight cases the attack was more severe in 1919 than in 1918; on four varieties it was of about the same intensity, and in one case it was slightly less severe. This and other evidence support the behef that the abnormal weather of 1919 was partly responsible for the more pronounced attack in that year. (h) Food supply. It was, however, clear that the greater intensity of attack on the F^ cultures in 1919 was not due solely -or even chiefly — to tlu' different weather conditions. The increased severity of attack was far moie pro- nounced on these than on the standard varieties growing alongside. S. F. Armstrong 8:1 while the only known difference in environmental conditions was that of food supply. The rust-cultures had all received a heavy dressing of nitrate of soda, whereas the other plots were unmanured. The nitrate was carried to the roots of the plants when their Very existence was in the balance at the end of the drought (June "20th). Thus the nece.s.sary supply of moisture, and also a large quantity of available nitrogen, were simultaneously presented to the plants. It is very important to note, therefore, that the change from a critical condition to a state of active growth was as sudden as it was pos.sible to be. The "second growth" — which also occurred on the unmanured wheats — was stimulated to an enormous extent on the F^'s, and the cultures assumed the very dark green colour characteristic of plants receiving an excess of nitrogen. The maturation of the plants was also considerably delayed. These conditions evidently in some way afforded a greater opportunity for rust attack, for it was precisely during this period of delayed maturation that the great epidemic of the season developed. Nitrogenous manures, especially when greatly in excess of other fertihzers, are well known to be very effective in increasing the severity of rust attack. Biff'en(3) has pointed out that, on the Rothamsted wheat plots, rust attack is invariably encouraged where ammonium salts or nitrates are continuously applied in heavy doses. Spinks(i2) also showed that susceptibility to Yellow Rust is increased by the use of large quan- tities of available nitrogen, while plants which are semi-starved as regards nitrogen may exhibit a considerable degree of resistance. Further, as Spinks and others have shown, some salts, e.g. salts of potassium and especially lithium salts, may markedly reduce susceptibihty. The question of food supply is therefore certainly of great importance in connection with the observed fluctuations in susceptibility. In connection with this question of food supply, it may be noted that during the present experiments there appeared to be a difference in susceptibility between "interior" and "exterior" plants in the same pure susceptible cultures. Records of the first plants to be attacked were made on 23 of the homozygous susceptible cultures given in Table VII. At the time of "first infection" these cultures contained 642 "interior," and 124 "exterior" plants. By June 7th the number of infections noted on "interior" plants was 35, i.e. 5-4 per cent., while at the same date the infected "exterior" plants numbered 20, i.e. 16-1 per cent. Similarly, in 12 other pure susceptible cultures, by the end of May it was found that 7-3 per cent, of the "exterior" plants were attacked, but only 2-3 per cent, of the "interior" plants. The number of early infections were 6—3 84 Mendiiiun InlKritance tniii Yrlhnr Rust In Wlniif therefore relatively three times as numerous on the "exterior" as on the "interior" plants. An experiment was planned in order to find out if possible to what relative extent such "circumstances as wide spacing and large amounts of available nitrogen were responsible for the increased severity of attack observed. The varieties included in the test were tlie two parent wheats Wilhelmina and American Club and the following extracted F^ tjrpes descended from these parents: GG/'djd, a rust-resistant type, with white chaff and beardless medium- lax ears (a resistant Wilhelmina type). 82/14/a, a rust-resistant type, witli red chaS and beardless medium- lax ears. 75/7 /c, a susceptible type, with wliite chaff, and dense bearded ears (a white chaffed susceptible American Club type). 75/1 1/(^, a dense eared, bearded, susceptible type with red chaff'. (This was practically a susceptible American Club.) Four beds, A, B, C, and D, each 4 feet wide, were laid out, and each variety was sown across these in a single row 16 feet long. A space of () inches was allowed between each row, but the grains were sown 2 inches apart on beds A and D, and 12 inches apart on B and C. The grain was sown on November 2'lth, 1919, and a fairly uniform plant was obtained. On May 4th, 1920, beds A and B were top-dressed witli nitrate of soda at the rate of 7 cwt. ))er acre. It should be noted that, although beds C and D received no nitrate, the soil was in good condition after the potato crop of the previous year; the comparison to be made was not between starved and overfed plants, but between plants grown under normal and abnormal conditions as regards space and nitrogenous manuring. Yellow Rust was first seen on May 7th on one or two plants of 75/11/rf and Wilhelmina (exterior row) in each case on bed A. By May 12th rust attack had become general on all the susceptible varieties, the only difference being that it was of a less pronounced character on Wilhelmina. At that date no sign of attack could be found on any of the resistant varieties although their foliage was literally ])owdered daily with nredo- spores from their susceptible neighbours. The condition of the plants on May 21st and June 2nd is shown in Table XII in which the numerator of each fraction indicates the number of plants rusted out of the total surviving plants of each variety on each bed (given as the denominator). 8. P. Armstrong 85 General conditiov of the plants on May 2\st, 1920. The plants on bed A were of a dark green colour and growing vigorously; they were about 2 feet high, and owing to the close planting and strong growth the shoots were densely crowded together. Mildew was abundant, especially on Wilhelmina and 6fi/9/rf. Table XII. Effects of wide planting and heavi/ application of Nitrate upon wheats susceptible and resistant to Yellow Rust. Grain soivn Nov. 2itk, 1919. Nitrate of soda applied May 4th, 1920, at the rate of 7 c.wt. per acre. ^ fc. r^ !i, o a> &7 S 3 &. 3 ^:2 c '3 c ■^ o -3 -2 +> |§ ^ .2 © fe .2 o eS t; =0 P P CJ O a, £ o .a 1 31 0- ^ sf Is 3 1 rt o a i^ 1 -^^ 1 c - 'S ^ Si 'C II 5 ^ si g ^2 si ^1 11 "ail 't^ No. of row 1 2 3 4 5 t; 7 8 9 10 11 12 Bed r ^ Planted 13 0 14 () 19 0 14 0 15 0 10 20 2" apart 18 18 14 19 19 22 ri 19 17 15 20 22 B '1% Planted 3 0 4 0 3 0 4 0 3 0 4 0 S & 12" apart 3 4 4 3 3 2 4 2 3 4 4 1 C t. Planted 2 0 3 0 5 0 2 0 4 0 2 3 12" apart 4 2 3 5 5 3 2 2 4 2 2 3 |l D S s Planted 2 0 10 0 7 0 17 0 11 0 10 16 l'^ 2" apart 15 15 Yi 18 13 IS 17 19 19 13 12 n , r ''^ 3* _^ 4 ■3 6 T! 0 •a 4 s 4) OJ fD -3 18 19 22 CO 3 19 CO 1.5 1 1 11 B t.1 2 4 ■a 2* 3 1 2 2 2 1 0 4 2 -4^ 3 gi^ 9 ^ ^ ^ Xi c C > 0 2 0 5 1* 3 -4^ 0 2 'ft >-> 0 2 'ft > 5 "ft ^ u D H 0 > 0 0 3> 0 a. 0 H H ll ^ 15 w 18 w 18 W 19 W 13 _ ^ * In these cases the plants were only "flecked.' 86 Meriflelian Inheritance and Yelloir Jinst in WJimf On beds B and C the plants were also makinj; a very strong growth. Owing to the large amount of space available, numerous tillers were formed and the shoots were spreading out obliquely. On bed D the plants were about 18 inches high, erect, and of a light green colour. They had tiliored much le.ss freely than the plants on the other beds. Table XII shows that on this date all the plants of American Club and of the two extracted resistant types (altogether '20F> plants) were still entirely free from infection. Taking the proportion of plants infected and the severity of attack into consideration, Burbank's was distinctly the most susceptible type, 08 out of 72 plants being badly attacked. The proportion of rusted plants in Wilhelmina and the two extracted sus- ceptible types were very similar, hut the attack ou'Wilheimina was less severe; the actual numbers were: in 7."j/7/f, .34 rusted out of 40 plants; in 75/11 /(Z, 33 rusted out of 43 plants; and in Wilhelmina. 91 rusted out of 121 plants. Confining our attention for the moment to the susceptible varieties, we get some idea of the e.xtent to which "wide spacing,"' etc., apparently favour rust attack. The effect of "wide spacing'" is seen by comparing beds C and D. On bed D, out of 107 plants of the susceptible varieties, 73 were rusted, i.e. 68 per cent. On bed 0, 21 out of a total of 23 susceptible plants were rusted, i.e. 91 per cent. The effect of the heavy dose of nitrate is seen from a comparison of beds A and D ; on the former, out of 1 24 plants of the susceptible varieties, 111 were rusted, i.e. 89-5 per cent, as compared with 68 per cent, on D. The combined effect of "spacing'' and "nitrate" is seen by-comparing beds B and D. On the former, 21 out of 22 susceptible plants were rusted, i.e. nearly 100 per cent, against the 68 per cent, on D. It is clear from these and other observations that very widely spaced plants are more hable to an early infection than are closely planted individuals — other conditions being equal. A partial explanation for this possibly lies in the fact that widely spaced plants tiller more freely, and so ofltcr a much greater area of leaf surface to infection than more crowded individuals. On the other hand the percentage of early in- fections on the closely planted bed A was as high as on the widely spaced bed C, and it is unlikely that this effect was brought about by the in- creased tillering of the plants on A. The increased susceptibihty of plants receiving heavy doses of nitrogenous manures or extra space for growth appears to depend rather upon the increased or modified food-supply offered. 8. F. Armstrong 87 Condilion 5 mixed stock of "Jap." On May 12th, 1920, out of 57 plants in culture 3/16/29, 41 were attacked by Yellow Rust; on the same date the 66 plants in culture 1/4/10 were free from infection. Culture 3/16/29 was about 7 days earlier than the other, and all the plants in it were rusted by June. Only a few slight traces of attack and some " flecked" plants were found on culture 1/4/10 before harvesting. The original intention in growing these cultures was merely to see whether the extracted types bred true to resistance and siisceptibihty respectively, but, at harvest-time, they seemed to offer very favourable material on which to determine the loss directly due to Yellow Rust. In the first place, their identical — or almost identical — morphological features appeared to afford safe ground as a basis for such a comparison. Again each culture had remained free from attack by Brown Rust or other fungi up to the time the plants were pulled — July 23rd — so that the difference in yield was chiefly, if not entirely, due to the presence or absence of Yellow Rust. Further, each plot was sown on the same day. grown under identical conditions side by side, and no loss of grain had occurred at harvest-time. In one respect only there was a sUght cUiTerence between the cultures, viz. in the number of plants per row, this being due to the rather poorer germination of the grain in the susceptible cultures. The results were as follows: Culture of Culture Culture "Original Mixed 1/4/10 3/16/29 Jap" "Resistant" "All susceptible Rows 3 4 4 Plants 4.5 66 57 Plants per row- 15 16-5 14-2 Total weight of ears (grams) 310 555 245 Weight per ))lant (grams) ... 6-9 8-4 4-3 Percentage relative weights: (a) Per plant 821 100 51-2 (b) Per unit area ... 741 100 44 It will be seen that the weight of ears in the "mixed" culture was reduced by 18 per cent, per plant, and by nearly 50 per cent, per plant on culture 3/16/29. The reduction per unit area was still greater in each case. Had the grain been threshed out and weighed separately, it is certain that the differences would have been still more striking, but this unfortunately was not done. These figures indicate that a variety of wheat having on the average a moderate degree of susceptibihty (comparable with, say, Square Head's Master) may give a yield at lea.st 25 per cent, below that obtainable from almost precisely the same form when rendered rust-resistant. 96 Mcndefldii hiln ritdiia ftml Ydlmi- /'ii.sf ill win lit LIST OF I'APERS REFEHREI) TO I.\ THK TKXT. (I (2 (3 (4 (6 (7 (8 (f» (10 (11 (12 (l.-i (14 (15 (16 (17 (18 (19 (20 (21 BiFFKN, K. H. (IIK).")). -Menders Laws of Inheritance and \\ lieal llreediiig. Joum. Agric. Sci. 1, 1. (1907). Studies in (lie liiheiitniiee of Disease Resistance. Joum. Agric. Sci. 2, 109. (1912). Studies iiil he I iilieritaiue of Disease Resistance. Part II. Jourii. Agric. Sci. 4, 421. Eriksson, J. (1896). Die Oelreiderotile. Evans, I. B. P. (1907). The Cereal Rusts. Annah of Bolutiij, 21, 447. (1911). South African Cereal Rusts, with observations on the Problem of Breeding Ru.st -resistant Wheats. Jourii. Agric. Sci. 4, 9;'). (iiBsoN, Miss ('. M. (1904). Notes on Infection Experiments with various Tredineae. New Plii/lnlogi.il. 3, 184. Haves, H. K. and others (1920). (Jenetics of Rust Resistance in crosses of varieties of Tritieum vulgare with varieties of T. durum and T. dieoccum. Joum. Agric. Res. 19. 523. Keebi.e and Armstkonc; (Hti.'J). The Role of Oxydases in the Formation of tlie .Anthocyanin Pigments of Plants. Joum. Oeitelic.t. 2, 277. Marrvat, DoROTirEA ('. 10. (1907). Notes on the Infection and Histology of two Wheats immune to the attacks of Puccinia glumarum. Joum. Agric. Sci. 2, 129. NrL.ssoN-EHLE (1909). Kreuzunguniersiichungen an Haje.r und Weizen. SpiNK.s, (J. T. (19i:5). Factors affecting Susceptibility to Disease in Plants. Joum. Agric. Sci. 5, 231. Stakman, E. ('. and others (1918). Plastieity of Biologic Forms of I'uccima graminis. Joum. Agric. ffp.s-. 15, 221. Vavilov, N. I. (1914). Immunity to Fungous Diseases as a Physiological test in Genetics and Systematics, exemplified in Cereals. .Joum. Oenelic.i, 4. 49. Ward, H. xMarshall (1890). Croonian Lecture. Proc. Roy. Soc. 47. '1902). Proc. Rnij. Soc. 69. 4.-)l. (1902). Proc. Roy. Soc. 71. l;i8 (1902). Proc. Camb. Phil Soc. 11, part o. (1902). On the relations between Host and Parasite in the Bromes anil their Brown Rust (P. disper.'ia). Annals of Botany, 16. 233. (1905). Recent researches on the Parasitism of ['"ungi. Annals of Botany, 19, 1. Wheldale, Miss M. (191 I). On the Formation of .Anthoevaiiin. .Joum. Genetics, 1, 133. {Received October 8th, 1921.) NOTE ON THE COMPOSITION OF A FLUID OBTAINED FROM THE UDDERS OF VIRGIN HEIFERS By HERBERT ERNEST WOODMAN, Pji.D., D.Sc, AND JOHN HAMMOND, M.A. {From fJie Institute, for the Study of Animal Nutrition, School of Agriculture, Cambridge Universitg.) The secretion of milk by the mammary glands normally follows a period of pregnancy, but numerous cases have been cited in the literature of secretion which takes place in animals that have never borne young. Non-pregnant bitches frequently secrete milk several weeks after oestrus (1), and a secretion, apparently similar to milk, takes place after pseudo-pregnancy in rabbits (2). No quantitative analyses have been made of these secretions to show whether they possess the characteristics of true milk or colostrum. The origin and significance of colostrum itself, as distinct from milk, is also in doubt, and its formation has variously been attributed to: (1) The break up of the central cells of the alveolus, or as a result of their initial activities; (2) The filtration of lymphatic secretion through the walls of the alveoli mixing with the milk formed by the cells. Recent work on the proteins of colostrum (3) has shown that although the caseinogen and albumin must be elaborated by the mammary gland, yet the globulin, which is present in large amount in the colostral secretion, is in every respect identical with the globulin of blood serum. There also exists the possibility that colostrum results from the partial absorption of the more diffusible milk constituents, the secretion of which has been taking place to a small extent some time previously. It was therefore of considerable interest to find, during a study which is being made by one of us (J. H.) of the development of the udder of the cow, that the galactophoroiis sinuses and ducts of virgin heifers of some 1| to 21 years of age contained very frequently a fluid which could often be expressed from the nipples in quite considerable amounts. The object of the present work was to investigate the composition of the secretion and its possible relation to colostrum and to milk. It has been suggested by several writers (i) that the mammary gland Journ. of Agric. Sci. xn. 7 98 Fluid obtained from. Udders of Virgin Heifers undergoes a cycle correlated with that existing in the ovaries and culminating at oestrus by a swelling of the mammary gland. Since heifers have periodic corpora lutea and a well-marked ovarian cycle, it might be supposed that the former acted in much the same way as those of pseudo-pregnant rabbits, but to a smaller extent. It was not unreasonable therefore to suppose that this Ouid was secreted just before the period of oestrus in heifers, but attempts to prove this so far have been without success. The fluid can be drawn off from tlie udder at any period of the cycle, and the amount varies considerably according to the individual. In animals killed at various periods of the oestrous cycle, there was an indication that more existed just before the oestrous period than at any other time. But individual cases were also found which did not show this behaviour. The secretion of milk has been attributed to the removal of the stimulus which caused the mammary gland to develop. Thus Lane- Claypon and Starling (-5) and others have attributed this action to the foetus, but Marshall and Hammond (2) found that the corpus luteuni caused the growth of the mammary gland in the rabbit, and that milk was produced whenever the gland had developed to a sufhcient extent. The removal of the causative stimulus in heifers cannot, however, be the cause of the secretion, as the fluid has been found in the udders of previously virgin heifers pregnant 'i-A montlis, at a time when the corpus luteum is still large and active. The problem thus presented itself: Is the secretion similar to that of colostrum and milk, resulting from the activity of the cells of the alveoh, or is it merely an exudation of lymph filling up the galactophorous sinuses, which have been formed by the development of the mammary gland? If the latter hypothesis be found correct, then at what stage in the development of the mammary gland do the typical constituents of milk first appear? Chemical investigation of the fluid. In all, four samples of the fluid were examined. The first two samples were obtained from heifers which had never been served, whilst the third and fourth samples were taken from the udders of the heifers during the first three weeks of pregnancy. The amount of fluid was only small, ea(^h heifer contributing on an average about 7 c.c. to the supply. The liquid was slightly opaque in appearance, though not possessing the dense opaqueness of milk. A slight sediment settled out on standing, but the fluid passed fairly H. E. Woodman and J. Hammond 1H) readily tlirough a filter paper, the filtrate tlieu being almost clear aud possessing the appearance of a protein solutiou. It was slightly viscous, gave a foam on shaking and did not show the amphoteric reaction of fresh milk, but was very faintly alkaline to both litmus and phenol- phthalein. A preliminary investigation of the first two samples was carried out as follows. A' small volume of the fluid was diluted with distilled water and one drop of acetic acid was added. On shaking, a white flocculent precipitate settled out in a manner characteristic of the separation of caseinogen from milk by this method. This was filtered off and was shown to be a protein by the usual tests. Furthermore, it answered to all the characteristic tests for caseinogen. (1) It was insoluble in water but dissolved readily in dilute soda and in lime-water aud was rejireciiji- tated by the addition of a drop of acetic acid. (2) When fused with a mi.vture of KgCOg and KNO3, the solution obtained on extracting the residue with water gave a strong phosphate test with nitric acid and ammonium molybdate. (3) The freshly precipitated protein was com- pletely soluble in excess of acetic acid. Further tests were made in order to ensure that the substance was not a nucleoprotein, which class of bodies gives most of the general reactions of the phosphoproteins. The presence of nucleoproteins in such a fluid would, moreover, not be at all surprising, especially if it resulted from the filtration of lymph through the walls of the alveoH. That the protein was caseinogen, however, and not a nucleoprotein was shown by: (1) The readiness with which it sjilit off its phosphorus as inorganic phosphate by mild alkahne hydrolysis. Under these conditions, nucleo- proteins do not yield phosphoric acid, the phosphorus remaining bound up in the nucleic acid group (li). (2) The behaviour of the fluid towards rennet. To 5 c.c. of the fluid were added 2 c.c. rennet extract. On placing in a bath at 10°, only a slight turbidity resulted after two minutes. On the further addition of two drops of calcium chloride solution to the mixture, however, the turbidity increased aud a curdy precipitate settled out. A clear mixture of 5 c.c. fluid, 2 c.c. rennet extract and 1 c.c. calcium chloride solution was placed in a bath at 40°. In a few seconds, a turbidity spread throughout the hquid, followed by a rapid separation of curdy precipitate. This behaviour was consistent with the presence of caseinogen in the fluid, the rennet producing, in the presence of calcium salts, a precipitate of insoluble calcium caseate from the soluble calcium caseinogate. (3) The absence of any insoluble nuclein residue when the protein was submitted to peptic digestion. loo Fluid ohlaincd J'roiii Udders of Viryln Ilfifira The filtrate rcmaiuing after separation of the caseinogen was heated a short time in boihug water. This caused coagulation and the substance which separated in this manner was shown to be protein by the ordinary tests. The nature of the coagulable ])roteins was investigated in the following manner. After separating the caseinogen from another portion of the fluid, the clear filtrate was made exactly neutral by means of NjlO NaOH and was then saturated with inagnesiuiii suljjhate. The protein which was salted out by this ])n)cess was filtered olV and re- dissolved in water. On saturating this solution with magnesium sulphate, the protein was again throw n out. It possessed therefore the character- istics of a globuhn. The filtrate from the globulin was next acidified with acetic acid. A further precipitate of protein was obtained, its mode of isolation proving it to be an albumin. As far as it was po.ssible to judge from the tests, the fluid contained rather more globulin than albumin. The liquid remaining after removal of caseinogen by acetic acid and albumin and globulin by coagulation still answered to the biuret test and gave an appreciable precipitate with tannic acid. It still contained nitrogenous bodies of a simpler type, amongst which was a proteose, since a further precipitate was obtained on saturating the liquid with ammonium sulphate. The existence of non-protein nitrogen in the fluid was borne out by subsequent analysis. The fluid, after removal of proteins, was found to exert a slight reducing action when boiled for some time with Fehling's solution. The osazone, which was only obtained in small amount and with difficulty, had the characteristic appearance of lactosazone under the microscope. A sample of milk, tested similarly, showed ready and copious reduction of Fehling's solution and gave a good yield of lactosazone without diffi- culty. In all the samples of fluid examined, it was only possible to demonstrate the presence of lactose in small amounts. A prehminary extraction of the fluid with ether showed it to contain a small amount of fat. Sample 4 was found, by the Paper Coil method, to contain about 0-6 per cent, of a sohd yellow fat. This was saponified by boihug a short time with caustic soda. After acidifying, extracting with ether and allowing the solvent to evaporate, the residue possessed the characteristic odour of butvric acid. It followed that tlie fat in the fluid was the product of mamuuiry synthesis. The fluid did not contain any detectable amount of mucin, since this would have shown as gelatinous strands on the addition of acetic acid. The precipitate thus obtained w^as, however, perfectl}- flocculent H. E. Woodman and J. Hamjmond 101 and was, moreover, completely dissolved by the addition of excess of acetic acid. Mucin remains undissolved under these conditions. The per- fectly flocculent character of the precipitate obtained on adding alcohol to the fluid and stirring precluded the possibility of the presence of pseudo-mucins in the fluid. No other glucoproteins were found in the fluid, since on hydrolysing the total protein with hydrochloric acid, no reducing sugars were formed. The main constituents of the fluid were, therefore, caseinogen, globulin and albumin, with traces of fat, lactose and simpiler nitrogenous substances (proteoses). In view of the obvious importance of the fluid in its relationship to the ultimate colostrum and milk secretions, it was advisable, as fat as was possible, to conduct a quantitative enquiry. Unfortunately, no one sample obtained was large enough to permit of a complete analysis being carried out on it. The following figures were, however, obtained and give some idea of the quantitative composition of the fluid. The data for milk and a sample of colostrum containing roughly the same amount of protein are appended for comparison. Sa. nipl es 1 and •> Total jjrotcin Caseinogen Coagulable pi (total N y. 0-37 'oteins ) 5-82 ... 1-70 ... 3-29 0/ /o ,Sa III pie 3. Fluid Colostrum' Milk= Total protein (N-. 0-37) Caseinogen Globulin' Albumin 0-37 % 2-50 „ 2-2S „ 1-00 „ (MS % 3-2.5 „ 1-97 „ 0-70 „ 2-90 % Trace 0-60 „ >Sa ^itple 4. Fluid Milk' Colostrum^ ypecific gravity Total solids ... Ash^ Fat' 1-014 8-80 % 0-80 ., 0-63 .. 1-032 1- -«J /o 0-75 „ 3-40 „ 1-050 14-19 % 0-96 „ 4-21 „ ' Crowther and Raistrick, Bloch. J., 10, 435. 1916. The sample represented the fourth milking after parturition. - Fleischmann. ' The globulin was separated from the liquid, remaining after removal of caseinogen with acetic acid, by neutraUsing with N/10 NaOH and saturating with magnesium sulphate. After filtering, tlie precipitate was well waslicd witli saturated magnesium sulphate solution and its amount estimated by tlie Kjcldalil method. The albumin was isolated from the filtrate by acidifying with acetic acid and standing in boiUng water for a short time. " Fleischmann. ^ Eugling. Sample taken 48 hours from parturition and contained about 6 % protein. " The ash gave the reactions for phosphate .and calcium. ' Paper Coil method. 1()"2 Fluid ubtdiiicd Jroiii Uddtrx uj Vinjiii lleij'ers Conclusions. A secretion lias hocii found to occur in small amount in tlie udders of virgin heifers which contains the characteristic proteins of colostrum, together with sUght amounts of fat, lactose and proteose. It follows that the initiation of mammary gland activity in the dairy cow is not necessarily dependent on pregnancy, but may be associated with the occurrence of the oestrous cycle. The processes which result in the elabora- tions of proteins would appear to be the most active iu this early stage of the glands activity, since only small amounts of fat and sugar seem to be formed. The fluid bears, on account of its globulin content, a closer relation to colostrum than to milk. The question arises as to whether the secretion continues to accumulate throughout pregnancy, or whether it undergoes gradual reabsorption. It is hoped to gain information on this and other points by following the changes in the yield and composition of the fluid at regular periods throughout the progress of pregnancy. REFERENCKS. (1) Mausilvll and HaLNAn. I'roc. Jtoyal Sue. 15., 89, lillT. (2) Hammond iind Maushall. Proc. Royal Hoc. 15., 87, lit! 1. (3) Woodman. Bioch. J., 15, 187, 1921. (4) Marsh.\ll. The I'hi/nioloyi/ oJUeproduclion, Low\on. I!(l(). (5) Lane-Claypon and Stakuno. Proc. Royal Soc. ii., 77, I'JOU. (0) Plimmek. Practical Urganic and Blo-Cliemislry, p. 4(10. {Received Januari/ ]''>/li, 1922.) Cambridge University Press Fungi. Ascomycetes, Ustilaginales, Uredinales. By Helen Gwynne-Vaughan, D.B.E., LL.D., D.Sc, F.L.S. With a frontispiece and ig6 text-figures. Royal 8vo. 35 J', net. The author's intention in this book is to present the fungus as a living individual; the scope is mainly morphological, but, in dealing with objects so minute, morphology passes insensibly into cytology. The introduction deals with fungi in general; the special part of this volume is limited to the consideration of the Ascomycetes, Ustilagi- nales, and Uredinales. Cattle and The Future of Beef-Production in England. By K. J. J. Mackenzie, M.A., with a preface and chapter by F. H. A. Marshall, Sc.D., F.R.S. Demy 8vo. is. 6d. net. In this book the author makes a strong ■plea for agricultural reform ; he drives home the lesson, taught us by the war, of the folly of relying upon other countries for food that might be produced at home, and specially urges the need of co-operation amongst farmers, of the development of "dual-purpose" cattle, and, above all, of systematic research on the part of scientists into the many problems of modem agri- culture. Dr Marshall contributes a chapter on the physiology of cattle. A Course of Practical Physiology for Agricultural students. ByJ. Hammond, M.A.,andE. T. Hai.nan, M.A. DemySvo. With blank pages interleaved. Cloth, 65. dd. net; paper covers, 4J-. dd. net. A Course of Practical Chemistry for Agricultural students. By H. A. D. Neville, M.A., F.I.C, and L. F. Newman, M.A., F.I.C. Demy 8vo. Vol. I, 10s. 6d. net; Vol. II, Part I, 55-. net. Designed primarily for the use of students taking the course for the Degree in Agri- cultural Science at Cambridge University. Volume I deals with the chemistry and physics of the soil. Volume II covers the second year's course on the Chemistry of Foods, Part I dealing tvith such exercises in Pure Organic Chemistry as are essential for agricultural students. Composition and Nutritive Value of Feeding Stuffs. By T. B. Wood, CB.E., M.A., F.I.C, F.R.S., and E. T. Halnan, M.A. Second edition. Demy Bvo. Paper covers. 15. net. British Grasses and their Employment in Agri- culture. By S. Y. Armstrong, B.A., F.L.S., of the School of Agriculture, Cambridge. Demy Bvo. With 175 illustrations. 75. net. Manuring for Higher Crop Production. By E. J. Russell, D.Sc, F.R.S. Second edition, revised and extended. Demy 8vo. With 17 text-figures, ^s. 6d. net. The Fertility of the Soil. By E. J. Russell, D.Sc, F.R.S. With 9 illustrations. i6mo. Cloth, y. ?iet; lambskin, /^s. net. Cambridge Manuals Series. A Short Manual of Forest Management. By H. Jackson, M.A., School of Forestry, University of Cambridge. Demy 8vo. 7^. net. A Guide to the Identification of our more useful Timbers. By H. Stone. Demy 8vo. With 3 plates, "js. 6d. net. The Production and Utilisation of Pine Timber in Great Britain. Part I. Production. By E R. Burdon, M.A., Investigator in Timber, and A. P. Long, B.A., Assistant Investigator. No. I. Sample Plots of Scots Pine at Woburn. DemySvo. \s. dd. net. No. ■2. Sample Plots of Scots Pine at King's Lynn. DemySvo. <)d.iiet. Cambridge University Press Fetter Lane, London, E.G. 4: C. F. Clay, Manager CONTENTS. (All RigtUs reserved.) PAOB 1. Newton, Robert. A comparative study of winter wheat varieties with especial reference to winter-killing . . . . . 1 2. GOODEY, T. On the susceptibility of clover and some other legumes to stem-dise!\se caused by the eelworm, Tylenchus dipsaci, syn. devastati-ix, Kuhn. (With Plate I) 20 3. Salaman, R. N. and Lesley, J. W. Genetic studies in potatoes; sterility. (With Plate II) 31 4. Crowtheu, Charles and Woodman, Herbert Ernest. A study of nitrogen metabolism in the dairy cow 40 5. Armstrong, S. F. The Mendelian inheritance of susceptibility and resistance to yellow rust {Puccinia glumarum, Erikss. et Henn.) in wheat ........... 57 6. Woodman, Herbert Ernest and Hammond, John. Note on the composition of a fluid obtained from the udders of virgin heifers 97 The Journal of Agricultural Science is issued in quarterly parts of about 100 pagee with plates and figures, four parts constituting a volume. Volumes I — XI (1905 — 21) now ready. In four parts, paper covers, 30*. net per volume. Single numbers lOi. net each. Quotations can be given for Buckram binding cases and for binding subscribers' sets ; also for bound copies of back volumes. Papers for publication may be sent to Professor T. B. Wood, C.B.E., M.A., F.I.C., F.RS., School of Agriculture, Cambridge, or to the associate Editors. Other communications should be addressed to the University Press, Cambridge. Contributors will receive twenty-five copies of their papers free. A limited number of additional copies may be purchased if ordered when final proof is returned. In order to meet the increased cost of printing, contributors will be asked to pay a proportion of the cost of setting up tabular matter and of preparing blocks where these exceed a certain proportion of the whole communication. The subscription price, payable in advance, commencing with Vol. X is 30«. net per volume (post free) ; single numbers 10». net. Subscriptions may be sent to any Bookseller, or to Mr C. F. CI/AY, Manager, Cambridge University Press, Fetter Lane, London, E.C. 4. The Cambridge University Press has appointed the University of Chicago Press agents for the sale of The Journal of Agricultural Science in the United States of America and Canada and has authorised them to charge the following prices : — Subscription, $6.00 net per volume ; single copies, $2.00 net. PRINTED IN ENOLAMD AT THE CAMBRIDOE UNIVEB8ITY PRESS Vol. XII. Part 2 April, 1922 THE JOURNAL OF ' AGRICULTURAL SCIENCE EDITED FOR THE PLANT BREEDING AND ANIMAL NUTRITION RESEARCH INSTITUTES AT CAMBRIDGE, AND THE ROTHAMSTED RESEARCH INSTITUTES BY Professor R. H. BIFFEN, M.A., F.R.S., Cambridge Sir a. D. HALL, K.C.B., M.A., F.R.S., Ministry of Agriculture, London F. H. A. MARSHALL, Sc.D., F.R.S., Cambridge E. J. RUSSELL, D.Sc, F.R.S., Rothamsted Experimental Station, Harpenden Professor T. B. WOOD, C.B.E., M.A., F.LC, F.R.S,, Cambridge IN CONSULTATION WITH B. C. ASTON, Department of Agriculture, Wellington, New Zealand Dr C. a. barber, C.I.E., School of Agriculture, Cambridge. . Professor B. T. P. BARKER, M.A., National Fruit and Cider Institute, Long Ashton, Bristol W. BATESON, M.A., F.R.S., John Innes Horticultural Institute, Merton, Surrey J. R. CAMPBELL, B.Sc, Department of Agriculture, Dublin I. B. POLE EVANS, Department of Agriculture, Pretoria, South Africa F. B. GUTHRIE, Department of Agriculture, Sydney, N.S.W. Professor J. HENDRICK, B.Sc, Marischal College, Aberdeen Sir T. H. MIDDLETON, K.B.E., C.B., M.A., The Development Commission, London Dr frank T. SHUTT, F.I.C, Experimental Farms, Ottawa, Canada Professor W. SOMERVILLE, M.A., D.Sc, School of Rural Economy, Oxford Dr a. C. true. Department of Agriculture, Washington, D.C., U.S.A. < Sir FRANCIS WATTS, K.C.M.G., Barbados Dr H. J. WHEELER, American Agricultural Chemical Co., Boston, Mass., U.S.A. Cambridge University Press C. F. Clay, Manager London : Fetter Lane, E.C. 4 also H. K. Lewis & Co., Ltd., 136, Cower Street, London, W. C. t and Wheldon and Wesley, Ltd., 28, Essex Street, London, W.C. 2 Chicago : The University of Chicago Press (Agents for the United States and Canada) Bombay, Calcutta, Madras: Macmillan and Co., Ltd. Tokyo: The Maruzen-Kabushiki-Kaisha ^rice. Ten Shillings net {Issued 30 May, 1922.] Messrs Longmans & Co.'s List ENGLISH FARMING PAST AND PRESENT By the Right Hon. LORD ERNLE (Rowland E. Prothero), President of the Board of Agriculture and Fisheries, 1916-1919. Third Edition. 8vo. 12s; 6d. net. A chapter has been added to the third edition of this book on agriculture during the years igi4-i8. Historically, the interest of the period lies in the experiment of Slate Control, the form which it assumed, the results that it attained. Practically, the exceptional conditions which necessitated the adoption of the experiment must to a great extent neutralise its utility as a guide to action in times of peace. The appendices on.Corn Prices and Agricultural Wages havj aUo been brought up to 19JO. WEEDS OF FARM LAND By WINIFRED E. BRENCHLEY, D.Sc, F.L.S., Botanist, Rothamsted Experimental Station. With 41 Illustrations. 8vo. 12s. 6d. net. "A volume of surpassing interest and value. .. . A soand guide on all aspects of the subject of farm weeds. " — Tht Field THE ROTHAMSTED MONOQRAPHS OF AGRICULTURAL SCIENCE Edited by E. J. RussELt, D.Sc, F.R.S. SOIL CONDITIONS AND PLANT GROWTH By EDWARD J. RUSSELL, D.Sc. (Lond.), F.R.S. Fourth Edition. With Illustrations. 8vo. 16s. net. LONGMANS, GREEN & CO., 39 Paternoster Row, London, E.C. 4 E. & S. LIVINGSTONE. iii|>uBition n{ nun-fatty matter in 1 1 1 group weight weight body Ash 1 I'rf'toiti W.-.t.r I'lA lb. lb. % U ft U 5 107 900 86-4 4-00 4-44 20-72 74-84 4-69 4 207 163-2 153-5 6-01 4-71 20-36 74-93 4-33 4 301 2461 218-5 11-19 4-83 21-13 74-03 4-37 5 416 339-5 303-8 10-55 4-86 21-59 73-55 4-45 5 504 418-1 360-7 13-73 4-89 22-19 72-91 4-55 3 614 493-7 424-5 13-97 5-33 22-58 72-08 4-46 4 708 587-8 490-2 16-57 5-36 22-31 72-33 4-15 3 815 692-1 563-8 18-52 5-20 23-08 71-72 4-43 3 905 774-5 587-6 24-08 5-48 23-27 71-25 4-24 4 1008 880-7 643-6 26-91 5-41 23-41 71-18 4-31 3 1108 976-3 663-8 3203 5-62 24-13 70-25 4-29 3 1204 1077-0 728-8 32-32 5-46 23-66 70-88 4-:»3 I 1302 1 150-5 776-6 32-50 5-60 23-40 7100 4-18 I 1413 1237-0 834-1 32-58 5-21 23-95 70-84 4-60 1 1508 1352-9 843-9 37-59 5-14 25-19 69-66 4-90 79 78 — \ ^ fo77 3} '' ■1 76 g S 0) 1 75 X \ ■ ^ f o 74 f a 0) u S 73 \^x Cu 10000 1 0001 Hefiise ... 3-2 8-2 lid Fasted live weight (lb.) 2r)9 1232 Ul'.l The coiupo.sition o( the lattciiing increa.se (oxen) calculated from tlie data actually recorded by Lawes and Gilbert is as follows: Fat 94-54%, Water .j-49 %, Protein O-G %, Ash - ()-«4 %. The percentages of water and protein in the increase might be regarded as an indication of growth, but as the animals were four years old it is improbable that any increase in size occurred, or that such increase, if it did occur, would beaccom])anied by a loss of 1-71 lb. of ash ingredients. Further, as the ratio of water to protein in the increase is 9/1 whereas in the whole body it is only 3/1 it is more reasonable to attribute the apparent increase in non-fatty constituents to difference in composition of the two animals. The individual variations revealed by Haecker's data are more than sufficient to justify this assumption. The inference therefore is that, in fully grown animals the fattening increase consists entirely of fat. Tiiis, of course, is acceptable on other grounds, but those who have maintained it have had to do so in the face of numerical data which could lie (|uoted against it and i>f w hich. liitlicrto. lu) satisfactory explanation was fcuthcomiiig. It is to be expected that iu other animals I lie compo.siiioii of t lie uou- fatty matter will alter with age iu much the same manner as in cattle, and the author anticipated no difficulty in tracing this relationship in pigs from Swanson's data. A preliminary survey, however, indicated that these results might be affected by the differences in feeding as well as by the age and individuality of the animals, and that the variations due to these combined causes would make it difficult clearly to distin- guish the influence of any one. The best that could be done, it seenu'd, was to divide the data into three groups as shown in the table below. It appears that the ratio of protein to ash in pigs is higher than in cattle, but the percentage of water in the non-fatty matter alters with age not only in the same manner but even in the same degree. The latter J. Alan Murray 107 inference, however, must be to a large extent discounted in view of the magnitude of the probable errors. It could not have been deduced from Swanson'.s data alone and until confirmed liy further evidence it must be regarded as hypothetical. Table III. Influence of Age (Pigs). (.'(imposition of non-tatty matter No. of animals ^ empty weight Ash Protein Water PIA Under 60 Ih. 10 4-15 + 0-87 19-03^2-55 76-834312 4-60 6U-1(X» „ 10 3-91 ! 0-36 19-89 i 1-35 76-02 L 1-41 5-09 UIO-170 .. 11 4-01 0.30 20-90 (1-96 7509 1-04 5-21 The test of comparison with Lawes and tiilbcrt"s data is not con- clusive, but, so far as it goes, it indicates that the inference is perhaps more reliable than might be supposed. For this purpose the percentage of water in the non-fatty matter was calculated, as before, by the formula w = 90y;r'°^*"^; but owing to the difference in the ratio of protein to ash different coefficients must be used for these ingredients, thus P = 0-83 (100 -^W);A= 0-l('i (100 - W). The results are as follows: Store pig Fat pig Calculated Observed Calculated Observed Water Protein . . . Ash Fat . 58-28 . 14-28 2-86 . (24-.59) 58-15 14-45 2-82 24.59 42-71 11-13 2-22 (43-94) 43-01 11-35 1-72 43-94 Live weight (lb.) — 94 — 185 In the following table the data are grouped according to the kind of food consumed by the animals. Table IV. Influence of Food (Pigs). Food No. of animals Empty weight Fat-free empty weight Fat in live weight Composition of non-fatty matter Ash Protein Water PIA Corn alone 5 lb. 80-27 lb. 47-45 o/ 36-69 o/ o 3-56 0 17-93 o, 78-51 6-03 Corn and ash ... 5 130-74 68-25 37-03 4-16 17-39 78-45 4-17 Corn and protein 10 217-94 109-31 48-01 3-54 21-34 75-12 li-07 Corn, ash and | protein ...\ 5 226-74 118-57 45-5.5 4-41 21-17 74-42 4-82 In order to interpret these results correctly it is necessary to make allowance for the difference in size, i.e. to compare them with data 108 The Chonical Comjiosition <>/ Aniinnl Jio(/ies calculated for animals of the same size by means of the formula above, as follows: Corn aliiiit' Com and ash Com and protein Corn, asli and protein Cald Obseril Card ObserM Car.l Obser'd Cal'd Obserd Water Protein Ash /O 78-25 18-12 3-63 0 78-51 17-93 3-58 77-21 18-98 :! SI 78-45 17-39 lie. 75-91 20-07 102 75-'l2 21-34 3-54 o 75-68 20-26 406 o/ /o 74-42 2117 4-41 The correspondence between the observed and calculated data is, throughout, as close as could be expected in any circumstances. The inference therefore is that the differences which do occur should be attributed to individual variation and that the effect of the additions to the food was nil. It is to be observed, however (Table IV), that the addition of ash to the corn has reduced the ratio of protein to ash whereas the addition of protein has produced no appreciable alteration in that ratio. When ash and protein are added together the ratio is reduced only about half as much as by addition of ash alone. When the results are stated as percentages of the several ingredients these effects are obscured by the large amount of water present. Swanson remarks that the addi- tion of protein and ash accelerates the growth of the animal but does not affect its composition. The only data relating to sheep at present available are those of Lawes and Gilbert and. as these refer to only five individuals, they are inadequate for purposes of the present investigation. They may, however, be used to test the applicability of the formula deduced for cattle. Thus it is found that the ratio of protein to ash (-1-3) is practically the same but the coefficient in the formula for water must be reduced, i.e. w = 87»r "'*'''. Calculated on this basis the results are as follows: Fat-free live weight (lb.) 50-63 70-82 73-84 74-20 123-80 Fat (per cent, of live weight) ... 28-50 23-50 18-70 .35-60 45-80 Ratio of protein to ash 4-43 4-34 4-67 4-34 3-75 W'ater(percent. of non- 1 observed 75-91 74-42 76-10 74-63 71-80 fatty matter) ... (calculated 75-46 74-56 74-40 74-42 73-07 Amount of Body Fat. In farm animals generally the ratio of fat to non-fatty matter varies within wide limits. It depends lai-gely upon the quantity and quality of the food and, to that extent, it can be controlled. Elsewhere(i) the author has inferred that the maximum amount of body-fat is about J. Alan Murray 109 liO per cent, of the live weight. This estimate was based on the increase ill basal katabohsm, determined by Armsby("'), in an ox. Pigs probably have a smaller capacity for food than ruminants of the same size; but, on the other hand, the food which is customary and suitable for them has a much higher productive index (ratio of dynamic to total energy). It was recognised that the result depends upon these factors and that any animal would require an indefinite time to attain the maximum degree of fatness of which it is capable under any given conditions. The fact that over 60 per cent, of body fat was recorded in two cases in Swanson's data indicates that the theoretical maximum is probably higher for pigs than for ruminants. At any rate it disposes of the objec- tion that the estimated maximum (60 per cent.) is far beyond what had hitherto been recorded. The theoretical minimum of body fat is zero. The lowest recorded in the data under consideration is 5-27 per cent, in the pigs and 3-64 per cent, in the ruminants. There is no reason to believe that these records are the lowest attainable in living animals. Attention may be called to the fact that the diminishing percentage of water in the non-fatty matter coincides with diminishing rate of growth and to the possible connection between these phenomena. The biological significance of the constants in the formula for water may also prove a matter of interest to pliysiologists. Summary. Animal bodies are composed of fat and non-fatty matter. The relative proportions of these two ingredients vary within wide limits but can be controlled by food. The non-fatty matter consists of water, protein and ash. The percentage of water varies with the age of the animal in a definite manner. This has been determined with tolerable certainty for cattle. The available evidence indicates that the same formula is applicable to pigs and, with .slight modification, also to sheep, but these inferences require confirmation. The ratio of protein to ash is the same in sheep as in cattle but in pigs it is higher. In any case it does not alter with the age of the animal but it may be influenced to a certain extent by the food. Individual variation is wider in pigs than in ruminants. The average composition of the whole body at any stage can be calculated when the live weight and percentage of fat in it are known. 110 The Chemical Composition of Animal Bodies i:k1'KRFA'('rs. (1) MruKAV. J. A. lOlfl. Meat Produclion. .I„iir». Ai/rir. ,SV,. 9. I>art 2, 174. (2) I.AWKs and Gilbkrt. Rothcim-slfil MemoirK, 3. (3) Hakckkk, T. L. 1920. Investigations in Beef I'rodiiclion. liuU. 193, I'niv. Minnesota Agiie. Ex per. Sta. (4) SwANSON, r. 0. 1921. Krteet of Rations on Developnu'nt of I'igs. Joiirn. Aijric. lies. 21. .5. (.')) Arm.sby, H. p. 1917. InHuenee of the Degree of Fatness of Cattle ujioii tlieir Utilisation of Peed. Jowrn. Agric. Rex. 11, 10. {Received 18 Janiuiri/, 1922.) THE EFFECT ON THE PERCENTAGE COMPOSITION OF THE MILK OF (a) VARIATIONS IN THE DAILY VOLUME AND {!>) VARIATIONS IN THE NATURE OF THE DIET. By WILLIAM TAYLOE, M.D., D.P.H. {Carnegie Research Scholar), AND ALFRED D. HUSBAND, A.I.C. (From the Rowett Research Institute, Aberdeen.) (With 4 Text-figures.) The study of lactation has ahiiost always been undertaken from one of two very diverse points of view, the physiological and the commercial. The chief aim of the physiologist has been to determine, either the origin of the various constituents of the milk and the method of their elabora- tion, or the extent to which internal secretions affect the flow of milk by their initiation, inhibition or stimulation of milk secretion. From the commercial or agricultural point of view, on the other hand, the chief interest has centred round the problem of the production of butter fat, and a very large number of experiments have been conducted to deter- mine by what method of feeding a cow could be caused to give the maximum yield of milk with the highest percentage of fat. Much of this work is of great value, but in many cases the fat is the only constituent of the milk which has been determined separately, the other constituents having been estimated together as milk solids other than fat. With regard to the diets, unless the calorific consumption is stated, a simple addition of extra protein or fat to a ration does not allow of a judgment being formed as to whether the alteration in the daily volume and percentage composition of the milk is due to protein or fat per se, or simply to an increase in the calorific value of the ration. This considera- tion would appear to have been lost sight of in some cases. The most complete investigation of the influence on the percentage composition of the milk of feeding with an excess of one of the energy- yielding constituents of the food is that of Voit(i), who worked with a bitch. He found that, while an excess of one constituent in the food 11 "2 PerecntfUfe Composttioit of Milk teiulod to uive ;i slijilit iucrease in tlic percentage of that constituent in the milk, the deviation from the normal was comparatively slight. The work done since V'oit's time has produced rather contradictory results. For example, Ingle (2) found that a protein rich diet increased both the yield of milk and the percentage of fat, while Crowther(3) found that a food rich in protein gave a decrease in the yield but an inc-rease in the fat content. Again, Jordan and Jenter(4) show that the amount of fat in the food is without influence on the percentage of fat in cows" milk, while Morgan, Berger and FingerUng("i) found that a fat poor diet produced in goats a milk with a low percentage of fat. There is difficulty in correlating contradictory results such as these, and the difficulty is increased where there is uncertainty as to the calorific value of the food digested and absorbed, and, further, when no account has been taken of the probable influence on the percentage composition of a change in the daily volume of milk secreted. In the voluminous literature on milk there is very Uttle reference to the influence of volume on composition — except in the case of the fat — and it is possible that this may explain some of the contradictory results obtained by different workers on the question of the influence of ditt'erent diets, PRESENT IXVESTIGATIOX. The present investigation was undertaken to determine : A. To what extent variations in the daily volume (or rate of secre- tion) of the milk were accompanied by variations in its percentage composition. B. To what extent the percentage composition of the milk could be influenced by diet. The experimental animal was the goat, and throughout the investiga- tion the milking was done at 9 a.m. and 5 p.m. The total volume of the 24 hours' secretion of milk was measured, mixed and anal)'sed daily, the percentages of total protein, casein, albumin plus globulin, non-protein nitrogen, fat, lactose and ash being determined. METHODS OF ANALYSIS. Total Protein. The total nitrogen was estimated by the Kjeldahl method, the protein factor G-38 being used for the determination of total protein. Casein. The casein was precipitated with acetic acid and filtered off, after which tlie nitrogen was determined by the Kjeldahl method, the same protein factor. (!-38, being made use of. W. Taylor and A. D. Husband 113 Albumin and Globulin. These were precipitated by tannic acid from the filtrate obtained in the estimation of casein, after which they were estimated together as described above for casein. Non-Protein Nitrogen. This was estimated by the difference between the percentage of total protein and the sum of the percentages of casein, albumin and globulin. Fat. By the Soxhlet method. Lactose. Fehhng's method and Benedict's method were both made use of; but in the individual experiments one method was adhered to throughout. A.sh. By the ordinary method of ignition. EXPERIMENTAL. A. The Effect of Variations in Volume on the Percentage Com- position OP THE Milk, with an Animal on a Diet of Constant Composition. Experiment I. (Analytical Data, Table I.) In this experiment a series of milk analyses was carried out from day to day in the case of a goat fed on a diet of hay and a mixture of meals made from locust beans, earth nuts and oats, this mixture being sweetened with sugar. The animal was allowed to eat to its appetite, but the nature of the diet was constant throughout. Owing to the con- finement, the monotonous diet and the advanced stage of lactation, the volume of the milk fell steadily throughout the experiment with, as a rule, but slight variations from day to day. Table I. Alb jmin Non- + protein Volume Protein Casein globulin nitrogen Fat Lactose Ash c.e. 0' /o o/ /O o- /o % o/ /o o/ 1 310 4-50 307 1-20 0-23 4-54 4-26 0-98 2 290 4-73 3-38 14 0-21 4-85 4-28 0-97 3 320 4-.50 3- 12 15 0-23 5-24 4-24 0-98 4 310 4-39 301 15 0-24 5-80 4-21 0-94 5 2.55 4-59 319 I(i 0-23 5-55 419 0-94 6 250 4-72 334 20 0-18 5-12 4-24 0-90 7 245 4-82 3-37 24 0-21 5-56 4-24 0-98 8 240 4-80 3-41 24 0-21 502 4-35 0-97 9 225 500 3-,54 28 019 5-38 4-30 100 10 235 4-89 3-39 28 0-21 5-76 4-26 0-98 11 185 5-25 3-70 35 0-20 5-56 4-21 100 12 190 5-17 3-63 36 019 6-01 4-30 102 13 220 4-82 3-39 25 0-18 0-86 4-26 0-97 14 240 4-.54 314 24 010 6-37 4-26 108 15 170 501 3-52 25 0-25 0-30 4-20 0-98 10 145 5-40 3-79 30 0-24 5-94 4-30 104 114 Percentage Composition of Mill: Haiiiiniiii(l and Ha\vk(ii), ami otluTs, liavc shown that an inverse relationship exists between the percentage of fat and the daily volume. This was well brou<;ht out, but it may be seen from ¥\'^. 1 that the percentage of protein also varied inversely as the volume with a regu- larity at least equal to that of the fat, though the extent of the variation was less marked. The percentage of lactose maintained a very constant level, varying between M!) per cent, and A-'M) ))er cent. Tiic percentage of ash, which varied from 0-94 per cent, to 1-(I8 per cent., may l)e seen from the analytical data to have shown a tendency to rise with the fall in volume towards tlie end of the e.xperuiient . 10 11 12 13 14 15 16 Days 6% -^ 300i-p-> ■ Fat riiitrui Lactose Volume lOOc.c.-* Fig. 1. Showiiii-' the effect of variatiims in voliiiiie oi\ tlie percentage composition of tlie milk. Experiment II. (Analytical Data, Table II.) The fluctuations in volume from day to day having been compara- tively slight in the preceding experiment, advantage was taken in this case of the decrease in yield brought about by an absence of food, and the increase which follows the resumption of feeding. A goat was fed for eight days on an unrestricted diet of hay and oatmeal, and on the 9th and 10th days no food was taken. PVom the 11 til dav onward the feeding was resumed. W. Taylor and A. D. Husband 115 Table II. Volume Protein Fat Lactose Ash n p o/ 0/ O 0/ 1 500 2-97 3-95 4-33 0-82 o 480 309 4-32 407 0-84 3 460 307 4-08 412 0-88 4 470 3-00 4-09 4-00 0-87 5 470 2-91 4-45 4-00 0-89 6 460 2-95 4-47 4-07 0-79 7 460 2-98 4-53 4-07 0-83 8 460 2-93 3-95 4-07 0-79 9 350 3-02 5- 17 3-80 0-80 10 50 9-24 1016 2-31 1-37 11 340 2-88 5-40 4-00 0-98 12 350 2-52 4-81 4-23 0-91 13 320 2-93 5-53 4-20 0-94 14 230 3-23 5-98 4-21 0-96 15 275 2-93 5-71 405 0-89 l(i 300 2-77 5- 16 3-82 0-90 17 275 2-97 4-97 412 0-92 The tollowinji extract from tlie analytical data brings out very clearly the inter-relation.shij] of volume and composition. Volume Protein Fat. Lactose Ash ,. ,. o o O O' *^-^- o O O n Last dav before starvation ... 460 2-93 3-95 407 0-79 Second dav of „ ... 50 9-24 1016 2-31 1-37 after 3.X) 2-52 4Sl 4-23 O'.ll It will be seen that with the great fall in volume on the second day of starvation there was a rise in the percentages of all the constituents of the milk, with the exception of the lactose, the percentage of which came down with the volume. The percentage of fat, as was e.xpected, showed a marked rise, but the percentage of protein was found to rise markedly also ; and the percentage of ash, although affected to a less extent, distinctly increased with the great and sudden fall in the volume of milk. The lactose, a constituent of the milk the percentage of which is normally very constant, showed, on the other hand, a percentage de- crease as definite as the percentage increase in the protein, fat and ash. With the renewed consumption of food the daily volume of milk increased, with, at the same time, a close approximation to its previous percentage compo.sition. Experiments III and IV. (Analytical Data, Tables III and IV.) These experiments were conducted to determine whether the inter- relationship of volume and composition, maintained during Exps. I and II, would still hold good during the physiological increase and decrease 116 Percentage Conipositum of Mill, in volume wliich takes place at the commencement and cessation of lactation respectively. Table 111. Alhumin Non- ■f protein \(iliiiue Protein Casein "lobulin nitrOKcn Kat Lactose Ash c.c. /o o/ /o % % % % O' n 1 1.30 508 302 2-66 0-30 807 3-65 0-81 3 250 411 2-.i5 1-37 019 5-56 4-39 0-79 5 330 407 2-59 1-31 017 5-91 4-70 0-81 7 3t)5 3-88 2-45 1-25 0-18 5-70 4-72 0-78 i» 400 3()5 2-44 1-05 Olf) 608 502 0-68 11 500 3-81 2-49 1-08 0-18 5-18 5- 10 0-76 13 430 3-62 2-29 1-03 0-30 5-22 506 0-75 Table IV. Albumin Non- ■¥ protein \'()liime Protein Casein globulin nitrogen Fat Lactose Ash o/ o/ o/ o/ O- 0/ O' c.c. ,0 /o /o .0 /o /o 1 400 5-36 4-26 0-93 0-17 5-53 406 0-92 •> 340 5-65 4-55 0-98 012 6-02 406 0-96 3 275 (il7 4-91 1-05 0-21 6-53 4-24 0-96 4 250 0-52 5-25 114 0-13 7-83 4- 16 0-99 o 210 0-60 5-30 1-20 0-16 7-20 4-05 101 (i 185 7-14 5-59 1.39 0-16 7-96 4-05 1-02 7 110 8-08 6-38 1-.55 0-15 8-86 4 16 112 8 80 801 0-84 1-66 0-11 8-63 3-71 M3 9 55 8-47 (i-61 1-73 0-13 9-70 3-(l2 108 10 35 9-96 7-5(> 2-28 0-12 11-36 3-07 1-30 U 45 10-46 7-64 2-61 0-21 10-94 2-27 1.33 12 35 8-96 6-50 2-45 0-01 9-31 2-li2 119 13 30 9-22 6-68 2-57 y 8-40 2-89 1-20 14 20 9-20 4-90 4-20 0-10 9-44 2-35 1-36 15 25 906 5-88 306 0-12 9-06 2-27 1-18 16 15 9-88 6-64 3-21 003 9-08 2-20 1-36 17 15 10-88 7-28 3-56 0-04 9-49 2-20 1-42 18 .5 11-20 7-64 3.16 ? 7-65 — 1-41 It may be seen from Figs. 2 and 3 that tlie changes in the percentage composition of the milk which accompanied alterations in volume, due to these causes, were the same in nature as the changes which accom- panied the fluctuations in volume brought about in the preceding experi- ment by the two days' starvation. (During E.xp. Ill, commencement of lactation, a milk analysis was carried out only on alternate days.) As a result of these and other similar experiments, not described here for reasons of economy of space, it seemed possible to formulate the following general principle : That, on a diet of constant composition, the percentages of all the (-oustituents of the milk, with the exception of the lactose, tend to vary inversely as the daily volume of milk secreted ; AV. Taylor and A. D. Husband 117 auil that the percentage of lactose, whih^ uoriiially very coustant, tends to vary directly as the voliune, this tendency being particularly apparent at the beginning and end of lactation. 11 13 Doys Prut fill 100c.c._^ Fig. 2. Begiiiniiig of Lactation. 1 2 3 4. 5 6 7 8 9 101112131415161718 Days PioteiD 1 0°;. 400c.e.-> Fat, 200c.c.-> Oc.o.-^ Lactose Volume Fis. 3. End of Lactation. 118 Percentage Composition of Milk B. The Influence of Diet on the Volume and Percentage Composition ok tiik Mii.k. Experiment V. (Analytical Data, Table V.) The above prin(i])le haviiii: been established on a diet of constant composition, it was desired to see to wliat extent it niiyht be departed from on diets which were abnormally hi 0-93 Normal (.'> day.s) 380 4-42 3-43 0-99 (l-2(i 4-40 3-91 108 High r-arbohydratel (13 day's) )' 439 4-42 334 l-OS Oil) 3-52 4-17 0'92 (In ciilculatiug the average percentage of protein the iioii-proteiii nitrogen was not included.) To do this the normal calorific intake of the goat was first estimated by feeding her for 20 days on an unlimited and congenial diet of known composition, weighing the amount of food left over at the end of each day, and finally calculating the average daily calorific consumption. On this basis each of the experimental diets was made up. In this experiment, a high fat, a high ])rotein and a high carbohydrate diet were given, a period of normal diet intervening between each period of special diet. The average daily volume of the milk has been worked out for each dietary period, and against this has been set out the average percentage composition of the milk for the corresponding period. The first two days after each change of diet have, however, been omitted in the calculation of these averages. For reasons of space the results of the individual milk analyses from day to day have not been given in the tables, hut only the average daily volume and average percentage composition for each dietary period. (For details of diets, see Table VI.) W. Taylor and A. D. Husband 119 Table VI. High Fat Diet. Hay Earth nuts 243 Oatmeal 102 Locust bean meal 2ti(j Sui,'ar 7 1 )gms.' (Protein 121-7 gms. Fat 136-2 „ I Carbohydrate .599-9 „ Hiijli Protein Diet. Hay 560 gms. ) (Protein 305-1 gms. Plasmon 300 "„ [ = -^ Fat 41-2 ., Oatmeal 335 .. ] (Carbohydrate .504-5 „ High Carbohydrate Diet. Hay 586 gms. Locust bean meal 374 „ Maize meal 531 „ Oatmeal 9.S „ (Protein 105-3 gms. J Fat 40-3 „ (Carbohydrate 933-2 „ = 4225 calories. = 3702 calories = 4633 calories. The result of this experiment can most clearly be seen by consulting Fig. 4. In this graph the horizontal lines represent average volumes and average percentages, the sloping lines connecting the various horizontal levels representing the two days at the beginning of each dietary period which have been left out in the calculation of the averages. 5% 700 e.c. 300 c^ Fig. 4. Showing the effect of diets of different composition on the volunie and percentage composition of the milk. Journ. of Agric. Sci. xii 120 Percentage CoinposUioit of Milk It will be seen that jn no case was there a direct increase in the per- centage of that constituent of the milk corresponding to the constituent of the diet which was present in excess. The protein, fat and ash all tended to vary, albeit to different degrees, inverseh-, and the lactose directly, as the daily volume of milk secreted. It is true that with the different diets variations in the percentage composition of the milk were produced, but these variations were of such a nature, and took place to such a degree, as could be accounted for by the variations in the rate of secretion, and, therefore, in the daily volume of the milk. Fig. 4 brings out very clearly the direct relationsiiip existing between the percentage of lactose and the daily volume of milk secreted. Several different feeding experiments were carried out with, in every case, results similar to those stated above. In one of these it was con- sidered that a high protein diet had markedly stimulated the rate of milk secretion. In this experiment, tliercfore, it is worthy of note, the high protein diet was made up in such a way that it was of lower calorific value than either the high fat diet which preceded it, or the high carbo- hydrate diet which followed it. The greatest secretion of milk, however, was got on the diet of lowest calorific value, viz., the high protein diet. The following are the main features of the diets set out against the average daily volumes of milk obtained. Volume of C'aloritiu value of diet milk High fat diot ... 4230 calories (containing 136 gms. 400 c.c. of fat) High protein diet ... 3700 calories (containing 305 gnis. 540 c.c. of protein) High carbohydrate diet 4630 calories (containing 933 gms. 440 c.c. of carbohydrate) There is one constituent, however, not of the milk but in the milk, the percentage of which was found to have a direct relationship to diet, viz., the non-protein nitrogen. This was estimated by the difference in the figures for the percentage of total protein and the sum of the per- centages of casein, albumin and globulin. Wliilc not protein, therefore, it is expressed separately in terms of protein, in view of the fact that it was included in the estimation of nitrogen from which tfie percentage of total protein was calculated. It was found to vary from about 0-1 per cent, to 0-4 per cent, with the amount of protein taken in the food. (This is being dealt with in another })ubUcation.) W. Taylor and A. D. Hitsband 121 DISCUSSION OF RESITLTS. It seems justifiable to conclude from the results of these experiments, that with the exception of the non-protein nitrogen, which is not a product of the mammary gland, diet has no direct influence on the percentage composition of the milk. It has. however, an indirect influence by reason of its effect on the daily volume, but the percentages in which the various constituents of the milk will be present in any particular daily volume, obtained by some special method of feeding, will be the same as those in which they would be present in the same daily volume were it obtained in the ordinary course of lactation. Indeed, the inverse relationship of percentage of fat and daily volume is so unvarying on the average, as Tocher (7) points out in the case of the cow — in his statistical analysis of milk records, where he deals with the milk of thousands of cows — that it may be stated, and illustrated by a graph, with the mathematical precision of a proposition in Euclid. He says: "It will be seen that there is a direct proportional relationship between quantity and average quality. "Given two known values of average quahty for any two types of quantity, a straight line joining the two values of average quality and extended on either side will give the average quahties for all other types of quantity, average quahty being represented in value by a straight line proportional to its value standing at right angles to the base line of quantity at a point corresponding to its appropriate type of quantity." By the term " quality "' in the above statement is understood, of course, the percentage of fat. While this is true of the fat it is probably no less true of the protein. Throughout the present investigation the percentage of protein in the milk and the daily volume varied inversely with a regularity at least equal to that of the fat. It is true that these percentage variations were not so marked as in the case of the fat, the inverse rise and fall with each variation in volume having been less, as a rule, but they took place from da)^ to day with an equal consistency if to a less degree, even during periods of special dieting. The probabihty is, therefore, that it will be found from a statistical analysis of milk records — when the requisite data become available — that the relationship of average percentage of protein to each particular daily volume can be formulated as clearly and precisely in the case of this constituent of the milk as has been done by Tocher in the case of the fat. It is true that in a graph the hne representing the fat and that 9—2 122 Percentage Comjfosition of Mil h representing the protein would be at different angles, as the protein neither rises nor falls with variations in volume to the same extent as the fat, but the likeUhood is that where Tocher uses the word "quality" the words "percentage of protein" might be substituted, without im- pairing the accuracy and applicability of the statement. During the present investigation the individual protein constituents were estimated as casein on the one hand, and albumin plus glol)ulin on the other, and it was found that, from day to day, the casein and albumin plus globulin were present in varying proportions. Sometimes, when there was a rise in the percentage of protein, it was found that this was due chiefly to the casein; and again, on some other occasion, the albumin and globulin would be found to have contributed largely to the rise. Despite these irregularities, however, it was found that there was a tendency for the casein on the one hand, and the albumin plus globulin on the other, to maintain individually, as well as in combination, an inverise percentage relationship to volume ; and, consequently, these protein constituents of the milk were found present in larger percentages at the beginning and end of lactation than during the period when it was at its height. With regard to the ash, this was found to he the least variable of the constituents of the milk, but the analytical data for E.xps. II and V show that where there was a marked fluctuation in volume the percentage of a.sh .showed an inverse relationship to it. The relationship of lactose to volume is not so apparent as that of the protein or fat. Crowther and Euston(S) say: "It will be seen that the percentage of sugar, after rising a httle with the early stages of lacta- tion, fell steadily throughout the rest of the period." While due credit must be given for this observation it nuiy be pointed out that the per- centage of lactose is only indirectly connected with the period of lacta- tion. Its real relationship is to volume, and the accuracy of the words "steadily throughout" would seem to depend on the evenness with which the volume rises at the commencement of lactation, the evenness with wliich it is maintained, and the evenness with which it falls at the end of lactation. During the present investigation, where there were wide variations in volume at the height of lactation, due either to dieting or starvation, the percentage of lactose was found to vary also, and showed a direct relationship to it — as may be seen in Fig. 4. The inter-relationship of vohime and com])osition suggests a theory of milk secretion which may be put forward here, and on wliich further work is being carried out. W. Taylor and A. D. Husband 123 Vau der Laan(9) shows that a state of osmotic equihbrium exists between the blood and the milk, and that this equahtv of osmotic pressure persists even in diseases of the udder. As lactose is the substance which is likely to play the most considerable part in determining the osmotic pressure in the milk, the osmotic pressure in the blood is likely to be expressed to a very great degree by the percentage in which this con- stituent is present in the milk ; and as, normally, the osmotic pressure in the blood must be considered to be a constant, with but slight varia- tions from day to day, it is to be expected that an equal constancy will prevail in the percentage in which the lactose is present in the milk. It has been found that, while the percentage of lactose tends to vary directly as the daily volume, at the height of lactation these variations are normally so shght that an approximately constant percentage is maintained. The conclusion to be drawn is not that the lactose is elaborated till it reaches this practically constant percentage, but that it is kept down to this constant level by the rate of secretion of the milk. The suggested theory of secretion is, therefore, that the elaboration of lactose produces a flow of milk by a process of osmosis. This would explain why. at the height of lactation, the percentage of lactose varies but slightly and tends to do so directly as the daily volume, at the same time reducing the percentages of all the other constituents of the milk by a process of dilution — hence their inverse percentage relationship to volume. coxcLrsioxs. 1. The percentage composition of the milk seems to be determined by its rate of secretion. 2. The percentages of protein, fat and ash vary inversely, and the percentage of lactose varies directly, as the daily volume, the greatest variation being .shown by the fat and the least by the inorganic elements. !5. There is an inverse relationship between the percentage of lactose and the percentages of all the other constituents of the milk, this being particidarly apparent in the case of the fat. 4. Diet has no direct influence on the percentage composition of the milk, except in the case of the non-protein' nitrogen which is not a product of the mammary gland. Diet has, however, an indirect influence by reason of its efiect on the daily volume. 5. A high protein diet would appear to stimulate the rate of secretion of the milk. 124 Percentage Composition of Milk 6. It is suggested that the quantity of lactose elaborated hy the uiainmary gland controls the daily volume of the milk, and that, there- fore, the rate of its elaboration controls the rate of milk secretion. We wish to express our indebtedness to Dr .1. B. Orr, who suggested this research, for constant encouragement and advice during the progress of the work. KF.I'i:i!KX(i:.s. (1) VoiT. Zeilschnjl far B'whijk, 1 8(19, 5, p. 122. (2) Ingle. Bull. Yorks. Coll., Leeds, No. 2.'). Mtdl. Cited by Hainnioml and Hawk, Joiirii. A(inr. Sci., 1917, 8, p. 139. (:i) C'rowther. Univ. of Leeds, Bull. Xo. 44, 1903. Cited by Hammond and Hawk, ./onni. Agric. Sci., 1917, 8, p. 139. (4) Jordan and Jenter. New York Agric. Kxper. Station. 1901. Bull. No. 197. Cited by Lusk, The Science of NiUrition, 1917, p. 393. (5) Morgan, Beroer and Fincjerling. LandwirUchaJt. Venuchaatal, 1904, 41, p. 1. Cited by Lusk, The Science of Nutrition, p. 393. ((!) Hammond and Hawk. Studies in Milk Scciction. .Jottrn. of Aqric. Sci.. 1917, 8. J). 139. (7) TooHER. Investigation into. Milk Yield of .Ayrshire Cows. Trans. High.and Agric. Soc. of Scot., 1919, 31, p. 237. (8) Crowther and Ruston. Variation in the Composition of Cows' Milk witb the Advance of Lactation. Trans. High, and Agric. Soc. of Scot , 191 1. 23. p. 93. (9) Van der Laan. Biochein. Zeil.tch., 1919. 73. p. 313-2.5. {Received 6 January 1922.) THE CITRIC SOLUBILITY OF MINERAL PHOSPHATES By J. F. TOCHER, D.Sc, F.I.O. Consulting Chemist to the HiyhlunA and Agrieidtural Societij oj Scotland; Lecturer on Statistics, Vnirersity of Aberdeen. (With 8 Diagrams.) I. INTRODUCTORY. The Fertilisers and Feedin.ii Stuffs Act 1906, Section 10, defines the expressions "soluble" and "insoluble" to mean that the fertilising con- stituent is soluble or insoluble in water, or, if specified in the invoice, to mean that the fertilising constituent is soluble to the extent guaranteed in a solution of citric acid, or other solvent, of the prescribed strength. In particular the section defines the percentage of soluble phosphate and the percentage of insoluble phosphate to mean respectively the percentage of tribasic phosphate of Ume equivalent which has been, or that which has not been, rendered soluble. "Citric solubihty" under the Act is further and more definitely defined in the FertiHsers and Feeding Stuffs General Regulations 1906 as follows : "When, in an invoice relating to hasic slag or basic superfhos-phate, it is specified that a certain percentage of the phosphate contained in the basic slag or superphosphate is soluble in citric acid, this shall be taken to mean that it is capable of being dissolved to the extent of such percentage when 5 grams of the fertiliser and 500 cubic centimetres of water, containing 10 grams of citric acid, are continuously agitated in a flask or bottle of about 1 htre capacity for the period of half an hour at the ordinary temperature." It is clear from these regulations that "citric solubihty" refers to ba.sic slags and "basic superphosphates "^ and that the citric solubihties of basic slags and of basic superphosphates have to be determined at "room" temperature by means of a 2 per cent, solution of citric acid, the duration of contact being limited to half an hour in a quantity of ^ The seller need not, unless he chooses to do so, f^ive any (guarantee of citric .solubility. 126 The Citric SoJiihility of Mi ii end FJio.yjJuifes the fertiliser exactly one half in amount of citric acid present. The basis of this test is evidently the work of Wa<;ner and appears to be of a wholly empirical character. .Since Wagner's method has been adopted as an official test, citric solubiUty has been studied by a number of different workers. Stead^ found that for normal basic slags the solubihty increased with the amount of silica present. He also .showed that citric solubility was associated with the degree of fineness of the powder. Kobertson- .studied the degree of solubihty of mineral phosphates in citric acid using the official test. He shows that mineral phosphates arc completely soluble iu 2 per cent, citric acid solution if a sulticient number of extracts are made by successive lialf liour contacts. Robertson in his conclusions states that "Even a small amount of free lime or calcium carbonate decreases substantially the solubihty of mineral phosphates as judged by the citric acid test. When a large amount of calcium carbonate or free hme is present, the citric acid test as commonly practised, is a test for hme and not for phosphates. It is important in this respect to dis- tinguish between free hme and calcium carbonate, and hme actually entering into the composition of the phosphate. The higher the percentage of hme actually entering into the phosphate compound, the higher the citric solubihty of the phosphate." He has also shown that fluor-spar greatly decreases citric solubility in slag and concludes that the official test gives no true idea of the solubility of the pho.sphate in slag. He states that "one of the effects of fluor-spar is to cause the formation of a phos- phate which does not contain silica in combination as is the case with high citric soluble slags." The effect of fusing with fluorides apparently is that a compound of silicon and fluorine is formed leaving lime and phosphorus in combination. Dixon^ made a study of the citric solubihty of various bone phosphates. In his case he varied the citric acid con- centrations and with a constant weight of fertihser he found that in every case the stronger the citric acid solution the greater was the amount of phosphate dissolved. Ramsay* prepared pure tricalciuni phosphate by mixing three equivalents of CaO with one equivalent of PoOj and show-ed that 91 per cent, of the total phosphoric acid content of the pure tricalciuni phosphate was soluble in the prescribed 2 per cent, citric acid solution in 30 minutes. He also showed that the simple addition of calcium carbonate reduced citric solubility. Russell and Prescott^ studied the (•itri(' solubihty of the phosphates of the soil. They ' Trans. Faraday Soc. 16, I'art 2. - Soc. Cliem. Ind. No. 4, 35. ' Journ. ofAgric. Sci. 1906, Part 4. ^ Ibid. June, 1917. ' Ibid. Sept. 191U. J. F. Tocher 1-27 found a greater solubility for short periods of contact when compared with long periods and established the fact that adsorption of phosphate took place during long periods of contact. II. CITRIC SOLUBILITY— OFFICIAL AXD OTHKRWLSE. The writer was attracted to this subject by the fact that the citric solubihty of certain commercial phosphates was determined in a dilute citric acid solution ((»-"2 per cent.) and not by means of the 2 per cent, solution prescribed by the official test for slags. Samples of ground mineral phosphate are occasionally guaranteed to contain as much as 50 per cent, "citric soluble" phosphate. To the un- wary this might be taken to mean that the sample contained 50 per cent, "citric soluble" phosphate as determined by the official test for slags and "basic superphosphates" prescribed in the Regulations. As a matter of fact, however, the citric solubility in this case was determined for the sellers by agricultural analysts by agitating for half an hour 5000 parts of a solution of citric acid (0-2 per cent, strength) with 1 part of the sample. That is to say the amount of the citric acid employed was one-tenth of the amount officially prescribed while the proportion of fertihser was 50 times less than the proportion prescribed as may be seen from the following table (Table I). Table I. Test Official quantities ... Quantities for private test i.e. ... Citric acid 10 grams 1 gram 10 grams Fertiliser .5-0 grams O-I gram 10 ,. Total volume .500 c.c. .500 „ 5000 „ The following results (Table II) were obtained on using the official citric solubihty test on samples of ground mineral phosphate and basic slag: Table II. Basic slags A • Mineral , phosj^hate Percent. ' Percent. Citric of total Citric of total sol. Total phos. sol. Total phos. phos. phos. Fineness dissolved phos. phos. Fineness dissolved 27-94 30-59 84-8 91-3 21-92 04-80 64-7 33-8 28-21 31-00 79-0 91-0 18-26 55-86 85-1 32-7 4-82 19-35 76-0 24-9 20-39 57-84 84-5 35-3 6-36 25-55 74-1 24-5 19-72 57-98 82-9 340 22-68 23-60 93-6 96-1 21-53 58-51 98-0 36-8 28-54 39-10 85-4 73-0 19-30 01-92 88-0 31-3 128 The Citric Soliihilifi/ of Mineral PJwsphates The results of this table show that when mineral phosphate is as finely ground as basic slag a fair amount of phosphate is rendered " citric soluble"' by the official method but the "citric solubility" of mineral phosphates appears generally to be much less than the citric solubility of slags, when this test is apphed. irr. TITK SCHK^rK of KXrKRTMKXTAI. WORK. The citric solubility uf mineral phosjihates was accordingly studied at "room" temperature over a period of 30 minutes agitation: (1) In varying dilution, the quantities of acid and mineral phosphate being constant. (2) In varying concentrations of acid, the volume of fluid and the weight of mineral phosphate being constant. (3) With varying amounts of the phosphatic fertihser, the volume of the fluid and the concentration of acid being both constant. In order to compare the citric .solubility of mineral phosphates under the above conditions, with the citric solubiUty of a pure phosphatic compound, a fourth series of experiments was conducted, namely, (4) the solubihty of dicalcium phosphate in dilute hydrochloric acid. It is evident that whatever be the value of determining the j)ro- portion of citric soluble phosphate in a phosphatic fertiliser, it is neces- sary, for comparative purposes, if other conditions are similar, that the test should be applied in the same way and with the same proportions of citric acid or of other acid and of fertiliser as prescribed officially, for basic slags and basic superphosphates. Other conditions are, of course, open to study. For example we could have two of the above factors varying differently with a series of constant values for the third factor. The citric solubility of tricalcium phosphate and other pure phosphatic compounds could be studied and the results compared with the results from the above four series. The writer has not been able to carry out these latter experiments. He therefore sub- mits the results of experimental work under the above four heads'. ' The writer has to acknowledge his indebtedness to Mi .lnhn I']. Ritchie, M.A., B.Sc., A. I.e., who has performed the necessary analytical determinations in the exi)eriments on mineral phosphates, and has also given valuable assistance in preparing the memoir. He has also to thank Mr W. T. H. Williamson, B.Sc., A.I.C., for the determinations in the case of dicalcium phosphate. J. F. Tocher 129 IV. EXPERIMENTS WHERE THE QUANTITIES OF MINERAL PHOSPHATE AND CITRIC ACID USED WERE CONSTANT AND THE DILUTION WAS VARIED. In order to determine the variability in citric solubility of mineral phosphate with varying dilutions of citric acid, a series of experiments was conducted with different dilutions shaking for half an hour, the time prescribed in the official test. The quantities of citric acid and of mineral phosphate used are indicated in the following table (Table III). Let m-^ = amount of citric acid used, itio ^ amount of mineral phosphate used, and ni^ = volume of fluid used. In tliis series the ratio rn^jm^ was made constant (and equal to 2) in order to secure that the effect of the presence of Ca(0H)2 and of CaCOj and other hydrates and carbonates, on the citric acid concentration was of a constant character. Throughout this series of experiments 10 grams citric acid and 5 grams mineral phosphate were used, the volume being varied as shown in Table III. Table III. Quantities of ininend phosphate and citric acid used are constant — dilution, i.e. degree of concentration, varies. 5 grams mineral phospliate 10 grams citric acid 1 Exp. 2 Vol. of sol. 3 4 Wt. of Phosphate dissolved. Grams per volume stated 5 Pliosphate dis.solved per cent, of sample as Ca,(PO,), 6 7 Percentage dis- solved of total phosphate con- tent 8 9 :Molecular concentration of phosphate at end of 30 minutes. Gram-mols. per litre Observed as Ca,(PO,), Tlieory Ca3(P0.), Observed Theory Observed Theory 1 2 3 4 5 500 625 833 1250 2500 0-9620 1-0210 1-0875 1-2080 1-4070 0-9697 1-0215 1-0907 1-1990 1-4030 19-24 20-42 21-75 24-16 28-14 29-8 31-7 33-7 37-5 43-6 30-1 31-7 33-8 37-2 43-5 0-006206 0-005270 0-004210 0-003117 0-001815 0006256 0-005272 0-004222 0-003094 0-001810 60 i oo gp;' a a §■3 20 in (.'iihio ("'clitimetrcs 500 1500 2500 Diagram 1. See columns 8 and 9, Table III. sag S^-2 24 Sfea 22 "So- Voiunic in ('nl>ic Centimetres 500 1500 Diagram 2. See column 5, 2500 Table III. 130 The Citric SolnhiUty of Mineral Phosphates It is seeu from these results that the hi^'hcr the dilution of citric acid, with mjin.^ kept constant, the greater is the proportion of citric solubility of tho jihosphato expressed as a percentafze of the weifiht of the sample (coluiim 5. Tal)le III). Bassett' considers that the compound usually present in mineral phosphates is hydroxyapatite which may be written |Ca3(P04)2l3C'a(()H)^. Ho also thinks it probable that hydroxy- apatite is the only calcium phosphate that can permanently exist under normal soil conditions. It forms the stable solid phase over a ran^'c of acidity of great practical importance, as it can exist in contact with faintly acid, neutral or alkaline solutions. An attempt has been made to fit a theoretical curve to this series of experiments on the assumptions that there is equilibrium at the end of the experiment and that the following equation represents the reaction : Ca03Ca3(P04)2+ .'^HaCeHsO,, H20i^CaH,(PO,)2+ SCaHCgHsO, + 2Ca3(POA+H,p (1) Of course other equilibrium equations including dicalcium phosphate can be written from which the same mass action equation can be deduced. The above equation is merely given as a suggestion. The sample of mineral phosphate contained a proportion of COj equivalent to 0-G5 gram of calcium carbonate in 5 grams of the sample. Hence 1-36 grams of citric acid would be used up in the formation of citrates, leaving 8-64: grams of acid available to attack the phosphatic compound. If the original acid concentration is taken to be proportional - to the concentration at ecpiilibrium, a constant should be obtained on applying the law of mass action. A good agreement between theory and observation is obtained on this hypothesis. If w = molecular concentra- tion of acid after alkaline lime has been neutrahsed ; m = molecular con- centration of phosphate (expressed as tricalcium phosphate) at the end of 30 minutes agitation then we should have y^jic^ = k. The last two columns (columns 8 and 9, Table III) show the observed molecular con- centrations and the theoretical values on the basis of above equation. A good fit is also obtained using the equation u*l{w — u)\w — 2m) wIumi the theoretical values are found by Horner's method. Diagrams 1 and 2 show in graphical form the results of Table III. ' Trans. Chem. Soc. 1017, 111. - The values of u from the equation m*'(wi - 3«)' show ureater cliver<;ence.s from observa- tional values than the values obtained from either of the equations civen in the text. Tho ratio ^ varies from -72 to -SO, showint; that the original concentration is nearlv w proportional to concentration at the end of 30 minutes shaking. Whatever the reason, the formula u'jiv^ gives by far the best fit to the results. J. F. Tocher .181 A second set of experiments was carried out in which the ratio mjm^ = 10 instead of 2, that is, in each case 10 times more citric acid than phosphate was used in each separate experiment. The undernoted table (Table IV) shows the quantities used and the results obtained. This set of experiments when plotted against the theoretical curve ex- pressed in equation (1) was found to be a very bad fit. In order to deter- mine the best fitting equation the ratio of the exponent x to the exponent c in the general equation C ■ l\u'' k.M - was determined where C, l\ and k^ are constants. The ratio — was found to be egixal to 1-0669 mean 2: 16/15. The following table (Table IV) and accompanying diagrams (Diagrams 3 and 4) show the observed results and those reached from the ecpiation, C = --j^ , or log C = 1 6 log u — 1 5 log w. Table IV. Quantities of mineral phosphate and citric acid constant — dilution, varied. 1 10 grams citric acid ( 1 gram mineral phosphate 1 Exp. 0 Vol. of solution 3 4 Weight of phosphate dissolved as GajPjOg 5 Percent, dissolved of total Ca,,P20, content 6 Phosphate dissolved per cent, of weight of sample 7 8 Molecular concentration at end of 30 minutes Observed Theory Observed Theory 1 2 3 4 5 6 500 625 833 1250 2500 5000 0-5376 0-5472 0-5610 0-5708 0-59.55 0-6180 0-5388 0-5464 0-5565 0-5705 0-.5960 0-6220 83-3 84-8 87-0 88-5 92-3 95-S 53-76 54-72 56-10 57-08 59-55 61-80 0003468 0-002824 0-002172 0-001472 0000768 0-000399 0-003476 0-002820 00021.54 0001473 0-000769 0-000401 c.-, 30 oO 20 £•3 10 N'uluinc in Cubic Cclitiiiicdcs ;-S 500 2500 5000 Diagram 3. See columns 7 and 8, Table IV. 03 s a ^ "" o 111 COS %^ -SO 54 / Volume in Cubic Centiuietres 500 " 2500 5000 Diagram 4. See column 6, Table IV. 132 The Citric Solxbility of Mineral P/iosphates Experiment 6 of Table IV is an experiment identical in character with the private test already mentioned (see Table I). The results of experi- ments parallel to this experiment were used by certain sellers to describe the citric solubility of the mineral phosphate put on the market. It will be seen that Exp. 6 falls naturally into its place among the experi- ments given in Table IV. The amount of phos])hate dissolved amounts to (52-2 per cent, of the weight of mineral pliospliatc tested while the same phosphate, analysed in accordance with the official test (Table III, Ex]). 1) shows a citric solubility of 19-24 per cent. If we consider the proportion dissolved in relation to the total amount of phosphate present expressed as tricalcic phosphate it is found that, by the official method 29-83 per cent, of the total Ca^P.^Og content is dissolved, while by the private test (at great dilutions, see Table IV) 95-8 per cent, of the total CagP-^Og content was dissolved. These results show what might naturally be expected, namely, a much higher solubility of mineral phosphate in the experiments where the constant ratio m^jni^ = 10 (Table IV) was used than where mjm.^ = 2 (Table III) was used. In other words if mjm2 is nuide large enough we should reach the limit of 100 j)er cent, citric solubility for all very high dilutions. The results in Table II show that, using the official test on both slags and iniueral phosphates, slags generally show a higher citric solubility tliaii mineral phosphates. If the uiiodicial test (Exp. (1. Tal>le IV) was universally applied to slags we should have similar liigh solubility figures, in other words the citric solubility at high dilution would be practically 100 per cent, and the only item of information whicii would be valuable to the purchaser would be the actual proportion of phosphate, expressed as tricalcium phosphate, present in the fertihser. In some slags the citric solubility in terms of total phosphate content is as much as 90 per cent. (see Table I) and therefore the actual increase on dilution must be necessarily small compared with the increase in citric solubility on dilution of any mineral phosphate. The undernoted table (Table V) shows the average composition of five commercial mineral pho.sphates^. The sixth (Egyptian) was analysed in my laboratory. Since calcium carbonate is present in varying ))roportions in com- mercial mineral phosphate it is clear that with a constant initial molecular concentration of citric acid, varying quantities of citric acid will be available to attack the insoluble phosphate. For example, suppose we selected two dift'erent varieties of mineral phosphate, ground to the same 1 Robertson, J.S.C. I. ,35, p. 218. Makatea Florida Island Pebble Alfrerian Gafsa Tunisian Egyptian 52-38 47- 10 40-13 43-30 48-40 ^46-81 38-24 31-50 27-27 25-35 26-13 29-52 1-69 3-04 0-70 5-50 9-03 9-42 1-46 1-0(1 0-72 3-26 0-98 — 3-30 2-41 3-28 4-39 3-21 2-0(i 1-01 Mill 2-50 4-08 4-25 2-78 1-35 1-itO 2-02 1-72 0-90 0-77 0-28 7 03 8-12 7-56 4-65 7-73 0-20 4-42 3-17 3-94 2-45 0-91 .1. F. TociiER 133 degree of fineness, and containing the same proportions of liydroxy- apatite but quite different proportions of calcium carbonate. The sample wliich contained the smaller (|uantity of calcium carbonate would show a higher citric solubility than the second sample which contained a higher proportion of calcium carbonate, due to the presence of a relatively large proportion of free citric acid in the former. Table V. Composition of Mineral Phosphates. Cak-iura o.xide Phosphoric acid ... Carbon dio.\ide Moisture ... Combined moisture and) organic matter ... \ Ferric and ahiminium j o.xides ... ... ( Magnesium oxide ... Sand Undetermined Citric solubility is not necessarily a te.st of the availability of the phosphate to the plant in the soil. If the sample is finely ground and has a low citric solubihty, the lowness of the citric solubihty in the case of mineral phosphate would mainly be due ( 1 ) to the presence of alkahne material which would neutralise a large proportion of the citric acid, leaving the residue to act on the phosphate and (2) to the chemical constitution of the phosphatic mineral. In the case of slags the citric solubility would be mainly dependent on ( 1 ) the compounds of fluorine as shown by Robertson, (2) the presence of alkaline lime as shown by Ramsay, and (3) the chemical constitution of the phosphatic compound in the slag. It has yet to be shown that the phosphate in mineral phos- phate is not utiUsed by the plant as readily and as efiiciently as the phosphate from slags. In other words the exact chemical composition of the phosphatic compounds in the various mineral phosphates and slags has, in each case, to be demonstrated. It appears to be necessary to test the citric solubihty of phosphates of knoivn composition against their availability in the soil as shown by yield of crop. The results would then show how far, if at all, citric solubihty is a measure of availability in the soil. Since the commercial fertilisers tested contain varying quantities of alkaline Hme, fluorides and other interfering substances, and since the chemical constitution of the fertihsers is incompletely known, the writer can see no scientific validity in the use of citric solubihty as a measure of availability. The three practical tests appear to be : 134 The Citric Solahility of Mineral Phosphates (1) Total phosphatic content. (2) Degree of fineness of the powder. (3) Presence or absence of substances capable of inhibiting growth. A fourth test of scientific, as well as practical, value would be a test demonstrating the constitution of the iihosphatic compounds in the fertiliser. If citric solubility cannot be regarded as a useful and practical test for mineral phosphate it is still less valid as a comparative test for slags and mineral phosphates alike owing to the varying composition of slags and to their widely different chemical compo.sition and constitution when compared with mineral phosphates. In regard to the chemical constitution of phosphatic fertihsers it has already been noted that hydroxyapatite is considered by Bassett as the chemical compound probably present in mineral phosphate. Alorison states^ that the molecular ratio of phosphoric anhydride to calcium oxide P-jOj/CaO is 1/.5 in slags and supports Stead's conclusions that the phosphatic content in basic slags consists of a chemical union of tetra- calcium phosphate and monocalcium silicate (CaO)4P205CaOSi02. On the other hand the ratio of phosphoric anhydride to calcium oxide in Bassett's hydroxyapatite is 'ij\0 — 1/3 J. Morison also deals with the effect of free lime on the citric solubility of slags and shows that the greater the amount of free lime in a slag the greater is the total solubility after three extractions. On the other hand Ramsay- shows that about 91 per cent, of the total pho.sphoric acid in pure tricalcium phosphate is soluble in the prescribed 2 per cent, citric acid solution. This degree of solubihty is very similar to the degree of solubility of the best grades of slags (see Table I). He also shows that by the simple addition of calcium carbonate to pure tricalcium phosphate the citric solubihty is reduced fmm 91 to 84 per cent. This is naturally to be expected and the apparent greater solubility of phosphate with increase of lime content found by Morison must be due to other causes. It should be noted that the quantities of free lime present in Morisons samples are relatively small. The increase in solubility with an increase of silica and the decrease in solubihty in the presence of fluorides have already been mentioned. Robert.sou finds that calcium carbonate decreases substantially the solu- bility of phosphates as judged by the 2 per cent, citric acid test. The exact effect of the presence of calcium carbonate or calcium hydroxide on the solubility of phosphate of a known composition can of course be found ' Juiini. Ayil. Sci. 1909. = Ibid. 8, p. 277. J. F. Tocher 1:35 a jwiori. The difficultv arises when the composition of the phosphates is unknown and when other interfering substances are present. Since the nature of the substances present are unknown, equations expressing the law of mass action cannot be written down in these cases. The following table (Table VI) shows the average compo.sition of some commercial slags as given by Collins {Chemical Fertilisers, p. 122). Table VI. Composition of Slags. 12 3 4 5 Total P-A IL'-IJO 20-49 O-O'J 17-57 19-35 SiUca 17-(i9 10-12 13-49 7-77 12-12 Lime 3S-02 46-81 40-43 52-22 44-75 Magnesia 4-24 2-92 5-01 1-94 0-11 Mansanese o.xide ... 7-39 4-38 5-41 9-37 4-68 Iron" 12-89 9-98 13-83 8-13 9-10 These latter tables (Tables V and VI) show in a general wa)' the differences between slags and mineral pho.sphates. These two classes of fertihsers contain non-jihosphatic residues differing in chemical com- position, and residues which are common to both in different proportions. The results of tliis section show that citric solubility is merely a special case of the law which has been proved to hold for the solubihty of a definite chemical substance in dilute acids and it can always be stated a prion when the conditions are known for a definite substance in a definite dilution. When, however, we pass from a single substance to mixtures of varying composition citric solubility cannot be descriptive ■of available jihosphate of definite composition. The reason for this lies (1) in the unknown changes which take place in the initial molecular concentration of the citric acid, due to the formation of calcium and other citrates from the carbonates and hydrates present in the fertiliser, (2) in the unknown changes which take place on agitating phosphatic fertihsers of varying composition and (.3) in the known effects produced by the presence of fluorides and of silica. V. EXPERIMENTS IN WHICH THE VOLUME OF FLUID AND THE WEICiHT OF MINERAL PHOSPHATE ARE BOTH CONSTA.XT. THE VARYIN(i FACTOR BEINC4 ACID CONCENTRATION. We shall now consider condition (2) namely, where the amount of mineral phosphate (m.^) and the volume of fluid (iii.^) are both constant, i.e. jWg/Wg = constant, while the amount of citric acid is varied. The following series of experiments was carried out with a constant weight of mineral pho.sphate (5 grams) in varying concentrations of citric acid in a constant volume of 500 c.c. The undernoted results (Table VII) Journ. of Agrio. Sci. xii 10 V,]i) The C'dric Sohihiliti/ of Mineral J'/iosp/Ki/cs show that if the acidity is expressed as the molecular conceutruliuu ol citric acid at the end of 30 minutes shaking, K — u^jifi fairly accurately describes the solubility of the phosphates at the acid concentrations named. Table VII. Amount of miner nl phosphate present and voluine constant — aci(Jit>i varies. m^j)iu = 100; vol. =.'500 c.c. =m^: m., -miiu'ial iiliuspliiite ^5 grums I 2 3 4 6 7 8 9 Exp. Wt. citric acid in 500 c.c. Aciility lis citric acid at end of 30 minutes Wt. of plio.spliate dissolved as CajPjO, in 500 c.c. Per centago dissolved of total CajP.O, content ea,P.O. found percent. of phos- pliute taken Molecular concentra- tion of phosphate at end of 30 minutes Observed Theory Observed Theory 1 10 grms. 8-87 grms. 0-9675 0-9492 300 19-35 -006242 -006124 9 8 „ 7-02 „ 0-8119 0-7958 25-2 16-24 -005238 -005134 3 •' - 5-11 „ 0-6743 0-6250 20-9 13-48 -0043.50 -004032 i 4 „ z-2r, „ 0-449r) 0-4470 13-9 8-99 -002900 -002884 5 •) 1-34 „ 0-2176 0-2300 6-7 4-35 -001404 -001484 6 1 ., 0-49 „ 0-1020 0-1082 3-2 2-04 -000658 -000698 oo 60 lo:=°- gco 30 II 20 If '0 .^' ^' If '0/' oja Wiijjht of Citric Acid in .">( 1 2 3 4 5 6 7 8 rt:»»......» -: vj„.v «,,l «., w ..., ."iOOc.c. "4 5 6 7 8 9 10 ;rain 5. See columns H and 0, n 15 Sfl §fe=. "Sob W(i"lil of ('ilric Acid in .')(X)c.c. 2345678 9 10 Diagram (>. See column 7, Tal>le VII. Oiagi Table VI] . The acconipanyini; diagrams (Diagrams 5 and (i) sliow the results of Table VII. These results show clearly increasing citric solubility with increasing acid concentration but with constant weight of mineral phos- phate in a constant volume. In a citric solubility test, used as a measure of availability, it seems necessary therefore to show that a two per cent, solution is the concentration of citric acid best suited for the utilisation of phosphate by the plant. Would it not be more in accordance with scientific practice firstly, to a.scertain the constitution of the ])liosphatic fertilisers, and secondly, to dptermine wliat rolo concentration has in the life history of the plant? J. F, ToCHEH VI. Ii;XPERBrEXTS WITH CONSTANT VOLUME AND CONSTANT CONCEN- TRATION OF ACID BUT WITH VARYING QUANTITIES OF MINERAL PHOSPHATES. In the first set of experiments fairly high constant vakies of mjnu were used. We shall now consider the effect of making mjm^ small, nij^ and iii^ being in this case constant and m.^ the variable. If m^Jm^ is made small enough we should have a citric solubility practically zero for all dilutions. An example indicating the approach to the latter con- dition is given in the following table (Table VIII) where the third set of conditions is observed. (See also Diagrams 7 and 8.) Table VIII. Amount of citric acid and volume constant — amount of mineral phosphate (»(.,) used varied. Exp I 3 4 Wt. of min- eral phos- phate taken 5 grams 10 ., 20 „ 40 „ ■-= = 50 (10 grains citric aciil = mj ^500c.c. volume = Hi, Ratio 1 0-.5 0'25 Acidity ex- pressed as citric acid at end of 30 minutes 8-87 8-50 7-6.5 5-98 Amount of pliosphate dissolved as CajP.O, 0-970.5 0-86(30 0-6070 0-4110 Mol. cone, of phospliate dissolved at end of 30 minutes -UU(i2Ul -00.5.5S7 -003916 -002652 Citric solubility expressed as Ca3P3*^)a per cent. of pliosphate taken 19-41 8-66 3-04 1-03 .a a 60 3 cd go a « o > Wei;;lU of Miner; taken — < Jciini-- 1 l'lios|iluite 5 10 IE Diasrram 7. 20 25 30 35 See Table VIII. 20 l-i ^ 3 fl IZ 10 2S.& ■son St 5- ;lit of Mineral ' ^|d^ate taken — t^ranis 5 10 15 20 25 30 3£ 40 Diagram 8. See Table VIII. If the mineral phosphate contained merely tri- or di-calcium phos- phate and was quite free from Ca(OH)a, CaO or CaCOj and also was in excess, mere variations in the quantity taken would have had little or no effect on the amount of citric acid present at the end of the period of shaking. The presence of Ca(0H)3 naturally reduces the acid concentra- tion with the result that, while 19-41 per cent, of the 5 grams mineral phosphate was dissolved only 1-03 per cent, mineral phosphate was dis- 10—2 138 The Citric Solubility of Mineral Phosphates solved wlu'ii 40 gnuns of the fertiliser were taken. It is thus shown, whicli is also apparent a jyriori, that citric solubility depends on the quantity of mineral pliosphato or slag used. With iiuantities like .50 grams or 100 grams of mineral phosphate the citric solubihty would be extremely small. On what theoretical grounds are quantities like 5 grams (official test) in the case of slag and 1 gram (private test) in the case of mineral phosphate used to determine the availability of pho.sphate in the soil? Hi \'II. JHK l.VCOMPLETEXESS OI' 'IHE REACTIONS. All the foregoing results have been considered from the standpoint of the amount of phosphate dissolved per cent, of the weight of mineral phosphate taken. It is desirable, however, to consider the amount of phosphate dissolved in relation to the amount theoretically obtainable if the reaction were complete. We cannot say definitely what the reaction is but we know that the equation in Section 1 V' gives a suitable prediction formula. Suppose we put the question : What proportion of the amount indicated l)y this equation is obtained in each separate experiment? We know from theory that the hydrogen ion concentration of an acid is increased by dilution and we should therefore expect (/realer proportions of the. possible total amounts at higher dilutions than at lower dilutions. Let us consider the first series of experiments from this standpoint. If the citric acid concentration at the beginning of the reaction was 3w gram- molecules and the reaction went completely, according to the equation given in Section IV we should have w gram-molecules of monocalcium phosphate formed, equivalent to ic gram-molecules of tricalcium phos- phate. Hence for the completed reaction we should have iv gram- molecules of citric acid giving 1/3 «' gram-molecules of tricalcium phos- phate. Suppose we find only y gram-molocules of tricalcium ])hosphatc then -^-^ is the percentage of tiie theoretical amount which has been found. Tlie following table (Table IX) shows the concentration of the citric acid, the theoretical amount of phosphate obtainable in a complete reaction according to the equation, the actual amount of phos- phate obtained, as gram-molecules of tricalcium phosphate per litre, and the percentage of the theoretical amount of phosphate obtainable. It is seen that this percentage is practically identical with the per- centage in col, 5 of Table III. Indeed the numbers in col. 5 of that table multiplied by 1-016 give the percentage in col. 4 of Table IX. This arises from the following considerations. If a^ be the weight of phosphate J. F. Tocher 139 dissolved aad «2 = weight of mineral phosphate taken for the particular dilution v then 100 — = Pi= amount of phosphate dissolved as a per- centage of amount of mineral phosphate taken. The number of gram- ,,...,,.. , , . 10 1000 molecules oi citne acid per litre in each case may be written -rr-^r x = ic ^ ■' 2]0 r where v is the dilution. The amount of phosphate theoretically obtainable expressed as tricalcium phosphate is therefore «'/3 = x gram-molecules per litre. The amount of phosphate found also expressed as tricalcium phosphate in gram-molecules per litre is 310 1000 V y- Table IX. Table .showing atnount of pliosphale found per cent, of amount theoreticaUij obtainable, at various dilxtionfi on basis of equation, Section IV. Gram-molecules per litre ;x|K Concentra- tion of citric acid Pliospliatc theoretically obtainable in complete reactions Plios]iliate actually obtained Phosphate founil per cent, of amount theoretically obtainable 1 •09.52 •0317 •0186 1955 2 ■07G2 •02.5-i •01.58 20-75 3 •0572 •0191 •0126 2210 4 •0381 •0127 •0094 24-55 n •0190 ■1)003 •0054 28^59 Now the number of gram-molecules dissolved expressed as a per- centage of the number of gram-molecules theoretically obtainable is clearly y X 100 = p, X 63a, P2- X ^ ' 310 In the series under consideration a, = 5 ^^^ thus we have 63 Ih = Pi 62 10 16 7^1. With increasing values of a^ it is known that a smaller series of values of jjj would be obtained (see Table VIII). For example, if 40 grams of mineral phosphate were taken then p.^^ = Sjjj (approximately). The results in Table IX merely illustrate the well-known fact that solubihty depends on the hydrogen ion concentration. They further show that it is mis- leading to adopt a particular set of weights to test citric solubility and to express the solubility in terms of the weight of mineral phosphate taken for the experiment, if it is intended to judge the equality of a phosphatic fertiliser by the result so obtained. 140 The Citric Sofiibilit!/ of Mineral P/iosp/idtea Vril. DICALCIUM PHOSPHATE EXPERIMENTS. The observed amount of phosphate dissolved by citric acid per cent, of the theoretical amount for a completed reaction found in these experi- ments may be contrasted with corresponding figures obtained from the interaction between a pure substance likedicalcium phosphate and hydro- chloric acid. A series of experiments was conducted with this end in view. In each experiment a constant ([uantity (25 grams) of precipitated dicalcium phospliate (Calir042H20) was shaken for half an hour in a constant volume (500 c.c.) of water. The proportion of hydrochloric acid was varied as shown in Ta])k' X. Tlie following results were obtained : Table X. Volume constant = 500 c.c. ; dicalcium pho sphate = 25 grams I 11 III IV V Original concentration of 11 CI. Gram niols. per litre Concentration of phoj^plmte after shaking. (granvmols. raH,{P(>4)2 per litre) Theoretical concentratiou of phospjiate 't ~ : — 7:7 X )'• l+2t Grams of dicalcium phosphate dissolved JXT litre xi>. Observed Theory 1 ' -2856 ^ •1299 ■1325 44^69 45-58 2 •2285 •1058 •1060 3640 3646 :i •1713 •0792 •0795 27^24 27-35 4 ■1142 •0547 •0530 18-82 18-23 'i ■0571 •0265 •02G5 912 912 (> ■0286 •0124 (li:!:! 4-27 4-58 These results show the decreasing percentage of phosphate di.s.solved with decreasing acid concentration as usually obtained in such experi- ments. The result, however, to be contrasted with the corresponding previous results, is the relative completeness of the reaction as seen in cols. I and II of the above table (Table X). Since there is excess of dicalcium phosphate the undernoted equation may be taken to represent the reaction namely, 2 CaH PO4 ! 2 1 1( '1 1; ( 'aH4(P04)2 + CaCl^ (2) In this equation two molecules of hydrochloric acid are used up to produce one molecule of monocalcium phosphate and one molecule of calcium chloride. Hence if we start with w gram-molecules of hydro- chloric acid we shall have, for a completed reaction, liv gram-molecules of monocalcium phosphate, the substance determined. If iv = molecular concentration of iiydrochloric^ acid per litre at the beginning of the experiment and u -^ molecular concentration of phosphate expressed as Ca3(F04).2 per litre at the end of 30 minutes shaking when equilibrium may be presumed in this case to be established, then according to the J. F. Tocher 141 law of mass action we should expect k = ujw — 2m to describe the ex- perimental results. An inspection of col. Ill (Table X) will show that the formula reasonably fits the data. No account has been taken of the degree of ionisation in the above equilibrium equation. The extent of the reverse action, CaH,(P04).l;CaHP0, + H3PO4, has also been neglected. If the degree of ionisation of liydrochloric acid is taken into account together with the fact that a minute Cjuantity of calcium hydroxide was found to be present in the dicalcium phosphate used it is found that 2m is slightly greater than hi as may be seen on inspection of Table XI (col. IV). The value of 'liijw varies from 1-008 to 1-039, the average value being 1-0228 and theory requires this ratio to be equal to unity for a completed reaction. Table XI. I II III IV Concentration of dissociated HCl Gram-mols. (gram-mols. per litre) CaH,(P04)2 2m = w = M 2« w ■2576 •1299 •2.598 1^0085 •2087 ■ 1058 •2116 b0139 •1573 ■0792 •1584 b0070 •1053 ■0.547 •1094 1^0389 •0514 •0265 •0530 10311 ■0239 ■0124 •0248 10377 The work in this section is intended merely to illustrate the approxi- mation to a completed reaction in the case of a known phosphate. More detailed work is necessary and a fuller con.sideration of the theory is also necessary in order to give a complete physico-chemical explanation of the phenomena. IX. f«i\CLUSIO\S. (1) If constant weights of sample and of citric acid are used in a series of experiments, the only quantity varied being the volume, the quantity of phosphate dissolved per cent, of weight of sample taken increases with increasing volumes. With 5 grams of sample and 10 grams of citric acid, the citric solubihty varied from 19-24 to 28-14 per cent, of weight of sample. With 1 gram of sample and 10 grams of citric acid, the citric solubihty varied with increasing dilution from 53-8 to 61-8 per cent, of weight of sample. The effect of the presence of alkahne lime was ehminated by maintaining mjm^ at a constant value and with increasing dissociation as a result of increasing dilution increasing percentages of 142 The Citric Sohihility of Mineral PJioxphnteii phosphate were dissolved (i.e., percentages of mineral phosphate taken). The molecular concentration on the other hand (i.e. number of gram- molecules per litre) decreased with increasing dilution. It is not held that the equations describing the results are valid for concentrations or dilutions outside those used in these experiments. (2) If a constant weight of sample and a constant volume are used in a series of e.xperiments, the concentration of citric acid being varied from 1 gram to 10 grams, the citric solubilities varied from 4-4 to 19-4 per cent, of weight of sample taken. That is to say, citric solubility in- creased, as we should e.\pe(t, with increased acid concentration at con- stant volume and with a constant weight of hydroxyapatite. (3) If a constant weight of citric acid and a constant volume are used in a series of experiments, the quantities of sample being varied from experiment to experiment, the citric solubilities decreased from 19 to 1 per cent, of weight of sample taken. These decreases are due (1) to the presence of Ca(0H)2 in the molecule of hydroxyapatite, {'!) to the presence of Ca(0H)2, CaO or CaCOg in the free condition and (3) to the fact that the results are expressed in terms of mineral phosphate taken for analysis. (4) It is seen that, other conditions being constant, citric solubility depends upon an unlimited choice of constant values of any two factors, together with an unlimited number of values of the third varying factor. If, therefore, a citric solubility test has to be adopted, the theoretical condition determining the relative quantities of sample, acid and volume must be found, otherwise the test has no practical value. Wagner has not supplied any theoretical basis for selecting the specified constant quantities of the three factors. We can, therefore, at will select for slags suitable values of sample, acid and volume to secure high citric solubility values. We can, however, select at will quite different values of sample, acid and volume which will give equally high citric solubility values for mineral phosphates. Further, we could select for both slags and mineral phosphates suitable values of sample, acid and volume which would give, on the one hand, perfect citric solubihty (100 per cent.) or, on the other hand, no citric solubihty at all. (5) The citric solubility of mineral phosphate has been contrasted with the solubility of a pure substance, dicalcium phosphate, in hydro- chloric acid. It is shown that with the dilutions used and with an excess of dicalcium phosphate the reaction is practically complete. Since there is an excess of dicalcium phosphate it has been assumed that the main substance present at the end of the reaction is monocalcium phosphate. (6) Citric solubility, if apphed to fertilisers may in a certain degree J. F. Tocher 143 be a measure (1) of fineness of grinding as already pointed out by other workers, but it seems necessary also to postulate similarity of composition in comparing degrees of fineness in practice, (2) of tlie presence or absence of alkaline substances in fertilisers approximately of the same composition and ground to the same degree of fineness, (3) of the presence or absence of fluorides as well as alkaline substances in slags and (4) of the differences in the constitution of the phosphatic compounds in finely ground fertilisers containing approximately the same proportions of ex- traneous substances. None of these conditions are realisable in practice because the propor- tions and the actions of extraneous substances are generally undeter- mined and because the nature of the phosphatic compound present is either unknown or is not given. Further we can with any phosphatic fertiliser irrespective of its composition and constitution, use concentra- tions or dilutions to get any value of citric solubihty we please. Citric solubihty is therefore an unreliable empirical test of the agricultural value of mineral phosphates and slags. It has yet to be .shown that high citric solubihty is a measure of the presence of phosphate in a readily available condition for plant growth, or is an indication of the presence of a highly citric soluble phosphatic compound of a known chemical constitution. We are driven, therefore, to the conclusion that the only practical tests of value of phosphatic fertilisers from the agricultural standpoint are : (1) Total phosphatic content. (2) Degree of fineness of grinding. (3) Freedom from injurious substances and of substances inhibiting plant growth. The eft'ect of using dift'erent kinds of pho.sphatic fertilisers on yield of turnip crop, where the same amount of fertihsers, expressed as tri- calcium phosphate, is apphed in each case forms the subject of a separate communication. {Received December 23rrf, 1921.) COMPARATIVE DETERMINATIONS OF THE DIGES- TIBILITY AND METABOLl SABLE ENERGY OF GREEN OATS AND TARES, OAT AND TARE HAY AND OAT AND TARE SILAGE. By HERBERT ERNEST WOODMAN, Ph.D., D.Sc. {From the Institule for the Study of Aninud Nutrition, School of Agriculture, Cambridge University.) In a recent comnniiiicationi, the results of au investigation into the digestibiUtv of oat and tare silage were recorded. The desirabihty, how- ever, of extending the scope of this initial work was recognised, in view of attempts which are being made to re-establish on a large scale the practice of ensilage in this country. Before incurring the initial e.xpense involved in the setting up of a silo, the farmer is justly entitled to ask and receive answers to such questions as the following : (1) What are the precise conditions under which the production of a palatable silage of good quahty can be guaranteed ? (2) What is the magnitude of the losses of nutrient matter sustained by the crop when stored in the silo? Are such losses greater or smaller than those which accompany storage in the haystack? (3) Does the green forage suffer any marked diminution in digesti- bility and nutritive value during its conversion into silage, and what are the relative merits of the silo and the haystack in this respect? (4) What are the best kinds of forage to be grown in this country for the purposes of ensilage? In America, where the practice of ensilage has been well established on a very large scale for many years, numerous trials have been carried out at the various E.vperiment Stations with a view to throwing light on these questions. The conditions obtaining in that country, however, are markedly dilTerent from Enghsh conditions. The maize plant is the forage which is widely grown in America for the production of silage, and it has been found to be in every respect excellently adapted to this purpose. In this country, however, httle success has so far attended the efforts to * Wood and Woodman, Joum. of Agric. i'ci. 11, 304, 1921. H. E. Woodman 145 grow and utilise maize forage for ensilage, and consequently the attention of agriculturists has been directed towards the discovering of other crops which are suitable for being grown for silage. A large measure of success has attended these efforts. Aniosi, for instance, has shown that a mixed crop of oats and tares jjossesses all the characteristics for the successful inclusion of the jDractice of ensilage in the ordinary farm routine. It would, therefore, be unsafe to assume that the results obtained in American trials with maize forage apply with equal force to the methods of ensilage which are being adopted in this country. Concurrently with the extension of the practice, it is necessary to prosecute enquiries with a view to obtaining satisfactory answers to such ci[uestions as those enumerated above. Investigations are proceeding in all these directions at Cambridge, and the work to be detailed in the present communication was undertaken in order to obtain a direct comparison of the digestibility of oat and tare silage not only with oat and tare hay, but also with the green oats and tares from which both hay and silage had been produced. Such information must be taken into account when coming to a decision as to the relative merits of ensihng and hay making. From prior considerations, it would be natural to presume that the processes of hay and silage making would result in a diminution of the digestibility of the ingredients of the green forage, since it seems reason- able to assume that whatever fermentative or bacterial changes take place, do so mainly at the expense of the more readily assimilated con- stituents. The characteristic change which occurs when green forage is packed into the silo involves the destruction of carbohydrates with the formation of organic acids, hke lactic and acetic acids. This is the result of fermentation mainly brought about by the enzymes of the plant cells and by bacteria. Since the more ea.sily assimilated carbohydrates are hable to be used up in this process, it would be anticipated that the residual carbohydrate ingredient of the silage would have a diminished digestibiUty. On the other hand, the percentage of digestible ether extractable material in the forage will be augmented by the formation of the organic acids, although the net result of the change must involve loss of nutritive value. This loss, however, need not necessarily, on theo- retical grounds, be considerable, since these changes are quickly arrested after the development of a sufficient degree of acidity. In any case, it should be remembered that all types of roughage suffer similar destruction of soluble carbohydrates during the time they stagnate in the rumen of the animal. 1 Amos, Journ. of the Farmers' Chih, Part 2, 1920. 146 Oat and Tare Silage Similar considerations arise when forage is dried in the field and stored in the stack. If this could be carried out without undue fermentation takiiij; ])la(p. then there seems no particular reason why tlie mere drvinp clown should result in any appreciable depression of the dij^estibility. In the usual practice of hay making, however, field fermentation and heat- ing in the stack tend to deprive the forage of quite considerable amounts of its more easily assimilated organic matter, and the usual effect is to cause a decrease in the digestibihty of the protein and the nitrogen- free extractives. The pos.sibility must not be lost sight of that the changes which occur in the silo and the haystack may be accompanied by an actual increase in the digestibility of the less easily assimilated ingredients of the green forage. For instance, evidence is not lacking that the heating which occurs both in the silo and the haystack leads to an increase in the digestibility of the crude fibre constituent. The changes which affect the protein ingredient of ensiled forage, resulting in an increase in the amount of amino acids, may be expected to lead to an increase in the "ease" of digestibility of the material, rather than to an increase in the actual protein digestion <-ocfficient, since the same type of change is readily brought about in the digestive tract of the animal. The palatable and succulent nature of good silage sliould give it a distinct advantage over the dried fodder in regard to the ease ^\-ith which the animal is able to digest it. Like other succulent foods, silage is reputed to have a beneficial effect on the digestive organs, although, of course, a poor quality of silage possessing undue acidity may have the reverse effect. The available data in connection with digestibihty trials on corn forage agree substantially with the foregoing considerations. The follow- ing table gives the average digestion coefficients obtained in a large number of American trials for corn silage and green and cured corn fodder^. Forage Dry matter A.sli Protein Crude (ibre N-free extractives Ether extract Green corn fodder, ",„ Cured corn fodder, "„ Corn silage, % 68 66 66 35 34 31 61 55 53 61 66 67 74 69 70 74 72 81 These figures afford some idea of the relative digestibilities of the different ingredients of the three types of fodder, although it is not certain that in every case the results were obtained by strictly comparable trials. It will be noted that there is no appreciable difference in the 1 Henry, Feeds and Feeding, 1902, p. 248. H. E. AVOODMAN 147 digestibility ol: the com silage and dry corn fodder, the most marked difference being in the case of the ether extract, as would be anticipated. Both stored fodders are somewhat less digestible than the green fodder, although it is of interest to note that in the case of the crude fibre, the percentage digestibility shows an increase in the case of both the silage and the dried corn fodder. It is also noteworthy that drying and ensiling do not appear to depress the digestibihty of the nitrogen-free extract to any marked extent. General Arrangement of Experiment. For the purpose of the digestion trials, a plot of about 1600 square yards was measured off from a large field of oats and tares situated on the Howe Hill Experimental Farm at Cambridge. Within the hmits of the experimental plot, the growth of the crop was, to the eye, of a reasonably uniform character. In order to bring out in the sharjaest manner possible any differences in the digestibilities of the three types of foodstuff', it was decided that the experiment should consist of three main periods in which ample rations of the green forage, hay and silage, unmixed with concentrates, should be fed successively and their respective digestibihties determined directly. In the earlier work on oat and tare digestibihty, doubt was felt as to the desirability of keeping the sheep for a period of three weeks on an acidic food Uke silage alone, and the possibihty of digestive disturb- ance was avoided by feeding a smaller ration of silage together with a basal ration consisting of meadow hay and a httle linseed cake. This procedure necessitated the carrying out of digestibility measurements on the basal ration and the calculation of the silage digestibihty by difference. In this experiment, however, no difficulty was encountered in main- taining the sheep during the three weeks of experiment on an ample ration of silage alone. The quahty of the silage was good and the sheep from the outset consumed the ration quite readily without suffering the slightest discomfort. Indeed, it was noteworthy that less difficulty was experienced with the feeding during the silage period than in the green fodder and hay periods. When the green fodder was introduced into the diet, one of the sheep showed at first a tendency to be shghtly "blown," but by judiciously cutting down the amount fed for some days, this difficulty was overcome and a very satisfactory trial was obtained. In the hay period, one of the sheep displayed an inabihty to consume the entire ration, and consequently the digestibihty of the hay was deter- 148 Oat and Tare Silage miued on a smaller ration tliau was previously designed. Reference will be made to this point again at a later stage. It was thus possible, by avoiding basal rations, to measure the digestibiUty of the silage directly on a much larger quantity of material than could possibly have been fed in an indirect method of determination, and, moreover, the conditions obtaining throughout the whole trial were thus made comparable. The green oats and tares period began on .Tune IG. 1921. Samples of about 25 lbs. were cut from the plot daily by means of a sickle and were passed through the chaffing machine before being fed to the sheep. During the few hours over which it was necessary to store the material, it was spread out in a thin layer on a concrete floor to prevent " heating," which occurred fairly readily when the cut fodder was kept in bulk in bags. The .samples for analysis were taken at the .same time as the rations for the whole day were weighed out. Moisture determinations were carried out on the representative .samples every two days during the analytical period, the dry matter from such determinations being utihsed in the making up of the period composite sample. The gradual increase of dry matter in the maturing crop is illustrated by the following figures obtained during the fourteen days' experimental period : Days 1-2 3-4 5-6 7-8 9-10 11-12 13-14 Dry matter, % 2900 30-95 33-69 32-55 33-23 33-55 34-32 As the approach of the crop to maturity was probably accompanied by a gradual diminution in the digestibihty of the forage, it was necessary, in order to secure a fair comparison with the hay and silage, that the parts of the plot reserved for the liay and silage should be cut halfway through the green oats and tares period. This was accordingly done and the material to be converted into silage was carted without delay, passed through the usual cutting machine and then packed tightly into a small experimental silo reserved for this purpose. Further particulars will be given when the details of the silage period are discus.sed ; it is sufficient to note here that all the conditions for making the silage were .satis- factory and material of an excellent quaUty was obtained when the silo was opened at a later date. The making of the hay was not attended with similar good fortune. The first two days during which the forage was drying in the field were beautifully fine and sunny, but a sudden thunder- storm shortly before the time for carting rendered the sample quite unfit for use in a comparative trial with the silage. It was therefore necessary to use another sample of oat and tare hay which had been carted before H. E. Woodman 149 the storm, but whicli lia,d been cut at the same time as the forage on the experimental f>lot. It came from a neighbouring part of the field, but, as will be seen later, this circumstance detracted somewhat from the strictly comparative nature of the trials. The hay sample was finally packed tightly into a meadow haystack standing in a Dutch barn and a thick layer of straw was pressed compactly on top, so that the water vapour generated by the hea.ting in the stack should pass into the straw and not condense in the upper portions of the hay, thus causing it to mould. The experimental procedure in the trials was, in the main, identical with that adopted in the earlier work on oat and tare silage digestibility. The Hainan harness was again employed with success, and there can be little doubt that, from the point of view of convenience, safety and com- fort for the sheep, it is distinctly superior to the orthodox " funnel and bag" harness. The analytical period in all three cases consisted of fourteen days, this being preceded by a preliminary period of seven days. Bi-weekly composites of urine and faeces were made, nitrogen estimations being carried out on these samples. Determinations of dry matter were made on aliquot portions of the faeces samples, the dried residues being pre- served in air-tight bottles to be utilised in the making up of the period composite samples for complete analysis. Separate period urine com- posites were kept for dry matter estimations and the determination of the energy content by means of the bomb calorimeter. They were pre- served by the addition of chloroform. In the hay period, the daily rations for the whole period were weighed out previously into paper bags from the bulk of chaffed hay, the samples for complete analysis and moisture determinations being drawn at the same time. In the silage period, as in the green forage period, moisture determinations were made on representative samples every two days, the dried material being made up into the composite samples for com- plete analysis. It was not considered necessary to carry out determina- tions of nitrogen on the fresh silage samples, since in the earlier work with oat and tare silage, it was noted that no measurable loss of nitrogen occurred during the drying down of silage. In order to gain a trustworthy comparison of the digestion coefficients of the protein constituents of the three foodstuft's, it was essential to take into account the well-established fact that the faeces do not con- sist solely of undigested food residues, but that the latter are largely contaminated by nitrogenous metabohc products which have been 160 Oat and Tare Silage secreted into tlie alimentary tract and escaped re-absorption. Thus the protein digestion coefficients as determined directly represent minimum values, and in view of the possibility of this disturbinjr factor operating unequally in the trials with the difTerent fodders, it was deemed advisable to correct the protein digestibilities by basing them on the amount of pepsin-insoluble nitrogen in the faeces. Accordingly, determinations of the pepsin-insoluble nitrogenous constituents of the faeces were made in all three cases. The error occasioned by the presence of metaboUc pro- ducts in the faeces affects, of course, the accuracy of the digestion co- efficients of all the foodstulT ingredients, but it falls with especial weight upon the protein and ether extract. No satisfactory and reliable method for correcting the ether extract digestibihty has so far been evolved^. The sheep were weighed at the beginning and end of each period and account was kept of the nitrogen balance. The wethers employed were the same as were used in the earlier work on oat and tare silage. A study of the results shows an extraordinarily good agreement between the sets of figures obtained for the two sheep. Such agreement is not often met with in work with animals, and it would appear that the two sheep selected for the purpose possessed almost equal digestive capacity. The writer's thanks are due to his as.sistant, Mr V. .T. Aylett, for the skilful manner in which lie took charge of the animals and for tJie care with which he carried out the analytical work in connection with the digestibihty trials. Table I . Dclaih of ratian-s. Daily Dry matter Period ratinn HI ration t,'[n. nm. Gropn oats and tares 40(){) 1 2S)il-0 Oat and tare hay 1000 839-6 Oat and tare silage 3000 8190 In planning the rations for the trials, it was, for obvious reasons, intended that the sheep should receive roughly the same weight of dry matter per day throughout. This object, however, was not attained for the following reasons. The first moLsture determinations carried out on preliminary samples of the green oats and tares showed them to have a drv matter content of about 27 per cent. A diet of 4000 gm. of this fodder contained therefore about 1080 gm. of dry matter. By the end of the j)erio(l, however, the thy matter content of the crop rose to over 34 per cent, and the mean percentage for the analytical ])erio(l was ' For fuller information on this jmint, see Crowlher and Woodman, Jniini. of Agric. Sci. 8, 434, 1917. H. E. Woodman 151 32-47 per ceut. The dryness of the crop was probably connected with the droughty season during which it was approaching maturity. TJie con- sequence of this gradual increase in the percentage of dry matter was that the sheep received a larger allowance of dry matter during the period than was intended. Further, it was noted during the prehminary feeding of the oat and tare hay, that one of the sheep was unable to con- sume comfortably more than about 1000 gm. per day. There was no alternative, therefore, but to cut down the daily ration to this amount, and as a consequence, the amount of silage given in the daily ration during the third period had to be kept at the hay dry matter level, since it was desired, above all, to obtain a clean-cut comparison between the hay and the silage. The possible effect of these circumstances on the strict comparativeness of the trials will be discussed later. (Ireem Oats and Tares Period. Condition of crop during trial. The crop consisted of autumn-sown oats and tares grown upon clay land. Equal quantities of grey winter oats and tares were planted, but the tares predominated at the time of the trial. Owing to the exceptionally dry summer, the crop was short in the straw and .stood well. The oats were just coming into milk and the tares were in full flower. The condition of the crop was therefore almost ideal for cutting for liay, but was somewhat premature, according to the customary practice, for silage. It follows that the conditions of the trial slightly favoured the green oats and tares and the hay rather than the silage. Table II. Composition of green oats and tares composite samjile {calculated to dry matter). 0/ ■ /o Crude protein 10-83 Ether extract ... 302 Nitrogen-free extractives 50-22 Crude fibre 28-12 Ash 7-81 Average moisture content of green oats and tares = 67-53 °/g. Average amount of dry matter in green oats and tares per day = 1299-0 gm. Table III. Average weight and composition of faeces. Sheep I Sheep II gm. gm. Weight of fresh faeces daily 1483 1576 Weight of dry matter daily 469-37 474-85 Joum. of Agric. Sci. xii 11 152 Oat and Tare Silage Table III {cont.). Composition of dry maUer. Crude protein* ... Ether extract ... Nitrogen-free e.xtraetivcs ('rude fibre Asli Pcpsin-HCl insoluble protein * Crude protein as dcterinini-d on fresh faeces Tabic 1\'. Digesiihility of green ouls and lures. Daily ration: 4000 gni. green oats and tares. Sheep 1. HEEP I Sheep II o/ o/ /o ,o 11-52 9-54 4-01 3-99 31-93 3315 40- IS 40-84 12-30 12-48 5-24 5-17 3-65 3-15 (Jnnsumed (green oats and tares) Voided ... Digested Digestion (-oeHieicnts, "{, Total dry matter 129!)-(IO 409-37 829-63 63-80 Organic Crude matter gm. 1197-55 411-30 786-19 65-65 protein m. Ether extract 140-09 54-13* 86-56 01-53 39-24 18-82 20-42 52-04 N-free extrac- tives gm. gm. 652-35 305-27 149-87 188-59 502-48 176-68 77-03 48-37 Sheep II. * Calculated on nitrogen of fresh faeces. I'rulcin digcslibilil// cvrrcded for metnhnlic nitrogen. SuKKp I Sheep II Protein coiisuniod, i;ni. ... ... ... 140-69 Pepsin-insoluble protein \iiided, f.'m. ... 24-59 Protein digested, gm. ... ... ... 116-10 Corrected digestion coefficient, % ... 82-52 Mean corrected digestion eoeliieient, % 82-5 140-69 24-55 11614 82-55 Ash gm. 1(11-45 58-01 43-44 42-82 Consumed (green oats and tares) Voided 129900 474-85 1107-55 415-59 140-09 49-65* 39-24 18-95 652-35 157-41 365-27 193-93 101-45 59-26 Digested Digestion eoeliicicnts, "„ 824- 15 63-44 781-96 65-29 91-04 li4-70 20-29 51-71 494-94 75-87 171-34 46-91 42-19 41-59 Mean digestion coi^lf., "„ 03-7 65-5 63-1 51-9 76-5 47-6 42-2 The Siitisfactory a Av. daily oats and tares day In faeces In urine Total N balance gm. gm. gm. gm. gm. Sheep 1 22-.51 8-66 1100 19-66 +2-85 Sheep II 22-51 7-94 12-18 20-12 +2-39 Sheep I Sheep II St. 11). St. lb. June 16, 1921 9 4 10 1 July 7, 1921 9 2| 10 1 Change in weiglit - 1 ^ 0 Whilst it is unsafe to base conclusions in connection with the utihsa- tion of food protein from short period experiments, it is satisfactory to note that the forage diet provided a satisfactory maintenance ration for the animals. Oat and Tare Hay Period. Condition of fodder at time of feeding. Reference has already been made to the unfortunate circumstance which rendered necessary the rejection of the hay from the experimental plot and the employment in.stead of an unspoilt sample from another part of the field. The con- ditions under which the material was stacked have also been described. The small stack weighed about 5 cwt. ; the outer portions, which were shghtly spoilt by mould, were rejected and a shoe of about li cwt. was cut down the middle (October 7, 1921). The sample was chaffed and the sampUng and feeding were carried out in the manner already described. The hay was mainly of a nice green coh)ur, shghtly bleached in places by the sun. It had no smell of heating in the stack, but retained the natural aroma of the herbage, characteristic of green hay. The crop was cut at an ideal time for hay, the oats being just in milk and the tares in 11—2 154 (Jitt (did Tare Si Inge full flower. The leaf had been well saved and the fodder could be described as "a useful sample of hay, although perhaps uot of the very best quality ^" A mechanical analysis carried out on a representative sample of the hay showed it to contain rouyhly 30 per cent, by weight of oats and 70 per cent, of tares. Taljlc \ II . ( 'ttni position of oal and tare liai/ comjjosile sample (calculated to drij iixitlcr). rriidc protein ... ... KMM) Kthor extract 20!t N'itrogen-free extractives t.">SI Crude fibre 2".)07 Ash 9-13 Average muisture content of oat and tare liay - l(i'04 "/q. Average amount of dry matter ill oat and tare iiay per day = 839-() gm. Attention has already been drawn to the reasons why the amount of dry matter fed ])or (lav during this ])eriod was smaller than in the pre- ceding period. The liigh percrntage of protein in the hay, as compared with the green forage, must also be noted. It would be expected that the Jiay dry matter would contain rather more protein than the dry matter of the green oats and tares, since, as a result of field fermentation and lieating in the stack, losses of non-nitrogenous organic matter occur. Table VUl. Average weight and composition of faeces. Sheep I Sheep II Weiglit of fresh faeee.s daily, gm. 601 701 Weight of dry matter daily, gm. 291-73 290-80 Composition of dri/ mailer. Sheep I Sheep II o/ o/ O CI Crude protein* 12-13 12-.')-t Ether extract 3-77 3-77 Nitrogen-freo extractives ... 37-31 37-83 Crude film- 34-20 34-2.5 Ash \2r)9 11-0! PepsinHCl insoluble protein ... (i-21 <>-4(l * Crude protein as determined on /rc.«ft faeces ... ... ... 002 5-43 The increase in the percentage of protein cannot, however, be wholly accounted for in this way, and the explanation is probably to be found in the fact that the hay was not. for reasons already stated, taken from ' The writer's colleague. Mr .•Vrlhur Amos, M.A., very kindly cxj)ressed his opinion.^ regarding the quality of the hay and the silage used in these trials. H. E. Woodman 1 .15 the experimental plot, but came from another part of the held where the growth of the crop was not quite uniform with that of the experi- mental crop. That the crop was not quite uniform within the limits of the ex23erimeutal plot itself is evidenced by the fact that the silage was also shghtly richer in protein than would be anticipated from a study of the losses of carbohydrates in the silo. It is not considered, however, that these circumstances materially affect the main conclusions to be drawn from the experiment, though they illustrate the extra difficulties introduced into work dealing with a mixed crop. Table IX. DigestibilUy of oat and tare hay. Daily ration: 1000 gm. oat and tare hay. Sheep I. (onsunipd (oat and tare liay) Voided ... Digested Digestion coi'fticiiMita. "„ Consumed (oat and tarf liav) ... Voidi'd Digested Digestion ooeHicients, "„ Mean digestion coeff., % Total dr,y matter S39-00 291-73 -)47-87 (15 -2 5 Organic matter gm. 7G2!I4 2r)!5-00 507-94 (iO-58 Crude protein gm. 110-70 30-18* SU-52 09-00 Ether extract gm. 17-55 11-00 G-55 37-32 N-free extrac- tives gm. 384-G2 108-84 275-78 71-70 Crude fibre 244-07 99-77 144-30 .TO- 1 2 SuEEr II. Sheep I Sheep II 116-70 110-70 18-12 IH-99 98-58 97-71 84-48 83-73 84-1 Ash gm. 839-60 762-94 116-70 17,55 384-02 244-07 296-80 262-34 38-06* 11-19 112-28 101-65 542-80 500-60 78-64 (;-36 272-,34 142-42 64-05 65-62 67-39 36-24 70-81 58-35 65-0 06-1 68-2 36-8 71-3 58-7 76-66 36-73 39-93 52-09 76-60 34-46 42-20 55-05 53-0 * Calculated on nitrogen in fnsh faeces. Correction of ■protein d,igestibiUti/ for metabolic nitrofjen. Protein consumed, gm. ... Pepsin-insoluble protein voided, gm. ... Protein digested, gm. Corrected digestion coefficient, % Mean corrected digestion coefficient, % The results again show good agreement between the two sheep. They will be discussed in detail at a later staiie. 156 <><(t Olid Tare Sila(/e Table X. Percentages of digest ihle nutrients in oat and tare hay sample {calculated to dry mailer). Oudp protein* 9-48 Ether extract 0-77 Nitrogen-free extractives 32-66 Crude libre ... ... 1700 Production stareli equivalent (Kellner) per 100 lbs. dry oat and tare hay =43-24. * Ap))arent digestion coefficient used in calculation. Tnl)l(' XI. Nitrogen balance during period and weights of sheep. Daily ration N consumed N voided Av. daily 1000 gni. oat Av. per , * < N and tare hay day In faeces In urine Total balance gm. gm. gm. gm. gm. Sheep I 18-67 5-79 11-23 1702 +1-65 Sheep II 18-67 6-09 12-18 18-27 -(-0-40 Sheep I Sheep II St. lb. St. lb. Oct. 13, 1921 9 4 10 8 Nov. 3, 1021 9 1 9 12 Change in weight -3 - 10 It is scarcely a matter for surprise tliat tlie slieep during this period lost a little weight, since the ration was scarcely up to maintenance requirements, owing to the necessity which arose of having tg cut down the amount of hay to secure complete consumption. Both sheep, how- ever, showed a sUght retention of protein during the trial. Oat and Tare Silage Period. Quality of silage used in experiment. The condition of the crop at the time of cutting (June 23, 1921) has already been described. The material was carted within three hours of cutting, so that Uttle or no wilting was allowed to take place. The silage was made in a miniature wooden silo^, which was 4 feet in diameter and 6 feet high and rested on a foundation of gault clay. The forage was first cut by the usual chaif-cutter and then filled into the small silo, precautions being taken that the material settled down compactly. A thick layer of soil was placed on top. The silo was opened on Nov. 8, 1921, and after rejecting the small amount of waste ' A number of such experimental silos have been erected on the Howe Hill Farm in connection with work being carried out by Mr A. Amos and the writer on the making of silage mider controlled conditions, an account of which will be published shortly. H. E. Woodman 157 material on top, silage of excellent quality was encouutered. It possessed a good green colour and a pleasant "fruity" odour, the smell of butyric acid beinr;: entirely absent. The silage as fed contained very few tare pods and no tare seeds, nor did the oat husks contain any solid food material. The sheep consumed it readily; the surplus was fed to stock, and it was observed that they ate it with relish and throve upon it. Fresh samples of silage for feeding were taken every day and the top of the silo was covered by a tarpaulin during the experiment. As the trial proceeded, the quality of the silage fell oft' slightly, owing probably to the slowness with which it was being used up. The last portions were not quite so green and "fruity," hut still were of good quality and quite free from butyric acid. AiKihjf^is uffiiJiifie ciiract. In order to gain some insight into the nature of the changes which had occurred during ensilage of the green crop, aqueous extracts of the silage were submitted to analysis. A 2()0 gm. sample of the silage was submitted to extraction by shaking for four hours in a shaking machine with GUO c.c. of distilled water. The extract was filtered first through hnen, the residue being well squeezed out, and then through a filter paper. 1.50 c.c. of the aqueous extract were made up to 500 c.c. with alcohol. This occasioned the separation of a small amount of precipitate, which settled readily, and the resultant clear alcohol liquid was submitted to analysis by the Foreman titration method^. Fuller details regarding the analysis of silage extracts will be given in another communication. Table XII. Analijsis of silage extract. (The data roter to 100 gm. of the fresh silage — moisture content = 72-9 °,',.) c.c. iV/10 Total acid radicles (free and combined) ... 316'1 Amino acids and amides of asparagine type 104-2 Total organic acids of lactic and acetic type 211-9 Organic acids volatile in steam ... ... 56-6 Non-volatile organic acids ... ... ... 155-3 Volatile bases ... ... ... ... 18-1 Calculated as acetic acid, the percentage of volatile organic acids in the fresh silage works out at 0-34 per cent. This, of course, is reckoned as moisture by the customary method of determining dry matter, but allowance has been made for it in all the data tabulated here in connection with silage digestibiUty. To do this involved slight assumptions, which, however, could only possibly aft'ect the result to an inappreciable extent. 1 Foreman, Bioch. Joarii. 14, 451, 1920. 158 Out and Tare Silage Table XIII. ComjMsilion of oat and tare silage composite sample {calculated to moisture-free material). Crude i)rot<»iii ... ... 12-55 Kther e.xtiac-t 4-32 Nitrogen-free e.\trai:live.s 45-57 (,'rudc fibre ... ... 29-44 Ash 8-12 Average moisture content uf oat and tare silage -72-70 %. Average amount of dry matter in oat and tare silage per day = 819-0 gm. Table XIV. Average weight and composition of faeces. Weight of fresh faeces daily, gni. Weight of dry matter daily, gm. Composition of dry matter Crude protein* ... Ether extract Nitrogen-free extractives Crude fibre Ash Pepsin-insoluble jirotein ♦Crude protein a.s determined on fresh faeces Sheep I Sheep II 569 631 291-73 297- 10 Sheep I Sheep II /o % 1202 11-00 3-27 3-12 37-15 37-75 35-03 35-23 12-53 12-90 5-76 5-65 6-38 5-64 Table XV. Digestibility of oat and tare silage. Daily ration : 3000 gm. silage. Sheep I. Total dry matter Organic matter Crude protein Ether extract N-free extrac- tives Crude fibre Ash gm. gm. gm. gm. gm. gm. gm. Consumed (oat and tare silage) 819-00 752-50 102-78 35-38 373-23 241-11 66-50 Voided . 291-73 2,55-18 36-25* 9-54 108-38 102-19 36-55 Digested 527-27 497-32 66-53 25-84 264-85 138-92 29-95 Digestion coefficients, % 64-38 66-09 64-73 73-04 70-96 57-62 45-04 Sheep II. Consumed {oat and tare silage) Voided ... 81900 297- 10 752-50 258-77 102-78 35-56* 35-38 9-27 373-23 11216 241-11 104-66 66-50 38-33 Digested Digestion coefficients, "^ 521-90 63-72 493-73 65-61 67-22 65-40 26-11 73-80 261-07 69-95 136-45 56-59 28-17 42-36 Mean digestion eoeff., "„ 64- 1 6.-)-9 65-1 73-4 70-5 57-1 43-7 Calculated on nitrogen of fresh faeces. H. E. Woodman 159 The percentage of etlier extract i.s uot as large as would be expected from a study of the figures obtained for the organic acids in the silage extract. The difference is not wholly accounted for by the fact that a portion of the acids must exist in combination with bases, and it seems to indicate the occurrence in the silage of substances of an acidic nature which are not extracted by ether. Evidence that this might be the case was obtained during the titration of the extract with N/IO alkali. As the neutrality point was approached, a yellow colour developed in the originally almost water-clear solution. This point is being investigated further. Protein digestibility corrected for metabolic nitrogen. Sheep I Sheep II Protein consumed, gm. 102-78 102-78 Pepsin-insoluble protein voided, gm. 16-80 16-79 Protein digested, gm 85-98 8.5-99 Corrected digestion coetticient, % ... 83-66 8.S-66 Mean corrected digestion coefficient % 83-7 As in the other two trials, there is little fault to find with the agree- ment shown by the two sets of coefficients. Table XVI. Percentages of digestible nutrients in the oat and tare silage sample {calculated to dry matter). 0/ /O Crude protein ... ... 8-17 Ether extract 3-17 Nitrogen-free extractives 32-13 Crude Bbre 16-81 Ash 3-55 Production starch equivalent (KeUner) per 100 lbs. dry silage = 45-59. Table XVII. Nitrogen balance during period and weights of sheep. Daily ration 3000 gm. oat and tare silage N consumed Av. per day II N voided Av. daily N balance 1 faeces In urine Total Sheep I Sheep II gm. 16-45 16-45 gm. 5-80 5-69 gm. 9-88 10-40 gm. 1.5-68 16-09 gm. + 0-77 + 0-36 Sheep : I Sheep II Nov. 14 Dec. 5, Change , 192 1921 in w« St. 1 9 8 sight lb. 6 12 -8 St. lb. 10 5 9 8 -11 Both sheep lost weight in this period, but no conclusions can be drawn from this as to the relative feeding values of hay and silage. The 160 Oat and Tare Silage periods were too short and tlie sheep were not on controlled diets previous to the different trials. The sheep would readily have consumed a heavier diet of silage, l)ut it was desired to keep the amount of drv matter con- sumed per day rou^^hly the same as in the hay period. P>ot1i animals were rouijlily in nitrt)f^enous eiiuililnium during this peridd. Table Will. Siiiiiiiitiri/ of digestihllili/ n-sull.'^. 1. Oonipai'isdii of dijiestimi eoellieients. Oat and tare < Jreen oats Oat and Oat and silage and t-ares tare iiay tare silatie (1920-21) /O ",i .o Dry mattor 03-7 fW-O (ill iw-S Organic mattor (ir,r, (!(■• 1 (»■!» r.5-8 Cnnlr pidtciii (}i|)|)an'iu) (i:!- 1 tiSl' 0.5- 1 07-2 ('null- protein (i-orrcc-led) H-2-r, 84- 1 83-7 — Ether extriiet 51-9 3{i-8 73-4 78-9 Nitrogen-fret' extractives 76-) 71-3 70-5 52-2 Cnule HIjre 470 58-7 571 49-7 Ash 42-2 53-6 43-7 50-2 2. Percentages of digest ilile nutrients (calculated to dry matter) starch equivalents. Oat and tare Oreen oats Oat and Oat and silage and tares tare hay tare silage (1920-21) /o o o /o Crude protein 0-83 9-48 817 10-91 Ether extract ... l-.'57 0-77 317 3-35 Nitrogen-free extractives 38-42 32-(iti 3213 19-47 Crude fibre 13-39 17(11) 16-81 10-39 l'r(]dnction .starch eiiuiva-l lent jier KM) llis. of dry 44-92 43-24 45-59 33-4 fodder... ... ... ) Discussion of results. In the fourth column are given the ligures obtained from digestibiUty trials with the same sheep on the previous year's crop of oat and tare silage. These results were not corrected for the volatile acid content of the silage. They differ materially from the results obtained in the present investigation, the pre^^ous year's silage possessing on the whole a markedly lower digestibility. This difference comes out strikingly in the cases of the dry matter, organic matter, nitrogen-free extractives and fibre, whilst the protein and ether extract fractions possess similar digestibiUties. The previous year's silage was much richer in protein (16-2/5 per cent, on dry matter) than the 1921-22 crop (12-55 per cent.), and this is reflected in the table giviug tJie amounts of digestible nutrients. The wide difference in the amoimts of dige.stible carbohydrates in the two silage .samples is also noteworthy, whilst the lower nutritive value of the previous year's silage is evidenced by the difference between the values of the starch equivalents. (1) Black winter oata sown. (2) Crop out on July 12, 1920. It w'as quite mature. Oats had just passed milk stage and tares were well seeded. (3) Crop allowed to wilt one or two day.s before carting. (4) Silage made in commercial silo; maxi- mum temj^erature of fermentation wa.s 35° C. Silage brown in colour with somewhat pungent odour. (5) Seeds of both oats and tares in silage contained much solid food material. H. E. Woodman 161 It is of interest to examine the possible reasons for these wide variations in the results for the two silages, since obviously a practical point of some importance is involved. The differences in the procedures by which the two crops were produced are noted below in parallel columns. Sn,.u;E 1920-21. Silage 1921-22. (1) Grey winter oats sown. (2) Croji cut on June 23, 1921. Crop was immature and ideal for hay. Oats were just commg into mUk and tares were in full flower. (3) Crop carted within three hours of cutting. (4) Silage made in miniature ailo. Tempera- ture of fermentation did not exceed 25° C. Silage green in colour with pleasant fruity smell. (5) Silage contained very few tare pods and no tare seeds, and oat husks did not contain solid food material. The first set of conditions is customarily regarded as ideal for silage. The results of this investigation, however, indicate that early cutting and carting without wilting may lead to a great gain in palatability, digestibility and nutritive value of the silage, although, of course, the actual weight of forage carted may be somewhat smaller per acre. The chief results obtained in the present exjDeriment on the compara- tive digestibilities of the three types of fodder are, with the exception of the protein figures, in fair agreement with the results obtained in American investigations with corn forage. These results, which are the averages from a large number of trials, not necessarily strictly compara- tive, have been summarised earlier in the paper. It will be noted that the digestibilities of the total dry matter, total organic matter and crude protein are of a similar order in all three cases. The uncorrected figures for protein digestibihty indicated that the hay protein was distinctly more digestible than the protein of the silage and the green fodder, but on taking into account the metabohc nitrogen of the faeces, the in- equaUties are almost wiped out. It would be anticipated that striking differences would occur in the digestion coefficients of the ether extracts. The ether extract of the green forage is about half absorbed; in the case of the hay, the availability sinks to about 37 per cent., whilst with the silage ether extract, which contains the easily assimilated organic acids, the relatively high figure of 73 per cent, is reached. It is not to be assumed, however, that the organic acids of the silage possess the nutritive quahties of the soluble carbo- hydrates of the green fodder from which they have arisen during ensilage. 162 Oat miff Tare Silar/e Furthermore, it must Ix' borm- in mind tliat in not one of the three cases does the ether extract consist wholly of true fat. and also that the fat dijjostion coefficients are lialtlo to be subject to error, in view of the fact that the faeces always contain ether soluble material arising from meta- bohc products and not from actual food residues. The nitrogen-free extractives of the hay and the silage are approxi- mately of equal digestibihty, although in both cases the digestibility is lower than that of the corresponding fraction of the green forage. The depression of digestibility in this respect during the conversion of the crop into hay and silage is not, however, so great as has sometimes been supposed. It is interesting to note that the fibre constituent of the hay and silage is almost equally digested, whereas that of the green crop possesses an appreciably lower digestibihty. This finding confirms the supposition that heating in the stack and the silo leads to a definite increase in the dige.stibility of the crude fibre. In view of the fact that such fodders contain relatively large amounts of fibre, this increase of digestibility becomes of con.sidcrable significance. Attention should be called to the fact that whereas in the hay and silage periods almost equal amounts of dry matter were fed per day, yet in the green oats and tares period a much larger allowance of dry matter, for reasons already gone into, was consumed by the sheep. It is well known that animals tend to digest their food with somewhat less com- pleteness when the ration undergoes any marked increase in bulk. This variation is not necessarily very pronounced. Indeed, if it were, then the digestion coefficients based on feeding definite rations (usually sub- maintenance in such tests) could only possess a limited value. It is only fair, however, in comparing the green fodder digestion coefficients with those of the hav and the silage, to regard them as being minimum values, and to assume that if a ration of green fodder more comparable in dry matter content with the hay and silage rations had been fed. slightly higher values would liave been obtained. Tliis does not affect greatly the luain conclusion that the digestibility of the ]iay and silage dry matter compares very favourably with that of the green fodder dry matter. The findings outlined above are substantially confirmed by a study of the tabulated starch equivalents, which give the nutritive value of 100 lbs. dry fodder for production in terms of lbs. of starch. If anything, these results point slightly in favour of the silage. The results giving the actual percentages of digestible nutrients in the dry foodstuffs, whilst not strict!}- comparable owing to slight non- II. R. WOOD.AIAN 1H3 uniforniity iu the growth of the oat aud tare crop, show clearly that both hay and silage contain somewhat more digestible protein and fibre and rather less digestible nitrogen-free extractives than the green crop. The amounts of digestible ether extract vary within wide limits, the hay figure being very low and the silage figure relatively high. In commenting on the decrease of digestibihty which is assumed to occur when a green crop is ensiled, Henry and Morrison^ write: "The exceedingly favourable results from silage feeding are therefore due to the palatability of tiie silage, its beneficial eft'ect on the liealtli of the animals and the fact that less feed is wasted than when dry fodder is used." The results of this investigation indicate, however, that a con- tributory factor of great importance is the fact that the silage possesses a digestibility and a nutritive value which are only slightly, if at all, inferioi- to those possessed b}' the green forage from which it has been produced. Comparison of Metabolisable Eneruy of Greex Oats and Tares, Oat and Tare Hay and Oat and Tare Silage. In order to e.xtond the comparisons already outlined, it was decided to make determinations of the metabolisable energy of the three types of fodder. The metabolisable energy may be regarded as that portion of the gross energy of a foodstuff which is available for utilisation in the body of the animal; it does not, however, represent the true value of the foodstuft' for general production purposes, since further deductions are necessary in allowing for the energy used up in the processes of mastica- tion aud digestion. The metabohsable energy is ascertained by deducting from the gross energy of the food.stufi the losses of energy from the body in the form of the hquid, sohd and gaseous excreta; its determination involves, therefore, the carrying out of bomb calorimetric experiments on the foodstuft', dry faeces and dry matter of the urine. The course of the determinations will be gathered from a study of the following tables. Table XIX. Details of urine output during trials. Sheep I Sheep II Period Green oats and tares Oat and tare hay . Wheat l'..S. Wheat and Grass U.S tira=s and Wh^at Fig. 1. Comparison of types liy enmniation i\irvcs. of the mean diameters of the groups, the points so obtained being joined up by a smooth curve. The value -0001 mm. has been taken as the lower hraiting value of the dimensions of the clay group. This is purely an arbitrary proceeding. The proportion of "colloid" clay could here be V. L. Whittles Kjti shown with great advantage. In this connection the u.se of ultra-filtration methods might give valuable information. The curves for the fighter soils rise steeply and then flatten out, those for the heavier soils are flat at first and rise steeply later. Two particular appfications of these curves are of importance : {]) Two soils, analysed according to different systems of grouping of particles may be compared exactly. (2) The analytical results may be transformed from one system to another by reading oft' the values at the points at which the curve cuts the mean diameter of the selected classes. Baker (7) has devised two values for describing a soil from its mechani- cal analysis plotted in this way. This is of value for catalogue and descrip- tive purposes, but is not so well adapted for the preparation of drift maps as is the triangular method. In this method only three clas.ses of particles can be considered, which may be conveniently termed Coarse, Medium and Fine respectively. The question now arises as to which two limiting values shall be chosen for the separation of the Medium class from the Coarse and Fine grades respectively. Considering Fig. 1 we find that: U.S. Bureau of Soils take -0:3 and •()02!'). Wilsdon(58) takes -02.5 and -Odl. It is here suggested that, taking into consideration the known properties of the various grades, the curves are best characterised by their intercepts on the lines for -12 and -006. On this basis we have, on the usual English system: Coarse (fine gravel + coarse sand) Medium (fine sand + silt) Fine (fine silt -1- clay). Triangular Methods. In the method adopted by the U.S. Bureau of Soils the proportions of clay and silt are plotted along each of two axes at right angles. By joining the 100 points on these lines a right-angled triangle is formed. This is then conventionally divided up into compartments as shown in Fig. 2. The scheme of classification is given in Table III. Wilsdon has suggested a modification of this method whereby the proportion of the third constituent — the sand — may be read off directly from the diagram. The percentages of sand, silt and clay are plotted on an equilateral triangle. The detailed procedure is given in connection 12—2 170 Classification of Soils Table 111. U.S. Bureau of Soil's Classifitation. Coarse sand Mcdiuni sand Fine sand Sandy loam Fine sandy loam Loam iSilt loam Clay loam Sandy clay Sill clay Clay Fine gravel 2-1 Coarse " sand 1-5 mm. Medium sand More than 25 "„ (1+2) More than .W "„ (1+2 + 3) Less than 20 "„ (1+2) I More than 20% (1+2-t 3) Less than 20 "o (1+2 +3) More than 20'?;, (1+2 + 3) Less than 20% (1+2 + 3) Fine sand ■25-10 mm. Very fine sand ■10-05 mm. Silt •05-005 mm. Clay •005-0 0-15 \ 0-10 % Less than 20 "i, (G + 7) 0-15% I 0-10% Less than 20 % (6 + 7) 0-15% I 0-10% Less than 20% (6 +7) 10%-35%|5%-15", More than 20 "„ ((> H 7 l.*'S8 than 50 % (6 + 7) More than 20 "„ (6 + 7) Less than 50 "„ (6 + 7) Less than 15%-25",, 55 % (6) More than 50% (6 + 7) More than 55 % (6) •2.T o' _r,,-, q/ I>ess than 25 % (7) 25 %-35 More than 60 % (6 + 7 1 Less than 25 % (6) More than 20 % (7) Less than 60 % (6 + 7) More than 55 % (6) 25%- 35 % (7) More than 35 *^' (7) More than 60% (6 + 7) witli a furtlier inoditication. Soils arc clas.'iifipd into fmirtecii uroiips, th« liiiiiting 2)r()porti()ii.s are giveu iu Table 1\', and illustrated grajiliically in Fig. 3. Atteiitiou is directed to the following considerations in connection with this method : 1. Only a comparatively small area of the whole triangle is employed. 2. The crop types are not so well differentiated as in that of the U.S. Bureau of Soils. I i C L. Whittles 171 Table IV. Classification of Soils (Wilsdon). Percentage Sand Silt Clay DesiTiptirin (+004 mm.) (+0002 mm.) ( -0002 mm.) Sand + 75 -25 -25 Sandy Itiani I + B0 -40 -10 „ II + IJ0 -30 -20 Heavy sandy loam -75 -20 -30 Loam I + 45 + 30 -10 „ II -45 + 45 -10 „ III -60 + 20 -20 ,. IV -50 + 30 -20 „ V -40 + 40 -20 Silt -30 + 50 -20 Heavy silt loam -20 + 50 -30 Clay loam I -60 -40 -30 >— + 20 + 30 -30 Clay -70 -70 + 30 A + sign is placed before a minimum limit, and a - sign before a maximum. "'"/or 20° \ al-.S. GiaiK ^y ^ and Wheat \. \ Cuv \^ • U.S. Wieat \ ,. ,Wlwat ami Grass \ Ci.AV Loam \^ ^-1lS. Whea\ •^''"'^^- ""1" ^' • Orchard* SiT.TV C'LAV I.CIAM \^ .S.VNOY LO.VMS \ ,. ^Tobacco ^ •Sh.q. and BaiU (hemes \ Sm.T 1,0AM \^ »U.S .Ma. kit (iaidiii i.'ii,;. • .Maikil tiardeii Jl„[,^ \ •is Ka]lv .Market Garden SaSUS \ -'»°/o ^"'°/o Fig. 2. Comparison of types by U.S. method. 1110° /o Cr.AY Cr.AV Loam I Ci.AV Loam 11 \ • Wlieat •l.'K Wheat^l.oAM n'.LoAM V' Sandy Loam\ Loam \ ^Coai'^XHops Potato MI lOep and Barley^ _ ^.Market lialdell VS. Kail? Market (!arH>-ii To^iae,-.. -• -lolKie. ^ Cherried and Vi Loam I ine Hops \ Loam H SaMi.. II in nini .Silt XI 002 niui. Fig. 3. Comparison of t^'pes by Wilsdon's method. 172 Classification a/ Soils 3. It is difficult to associate each of the three clioseu groups with distinct properties in regard to their relation to the movements of water (49). 4. Luxmore(2!i) has shown that many properties of soils can be correlated witli the proportion of particles havinfj; a maximum diameter of 0-01 mm. 5. Takinf; into account the insensible degrees by which any given type merges into another, are we justified in laying down limiting per- centages for classes of soils? A soil with a given analysis behaving as a sandy loam in a district of deficient rainfall becomes a loam with a medium rainfall, and a heavy loam with a very high rainfall. The influence of the amounts of organic matter and of calcium carbonate cannot be ignored. 6. Different analysts, working on the same sample of soil frequently — as often as not — obtain discordant results for the clay and fine silt fractions, though they will agree as to the total amount of these two grades present. 7. The rapidity with wliich a large number of samples can be analysed if the determination of the clay is omitted is an important point. 8. The curves for the sunuuat ion percentages are better characterised by the proposed limits. The various limiting diameters of the three classes are shown in Table V. Table V. Classification of particles. Limits of diameters in millimetres . * , Coarse (sandl Medium (silt) Fine (clay) System Maximum Minimum Maximum Minimum Maximum Minimum American 20 005 005 OOOf) 0005 — Wilsdon 10 004 004 0002 0002 — Proposed 30 0-2 0-2 001 001 — [Or possibly — 0-2 0-2 0-02 002 — ] Atterberg (44) has suggested the limits -2, -02 and -002, and there appears to be fairly strong evidence that a geometrical progression is the type of classification required. As English analysts do not recognise the hmit 0-02, the nearest English point (0-01) has been adopted here provisionally. The result obtained wifh the types previously considered is showm in Fig. 4. C. L. Whittles 173 The actual procedure adopted for plotting the results can be seen from the following example : 8-9 Fine gravel u a , .,- „ ,, Coarslsantl 170 , -'■^- ^''^'^'^■ ^r '"""' III I 30-3. Medinm. Fine silt 24-5 1 .-,„ „ „. Clay 4-4 , ~^-^- ^""^ Fine ^OOl linn. Fine Siltl Clay ass\an(l Wheat Wheat and Wheat '''"^^ Ab Sheepand Barley* '^' ,«('oarse Hop Orchard .^tX ■^Vhe; ]5^ •T(d)acc Chenies and I'otato Kiin, Woj.s U.S. Alarkel (iaiden ^ ^ U.S. I'^arl^ aiarket (iardeii CoARSE>0'2uim. I Fine (iravelj [Coarse Sand| 0-2uim. Mkdu'im O'Ol mm. )Fine Sand| tSilt I Fig. 4. Comparison of types by proposed method. In Fig. 5 the side of the equilateral triangle ABC is 100 units, the apices represent 100 per cent, of the respective ingredients. From B along BA cut oft' BP = 25-9 units „ A „ AB „ AQ=30-3 „ „ A „ AC „ AR= 28-9 -„ From P, Q and R draw PA', QY and RZ parallel with BC, AC and AB respectively to form the triangle pqr. Then any point on the line PA' represents 25-9 per cent, of the coarse particles, on QY 30-3 per cent, of the medium particles, and on RZ 28-9 per cent, of the fine particles. 174 Clamficatioi) of Soil x Then tlie centre S of the triangle pqr, obtained by bisecting the base angles is the requirefl point. This construction avoids the necessity for raising the percentages arithmetically so that they total 100. In general they will be less than 100, for the losses on solution and ignition are not included in the amounts plotted. Soils with a high content of organic matter or of calcium carbonate cannot be compared by their position on the triangle alone. By the employment of an arbitrary colour scale, the organic matter content, or the acidity expressed in terms of titration values or of hydrogen ion concentration could be shown simultaneously with its approximate mechanical analysis. Rainfall and other climatic data arc obviously open to a similar method of treatment. In order that the position of a point on the diagram may be rapidly interpreted, a diagram showing the proportions of each of the three ingredients present, by steps of 10 per cent, has been prepared. By super-imposing this on any of the soil diagrams the limits of the groups can rapidly be read off. The amount of variation permissible in a mechanical analysis for survey purposes has been investigated by Robin.son(3;t). The analyses of the two yields, (a) uiiiforin, (h) too variable, are shown in Fig. 5. Atten- tion is drawn to this in order that an idea may be obtained as to the value that is to be assigned to any amount of scatter in a diagram. The general arrangement of the soil types is indicated in Fig. 4. The triangle has been divided up into three main divisions (marked in solid lines) according as one or other of the three ingredients predominates. Each division is subdivided into two classes (by dotted lines) according as one or other of the remaining two constituents is in excess. The inter- relationship between class and class is thus proportional to the length of the dividing line. They may merge into one another or diverge widely. For convenience of reference the classes nuiy be named A^, B^, Bq, Cfj, C^Y and A^,, respectively as shown. Fig. 6 shows a number of wheat soils from almost every geological formation. It will be noticed that they tend to be more or less concen- trated around those selected by Hall and Russell as typical. The lighter soils of Norfolk on which wheat is grown, though the soils are not par- ticularly well adapted to the crop, merge into the typical wheat group, which apparently lies near the boundary of B^. and C,, . The soils in Cg are on the whole more typically grass than arable (12, 16, n, 19,20,29, 31, 36, 40). Barley soils are illustrated in Fig. 7. The crop is grown on soils of the C. L. Whittles c; 175 Fine Silt] (Fine Gravel Coarse Sand \i Fine Sand| Silt I Fig. 0. Wheat soils. 176 Classification of Soils Fig. 7. Barley soils. Coavsor / V' > ^^^. varu'tios. \ / A,., • varieties % . \ • ®j • \ ''^ .. .- \ /' t \ Fig. 8. Hop soils. C. L. Whittles 177 ®TypiCcal Soils, quoted by Hall and Jfiisse • Other typical soils, from vaiions sources Soils on wliich crop is "irow n, tlioii"li soil not really topical oltlii' vv< ©Ditto (quoted by Hall and Hussell) Fi". 9. Orchard soils. Fig. 10. Potato soils. 178 Classificatio)} of Soils three classes Ajj, B;^ and B^., but it is only the ligliter soils of the latter class that can carry sheep. Hop soils are illustrated in Fig. 8, and Orchard soils in Fig. 9. The comparison of these two types made by Hall and iiiissell(i!i) is clearly illustrated. 'J'lie orchard soils in Cjj, apart from the extreme case quoted by Hall and liussell from the Weald Clay, are all grassed orchards, and according to the Bristol Reports (U, 14, 55) suffer from canker to a greater or less extent. Evidently these soils are really too heavy for fruit. In Fig. 10 a number of potato soils are shown. The limiting factor for a potato soil, low content of coarse silt, is not brought out very well by this method (1, lo, 2o, 41). -Market garden soils. The Biggleswade Market Garden soils (35) which are most typical are shown in Fig. 11. Those which have been utilised on account of economic reasons have been omitted. Hall and Russell's Merton Alluvial .soils a])pr().\imate more closely to the i^igglcswade type than does the Wey- bridge, but these are both more of the Market Garden type than the Bagshot Windlesham and the Thanets from Swanley and Greenhythe. C. I.. Whittles 179 Application to the Preparation of Maps. If each, of the tliree primary colours be taken to represent one of the three constituents (Coarse, Medium, and Fine), then to any position ou the triangle is a corresponding definite colour, produced by a combina- tion of the three primary colours in the same proportions. Any desired degree of differentiation may be attained by using a sufficiently large number of combinations. Steps of 5 per cent, would give 231 types, and this would be sufficiently accurate for all purposes, as will be seen from a consideration of Fig. 5. For the preparation of the drift maps of a district, stejjs of 10 per cent, giving 66 types would probably be sufficient. The proportion of stones and gravel, chalk, organic matter, etc., could be shown by dots or shading. Maps of the soil and subsoil prepared on these hnes would be of great value. In conclusion, the writer begs to tender his thanks to all those who have suppUed him with data and other help. A Bibliography on Soil Classification, with data for soil types. (1) AsHBY. .S. F. ( 190.")). A Contribution to the study of Factor.s affecting tlie Quality and C'(jinposition of Potatoes. Jour. Aijric. Sci. 1, 350. (2) Atterbeku, a. (1908). On Metliods of clay analysis. A'. Luiullhr. Ahul. Hniull. ocli Tidsl-r. 47, Nos. 5-0. (3) (1909). The constituents of mineral .soils; the analysis, classitioation and prmcipal properties of clay soils. Comjit. Rend. Conf. Inicnml. Aijrog^ol. [Buda- pest], 1. (4) (1912). Mechanical soil analysis and classification of iSvvedish mineral soils. A'. Landihr. Akad. Hand!, och Tidsh: 51, No. 0; Internat. Mill. Bodciik. 2, No. 4. (5) (1916). The classification of humus-free and huifius-poor mineral soils of Sweden according to their consistencies. Internat. Mitt. Bodenk. 6. (ij) Atteeberg, a. and Joiian.s,son (1916). The classification of the mineral soils of high humus content of Sweden. Internat. Mitt. Bodenk. 6. (7) Baker, H. A. (1920). Investigation of the mechanical constitution of loose arenaceous sediments by tlie method of elutriation. Geological Magazine, 57, Nos. 7-9. (S) BjjArlvkke, K. O. (1901-2). On the Classification of Soils. Berg. Norges. Laiidbr. Hoiskolen Virks, App. (9) Chulnoky, E. von (1909). Soil type as determined by climatic zones. Compt. Rend. Conf. Internat. Agroge'ol. [Bndapcst], 1, l(i3-176. ( 10) Coffey, G. N. ( 1916). The present status and future development of soil classification. Jour. Atner. Soc. Agron. 8, 239. (11) FiPPiN, E. 0. (1911). The practical classification of soils. Proc. Atner. iSoc. Agron. 3, 70-89. (12) Foreman, F. W. (1907). Soils of Cambridgeshire. Jour. Agric. Sei. 2, No. 2, 161. (13) CJiMMiNGHAM, C. T. (1912). Annual Report of the National Frnil and Cider Institute. 180 Classitication of Soils (14) GiMMlNCHAM, ('. T. and Gkovk (1019). Annual Report of llie National Fruit and Cider Institute. (15) Glinka, K. (1914). Dir. Typen dcr BodtnbiUlumj (Berlin). (16) Goodwin, W. The Snih of NoHinyham.ihirc. (17) Haooard, H. R. (1902). Rural England. (Maps uf soils and crops.) (18) Hall, A. D. (1912). The value of soil analyses to the farmer. Jour. Roy. Agric. Hoc. 73. (19) Hall, A. U. and Russell, E. J. (1911). Tin: Agriculture and Soils of Kent, Surrey and Sugscjc. (20) (1911). Soil Surveys and Soil Analyses. Jour. Agric. Sci. 4, No. 2, 182. (21) Hiloard (190(i). Soils. (22) Hope, G. 1). and ('AKrENTKR, P. H. (1915). Siujgcst ions for the inanurial treatment of tea soils. (Indian Tea Association. Calcutta.) (23) Inkev, B. de (1910). Nomenclature and classification of soil types. // Conf. Agrogeol. Internal. Sl(x-kh(>hn. Resume' (l). (24) KossowiTSCH (1912). Die Srhwarzerde. (2">) Kossovacn, P. (1900). The genesis of soils and the principles of a genetic soil classi- fication. Zhur. Opuitn. Agroti. {Russ. Jour. Kxpt. Landw.), 7, No. 4, 478. (20) (1910). The soil forming processes and the main principles of soil classification. Zhur. Opuitn. Agron. (Russ. Jour. Expt. Landw.), 11, No. 5. (27) Lagatu, H. (1903). Vetude des terres el les cartes agronomiques (Montpelier). (28) (190.5). Classification and nomenclature of soils according to mineralogical constitution. Compl. Renrl. Acad. Sci. (Paris), 141, No. 6, 363. (29) LuxMooRE. Soils uf Dorset. (30) MiiRc:o(;c'l. Die Bodenzonen Rumdniens. (31) Newman, L. F. (19121. Soils and Agriculture of Norfolk. Trans. Sarfolk- and Soru-tch Naturali.fts Society, 9, 349. (32) Pendleton, R. L. (1919). Are soils mapped under a given type name by the Bureau of Soils Method closely similar to one another? Univ. Col. Pubs. Agr. Sci. 3, No. 12. (33) Ramann, K. (1920). />V)./.'Hi-»HeT cent, more than that derived from sets of equal size without outgrowths. 3. The proportion of heav}- ware in a crop varied inversely with the size of the tuber-set, and was not materially affected by the existence or otherwise of outgrowths on the set. 4. The tendency to form outgrowths was not conveyed by tuber. The problem was re-investigated in 1921. As in the previous year, the Potato used was the variety Barley Bounty^, the crop of which had been grown in Scotland the previous season. The experimental plot was situated on the writer's farm at Barley, on a piece of ground which sloped gradually from east to west. The western end was shghtly more clayey than the eastern portion, but on the whole the soil was of a uniform type of loam and had been specially chosen on that account. The area of the ground on which the experimental sets were planted was 17-5 feet X 400 = 777 sq. yards. It was deeply cultivated: no stable manure was ' Salaimin, R. N., Journ. of the Ministry of Agric. Vol. xxvni, April 1021. - This is a wart immune variety raisi-il by the author in 1911 wliich is very resistant to roll and mosaic and partially so to Phytophthora. It is not yet on the market. R. N. Salaman 1S3 used, but it was sown with tlie following artiiiciai.s : superphospliate 84 lbs., sulphate of ammonia 2(1 lbs., and kainit 28 lbs. The guard rows were treated in like manner. Eleven classes of seed tubers were selected, boxed and sprouted, on the same date. In each class the seed was weighed, pound by pound, so that there was a check both as to size and individual weight. Thus, ehminating as far as possible, any source of error arising from ine(iuahty of seed within each class, an error which has vitiated some otherwise valuable work in past years. Class A. Tubers weishing 0-6 oz. or 2l> to 1 lb. B. ., „ 1-33 ozs. „ 12 C. „ ., I'-O „ „ 8 D- „ ,. -'-66 „ „ 6 E. „ „ 4-0 „ „ 4 F. ,. ,. .5-33 „ „ 3 U. Mixed unselecteil seed tubers. H. VVlicile tiiljers with outgrowtlis weigliiiig 2 ozs. eaeli. I. Tlie erown ends from tubers with lateral outgrowths w^eighiiig 1-66 ozs. each. J. Cut sets with outgrowths weighing 2 ozs. each. K. Qut sets with outgrowths, weighing 1'2.5 ozs. each. All tuber-sets presented short, strcuig sprouts and no blind sets were planted. Of these classes, B, C, D and G were present in sufficient quantity to plant five rows of 100 tubers each. A and E filled three rows respec- tively of 100 tubers each. F was sufficient for one full row and part of another. H, J and K were present only in sufficient quantities to plant part rows of each. It should be noted that in a standard plot, viz. a row 100 feet long, not only was the number of sets the same, but in respect to any class, the total weight of sets was e.xactlv the same. Th.e experiment was planned on the checker-board system (see Fig. 1 ), but in place of square plots, single rows were used. As the variety was the same throughout, no need was felt for guard rows of some neutral variety between the rows, but surrounding the whole experiment were rows of the variety Golden Wonder, each section containing 100 tubers. On the northern side two rows of Golden Wonder were succeeded by a crop of barley, on the south the two rows were followed by more potatoes. The rows were exactly two and a half feet apart and the tubers were sown by hand, the distance between the centres of every two consecutive tubers being one foot. The spacing between tubers was carefully con- trolled by actual measurement as each tuber was planted. The covering in, first hoeing, and earthing up were done by horse labour. Joum. of Agric. Sci, xn. 13 1air of 2-77, or for four jmirs of l-;58. There is next the difficulty of obtaining comparable mean yields for each variety. The data for B and C are : R. N. Salaman Row B C 1 19-5 21-5, 28-25 2 26 31-5 3 32-5, 39 35-5 4 56 51 187 If we average these results as they stand, C will be given two plots in the poor row 1, and B two plots in the good row 3. The average of B will thereby tend to be made greater than the average of C. This will not do. Where any variety had more than one plot in a row, the figures for these plots were therefore replaced by their mean, and the average calculated from the four resulting figures. In this way we get the com- parable or standardised yields for B, C, D and ( > : Standardised yield, B 34-3 C 35-7 D 41-6 G 40-6 The yields of B, C and D are in the same order as size of set, and the difference (7-3 lbs.) between the greatest and least is over five times the probable error of the dift'erence (1-4) — taking this probable error as based in effect on only four pairs of plots. But there remain A, E and F of the variety used for testing the effect uf size of set. A has plots only in rows ], 3 and 4: hence the averages obtained in the same way as above for A and for the plots in the same rows of B, C and D are not comparable with the standardised figures above; and we want a comparable figure. This was obtained in the following way: on the three pairs of plots available, the total yield of A was 76-5, and of B 111-2. The standardised yield of B is 34-3. An estimated standardised yield of A may therefore be put at 76-5 x 34-3 ^ 111-2 or 23-6. From the similar figures for C, D and G we obtain estimates of 24-5, 24-3 and 24-4. These figures are all fairly close together, and we may take their mean, 24-2, as a fairly close estimate of the standardised yield of A. Standardised estimates for E and F were similarly obtained, viz. E, 47-5. and F, 67-3. These figures increase without a break from the smallest sets A, to the largest sets F, so there can be no doubt about the result. For B, C and D, as already stated, the probable error of a difference is about 1-4 lbs. For A, E and F. owing to the method of estimation, we cannot state a precise probable error, But if the probable error of the difference 1S8 hijlin lire of S('<) 16 4 >j 18 5 j> 20 6 >> 24 7 )? 30 8 5) 36 0 under tliis weight 2 It was found that when this subdivision was made and the individual groups summated, they always came to a figure within a little of lOi lbs., the excess being due to error of the balance in weighing small lots. It is for this reason that the relative proportion of the different classes sliown in Figs. 3 and 4 is calculated to lOj lbs. It should be realised that in each of these classes into which the crop has been subdivided, the tubers are practically of exactly equal weight and size. The actual figures are given in Schedule II. Although in all the sani])les, division into all the 1 7 classes was carried out — as far as the material in each case allowed — it was found that for practical purposes it was better to re-group the findings into four classes: 1. Tubers of 3-33 ozs. and over. 2. ,, over 2 ozs. and under 3-33 ozs. 3. ,, „ I oz. „ 2 ozs. 4. ,, under 1 oz. As regards classes B, C, D and G, the mean value of each weight group in the crop was readily determined, always, however, taking the mean of the determination where two examples of a class occurred in one and the same of the four rows in the experimental plot. In regard to A, E and F. the same method in determining the mean ' Middleton, T. H. Guide. In K.r/ierinnnis conilurlcd nl liiiiyni/iic's Fartn, etc, (amb. Univ. Dept. (if Agriciilliiri', I'.KIT. R. N. Sal AM AN 191 value of each weight group was appHed as was used to determine their standardised mean yields. The results are shown in Fig. 3. Concentrating on the production of heavy ware (Class 1 of the weight series rendered sohd black in the figure) it will be seen that, except for a very shght increase, viz. -5 per cent., of this class in the B series over the A, there is a steady decrease of heavy ware with every increase of weight of the seed. 10-6 lb 10-5' 10- 7 6 5 4 3 2 1 ■5 0 C D Fig 3. A. Sets weighing -6 oz. each. E. Sets weighing 40 ozs. eaeli. B. „ ,, 1-33 ozs. „ F. „ „ 5-33 ozs. „ C. ,, ,, 20 ozs. ,, 0. ,, unselected. J). ,, ,, 2'()6ozs. ,, ^^^^^H Represents the weight of tubers of 3-33oz3. and over in every lOJlb. sample. ^^^^ ,. ,. ,. 2.0OZS. „ wzm. •' " '■ 1°^ „ less than loz. in wt. „ ,, 192 lojinence of Seed Weight, etc. on tin Potnto Crojt Turning to tlie production of the lighter tubers in the yields, two facts emerge. In the series A-D, the quantity of "chats" decreases in direct ratio with the weight of the set, and therefore in the same ratio as that of the " ware" in the same yield. On the other hand, in the E and F tuber-set series the amount of " chats " is much in excess of that found in any of the A-D groups. So that it would appear from these results to be very clearly demonstrated that heavy sets, weighing 1 ozs. and over, not only give greatly reduced ([uantities of useful heavy ware, hut also return in tlieir produce an altogether exce-^sive proportion of useless chats in comparison with the yields rendered by tuber-sets of smaller size. This result is entirely in accord with that obtained in the favourable season of 1920, when the sets weighing 4 ozs. and over produced roughly twice as many chats, and one-third or more less ware, than the sets of lighter weight. A similar analysis of the yields was made of the tuber-sets groups H, I, .1 and K, and the results are shown in Fig. 4, where they are placed in comparison with all the other groups represented in the same portion of the experimental plot. In this series the outstanding fact is the enormous proportion, viz. 4U per cent., of heavy ware produced by the K series. If reference be made to Kig. 1 it will be seen that series K and F were grown in the same line in the same portion of the experi- mental plot, and so may be fairly compared. The F series, however, are tuber-sets of 5-:53 ozs., whilst the K are outgrowths of 1-25 ozs. in weight, and the great disproportion, between 5 per cent, and 40 per cent, of the whole sample, the quantities of heavy ware produced by them res{)ec- tively, is confirmatory evidence of the previous deduction, that the Ughter the weight of the tuber-set, the greater the proportion of heavy ware produced. This relation of size of set to quantity of heavy ware was observed by Sir Thomas Middleton^, but liis results were unknown to the author till after this paper was written. No explanation is advanced either of this relation or of that between the weight of set and the weight of the total crop, but it is permissible to suggest that both phenomena may be related to the le.sser maturity of small tubers. In each of the analyses of crops made, — and many were duplicated — record was kept of the presence of secondary outgrowths on the tubers. Doubtless owing to the fact that this particular i)iece of land never ' Midilleton, T. H., "Potato Ex]K'rinu'ntH at l{iiii;o\ lie's Farm, Impingtoii, Ciiinbs.," The Pdldlo Year liuok, 1907. R. N. 8ALAMAN 1 90 received any of the late rains enjoyed by other parts, there were but very few, and they of very small size. It was found that their presence was not correlated with any feature such as size of tuber, nor were they any more frequent in the group H, I, J or K than in those of the series A-F. As in 1920, no evidence is forthcoming that such growths are conveyed by tubers from one generation to another. I A B C D F G H I Fi.L'. 4. A — G. As liefoie. H. Whole tubers with outgrowths, 2 ozs. eacli. I. Crown ends from tubers with outgrnwtlis. l-finzs. each. J. Cut sets witli outgrowths, 2 ozs. each. K. Cut sets with outgrowths. 1-25 ozs. eaeli. H^HjjjJI Represents the weight of tubers of 3-33 ozs. and over in e\erv lOi H>. sample „ ,, ,, less than loz. in wt. ,, ,, 194 Infuence of Seed WelgJtt, etc. an the Potato Crop The results attained in 1921, which so closely bear out those su{i;gested by the experiment of 1920, may be briefly summarised: 1. The total yield varies directly with the wcijjht of the tuber-set. 2. That small sets under 1 oz. in weight, although giving a great return in proportion to tlieir weight, and a high proportion of heavy ware, are unccononiical. 3. That taking into consideration the total weight of seed used, the proportion of heavy ware produced and the total yield, sets of 2 ozs. in weight are the most remunerative. 4. Cut sets consisting of secondary outgrowths weighing 2 ozs., and whole sets with similar outgrowths of the same weight to a lesser extent, produce considerably heavier crops than any other type of set, and at the same time produce a high quantity of heavy ware. 5. There is an inverse ratio between the size of the seed set and the percentage of heavy ware in the resulting crop. 6. The productivity of secondary growth, as well as 'the high pro- portion of heavy ware, yielded by small tuber-sets, may be correlated with immaturity of the seed tuber. 7. There is no correlation between the presence of secondary growth in the seed set and the existence of the same in the resultant crop. Schedule I. Showing the weights of crops in poiinrls of each seed class in the experi»wntal area, including the (luard Rows. Guard lOWH Guard rows 4 3 11-5 A B C D K G C 2 1 9-.'5 12-5 19-5 21-5 26-5 35-5 30-5 28-25 10-5 10-5 Ifio D E F C B D G 10 32-25 37 49-25 31-5 20 39 35-25 10 20 17-4 B C A G D E B 39 21 11(1 .32-5 35-5 27 39-25 48-25 48-25 10-5 C G H' D B G A^ 10 24 r>i 50-5 7-5 P 8 J» 42 50 50 58 18-5 27 Fs 10-5 K« 20 24-5 1.^ seta plan 20 ted. ' A. 6 F, 50 sets planted. 10 ' J, 05 » •K, ,40 *i 1{. N. Salaman 195 Schedule II. Analysis of crops of each of the seed classes as they occur in the experimental area, showing the amount in pounds in each weight grade excepting the 'chats^ under 1 oz. out of sample weighing l()-5 lbs. Tubers weighing over 3-3 ozs. A B C D E (J C 0 0 0 0 (1 1) 0 D E K c H U G 0 0 0 •5 0 0 0 B C A G i) E B ■i-2 2 1-25 •5 ■5 1 H D B G A 2-5 1 Ij 2-33 ;i-5 2 1 1-3 F ■5 J 1 K 4 Tubers weighing over 2 ozs A B C D E G C 2 1 2-5 1-5 1-5 1-5 2-5 D £ F C B D G 2-6 1 1-5 1-75 1-5 3 2-37 15 C A G D E B 4 3-5 4-5 3-25 3-75 2 3 C G H D B G A G GO 3!l I 3-3 J 4-5 61 5-83 7 4 F 4 K 6 Tubers weighing over 1 oz A B C D E G C 4-5 5-5 G 5-5 5 5-5 7 D E F C B D G 7-6 5-5 G G-5 6 8 6-75 B C A G D E B 7-75 7-5 7 8-25 7-75 5-5 7-5 C G H D B G A 9-5 91 7-3 I 8 J 8 9 9 10 8 F 7 K 9 i(() hitiucnce of Seed Weight, etc. on tin Potato Crop DESCRIPTION OF FIGURES. Fig. 1. ^5lio\v.s the plots as actually laid out. In the lowest row it should bo understood that Lots H, I and J formed one row, and A, F and K another. The blackened portion represents the yield j)er set in pounds, in each plot. fig. 2. Represents the standardised 3fields per set in each of the classes of seed weight A-G. Fig. 3. The crops from A-G are analysed as to the weight groui)s into which their con- stituent tubers fall. The figure represents the standardised mean of the various groups in each of the classes. Fig. 4. The analysis of croj)s of all -seed classes in the fourth or western .section of the experimental plot showing the relation of tlie special classes H, I, J and K to the normal. The writer has great pleasure iu acknowledjjiiifi his deep obligation to Mr Udny Yule, C.B.E., M.A., F.R.S., who most kindly worked out and elucidated the statistical data iu this paper. (Received Februur// llth, 1922.) AN INVESTIGATION UPON CERTAIN METRICAL ATTRIBUTES OF WHEAT PLANTS. By F. L. ENGLEDOW, M.A. PIkh/ Breeding liiiilituk', Sc/khiI of Agrioiltiire, Caiiihr'uhjc. AND J. P. SHELTON, Farrer Memorial Scholar, Sydiw//. ( 'ONTENTS. PAGE § T. Introdiu'tiun ........... 197 §11. Material and iMethod 199 § 111. The Ghime-Longtli : Raeliis-Leiigtii Ratio 200 § I\'. Concerning the Inter-relatiunshij) of the Tillers of a I'lant in regard to certain Measurable Characters ...... 202 § V. The Relation of Weight of Seed Sown to the Resulting Phint . . 203 Conclusions ........... 204 Bibliography 205 § I. Introduction. In dealing with the Inheritance of Ghmie-Leugth (Engledowii)] there were encountered certain problems relating to metrical characters. As it appeared that these problems must attach to all genetic work upon a metrical basis, they were made the subject of a separate inve.stigation. A simple account of some of the difficulties experienced in connection with Glume-Length Inheritance will serve to formulate the problems. It is necessary first to set forth the reasons which lefl to a genetic investigation upon glume-length. To the glume itself no intrinsic interest attached. The prime motive was to forge some sharper weapon than eye-judgment for the separation of "genetic types" mF^fi etc., and for more critical study of segregation. Rigid measurement, and classification solely by measurement, might, it was felt, provide such a weapon. Alike to eye-judgment and to measurement, "fluctuation" was certain to be an obstacle; but both observation and inference united to suggest that the length of the glume of the wheat plant was less liable to fluctuation than were most of its other observable attributes. There was the addi- tional advantage that glume-length could be measured accurately and 198 Metrical Attributes of Wheat Plants with facility. These two reasons, then, led to its adoption. Overlap of distributions is the form in which the difficulty of fluctuation makes itself felt. To counter overlap, two wndely differing parent form.< were selected, viz. the T. durum known as Kubanka (mean glume-length in England£il0-5 mm.), and the distinctive T. polonicum or Polish Wheat (mean glume-length in England £i 30-5 mm.). Greatly as these forms diverged in glume-length, their distributions nevertheless showed a slight overlap. In the F.,, the parental types reappeared in a "shifted" form. With them, and intermediate between them, came a heterozygote. Consequently the glume-length distribution of the whole F, took a trimodal form which, defying analysis on the basis of rigid measurement, enforced a reversion to eye-judgment. It was quite obvious that the wide "'fluctuation'" in glume-length was due to the occurrence in the parental and segregate populations of a certain number of very poorly grown plants. To reject such plants would have been disastrous and to include them was to confuse distributions and inhibit analysis. Plants of this kind were poor in every _way and, roughly speaking, the less the stature of a plant, the shorter were its leaves, its ears, its glumes, etc. This very patent fact suggested that instead of making, for every plant, an absolute measurement, there should be made some form of "com- pensated" measurement. In short, every plant should be "handicapped." "Length-of-rachis"' was selected as the basis of the "handicap." The working hj'pothesis was that a big, thri^^ng. plant had long ears (i.e. great rachis length) on which were borne proportionately long glumes. Per contra, small plants would be small "all round" and it seemed not unlikely that the ratio — length of glume : length of rachis —would exhibit a smaller plant-to-plant fluctuation than would absolute glume- length. To test the constancy of this ratio, then, became the first object of investigation. A phenomenon designated by the term "shift" was observed in the F2 of the Pohsh x Kubanka cross. It consisted in the appearance in F^ of two groups of plants, each in number about a quarter of the whole F2 population, and having respectively a complete eye-resemblance to the parental (F„) forms. In mean glume-length, however, the F^ Kubanka type .slightly exceeded the F^, while the F^ Polish type was 25 per cent, shorter than the Fq. Explanations of "shift" could be based upon "modification by cros.sing" [cf. Ruggles-Gates(3)] or upon "minor nmltiphnng factors," but yet another explanation seemed possible. The embryos and endosperms from which grew the Fo Polish-type plants, were nourished by F^ (i.e. heterozygous and intermediate) plants. F. L. EnCtLkdow and J. P. Shelton 199 Upbringing by such a ""foster mother'' might have some predetermining influence and might be the cause of "shift" [for a fuller consideration see Engledowd). pp. 127-8]. This explanation is based upon a belief that there is a fairly close relationship between the weight, composition, etc. of the mother seed, and the attributes (glume-length included) of the resulting plant. To test this relationship liecame a second object of investigation. Many morphological and economic attributes of plants may be metrically represented in a number of different ways. Cereal plants illustrate this point. As generally grown a jilaut has a main axis and a number of axillary shoots or tillers. To obtain an expression of the glume-length, rachis-length, ratio of grain to straw etc. for a population of any variety, it is possible to confine observation to one ear-bearing stalk per plant. The largest, the first formed, or a random ear may be chosen: or every ear of the plant may be included and the observations for the whole plant be averaged. Labour is minimised if only one ear per plant be observed but the available number of observations is increased by the inclusion of every ear. Differing sets of circumstances have led sometimes to the one practice and sometimes to the other and it is clear that the justification of each in its own circumstances, mu.st re.st upon the relationships which prevail among the tillers of the indi- vidual plant. These relationships formed the third part of the enquiry. § II. Material and Method. The strains of Polish and Kubanka were the ones used in the glume- length investigation. They are both old and carefully kept pure lines. The seed for each variety was obtained from forty plants of the 1919 harvest. The main ear of every one of these plants was measured for rachis-length and glume-length and all of the grains were weighed and separately labelled. Sowing (in 1920) was completed in one day and both germination and growth were good and of as great uniformity as is usually attainable under the conditions of careful experiment. At harvest (1920), every ear of every plant was separately collected, and later on its rachis-length and glume-length were determined. Previously it had been usual to measure only one glume per ear [see Engledow(i), pp. 111-2], but for this investigation both glumes of the "median" spikelet of each side of the ear {i.e. four glumes per ear) were measured. The average of these four measurements is, hereinafter, referred to as "glume-length." Journ. of Agric. Soi. xii 1'^ 200 Metrical Anribates of ^VlKat PlauLs § 111. The Glume Length : Rachis-Length Ratio. As a pifliininary to further work, the correlation between glume- length and rachis-h'ngth was evahiated from data obtained from the iyi4 crop of the two pure lines. There were 15U plants of each pure hne and the coefhcients of correlation (r) were: Polish = + 0-295 ± 0-050, Kubanka = + 0-469 ± 0-043. It seemed not improbable that the lowness of the correlations was due to the fact that the experimental plants were not alike in respect of the total number of ears (or tillers) produced per plant.~ Consequently tiie 1920 plants — the ones observed in the main investigation — were classified as "one-ear." "two-ear," etc. plants and the correlation between glume-length and rachis-length was evaluated for the separate classes. In the case of every class, the main ear only of every plant was dealt with in the correlations. Table I contains the correlation coefficients (r): Table I*. Coefficients of Correlation between Glume-Length at\d Rachis- Lenglh for one-car, ttco-ear, and three-ear Polish and Kubanka jdant^. (Only the main ear of the plant was observed.) No. of eai-s No. of Variety pel • plant j)lants r Polisli 1 310 0-918 ±0006 r.ilisli o HO 0-768 ±0020 I'.ili.sli ;s ;{.s 0-8.52 JO-OSO KubanUa 1 1'0.5 0-7,51 ±0-020 Kniianka 2 107 0-717:,. 0-032 Kllliauka :! o5 0-494 ±0-069 * All tho cori'olatioiis in llic labk- are jjusitive. The unit of nieasuroMiont throughout' is 1 mm. The values of the correlations are high, botli absolutely and in relation to their own probable errors but even a correlation of unity lietween two variables does not imply a constant ratio between them unless tlie regression lines pass througii the origin (0 . 0). This fact was well illustrated by the results obtained when the ratio glume- length/rachis-length was evaluated for individual ])lants. The ratio iluctuated as wildly as did the glume-length and to show this it is necessary to give no more than the coefficients of variation (viz. V = 100a/ 3/, where a = standard deviation and M = mean). These are given in Table II. F. L. Engledow and J. P. Shelton 201 Table 11. Coefficients of Variation for 01 Nine-Length, Baehis-Lengtli and the Ratio of these two quantities in Polish and Kiibanhi ivheats. CiiefRcicnt. of vaiiation of Variety iiiid No. of ears , — no. of plants ])er plant (ihinie-lengtli Rachis-length Ratio Poliali (:)1<;) 1 ]7-13±0-47 25-77±0-74 l.'-)-47 -l:0-4:? Polish (110) 2 S-5()±0-39 13-8.3±0-l!4 10-S:?:i 04!1 Kuhanka (20")) 1 10-I8±0-34 20-43±0-71 18-L'niO(>3 Kiihanka (107) 2 8-40 |:0-.3n Id-SSi-O-SO 12-.58-J:0-.^lt It is therefore to he concluded that this particular "compensated" ratio possesses no value either to the geneticist for a critical study of variation or to the plant breeder for the discrimination of closely re- sembling agricultural strains. Despite this record of failure, it is felt that "compensated" observa- tions {i.e. some form of ratio) ofi'er still the only alternative to "abso- lutely (unattainably?) uniform conditions of growth" as a means of removing the masking effects of "fluctuation."" It is not to be expected that the random choice of two characters — e.g. leaf-width and grain- length — will serve as a basis for compensation or handica]) and ])rovide a con.stant ratio. Manifestly, in seeking constancy, one should endeavour to find two "lengths" (or other attributes) which are determined by the same causes and during the same time-period. In glume- and rachis-lengths, it would seem, these requirements are as likely to meet fulfilment as in any pair of attributes of the wheat plant to which one could point. This granted, there follows the regrettable but not surprising conclusion that "lengths'" are such vague expressions of the real "nature'" of a plant variety, of its physiological activities, as to be of little value in attempts to determine accurately its modes of inheritance. The employment of "ratios"' by genetici.sts has usually been dictated not so much by a desire to achieve "compensation"" as by the necessity of devising representations for elusive attributes like leaf-shape, etc. The work of Martin-Leake (S) upon Cotton, of Balls (6) upon the same plant, and of Groth(7) upon the Tomato afford examples; but in all these cases fluctuation has played its customary havoc. In outline, the "handicap"" principle for plants is strictly analogous to that followed in flat-racing. Horses are handicapped so that they may afford an exciting "bunch" at the winning-post. It is required to "bunch"' metrical observations upon a plant population {i.e. to minimise fluctuation) but it seems probable that a really successful "handicap"' will have to be a complex one. It will, in fact, have to be applied not 202 Metrical Attributes of Wheat Plants simply to tin- "eiul-iJiotluct " (tlit- inature plant) but to the more iiu- portant life-stages by whicli tlie end-product is determined. OvWl j.0().">0 0-66.5:;; 0036 0-5l2±0-047 2 I'olish (38) Ears 1 and :! (1-42!) j;()0S9 0-439±0-080 0-428±0-089 3 Kubanka (107) Ears 1 and 2 0-,Wni0042 0-li.">!ti0037 0-699±0-033 2 The first inference from these iigures seems to be that, when dealing with the glume or with any similar ear-character, it is essential to confine the observations to one ear per plant. In that, as a very general rule, the ear of the main stalk is formed first and attains the greatest growth, it is the best ear of the plant for observation. This inference from the facts of inter-tiller correlation naturally leads one to ask whether, in furtherance, it would be well to limit the experimental population to plants which all possess the same total number of tillers or which all ripen the same number of ears. Upon this point further evidence is available and it is set forth in Table IV. The facts are quite clear and Table IV. Mean Values for the main tillers of plants grouped according to the number of ears -per plant. Mean value (main tillers only) of • " ^ -^ Olnme-length Rachis-length Ratio 26-.->!017 !)2-I±0-90 0-298±0-002 30-0±017 118-4±l-0.i 0-262±0-002 I0-9±00.5 58-0±0-50 0-196i;0-002 ll-7-i;000 68-5 + 0-7.5 0-17o±0-001 rather striking. Although only the main tiller was observed in the case of every plant, the one-ear and two-ear plants give very different results. Variety and No. of ears no. of plants per ))lan( Polish (316) 1 Polish ( 1 10) >> Kubanka (205) 1 Kubanka (107) '2 F. L. Englbdow and J. P. Shelton 203 Both for Polish, aud Kubauka, ghiiiie- and rachis-leugths are greater in the case of two-ear than in the case of one-ear plants; and since the dif^'erence is more marked for rachis than for glume, the "ratio" is lower for the two-ear plants. A difference is, perhaps, to be anticipated for, within limits, the more vigorous the plant the more tillers it produces: and the facts displayed in Table IV make it appear that greater vigour is evinced in "all-round"' form — not only more tillers but bigger ones (larger glumes and rachis). This "co-fluctuation" of the attributes rachis-length. glume- length, and number of tillers, encourages one to think that in such attributes is to be found a means of estimating degree of growth and of comparing and contrasting different pure lines. There may still be room for hope but no progress is possible until "fluctuation" has been dealt with and, as far as this case goes, the mathematical handling of attribute measurements (use of ratios, etc.) has been valueless. It has been inferred that observation should be confined to the main ear of the plant and this inference, if valid, is not without interest from the point of view of yield-investigations. When, in seeking higher- yielding forms, an F2, is raised from two parental strains, it is almost essential to cast out what are believed to be the "inferior" segregates. Even if there be no casting of Fj plants, the process must be applied to the resulting F^ families for otherwise available time and ground- space become inadequate. In actual practice " casting" by eye-judgment has attained some very conspicuous successes but there is now a tendency to try to substitute an accurate "method" in its place. If those attri- butes which are considered to govern yield behave in a way analogous to that described for glume-length, etc., then the best tiller and not the whole plant should be the basis of estimation. In the field, however, small tillers as well as large go to make a crop, and whatever form of judgment of F2, plants is evolved, it must of course, in some way, pay due regard to tillering power. § V. The Relation of Weight of Seed Sown TO THE Resulting Plant. It has been explained in the introduction that there existed a special reason for attempting to determine the influence upon the plant of the weight and composition of the seed from which it grew. Apart from this, in a great mass of literature [for a very full summary see Kidd and West (8)], there is evidence which justifies the suppo.sition that much of the troublesome "fluctuation" is induced by lack of uniformity of seed. 204 Metrical Affn'hntes of Wliiat Plants As a contributory test of this point, correlations were evaluated between weight of mother-seed and glume-length, rachis-length, and their ratio in the resulting plants. Preliminary investigation having shown an absence of conolatioii for a population of jiinnts having different numbers of ears, attention was confined to single-car plants. Talilc \' contains the results. Table \ . Correlation between Weight of Mother Seed and the Characters of the Resulting Plant. (Only one-ear plants are included.) Varii'tj' and Correlation between weight of mother seed and no. of plants Glume-length Rachis-length Ratio Polish (.316) 0-O(55±OO38 0043-0038 OOOOi — Kuhanka (20.5) 0-262i0-a44 0-1.57±0-04li 0003±0-047 The supposed relation between weight of mother-seed and glume- length, etc., is thus emphatically negatived. Mere weight of seed is not — save at the lower extreme — likely to be of great physiological importance and the belief in its importance (vide literature above mentioned) is, perhaps, largely due to the form, of ex])eriment often adopted — the removal of a portion of the endosperm, work with non-pure lines, fewness of observations and so on. No embryo during germination uses the whole of the reserve food with which it has been provided and reserves beyond a certain amount must be simply "surplus." That "quality" of endosperm may be important still remains a possibility and evidence exists to this effect. This possibility is the only remaining basis for the explanation of "shift" of which an account is given in the introduction. With the general principle that good, sound, seed must be sown to reap a good crop, there is common agreement, but beyond this, even in genetic work and in face of the danger of fluctuation, there seems no need to go. The weighing of a vast number of seeds in order that seeds all of the same weight may be .sown, seems, from the facts above recorded, not to be worth while. Conclusions. The conclusions which follow hold in strictness — failing further test — only for the year, the locality, and the wheat varieties concerned in the investigation. Since, however, the residts iwe fairly emphatic, they seem likely to prove applicable in principle to other circumstances. (i) Glume-length and rachis-length in both Polish and Kubanka "Wheats are very highly correlated. F. L. Engledow and J. P. Siielton 'JOS (ii) Nevertheless, the ratio of these two quantities has about as big a coefficient of variation as the absolute glume-length. (iii) Therefore, despite contrary expectations, the ratio appears to be of no special value in investigation. (iv) Among the tillers of any one plant correlations e.xist for glume- length, rachis-length, and ratio. Their general value is about + 0-5 and consequently when dealing with attributes of this kind observation should be confined to the main stalk of every plant. (v) It is desirable further, to restrict the experimental population to ])lants all of which produce the same number of tillers. (vi) Weight of mother-seed, for a reasonably good seed sample, seems not to determine in any observable degree the growth of the resulting plant as judged in general by glume-length, rachis-length, and the ratio of these two. (vii) The great labour of picking out for sowing a sample of seeds all of one weight in order to reduce "fluctuation"' among the resulting plants, seems not likely to be repaid. BIBLIOGRAPHY. (1) Engledow, F. L. (1920). The lulieritance of Glume-Length and Grain-Length in a Wheat Cross. Journ. Genelics, 10, No. 2. (2) Haklan, H. V. (1920). Daily Development of Kernels of Hanuchen Barley from Flowering to Maturity at Aberdeen, Idaho. Journ. Agric. Research, 19, No. 9. (3) Gates, R. Ruggles (1915). On the ModiKcation of Characters by Crossing. Am. Nat. 49. (4) Engledow, F. L. (1920). Inheritance in Barley, No. I. Joiirii. Ornelirs, 10, No. 2, ajid No. II. Journ. Agric. Sci. 11, 1921. (5) JIabtin Leake, H. Studies in Indian Cotton. Journ. Griirlics. 1. (0) Balls, W. L. (1909). Studies of Egyptian Cotton. (7) CJroth. Bulletins of the New Jersey Experiment Station, Nos. 228, 238, 239, 242, 278. (8) KiDD, F. and West, C. (1919). Physiological Predetermination; the Influence of the Physiological Condition of the Seed upon the Course of the Subsequent Growth and upon the Yield. Ann. App. Biol. 5, Nos. 3 and 4. {Received 1st June, 1921.) NOW READY Report of the General Discussion on PHYSICO-CHEMICAL PROBLEMS RELATING TO THE SOIL held by the Faraday Society in London. May 1921 (containing additional matter not presented at the meeting) Price 10/6 net in limp cloth, 156 pages Orders should be sent to THE FARADAY SOCIETY, 10, Essex Street, London, W.C. 2 BASIC SLAGS AND ROCK PHOSPHATES By GEORGE SCOTT ROBERTSON, D.Sc. (Dunelm), F.I.C, With a Preface by EDWARD J. RUSSELL, D.Sc. (Lond.), F.R.S. With 8 Plates and a Map. Small Royal 8vo. 14s net. Cambridge Agricultural Monographs The main puspose 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 1915-1920. Dr E. J. Russell says in his preface: " 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. His monograph contains a store of information about the new slags, and is a model of thorough and systematic investigation. It deserves close study by all who are interested." Cambridge University Press, Fetter Lane, London, E.C. 4: C. F. Clay, Manager THE UNIVERSITY OF CHICAGO PRESS Agricultural Economics By EDWIN G. NOURSE. A volume comprising several hundred carefully edited selections with valuable original introductions, which make of the material a connected whole. This book follows the plan, successful in so many recent cases, of combining in one volume the virtues of the usual text and the collateral book of readings. Particularly strong is the emphasis upon the interrelations of the technique and the economics of agriculture. Most economists have made little use of the data of agriculture in developing their exposition of economic principles. Chapter III of this book is a notable example of the attempt to found our agricultural economics upon a sound under- standing of the technical nature of the productive processes of agriculture. Many selections also show how changing economic or commercial con- ditions are modifying the practices of agriculture, and how at the same time technical considerations, such as the exhaustion of nitrogen from the soil or the growth of knowledge about plant or animal breeding, introduce new factors into the economic problem of agriculture, xxvi + 896 pages, 8vo, cloth ; 24s net, COMPANION VOLUME Outlines of Agricultural Economics 106 pages, paper; 4s 6d net. The Cambridge University Press Agents for the British Empire (except Canada) London, Fetter Lane, E.C. 4 CONTENTS. (All Eights reserved.) FAQB 1. Murray, Alan J. The chemical composition of animal bodies, (With 1 Text-figure.) 103 2. Taylor, William and Husband, Alfred D. The effect on the percentage composition of the milk of (a) variations in the daily volume and (6) variations in the nature of the diet. (With 4 Text-figures.) Ill 3. Tocher, J. F. The citric solubility of mineral phosphates. (With 8 Diagrams.) .......... 125 4. Woodman, Herbert Ernest. Comparative determinations of the digestibility and metabolisable energy of green oate and tares, oat and tare hay and oat and tare silage 144 5. Whittles, C. L. A note on the classification of soils on the basis of mechanical analyses. (With 11 Text-figures.) . .166 6. Salaman, Redcliffe N. The influence of size and character of seed on the yield of potatoes. (With 4 Text-figures.) . . 182 7. Engledow, F. L. and Shelton, J. P. An investigation upon certain metrical attributes of wheat plants ...... 197 The Journal of Agricultural Science is issued in quarterly parte of about 100 pages, with plates and figures, four parts constituting a volume. Volumes I — XI (1905 — 21) now ready. In foiu' parts, paper covers, 30j. net per volume. Single numbei's 10s. net each. Quotations can be given for Buckram binding cases and for binding subscribers' sets; also for bound copies of back volumes. Papers for publication may be sent to Professor T. B. Wood, C.B.E., M.A., F.I.C., F.RS., School of Agriculture, Cambridge, br to the associate Editors. Other communications should be addressed to the University Press, Cambridge. Contributors will receive twenty-five copies of their papers free. A limited number of additional copies may be purchased if ordered when final proof is returned. In order to meet the increased cost of printing, contributors will be asked to pay a proportion of the cost of setting up tabular matter and of preparing blocks where these ezceed a certain proportion of the whole commuuicatidn. The subscription price, payable in advance, commencing with Vol. X is 30». net per volume (post free); single numbers \0i. net. Subscriptions may be sent to any Bookseller, or to Mr C. F. Clay, Manager, Cambridge University Press, Fetter Lane, London, E.G. 4. The Cambridge University Press has appointed the University of Chicago Press agents for the sale of The Journal of Agricultural Sdence in the United States of America and Canada and has authorised them to charge the following prices : — Subscription, f 6.00 net per volume ; single copies, $2.00 net. PBJNTED IN ENGLAND AT IHG CAMBRIDGE UNIVERSITY PRESS Vol. XII. Parts July, 1922 THE JOURNAL OF AGRICULTURAL SCIENCE EDITED FOR THE PLANT BREEDING AND ANIMAL NUTRITION RESEARCH INSTITUTES AT CAMBRIDGE, AND THE ROTHAMSTED RESEARCH INSTITUTES BY Professor R. H, BIFFEN, M.A., F.R.S., Cambridge Sir a. D. HALL, K.C.B., M.A., F.R.S., Ministry of Agriculture, London F. H. A. MARSHALL, Sc.D., F.R.S., Cambridge E. J. RUSSELL, D.Sc, F.R.S., Rothamsted Experimental Station, Harpenden Professor T. B. WOOD, C.B.E., M.A., F.LC, F.R.S., Cambridge IN CONSULTATION WITH B. C. ASTON, Department of Agriculture, WeOington, New Zealand Dr C. a. barber, C.I.E., School of Agriculture, Cambridge Professor B. T. P. BARKER, M.A., National Fruit and Cider Institute, Long Ashton, Bristol W. BATESON, M.A., F.R.S., John Innes Horticultural Institute, Merton, Surrey J. R. CAMPBELL, B.Sc, Department of Agriculture, Dublin I. B. POLE EVANS, Department of Agriculture, Pretoria, South Africa F. B. GUTHRIE, Department of Agriculture, Sydney, N.S.W. Professor J. HENDRICK, B.Sc, Marischal College, Aberdeen Sir T. H. MIDDLETON, K.B.E., C.B., M.A., The Development Commission, London Dr FRANK T. SHUTT, F.LC, Experimental Farms, Ottawa, Canada Professor W. SOMERVILLE, M.A., D.Sc, School of Rural Economy, Oxford Dr a. C. true. Department of Agriculture, Washington, D.C., U.S.A. Sir FRANCIS WATTS, K.C.M.G., Barbados Dr H. J. WHEELER, American Agricultural Chemical Co., Boston, Mass., U.S.A. Cambridge University Press C. F. Clay, Manager London : Fetter Lane, E.C. 4 also H. K. Lewis & Co., Ltd., 136, Cower St., London, W. C. i and Wheldon and Wesley, Ltd., 2-4, Arthur St., New Oxford St., London, W.C, 2 Chicago : The University of Chicago Press (Agents for the United States and Canada) Bombay, Calcutta, Madras: Macmillan and Co., Ltd. Tokyo : The Maruzen-'Kabusbiki-Kaisha 'Price Ten Shillings net Sls&ued y November 1922.] naDDDDnaaaxonDaanaaaDQuaaLDDaDcxriDDoannDcxiaxixxx: '^**P*ttt4*****4^****-^***^4^t*¥*^**^9*^ft.^^*4-t'^*-¥********^**********»* 94 i **************** t^*tt**t**»******^*«**4*»4*9»*t**\j & O o :n ;n n n n : D n Cambridge University Press British Grasses and their Employment in Agriculture. : By S. F. Armstrong, B.A., F.L.S. With 175 illustrations. Demy 8vo. 10s. 6d. net. Manuring for Higher Crop Production. By E. j. Russell, D.Sc, F.R.S. Second edition, revised and extended. Demy 8vo. With 17 text-figures, ^s. 6d. net. The Fertility of the Soil. By E. J. Russell, D.Sc, F.R.S. With 9 illustrations. i6mo. Cloth, 3J. net; lambskin, 45. net. Cambridge Manuals Series. A Course of Practical Chemistry for Agricultural a\ Students. By H. A. D. Neville, M.A., F.I.C., and L. F. Newman, D M.A., F.I.C. Demy 8vo. Vol. I, los. 6d. net; Vol. II, Part I, 5$. net. 3' A Course of Practical Work in Agricultural Chem- istry for Senior Students. By T. B. Wood, M.A., F.R.S., n Drapers Professor of Agriculture in the University of Cambridge. Demy §: Svo. Paper covers. 2s. net. Cattle and The Future of Beef-Production in England. tjl By K. J. J. Mackenzie, M.A., with a preface and chapter by F. H. A. p Marshall, Sc.D., F.R.S. Demy Svo. ys. bd. net. Course of Practical Physiology for Agricultural R Students. By J. Hammond, M.A., and E. T. Halnan, M.A. Demy Q ; Svo. With blank pages interleaved. Cloth, 6^. bd. net; paper covers, 0 ^. bd. net. n Physiology of Farm Animals. By t. b. Wood, c.b.e., M.A., H| F.R.S., and F. H. A. Marshall, Sc.D., F.R.S. Part I, General. By F. H. A. Marshall. Demy Svo. ids. net. Microscopic Analysis of Cattle-Foods. By T. N. Morris, B.A. With 54 text-figures. Crown Svo. 35. 7iet. Insect Pests and Fungus Diseases of Fruit and Hops. L A Complete Manual for Growers. By P. J. Fryer, F.I.C.,F.C.S. Crown \^ Svo. With 24 plates in natural colours and 305 original photographs and " diagrams. 45^. net. Ri Agriculture in the Tropics. An elementary treatise. By J. C. n Willis, M.A., Sc.D., F.R.S. Second edition, revised. Demy Svo. With p}:; 31 plates. i2s. 6d. net. Cambridge Biological Series. Rubber and Rubber Planting. By r. h. Lock, Sc.D. Crown jtj : Svo. With 10 plates and 22 figures. 6^. bd. net. CAMBRIDGE UNIVERSITY PRESS I n : n; Ri FETTER LANE, LONDON, E.C.4: C. F. CLAY, MANAGER |[^ : n::^^^^ .,,,.m.i.. , , , , in •<^^ff**tt**4 ******** >»*»**-*t**«* ttt4*44^4tHtt»*t*** ♦■♦^■♦♦♦♦♦•♦♦X* »t*«t««»*««»»«4« *»*»**«♦♦»»»»♦«<»♦•<»♦»♦♦<♦# ♦♦t**^** ♦-»♦♦*■»♦ ♦■»<»^ Volume XII JULY, 1922 Part III tJOTANiCAR, SOME INVESTIGATIONS ON THE ELECTRICAL METHOD OF SOIL MOISTURE DETERMINATION. By THOMAS DEIGHTON, M.A., B.Sc. (School of Agriculture, Cambridge.) (With Six Text-Figures.) The investigations described in the following pages had their origin in a desire to find some method of moisture determination in the soil which would not require the taking of samples, with a view to some further experiments on the cultivation of the soil in root-crop pro- duction. Tentative work on these hues was commenced in America in 1887 by Professor Milton Whitney (i) who employed an earth battery con- sisting of alternate copper and zinc plates buried in the soil, connected through a galvanometer. Later in Bull. No. 6, "Division of Soils," U.S. Dept. Agriculture the same author further develops this method of moisture determination. In 1897 M. Whitney and T. H. Means (2) published an accoimt of some experiments in which they determined the specific resistance of various soils, making a correction for packing error, and showed that this varied considerably for different soils. The authors plotted the resistance against the moisture in various soils and found that the curves obtained "agree in the main with an hyperbolic curve except that they are more or less rotated," and they discovered that "it requires from 1 to 3 per cent, more water to produce the same change in the resistance of some soils than in others." To this paper is appended a table for the reduction of soil resistances to a standard temperature. In the experi- ;ments a compensating temperature cell was used. C2 The following year F. D. Gardner (3) dealt with the use of under- ground cables on ploughed land and gave results of an extended series ^ of measurements of soil moisture by the electrical method compared c.Jwith determinations by drying on the same plots. The results, though J.-J considered satisfactory by the author, show a mean difference of about 2 per cent, in the moisture content between the two methods of working. The maximum difference was as high as 4-3 per cent. Journ. of Agrio. Soi. xil 15 208 The Electrical Method of Soil Moistnre Determination In 1899 L. J. Briggs{4) described improved instruments and electrodes and introduced a condenser in parallel witli that arm of the bridge adjacent to the soil resistance, which would throw the small capacity found in the soil and the condenser capacity on opposite sides of the bridge with respect to the telephone receiver. I have found a reference to some work by R. 0. E. Davis (Trans. Arner. Electrochem. Soc. 17 (1910), 391-403), in which he is said to have found the resistance of soils within the limits 10-20 per cent, of moisture to be inversely as the moisture content. Unfortunatel)- the original of this paper is not to be obtained here, but the result thus stated is at variance with Whitney's results which experiments to be described in this paper fully confirm. The electrical method has never been popular on this side of the Atlantic for several reasons, notably the discrepancy in the results attained by Gardner using the two methods in a parallel series of experi- ments and the failure of the investigators to deal satisfactorily with the question of movement of salts in the soil. Preliminary Experiments. Apparatus. The experiments which follow were made on a small plot of ground on the south side of the School of Agriculture at Cambridge. It had been dug over a short time previously and was not ideal for the purpose as a thin layer of builder's refuse from the building of the school rendered it less homogeneous than cultivated farm land. It had, however, the merit of being conveniently situated with regard to the laboratory, and the results obtained upon it appear to justify the conclusion that no great error was introduced by this lack of uniform texture. For the measurement of the resistance an ordinary post-office pattern Wheatstone's bridge was employed, with an induction coil from which the condenser had been removed to avoid any possibility of polarisation of the electrodes, and a telephone receiver. As a source of energy two small storage cells were used. At first a large coil giving a frequency of something under 30 per second was put in, but this was afterwards changed for a smaller one with a frequency rather over 100 per second, as it was found much easier to estimate the minimum of sound with this. The condenser in parallel with the bridge arm used by Briggs was dispensed with, being found unnecessary for the attainment of the degree of accuracy aimed at. The electrodes employed were formed from cored electric arc light carbons 9" long and I" diameter. The tapered ends of these were ground T. DeiOxHTON 209 down to form roughly a paraboloid of revolution, a shape found most satisfactory by Briggs (I.e.) for maintaining contact with the soil. The core was then drilled out for about an inch from the end and the cavity tilled with melted caoutchouc and sealing wax, a mixture which main- tained its position in the cavity and rendered good service for insulating the end of the core throughout the experiments. The opposite end was then scored and an inch of the core drilled out. A thread was tapped on to the inside of the cavity and a copper wire sealed in with fusible alloy. The top was then covered with sealing wax. The whole of the outside of the electrode, except a band one inch wide round the para- boloidal surface at the bottom, which was to act as the electrode surface, was then insulated by painting with two coats of "Duroprene"" and thoroughly dried for two or three days at about 50° 0. As the experi- ments progressed this insulation was found un.satisfactory and in cases where it was necessary to leave the electrodes a considerable time buried another type was adopted which will be described in a later paper. There appears to be a theoretical objection to the use of metal electrodes of any kind in direct contact with the soil which has not hitherto been noticed by workers in this field : namely that polarisation is very likely to take place even when an alternating current is employed unless the frequency is very high, since the free metallic ions are not completely returned to the electrode on reversal; and in some cases rapidly form double salts which are not decomposed on reversal. In these cases polarisation may occur even at frequencies of 40,000 per min. Owing to its abiUty to absorb gases a carbon electrode is self-depolarising to a sufficient extent to nullify the first effect and the second does not arise. The phenomenon mentioned is well known to electro-chemists (5). Local Variations in Resistance. It was desirable before proceeding further to find whether the resistance was the same under like conditions in different parts of the plot. The method of experiment was simply to push two such electrodes as have been described into the soil of the plot to the desired depth at a definite distance apart; a preliminary test having shown that the contact error in this, while not negligible, did not seriously affect the results. The same pair of electrodes was then moved to another part of the plot and placed at the same depth and the same distance apart. The results for a depth of 7" at 3" interval are shown in Fig. 1, which may be taken to represent the plot, which was about '22' X 30' in extent with the observations entered in the place in which they were taken. Subsidiary experiments proved that the obser- vations were sufficiently close together to justify the insertion of the 15—2 210 The Electrical Method of Soil Moisture Deterimnation lines of equal resistance as shown. These results are quite in accordance with the observations of Gardner (I.e.), who in attempting to standardise a pair of electrodes on an experimental plot found, by actually drying samples, that the moisture content at any depth might vary over a small area by as much as 7| per cent. Although moisture contents were not determined in the present case it will be seen from the resistance- moisture observations given later in this paper (Table VII) that such variations are sufficient to accoimt for the different resistances observed, although it should be remembered that other factors may also be operative, such as variations in the soluble salt content — especially of nitrates — from point to point in the plot. 1- ^ \ \ V t 1 1 1 \ 4i(k} >> \ moo \ \ 500O \ 2600 \ \ V \ N \ \ 3650 \ -~, V \8500 \ \ 4300\^ 3606\ \ \ \ \ \ \ \ \ \ \ / t \ 3006 2000--- 2 ICO \ \ 6300 5400 \ V \ \ Scik of feet £_ Fig. 1. Local variations of resistance on plot at 7" depth. The Effect, of Distance apart of the Electrodes. Gardner (I.e.) made numerous experiments on this matter by burj'ing two large electrodes 15' apart, and a series of smaller ones at distances from J" to 18" apart, 15' from each of the larger ones. He then took the resistance between the large pair and between the smaller ones, singly, and each of the larger ones. He adds " From these measurements it was a simple matter to calculate the resistance due to each of the individual electrodes at T. Deighton 211 a distance of 15 ur more feet. By adding the values of any two of the small electrodes thus found, the resistance is obtained which they would have if 15 feet apart." A table is given showing the actual resistance between the small electrodes at their actual distance apart and the calculated resistance between them when 15' ajDart. He concludes that "when the electrodes are 15 ft. apart 98 per cent, of the resistance is encountered within 9 inches of the surface of each electrode and the intervening 13i ft. of soil causes only 2 per cent, of the total resistance." Briggs (I.e.) states that the resistance is practically confined to volumes of soil not exceeding 6" in diameter with the electrodes as centres. If this be so, there should be no important variation on increasing the distance apart beyond 6"; but both Gardner's results quoted below (Table I), and those of the present author on a more extended scale, show a quite appreciable change beyond this point. Table I. Excerpt from Gardner s table. Distance Resistance between standard apart electrodes* 5" 928 8" 934 12' 968 18" 979 * Gardner calls this column (the numbers being given as percentages 92-8 etc.), "Relation of observed to calculated resistance." By making the resistance 1000 ohms at 15' the standard and reducing the other resistances to this the same figures are obviously obtained. Gardner's experiments were all made at one depth and the author therefore carried out a series of measurements of the effect of distance apart of the electrodes on the resistance of the soil at varying depths, Table II. Effect of Distance Apart of Electrodes on Resistance at various Depths. Distance Resistances, in thousanc Is of ohms at apart of electrodes '. 1" depth 2" depth 3" depth 4' ■' depth 5" depth 6" ' depth 7" depth 2" 100 92 33 15,7 11,0 8,7 6,2 3" 60 27 17,o 14,0 11,9 11,9 8,5 4" 75 34,5 15,5 10,4 8,o 6,4 5,2 6" 85 25,5 14,0 9,4 7,o 5,3 4,0 9" 51 22,o 11,3 7,2 5,9 4,9 4,4 12" 45 15,5 9.1 7,o 5,8 5,o 4,6 18" SO 22 lO.q 7,5 5,5 4,5 3,S 27" 70 20,9 13,9 8,2 6,5 5,2 4,o the results of which are given in Table II. There appears to be a tendency for the resistance to fall to a minimum when the electrodes are about 212 The Electrical Method of Soil Moisture Detenu! nation ]"2" apart at shallow depths while in the case of depths about 0" the minimum is found when a distance apart of about 18" is reached. The Effect of Depth. The following experiment was made to determine the effect of depth unhampered by other considerations. A very dry soil was sifted through a 3 mm. sieve to extract .stones and packed as uniformly as possible into a pot. Two electrodes 4" apart were then sunk to different depths in it and the resistance measured. It was thus shown that tlie resistance between them was approximately halved by sinking to a depth of 4". A repetition of the experiment after .saturating the soil gave a result quite in accordance since in this case there must have been some moisture gradient due to the action of gravity on the water in the pores of the soil. The observations are sho\\Ti in Table III. Table 111. Effect of Depth in Uniform Soil. Depth of centre of Resistance in oiims elei^trode below , ■ -^ ^ .soil surface Dry Saturated 1" 12.5.000 550 2' 102,000 460 3" 72,000 260 4" 66,000 200 Electrode Values. A further series of experiments was undertaken to determine whether it was really pos.sible to assign values to the electrodes on a system similar to that employed by Gardner. For this purpose equal sized electrodes were set out at the four corners of a six-inch square and at a depth of G". The distance apart agrees witli Briggs' figure for the volume of soil concerned in the resistance measurements. Observations were then taken in all po.ssible ways between them. In Table IV the results are shown, — in the last two columns we have the resistances observed and parallel with these the resistances as calculated from the first four observed resistances by Gardner's method, e.g. the electrodes being equal, if R^ is the resistance due to the electrode A, Ri, that due to the electrode B, and R^ that due to the electrode C; and if Rat,. Rac ^ifl Rbc are the resistances observed between A and B, A and C, and B and C respectively, we have .'. -Ra + ^6 + ^c = Rab + Rac, whence R^ = R<^ + ^a^ -JR^ + ^q) ^ T. Deighton 213 but R, + F, = B,,. ^ab + "ac ~ ^ic Similarly R^ and R,. may be determined and R,i will equal i?^,, — R^- Obviously R^^ and R^^ can then be calculated. As will be seen no very good agreement is attained. Table IV. Resistances, observed and calculated, on 6" square. ^M Pea A Ixp. Ki P^.c «.c K, tihs. Calc. (.)bs. Calc. 1 210U 2300 2900 1300 2300 1900 2400 2100 -> 2710 3080 4000 1520 2970 2440 3180 2810 3 23SO 21)30 3500 1400 2670 2270 2800 2520 4 3400 3570 5550 1620 3570 3600 4000 3770 5 2800 3000 4370 1450 2970 2820 32.50 3020 6 2490 2900 3850 1420 2720 2370 2970 2780 7 2470 2770 3750 1390 2670 2370 28.50 2r,70 8 3020 3900 4950 1590 3490 2640 3650 3520 9 41G0 4250 sdOO 1490 4400 5240 5500 5330 This method of calculation gave results more in accordance with the observations when seven electrodes were set out in the angular points of a regular hexagon with the seventh in the middle, in such a manner that the distance between any adjacent pair was the same = 18" corresponding to (Gardner's limit: the value of an electrode being de- termined as the mean of the results obtained from two or three triangles. As these results are very length v and of no importance in themselves they are omitted. The calculated and observed results agreed to about 100 ohms in 3000. It will be seen later that in an infinite isotropic medium the calculation should give the result desired, the variations are therefore most probably due to local aelotropic conditions. Resistance at greater Distances. The question of what happens at distances much exceeding 18" was investigated by the author, once as a continuation of the 7" depth observations of Table II and once inde- pendently of these. The results, given in Table V, show a slow increase of resistance beyond 18" up to quite long distances. We are therefore driven to conclude that had Gardner actually measured the resistance between two of the smaller electrodes at 15' instead of merely calculat- ing it he would not have concluded that the intervening ISi' of soil accounted for only 2 per cent, of the resistance at 15', the rest being due to the 9" of soil about the electrodes. It will be observed that in Exp. 2 there is no minimum at 18". There is here no real discrepancy, the occurrence of a minimum being a possible 214 The Electrical Met hud uj' Soil Moisture Deteriitinatiun and not a necessary pheaomenon, as will be shown from theoretical considerations later on in tliis paper. Table V. Resistances at greater Distances. listanco E.\p. 1 Exp. 2 Distance Exp. 1 Exp. 2 4" r,200 2400 3' — 3200 6" 4000 2300 4' 4500 3700 9- 4400 2600 6' 5200 4100 12" 4G0U ao.w 8' — 2250* 18" ;!800 3000 9' 5400 — 2' — 3700 12' — 4300 27" 4000 — 10' — , 5400 * I am \ituible fully l(j account for this lower figure; several subsequent tests about the same place gave results in the neighbourhood of 3600. To summarise the results to this point, we find (hat the pot experi- ment appears to favour Briggs' theory, the fact that agreement between calculated and observed resistances, even when the calculation was made from a single triangle, was better in the hexagon than in the square speaks more in favour of (Jardner's limit being correct. On the other hand neither theory explains the increase of resistance at distances exceeding 18", the minima observed in Table II, nor why these minima should appear at greater distances as the depth increases. Thk Path of thk Current. Experimental. It seemed probable that a more satisfactory result might be attained by casting overboard the idea, really no more than a convenient fiction, of the resistance being due to the soil in the im- mediate neighbourhood of the electrodes and considering the whole mass of soil about them. An experiment was devised the object of which was to determine at what depth below the line joining the elet^trodes a conducting layer would appreciably affect the resistance between them. For this purpose a beaker of about 600 c.c. capacity (Fig. 2) was taken and four scales of inches were gummed up the sides of it. Inside this lay a clean per- forated metal plate of nearly the diameter of the beaker so that the curved-in sides of the vessel would just keep it off the bottom. Through the centre of this plate was passed a glass tube carrying a copper con- ductor which was soldered to the underside of the plate. At the top, the tube made connection with a funnel by means of about an inch of rubber tube, but the copper wire passed out at the side and the opening was rendered watertight with sealing wax. Tlie beaker was then filled up to the 2" mark with pure dry sand and five subsidiary electrodes of T. Deighton 215 copper wire were buried therein at the 11" level. The rest of the beaker was filled with the same saud very shghtly damped so as to have a resistance of about 11,000 ohms between the main electrodes, consisting of solder spheres on the ends of thick copper wires, placed 2^" apart in the sand at the 3|" level. The subsidiary electrodes were connected to a ring conductor surrounding the apparatus and this was joined through a galvanometer to a single dry cell the other terminal of which made connection with the wire from the metal plate at the bottom. The main electrodes were connected through the bridge to the secondary of the induction coil. The sensibility of the galvanometer was such that when Fig. 2. Apparatus and connections in Beaker Experiment. one subsidiary electrode was placed in conducting connection with the metal plate the needle moved only a fraction of the distance across the scale, so that all five electrodes would have to be in use to give anything like a complete cross swing. Hence on pouring a solution of an electrolyte (very dilute hydrochloric acid) down the funnel, it was possible to see if the surface of the solution rising by capillary attraction was sensibly plane or whether it was depressed or humped up at the centre; the condition for a plane surface being that a sudden complete cross swing of the galvanometer needle should be observed at the moment the water line visible on the 216 Tlie Electrical Method of Soil Moist lar Deteriiii nation outside reached the 1|" mark. In the experiment this was observed quite sharply and it was therefore assumed that the liquid surface in the capillaries was sensibly plane. The dampiu^j of the sand above had not been carried so far as to interfere with packing and it was therefore presumed that this condition was equally satisfied above. The resistance unplugged in the standard arm of the bridge was 10,000 ohms, so that the effect awaited as the solution gradually rose was a sudden cessation of sound in the receiver, followed immediately by a sudden increase. This also gave a very sharp indication when tin' water line read on the four scales round the beaker was s(jmething between }/ and \" below the line of the main electrodes, the readings were 3g", 3j", 3", 3^". Thus when the conducting surface reached a level of somewhat more than \" below that of the two electrodes the resistance between them fell by about 10 per cent. From this result it appears probable that the current density which is concentrated in the line joining the two electrodes becomes rapidly less on moving away from this part of the field, though probably not as quickly as the result appears at first sight to suggest, as the conductivity is only increased over a comparatively small part of the region surrounding the line of the electrodes. The following experiment, originally performed as a means of con- firming the above result, is described not so much for any value it has in itself, as on account of the justification it seems to supply for the assumptions made in the short mathematical investigation which follows. A tray was taken in which moist sand was laid out in a lamina l" thick. Electrodes were placed 61" apart in this and the width of the conducting layer was gradually cut down by the insertion of wooden partitions on each side of the line of the electrodes at varying distances. The increase in resistance with decrease in width of the conducting layer was quite in accordance with expectations, becoming much greater as the electrodes were approached, while remaining practically unchanged at considerable distances. The results are given later in Table VI. Theoretical. It is shown by Mascart and Joubert(fi) that the resistance between two electrodes immersed in an unlimited isotropic conducting medium depends only on the medium and the form and dimensions of the electrodes, and is doubled if the medium is limited by an infinite plane passing through the electrodes, i.e. if the medium extends on one side only of this plane. On account of the comphcated nature of the considerations involved they omit any consideration of the problem when the bounding plane passes to one side of the line of the electrodes. It will suliice lor our purpose however if we can determine in some T. Deighton •217 simple manner the volume of soil through which the practically important portion of the current flows. The method of attacking the problem adopted by Mascart and Joubert, viz. by electrical flux and tubes of How leads to complications when applied to cases other than the two selected. Since we are using an alternating current, we may probably treat the matter from the electrostatic point of view at any one moment, instead of considering it generall}' from electrodvnamic considerations, without any serious error. X ^^ ^ ^^ "~^^ __p ^^ "-. z / ^ -^ ^ %' y^ " ^^ ^"^^ .^-^-^^ ^y^ ^^\^ \ ^y Y \ y ^^^ \ / N / ^ N / ^-^ \ / ^^ \ / ^^ \ A i^e c \ \ p' Let A and B (Fig. 3) be two electrodes, supposed small, in an isotropic conducting medium. We may assume that at any moment an ion at any point P in the median plane PC'P' is subject to two forces, one acting from A, represented in magnitude and direction by PX, and one towards 2?, represented in magnitude and direction by FY . These wnll be equal at any point in PVP' and will have a resultant FZ which may be taken to represent the force tending to move the ion at the moment which will be proportional to the current, since assuming Ohms law, the mean velocity of the ions resolved in the direction in which the force acts varies as the force. Thus the current density at any point on the median plane FCF' will be proportional to FZ at that point. 218 Tite Electrical Method of Soil Moisture Determination Now if the angle BAP -= 6 we have in three dimensions px = py cc ^ Jp^, = ^Acle^er "" '"'" ^' if AC remains constant. Hence PZ - (PX + PY) cos 6 oc cos^ 6. Precisely the same result is obtained from a consideration of the equipotential surfaces. Since the electrification of the electrodes at any moment will be equal and opposite the median plane is the equipotential surface V = 0. On either side of this will be equipotential surfaces V = + dV and V = — dV and the current across the median plane therefore will be inversely proportional to the distance x along a line of force between these two surfaces^, since this is a measure of the resistance assuming the medium isotropic. It makes no difference in this consideration whether we take the momentary state or put V in virtual volts and the current in virtual amperes. We have then that the current across the median plane at the point P, is equal to K . Now considering spheres round the electrodes V = j^ where r is the radius of the sphere, whence V and since x is inclined at an angle = 6 to the direction of r X = dr sec d = — ,- r^ sec ddV = _ ^^ysec^edV. Hence when the potential difference at the ends of x is constant = 2dV we have X oc sec* 6, whence current across the median plane at P x cos' 6 as before. Thus the fraction of the total current passing outside any tore^ ' This may be considered a straight line. - The lines of force are known to be circular arcs vide Masoart and Joubert, Electricity and Magnetism, 1, p. 205. T. Deighton 219 formed by the revolution of a line of force AQPB about the axis AB will vary as: tan e cos3 9 (W (since the circumference of the tore in the median plane is 'IvCP oc tan 6) sin 6 COS" d (19 cos2 9 d (COS 9) = i cos3 9. Thus the fraction of the total current passing outside any tore oc cos^ 9 and the fraction passing inside any tore cc (1 — cos^ 9). It follows from this that 35-7 per cent, of the current passes outside the sphere whose poles are the electrodes. Proceeding in a similar way we have for a lamina 1 1 PX = PY cc oc cos 9 AP AC sec 9 if AC is constant. Therefore the fraction of the total current passing outside any portion of the lamina enclosed by two lines of force of equal length on opposite sides oi AB varies as (1 + cos2^)f/0 = cos- 9 d.9 "1„ 1 . 77 4 in 2^ Table VI. Results of Tray Experiment, Observed and Cahulated. Resistance Extent of sand on ^ Corrected for Calculated from each side of line Mean of three temperature and the above of electrodes actually observed evaporation formula oc — — 18,400 7i" 18,500 18,600 18,600 5i" 19,100 19,.500 19,800 4rV' 20,100 20,000 20,700 31" 20,800 22,000 22,200 2r 23,400 2.5.000 24,400 If" 27,100 29,700 36,800 As will be seen from Table VI this agrees very well with the results obtained in the experiments ^\'ith the tray of moist sand described above. 220 The Electrical Method of Soil Moisture Determination The observation at 1|" is doubtless aSected by the nearer a^jproach to a three dimensional condition and the fact that the electrodes were not small in comparison with the mdth of the field thus restricted. The correction for evaporation was made by interpolation from the original and a final observation at 1\". We see then that no grave error is likely to have been introduced by the assumptions made at the beginning of this section. This being so we are probably not far wrouj^ in inter])reting our soil phenomena in terms of the calculation made for the three dimensional case, more especially as it explains all the phenomena observed. Each electrode may obviously still be considered to possess a definite re- sistance value of its own at any stated distance, and calculations based on these values will still yield approximately correct results but slight differences will doubtless appear according to the direction in which the second electrode lies. Since with increasing distance apart local differ- ences in specific resistance will tend to cancel out in their effects it is clear why a better agreement was obtained in the hexagon than in the square. The result of the pot experiment is explainable on either theory, but it is noteworthy that the theory agrees with the halving of the re- sistance at 4" depth. Calculation actually shows an external residue of current of about 8 per cent, but since the surface is plane it follows that a considerable portion of this would, in the circumstances of the experi- ment, be included. The steady but slow increase of resistance as the electrodes are moved further apart beyond 18" is explained by the cutting off by the air layer of regions of increasing current density. Combining this with the fact of the existence of a moisture gradient in the soil we obtain an explanation of the minima observed in the plot trials (Table II) since, starting with the electrodes in close proximity and gradually moving them apart the fir.st effect is to cause a fall in resistance by causing the current to dip down into lower moister layers. After a time however, with increa.sing distance, the non-conducting air laver above begins to cut into regions of greater and greater current density till the resistance introduced by this more than compen.sates for the other effect. In the second series of experiments in Table V there is no minimum. These were made at the end of the long drought of 1921 and it appears probable that the second effect overbalanced the first from the beginning, owing to the moisture gradient in the soil being less steep. An explanation is also afforded of the increase of the distance apart at minimum resistance with increasing depth since it will in this case be necessar}' to move the electrodes further apart before the air T. Deighton 221 layer begins to affect the couduction between them to a similar extent. The result obtained in the beaker experiment is easily understood since the surface of the conducting layer being plane, only small segments of tores carrying an appreciable fraction of the current pass through this layer, and the segment is less the greater the current density in the tores. An attempt to solve the matter completely by mathematics led in all cases to expressions integrable only over restricted ranges or in very slowly converging series, putting this method of procedure out of court for practical purposes. A series of rough approximations led to results not out of keeping with those obtained. CONCLUSION.S FROM ABOVE ExPERI.MENT.S. These results are of practical importance as they show what may and what may not be expected from this method of moisture determination. It is clear in the first place that what is obtained is the mean resistance of a volume of soil which may for practical purposes be considered as of rather greater extent than a sjahere whose poles are the electrodes, something the shape of an apple with the electrodes at the calyx and stalk, since the fraction of current passing outside a selected tore falls off very sharply beyond d = 45°, and soon becomes negligible. The moisture content will also refer to this volume of soil and local and depth varia- tions will be obliterated. If therefore we desire to measure the moisture content at any depth we must place the electrodes at such a distance apart that the moisture gradient can be considered uniform in the volume of soil concerned when the deficiency above the level of the electrodes will be counterbalanced by the excess below this level. More- over if the distance is more than a few inches it will be necessary to take account of the tendency of the current to dip down into the moister layers, or to continue the simile above, of the downward bend in the core of the apple. Secondly the electrodes should really be of such a size that they may be considered small in comparison with the distance apart as otherwise the calculations do not strictly hold and the new conditions ought to be investigated afresh. The necessity of obtaining good contact with the soil puts a practical limit to the size of the electrode and it was found in consequence that 3" apart is about the lower limit for the electrode distance. The distance apart must moreover be chosen with a view to confining the current or such part of it as matters to the portion of soil concerned, e.g. a distance of 15'-20', while probably giving illuminating information on forest land would usually be quite useless in ordinary arable farming. Finally it seems possible that considerable •222 The Electrical Method of Soil Moixtiirc Deterinlnation use might be made of the method in arid sandy regions for determining the approximate depth at which the water table lies, e.g. supposing the depth of this were 50' — starting with electrodes 6" deep say and a short distance apart on increasing the distance the resistance would gradually increase to a point but on reaching a distance apart of about 200' the beaker experiment would lead us to expect a fairly sharp fall due to the water at 50' depth. A few observations would suffice to determine at what fraction of the distance apart when the resistance falls the water table would be found. It is perhaps worthy of note in this connection that the resistance of earth returns in telegraphy is very low; though of course, as large plates are usually buried for this purpose, we should not erpect anything like the resistances encountered in the soil experiments. Resistance-moisture Curves. Experiments were made in the laboratory to determine the character of the resistance-moisture curve. Artificial mixtures, natural sands and soils were employed and results obtained in accordance with those of Whitney (^c), except that at low moisture contents definite discon- tinuities were observed. The method of carrying out the experiments was alike in all cases. 500 grams of the soil or other substance used was placed in a wooden box whose internal dimensions were 2|" x 6" x 2^" and the electrodes, small carbon cones, were placed at each end. An extra amount of the same sample was used as a reserve from which samples removed for moisture determinations could be replaced, maintaining the original weight of dry soil. The hygroscopic moisture was found not to affect the resistance within the capacity of the instruments used; that is to say the air dry substance had in all cases a resistance under the con- ditions of experiment exceeding one megohm. As wetting a sample in such a way as to render it only slightl)- moist is difficult the following plan was adopted: the sample was wetted slightly with distilled water and then thoroughly mixed, after which it was warmed to about 30° C. and allowed to cool again once or twice. This seemed to distribute the moisture evenly through the mass and was preferred to the alternative of gradually drying out a saturated mass as its character is changed by this treatment. The damp sample was then placed in the box and packed down to a constant leveP — a thermometer placed in the middle of the mass enabled the temperature to be read and the resistance was taken ' Great clifTicultv was found in this in the case of the boulder einy; the observations were less satisfactory on this account. T. Dkighton 2-23 when this liad fallen to some point but shghtly above room temperature, the same temperature being employed throughout any one series of determinations. A sample was taken out from the middle of the mass after each resistance reading and the moisture in it determined by drying at 110° C. About 15 to '20 determinations were made in this way for each sample. It was not, however, found possible to carry the moisture content up to saturation since near this point the water tended to oose out from the bottom of the bo.x. The results obtained with one soil are given in Table VII. If plotted out they yield hyperbolas, but if, instead of Table VII. Result.s of Resistance Moisture Detenu 'mat ions on a tijpical soil (Greensand). Hygroscopic moisture 1-4 % Moi.stui'c "(J Resistani-c iluisture ",, Resistance dry weight in olims dry weight in ohms 2-3 1.100,00(( 71 16,800 2-9 :j!iU,OUO 7-6 11,700 3-4 257,000 8-9 8,920 3-6 208,000 9-6 6,200 4-4 130,000 10-7 5,050 4-9 114,000* 11-6 3,950 5-1 (>0,000 12-5 3,490 5-3 lll,.")0O* 13-0 3,270 5-7 :i!i,ooo 14-2 2,.50O 5-8 70,9110* 18-6 1,870 6-6 19,200 plotting directly, we plot the logarithm of the moisture expressed in percentage dry weight, less the hygroscoi^ic moisture, against the loga- rithm of the resistance observed we obtain the curves shown in Figs. 4, 5 and 6: whence it is clear that at low moisture contents the curve is discontinuous at any rate at one point, the existence of the first point of discontinuity is disputable owing to the small number of observations. It will be observed that the type of curve is different for the artificial mixtures to what it is for the natural soils and sands taken. In the former case the resistance falls sharply at first, then less so and finally sharply again, in the latter the reverse holds — slowly at first, then more quickly, and slowly at the end. It would appear that there ought to be no sharp angle between the different regions as shown in the curves. It was felt however that as the accuracy of the measurements was not sufficiently great to justify the use of them as a basis for the discussion of this matter greater clearness would be attained by drawing the fines as shown. The facts seem explainable as follows : — In the case of pure sand and artificial mixtures we may assume that there is no measurable amount Joum. of Affric. Sci. xii 16 224 The l-Jltctrical Method of Soil Moisture Detenu i nation of colloid present, while in the natural calcareous sand (coarse) a mechanical analysis showed 0-45 per cent, clay and the other sands and soils used contained considerably more than this. Now in the case where we have no colloid present we should expect little or no reduction of resistance until the 5()/x/x film thickness of Quincke (7) is passed since Terzaghi (8) has shown that there is no brownian movement and therefore presumablv but little if any ionic movement in this layer. Owing to evaporation it was found impossible to obtain any results in this region. Pure sand w. high resistance water „ w. low „ >. + 1 % marble w. high reeiBlance water „ „ w. low ,. + ij % marble 17 18 19 00 01 02 03 04 Log. ot % water over hygroscopic Fig. 4. Beyond this the addition of water might be expected to cause a more rapid fall than in the third stage where it vnW be sensibly inversely pro- portional to the moisture content up to the point at which surface tension begins to overcome the force of adhesion ^ When surface tension ^ For in this case the surfaoe of electrode wetted will be directly proportion.il to the thickness of the moisture layer, thus if H is the radius of a particle, supposed spherical, and IR the thickness of the water film upon it and if ti is the number of contacts between T. Deighton •22b begins to take effect however the tendency is to concentrate the extra moisture added in the interstices where one particle touches another or the electrode surface and therefore we may expect that the resistance will vary inversely as some function of the moisture content. Experi- mentally it was found to vary approximately as the inverse square of the moisture as had been found iireviouslv bv Whitney. COABSt CALCAHEOUS SAN3 18 19 GO 01 07 03 04 Log. of % water over hygroscopio Fig. 5. A tentative explanation of the second series of cases where the particles are covered by a colloid substance was originally elaborated on the theory that the conductivity of a substance in the gel stage would be less than that of the same substance under similar conditions as a sol. Recent work by Miss Laing and Prof. McBainO) however shows that the several particles and the electrode the area wetted a:n[(R + AiJ)^ -iJ^] =2ni?. AiJ neglecting higher powers of Ai?. If in is the moisture content nR.\R cctiRm. But for any electrode obviously n oc 1 Resistance = 1 R conductivity area wetted experimental results agree with this in the region dealt with. oc - . The 16- 226 The Electrical Mle VII. 228 The Electrical Metliotl of Soil Jfoisfiin J)efenninafioit under exceptional circuinstiURes of izetting a nietastable condition of some kind; possibly due, if the theory adumbrated is true, to something of the nature of a supersaturation of the coagulum — or, more probably, a time factor comes in, Lainj; and McBain having found that the con- ductivity of a newly formed curd falls oil for a long time after its forma- tion, thus we might expect a similarly retarded recovery if water is added too rapidly. The probability of this is increased by the fact that these workers found that the best way to obtain the gel was to warm the curd very slowly. This error would in this case be unlikely to appear in soil work. The agreement of the last portion of the curve as to slope in all cases seems to show that in general at moisture contents exceeding 10 per cent, or thereabouts the American formula holds good approxi- matelv and we mav therefore use it in the form given by Gardner, viz. : '^-y^' where W = per cent, moisture at time of standardisation, R = the resistance corresponding to W, i?j = the resistance observed, and Wi = the per cent, moisture corresponding to the resistance R^, provided always that we do not lose sight of the sources of error noted above and of the limitations of the method dealt with earlier in this paper. There remains the question of the effect of movement of soil salts in the soil which, if considerable, may either vitiate the method alto- gether, or render it too cumbrous for use by necessitating a too frequent standardisation of the electrode. Experiments are proceeding on tliis matter, but from the extreme slowness of the method adopted for this purpose, the full results are not likely to be available for some time. It may be stated however that the material so far to hand seems to indicate that where a soil is protected from rain these salt movements do not affect the moisture determination by as much as one half per cent. I take this o])])oi'tunit v of tiianking all those who by their unfailing interest, help and encouragement have enabled me to bring this in- vestigation to its present position. Among these I owe a special debt of gratitude to Mr J. W. Capstiek. O.B.E., M.A., D.Sc. for continuous advice and sympathetic criticism throughout the whole time that the experiments were in progress. T. Deighton 229 Summary. In this paper an examination is made of the processes operative and the limits of accuracy of the electrical method of determining soil moisture. The resistance over a small plot is found to vary under similar conditions and it is concluded that these differences are mo.st probably due to actual differences in moisture or other factors. The effect of the distance apart of the electrodes is investigated and a probability of a minimum resistance between two electrodes being observed under certain conditions is indicated. The use of electrode values in computing soil resistances is discussed and criticised. The path of the current in the soil is investigated mathematically and it is shown that the results obtained accord well with the observed facts. It is concluded that the method gives the mean water content of a volume of soil somewhat greater than a sphere whose poles are the electrodes. The practical limits of the method are indicated. Certain resistance-moisture curves obtained in the laboratory are discussed and it is concluded that while at water contents above 10 per cent, the relation found by the American investigators holds good — viz. that the resistance varies inversely as the square of the moisture content; at lower water contents than this one and possibly two dis- continuities appear in the curve. These discontinuities are reversed in the case of artificial mi.xtures not containing colloids. A tentative explanation of these phenomena is given. REFERENCES. (1) Whitney, M. Some Physical Properties of Soil in their Relation to Moisture and Crop Distribution. U.S. Dept. Agric, Weather Bureau Bull. No. 4, 1892. (2) Whitney, M. and Means, T. H. An Electrical Method of determining the Soluble Salt Content of Soils. U.S. Dept. Agric, Dimsion of Soils, Bull. No. 8, 1897. (3) Gardner, F. D. The Electrical Method of Moisture Determination in Soils; Results and Modifications in 1897. U.S. Dept. Agric, Division of Soils, Bull. No. 12, 1898. (4) Beigqs, L. J. Electrical Instruments for Determining the Moisture, Tem- perature, and Soluble Salt Content of Soils. U.S. Dept. Agric, Division of Soils, Bull. No. 15, 1899. 230 The Electrical Method of Soil Moisture Determination (o) Leblanc, AF. and Schick. Wechselstromelectrolyse. ZeiLichr. f. Phys. Chem. 46 ( I. (Received April IMi, 1922.) THE CHEMISTRY OF THE STRENGTH OF WHEAT FLOUR. By HERBERT ERNEST WOODMAN, Ph.D., D.Sc. {From, the Animal Nutrition Institute, School of Agriculture, Cambridge University. ) It is well known that different flours vary enormously in respect of the size and shape of loaf they yield on baking. The factor which determines the quahty of flour in this connection has been termed "strength" and the latter has been defined as "the capacity of flour for making large well-piled loaves "(i). Many views have been held from time to time regarding the ex- planation of flour strength from the chemical standpoint. The earliest view was that strength was determined by the gluten content of the flour, which by virtue of its tenacity was able to retain in the bread the carbon dioxide produced as a result of the activity of the yeast. Many cases, however, were investigated where flours possessing a high gluten content were not so strong as a flour with a low content of gluten. Furthermore, no accepted regularity has been found to exist between the strength of flours and the water-holding or gas retaining capacity of their glutens. Attention was next directed to the consideration of the individual proteins in the gluten of flour, namely, gliadine and glutenine. It was found that measurements of the absolute amounts of gliadine showed no correlation with strength. Neither was it possible to show any con- sistent relationsship between strength and the ratio of ghadine to glutenine in the gluten. It was suggested by Hall (2) that gliadine might not be a definite substance and that the gliadine contained in very strong flours might be different from that in weak flours. Wood (3), however, prepared and examined samples of gliadine from strong and weak flours and concluded, on the grounds of their content of amide nitrogen, that the proteins from both sources were identical. Indirect determinations of the amide nitrogen of the glutenines of weak and strong flours led to the conclusion that the glutenine protein in different flours was one and 232 The Chenihtrii of tin StniH/f/i of W/irat Flour the same substance. It was therefore not possible, on the available evidence, to explain the difference in the loaf-iiiakinti; qualities of different flours by reference to their protein content. A valuable contribution to the study of the subject was made, when Wood (3) resolved the conception of flour strength into two factors: 1. Tii(> factor of strength which determines the shape of the loaf. ^- J) ;• >j I! size ,, ,, The results obtained by this investigator justified the conclusion that the capacity of a flour for giving off gas wjien incubated with yeast and water is the factor which in the first instance determines the size of the loaf. The latter depends not so much on the amount of sugar present in the flour as such, but on the diastatic capacity of the flour, which gives rise to continued sugar formation and consequently con- tinued gas evolution in the dough. The same worker also showed that the properties of gluten in regard to coherence and elasticita^ were subject to considerable modification by the concentration of acid, alkali or salt in the solution with which it was in contact, and he suggested that these properties have an important bearing on the shape of the loaf. A knowledge of the acidity and soluble .salt content of a flour should therefore afford a clue to that factor of strength wliich decides whether the flour will make a good-.shaped loaf. The information obtained in the investigation referred to above has been applied with success in several pha.ses of the milling industry. The view has long been held, however, by Professor Wood himself, that in view of the improvement of tlie methods employed in protein research, the question of the identity or non-identity of the corresponding gluten proteins in weak and strong flours should be re-investigated. It is now recognised that two proteins may be quantitatively identical with regard to their amino-acid content and yet be two distinct proteins, by virtue of dift'erences in the order of hnkage of the amino-acids within the protein molecules. Such a case is furnished by the caseinogens of cow's and sheep's milk. Though it is possible to show that these proteins are distinct sub.stances(i), yet by the ordinary chemical methods of analysis they are indistinguishable. It is clear, then, that any chemical method which is to be emplo3'ed to decide on the identity or non-identity of related proteins must be such as to take into account the pos.sibility of differences connected with the order in which the constituents of the proteins are linked up within the molecules. Such a method is the Racemisation Method, which has been used recently in the investigation of the corresponding proteins of H. E. Woodman 233 blood serum, colo.stnun and milk (•'>). The method depends on the be- haviour of proteins in dilute ulkaline solution. When such solutions are kept at 37° C, they suffer a progressive diminution in the value of their optical rotatory power as a result of a keto-enol tautomerism of the = CH — CO — groups in the protein complex. If the specific rotations of the solution be plotted against the time in hours during which the reaction has been allowed to proceed, then the readings fall on a perfectly smooth curve. It is found that the rotation sinks rapidly at first, then more slowly and subsequently after about 250 hours attains a practically constant value. The process thus results in the partial racemisation of the protein, and the graphs thus obtained are referred to as racemisation curves. Since individual proteins display specific behaviour quantitatively when racemised with dilute soda, it follows that the method may be used to test the identity or non-identity of related proteins. Thus, if two proteins are to be pronounced identical, then if racemised under the same conditions, their solutions must show the same initial rotation, the same final rotation and the same rate of diminution of rotation. They must also continue to display identical optical behaviour if the concentration of the alkah or protein in the solution is varied. On the other hand, if two proteins are not identical, this will be revealed by their possessing distinct sets of racemisation curves. In the investigation to be outlined in the present communication, samples of gliadine and glutenine have been isolated from typical strong and weak flours and have been investigated comparatively by means of the racemisation method. The results obtained are very suggestive in relation to their bearing on the existing ideas of flour strength, since it has been shown that whereas the gliadines from weak and strong flours are identical proteins, yet the glutenines prepared from the same sources appear to be two distinct chemical individuals. Preparation of Proteins. 1. Gliadine and glutenine from strony flour. The flour used for this purpose was a typically strong flour milled exclusively from Northern Manitoba Wheat. The method used for the isolation of the proteins was in its essentials the same as that described by Osborne (fi). The bulk of the starch and soluble constituents was removed by enclosing the flour in a muslin bag and kneading the material in a stream of running water, the process being completed by thoroughly 234 Thr ('hriiilxtrij of the Sfrftif/f/i of Wlimt Flour kueadiiig the gluten under a large volume of distilled water. The re- sultant characteristically sticky gluten was broken up into small pieces and extracted with alcohol, the alcohol added being such as to give, with the water in the gluten, a solvent containing 70 per cent, of alcohol by volume. After standing for 48 hours with frequent shaking, the super- natant alcoholic solution of gUadine was filtered ofT and the residue was repeatedly extracted with successive portions of 70 per cent, alcohol until the amount of gliadine going into solution was inappreciable. The combined filtrates were then concentrated in vacuo at 50° C, care being taken to keep the gliadine in solution by adding small portions of alcohol from time to time. The concentrated solution was cooled and poured slowly, with constant stirring, into a large volume of ice- cold distilled water containLiig about 10 grin, salt per litre. A gummy mass separated out which collected on the glass rod. This was repeatedly washed with distilled water, dis.solved in 70 per cent, alcohol and the solution filtered until water clear. After concentrating the solution in vacuo at 50° C, the syrupy residue was cooled and poured into absolute alcohol. It was found that complete separation of the gliadine could only be obtained by this method after stirring a Uttle salt into the alcoholic hquid. The addition of ether also enabled the gliadine to separate completely. The process of dissolving the gliadine preparation in 70 per cent, alcohol, filtering, concentrating in vacuo and precipitating by means of absolute alcohol was carried out in all three times, the final precipitation being effected by a mixture of ether and absolute alcohol, this resulting in the gliadine separating in a flocculent condition. The protein was then dried by washing successively with absolute alcohol and anhydrous ether and was filially obtained as a white powder which dissolved completely in 70 per cent, alcohol to give a water clear solution. As a result of the method of preparation, it contained a little salt as impurity. The gluten residue, after extraction of the gfiadine with 70 per cent, alcohol, was allowed to dry at room temperature and then powdered. It was then further extracted with alcohol, the gliadine-free residue being shaken with successive portions of ether to remove fat. After air-drying, the material was shaken with sufficient 0-2 per cent. KOH to effect solution. The extract was filtered, great difficulty being ex- perienced in obtaining a clear filtrate. From the latter the glutenine was precipitated by means of very dilute hydrochloric acid, the amount of acid requisite for flocculent precipitation being determined by a pre- liminary test on a small bulk of the alkaline solution. The precipitate H. E. Woodman 235 obtained in this manner was exhaustively extracted with 70 per cent. alcohol to remove traces of gliadinn. It was then redissolved in the minimum amount of ()-2 per cent. KOH, filtered clear and the glutenine reprecipitated in the manner already described. The process of pre- cipitating the protein fnjm alkaline solution was carried out in all four times, and after each precipitation, the glutenine was extracted with 70 per cent, alcohol to ensure complete removal of gliadine. After the final precipitation, the glutenine was well washed with, distilled water and was obtained as a white powder bv successive washings with absolute alcohol and anhydrous ether. The preparation gave a water clear solution in 0-2 per cent. KOH and an exhaustive extraction of a sample of the material showed that it was entirely free from gliadine. It contained, as a result of the method of isolation, a small amount of potassium chloride as impurity. 2. Gliadine and f/hitcnine fmn) iveak Jlnur. The flour used as the starting point for the preparation of these proteins was one which had been milled exclusively from English wheat. The same methods were employed as those outlined in the preparation of the proteins from Manitoba flour. The gluten of the Enghsh flour differed materially from that of Manitoba flour in respect of its physical properties. Whereas the latter was sticky and coherent, the former lacked coherency and resembled putty in its consistency. Much bigger percentage yields of the proteins were obtained from the Manitoba flour than from the Enghsh flour. Method of investigating the behaviour op the protein.s WITH dilute alkali. The protein .samples were first finely ground up and then dried in vacuo over calcium chloride for several days. The amount of ash-free protein in each sample was determined by means of the Kjeldahl method, the nitrogen content of the gliadine samples being multiplied by the factor - and that of the glutenine samples by yyIiq ■ In the case of the gliadine .samples, the first series of determinations were carried out in the following manner. Exactly 1 grm. of the dry protein was weighed out into a 50 c.c. flask containing a little distilled water. 2-5 c.c. of N NaOH (or Nj2 NaOH as the case may be) were then slowly run in from a pipette. After mixing gently, the volume was 2;>6 The Clumistnj of the Sfniu/fh of Wheat Flour made up to the 50 c.c. mark with distilled water and the flask was placed in an incubator kept at 37° C. When the protein was completely in solution, the liquid was filtered quickly into a small flask, which was then stoppered and kept in the incubator. From time to time, the optical rotation of the alkaline solution was determined in a Ir/- polar! metric tube, using a Schmidt and Haensch instrument and sodium light. This procedure was continued for about 300 hours, when the value had become practically constant. Graphs were then con.structed .showing the progress of the racemisation, the ordinates representing specific rotation values and the abscissae the number of hours during which the reaction had been allowed to proceed. In carrying out the determinations in the above manner, some delay always occurred in eft'ccting complete solution of the gliadine in the alkali. The difliculty was satisfactorily overcome in the following way. The 1 grm. sample of dry protein in the 50 c.c. flask was first dissolved in 10 c.c. of 70 per cent, alcohol. The 25 c.c. of .standard soda were then slowly run in, the flask being gently shaken during the process. The volume was then made up to 50 c.c. with distilled water and the flask was placed in the incubator. A clear solution was obtained in a few minutes by this method. The racemisation data of the glutenine samples were obtained in a somewhat similar manner. 1 grm. samples of the dry protein were weighed into 50 c.c. flasks containing a little distilled water. The 25 c.c. of standard soda were then run in slowly and the flask was gently shaken. A jelly like mass was first obtained, which quickly liquefied when the flask was put in the incubator. The volume was then made up to the 50 c.c. mark with distilled water and the polarimetric readings were taken as described above. Investigation of the gliadine samples. 1. Specific rotations in 70 per cent, alcohol. 2 per cent, solutions of the gliadine preparations in 70 per cent, alcohol were examined by means of the polarimeter with the following results: Gliadine from Manitoba flour [ajy^ = — 93-60°. ,, EngUsh „ [a]/, = - 93-78°. Osborne (6) gives the value of — 92-28° for the specific rotation of gliadine in 80 per cent, alcohol. H. E. Woodman 237 It will be noted that the gliadine.s from the strong and weak flours display no difference in respect of their rotations in 70 per cent, alcoholiu solution. 2. Data obtained in the racemisation of the gliadme samples. Manitoba tlour gUadine 2 % in xV/2 NaOH English Hour gliadine 2 % in Nj-l NaOH Time in hours Specific rotation Time in liours Specific rotation 2 -110-5° 2 -111-0 5 - 106-8 5 - 106-3 9 - 104-1 9 - 1040 24 - 96-8 241 - 96-3 49 - 89-0 49 - 89-5 74 - 84-1 73 - 83-9 101 - 78-8 101 - 79-5 125 - 75-1 124 - 75-6 169 - 71-5 170 - 71-3 242 - G7-3 240 - 67-() In the series of determinations recorded under 3, 4, -5 and 6, the weighed out samples of protein were first dissolved in a measured volume of 70 per cent, alcohol as described before the addition of the standard alkali. 3. Manitoba fiour ghadlne 2 % in Nj2 NaOH (alcohol present) English flour gliaduie I in iV/2 NaOH (alcohol present) Time in hours Specific rotation Time in hours Specific rotation 1 - 109-4° 1 - log-o" 5 - 98-3 5 - 97-8 9 - 930 9 - 93-2 24 - 86-2 24 - 86-7 48 - 80-5 48 - 80-6 74 - 76-2 72 - 76-2 125 - 72-6 124 - 72-1 170 - 69-4 170 - 68-9 240 - 66-0 242 - 65-6 6. Manitoba flour gliadine 2 % in Nji NaOH (alcohol present) EngUsli flour gUadine 2 % in Nji NaOH (alcohol present) Time in hours Specific rotation Time in hours Specific rotation 1 -110-5° 1 -110-8° 5 -103-1 5 - 103-1 9 - 100-0 9 - 99-3 25 - 94-7 24 - 94-3 48 - 90-5 48 - 90-5 96 - 88-4 97 - 88-0 119 - 86-8 120 - 86-5 168 - 84-7 170 - 84-0 244 - 82-0 242 - 81-7 ■238 The Chninsh-ii of the Sfntif/f/' "J' W'/naf llmtr The above deteniiination.s wore carried out at least in duplicate and the results were confirmed by independent investigation of further samples of the two gliadines prepared from the same flours. It will readily be seen from the above data that the gliadines from the two types of flour display throughout identical optical behaviour when racemised by dilute alkali at 37°. This fact is brought out more strikingly if the graphs representing the course of raceniisation be con- structed. For each set of experimental conditions, it will be found that one smooth curve can be drawn to satisfy equally the two sets of readings for the Manitoba and English flour gliadines. Such slight discrepancies as may occur fall within the general error of e.xperiment. When racemised by means of Nj'l NaOH, each gliadinc solution possesses an initial s])ecific rotation of about — 114° (taken from graph ^). The rotations of both solutions diminish fairly rapidly and at an equal rate during the first 24 hours; tlie rate of diminution in each case then falls off equally and after 240 hours, when the value of the rotation has become practically constant, the two solutions possess an e(|ual specific rotation of about — (57°. The presence of the small ainoujit of alcohol in the second series of determinations e-xerts a striking elTec't on the rate of racemi.sation and affects both gUadines in an equal degree. The effect is best studied by constructing the graphs. The rate of racemisation during the first 24 hours for iV/2 NaOH is much quicker when alcohol is present than when it is absent. In the presence of the alcohol, the specific rotation during the first 24 hours sinks from — 114° to — 86°, whereas when no alcohol is present, the corresponding diminution is from — 114° to about — 97°. After about 50 hours, however, the two curves begin to come together again and almost coincide after about 250 hours. The presence of alcohol causes Nj^ NaOH to effect a greater reduction in specific rotation during the first day than does iV/2 NaOH without alcohol. These curves cro.ss after about .'}5 hours and from this point the rate of diminution with the weaker alkali is relatively slow. The reason for the effect thus produced by the alcohol, which resembles that of a catalyst, requires further investigation. For the purposes of this enquiry, it is sufficient to note that under the three different sets of conditions, the gliadines from the weak and strong flours display the same optical behaviour during racemisation. This fact is taken as evidence of the identity of the two gliadines. A final test on a mixed sample of the two ^ The graphs are init rrprnthu-ed lioro, as they are not necessary in enabling the essential conehisions tu he drawn. H. E. Woodman 239 gliadines gave readings which fell on the racemisation curves which the two proteins have been shown to possess in common. 3. Vombining weights of the gliadines. Foreman (7) has shown that aqueous-alcohohc solutions of certain amino-acids containing about 85 per cent, alcohol can be titrated accu- rately with alkali in the presence of phenolphthalein. The amino groups under these conditions display no basicity to phenolphthalein and the carboxyl groups can therefore be estimated by titration. On the basis of this observation, it seemed probable that the basic effect of the free amino groups in the ghadine molecule towards phenol- phthalein might be held in abeyance when the protein was dissolved in alcohol, thus permitting of the direct estimation of the free carboxyl groups in the molecule by titration with alkali. Accordingly, 2 per cent. solutions of the gliadines in 80 per cent, alcohol were titrated with iV/10 NaOH in the presence of phenolphthalein, and the following results were obtained : Manitoba flour gliadine 1 grm. dry, ash-free protein required 1-95 c.c. A7I0 NaOH. Combining weight = 5128. Enghsh ,, „ 1 grm. dry, ash-free protein required 1-99 c.c. iV/10 NaOH. Combining weight = 5026. It follows that both gliadines possess the same combining weight, the difference recorded being due to experimental error. This fact affords further confirmation of the identity of the gliadines. The data recorded above possess an additional interest when con- sidered in conjunction with the minimum molecular weight ascribed to gliadine by Osborne (8) on the assumption that the molecule contains five atoms of sulphur, viz. 15560. It would thus appear probable that the ghadine molecule contains three or a multiple of three free carboxyl groups. Investi(!ation of the glutenine samples. 1. Specific rotation in iV/25 NaOH. 0-5 per cent, solutions of the glutenines in iV/25 NaOH were examined by means of the polarimeter with the following results : Glutenine from Manitoba flour [a]^, = — 99-5°. ., English „ [a]^, = - 78-8°. 2. Similar differences in the optical properties of the glutenines were observed during the racemisation of the proteins by means of iV/2 NaOH at .37° C. Joum. of Agric. Sci. xn 17 240 The Chemist ri/ of the Strength of Wheat Flour Munitoba Hon r glutenine English ti lour glutenine 2 % in AT- I NaOH 2% in 5l 2 NaOH Time in hours 1 Specific rotation Time in hours Specific rotation 1 - 930° 1 -740'= 5 -850 5 -67o 9 -79-5 9 -62-9 24 -71-0 24 -57-0 48 -60-5 49 -500 72 -55-5 72 -45-9 96 -51-2 98 -41-5 143 -470 144 -370 190 -44-0 192 -350 240 -41-5 242 -32-5 ."5. The following data show the behaviour of the glutenines when racotuised with N/i NaOH at 37° C. Manitoba flour glutenine 2 % in A74 NaOH English Hour glutenine 2 % in A74 NaOH Time in hours Specific rotation Time in hours Specific rotation 2 - 940° 2 - 77-0° 9 - 85-5 9 -71-6 24 -78-6 25 -650 96 -67-8 96 - 57-5 144 -63-5 143 -540 242 -57-0 240 -48-5 The differences in optical behaviour displayed by the glutenines during racemisation with Xj'2 and iV/4 NaOH point to the conclusion that the glutenine of strong flour is a different protein from that con- tained in weak flour. Two objections, however, may be raised against the evidence on which this conclusion is based. 1. In view of the difficulty of effecting complete separation of two proteins from each other, it is possible that the samples of glutenine are still associated with small amounts of gliadine. and that the high rotation values obtained with Manitoba Hour glutenine as compared with English flour glutenine may be explained on the grounds that the former glutenine contains more of the high rotating gliadine than does the latter. 2. It has not been demonstrated that the glutenines are extracted from the glutens without change. It is possible that the 0-2 per cent. KOH used in the isolation of the glutenine may cause, even at the room temperature, a slow Taceinisati(m of the proteins and thus render un- certain any conclusions which may be drawn as a result of the optical behaviour of the final samples with dilute alkali at 37° C. The first objection cannot be sustained, liowever, since a study of the initial rotations of the proteins in alkali shows that the Manitoba flour glutenine would have to contain relatively large amounts of gliadine to account for its optical behaviour on this assumption. Ex- H. E. Woodman 241 hauistive extractions of the sample with 70 per cent, alcohol failed to reveal the presence of even traces of ghadine. In view of the second possible objection, the following tests were carried out. A 0-5 per cent, solution of Manitoba flour glutenine in 0-2 per cent. KOH, to which was added a drop of toluene, was allowed to stand at room temperature for about a month and the specific rota- tion was determined from time to time. The initial specific rotation was — 95-0", and during the period of the trial this value did not suffer any measurable diminution. It is reasonable to assume, therefore, that the optical properties of the glutenines were not affected as a result of their mode of extraction by means of 0-2 per cent. KOH. Moreover, the rate of diminution of rotation of a 0-5 per cent, solution of Manitoba flour glutenine in N/25 NaOH when kept at 37° C. was exceedingly slow, as is evidenced by the following series of determina- tions : Time in hours Specific rotation 2 - 990' 25 -97-0 98 -910 267 -79-0 It will be observed that the specific rotation of the Manitoba flour glutenine solution had barely fallen to the initial value of the specific rotation of the English flour glutenine even after standing 267 hours at 37° C. Both objections become untenable in view of the fact that similar investigations carried out on further samples of tlie glutenines from the same flours yielded confirmatory results. If the differences observed between the two glutenines arise from the operation of factors involved in the objections 1 and 2, then it would be a remarkable coincidence if such factors should operate with exactly equal eft'ect in the case of the independent samples which were examined. The results of the investigation may therefore be interpreted as de- monstrating the non-identity of the glutenines from strong and weak flours. If, as seems probable, the characteristic physical differences between the glutens of the two flours are related to the differences existing be- tween the glutenine fractions, then it would at first sight appear feasible to produce a "strong" gluten by preparing a moist mixture of English flour gliadine with Manitoba flour glutenine, or a "weak" gluten by mixing together Manitoba flour gliadine with English flour glutenine. Attempts to demonstrate this possibihty, however, met with no success, 17—2 24-_' The Chemistry of the Sfrcnfff/i of Whmf Flour the reasons being twofold : 1 . It is not possible by grinding.' the ])roteins together to effect the same intimacy of mixture as occurs naturally in the flour. 2. The physical properties of the proteins themselves have probably been considerably modified during the process of their isolation as a result of prolonged contact witii different reagents (alcohol, ether, alkali, etc.). The view has been put forward by Kosutany(!») that glutenine is derived from gliadine by the splitting off of water. The i)uantitative work of Osborne in connection with the hydrolysis of these proteins has shown beyond doubt that they are two absolutely distinct substances, and a comparison of the racemi.sation data for the glutenines with those of the gliadines confirms Osborne's view. The gliadines exhibit distinctly different optical behaviour during racemisation from that displayed by the glutenines. Summary and Conclu.sioxs. The gliadine and glutenine proteins from typical strong and weak wheat flours have been isolated and investigated by comparative methods. The gUadines from the two sources have been showTi to be identical proteins. This conclusion, which is in harmony with the earlier results obtained by Wood, has been arrived at on the following grounds: 1. The identity of their optical behaviour when racemised by dilute alkali at 37° C. under three different sets of conditions. 2. The identity of their specific rotations in 70 per cent, alcohol. 3. The identity of their combining capacities for alkali, as determined by titration in 80 per cent, alcoholic solution by means of iV/10 NaOH to phenolphthalein. The glutenines from the two types of flour have been showTi to be two distinct substances, this conclusion being based on their different optical behaviour during racemisation by dilute alkali. It is suggested that the existing ideas on flour strength require modification to include the facts recorded in this investigation. It is desirable to retain the dual conception of strength as put forward by Wood. The factor which determines the size of the loaf is most probably connected with the diastatic capacity of the flour, as was suggested by this investigator. On the other hand, the factor which determines the shape of the loaf and which appears to be directly related to the physical properties of the gluten of the flour, is possibly dependent on the par- ticular glutenine mechanism possessed by the wheat. The results of this investigation suggest that the strong wheat synthesizes one type of glutenine and the weak wheat a different type, H. E. Woodman 24:] whilst wheats of intermediate strength may contain varying proportions of the two glutenines. To a certain extent also, as was demonstrated by Wood, the physical state of the gluten will be conditioned by the acidic and soluble salt content of the flour. It is not claimed that the above explanation can be regarded as final, since there seems no particular reason to hmit the number of possible glutenines to two, nor indeed is it feasible to rule out, on the available evidence, the possibility of the existence of more than one gliadiue amongst the wheats. It is hoped to continue the investigation further along these lines. It has been shown that ghadine possesses a combining weight of approximately 5000, and from this the conclusion has been drawn that the ghadine molecule contains three or a multiple of three free carboxyl groups. It has been demonstrated that a solution of glutenine in 0-2 per cent. KOH undergoes no measurable diminution in optical rotation on standing over a long period at room temperature and that only a slow change occurs in iV/25 alkahne solution at 37°. It thus appears probable that proteins can be extracted by means of 0-2 per cent. KOH without suffering change, provided the alkahne extracts are kept cool. The writer would hke to avail himself of this opportunity to exjjress his thanks to Professor T. B. Wood, C.B.E., M.A., F.R.8., at whose instance this investigation was undertaken and whose advice throughout has been of material assistance. Also to Dr. A. E. Humphries, with whose help the writer was able to secure the samples of flour used for the preparation of the proteins. The sample of Manitoba flour was supphed by Messrs John White and Sons and the English flour by Messrs Soundy and Co. To both these firms the writer's thanks are due. REFERENCES. (1) HuMPHBiES and BrFFEN. J. Agric. Sci. 2, 1, 1907-8. (2) Hali,. Report of Home-gromn Wheat Committee, 1905-6. (3) Wood. J. Agric. Sci. 2, 139, 1907-8. .1. Agric. Sci. 2, 267. 1907-8. (4) Dudley and Woodman. Biochem. J. 9, 97, 1915. (5) Woodman. Biochem. .J. 15, 187, 1921. (6) Abderhalden. Handbiich der bioch. Arbeitsmethoden, 2, 320, 1909. (7) Foreman. Biochem. J. 14, 451, 1920. (8) Osborne. The Vegetable Proteins, p. 70 (1909). (9) Kosutany. J. Landw. 51, 130, 1903. [Received June 6th, 1922.) ON THE USE OF ARTIFICIAL INSEMINATION FOR ZOOTECHNICAL PURPOSES IN RUSSIA. By E. I. IVANOFF, Director of the Central Experimental Breeding Station for DonieMic Animals, Moscow, Russia. One of the greatest problems of Russia's present economic policy is the restoration and development of farming, and in particular, cattle- farming. The war and revolution have, together with other things, destroyed enormous numbers of cattle, horses, pigs, etc., and thoroughly undermined the meat industry, as well as the sources of supply of working animals. A decrease of 50 % below the former numbers of horses is the common state of things. The preservation of cattle-farms and studs is seldom met with, and the number of stock-producing animals is at least ten times less than formerly. The terrible drought threatens to bring into this sphere of national wealth even greater destruction. At the same time there can be no doubt that without the restoration and maintenance on a definite level of stud and cattle-farming, Russia cannot return to full economic activity. The present-day state of things demands that every effort be made, every possibility found and utilised, for increasing the number of domestic animals and for improving the methods of breeding and the breeds themselves. Mass-breeding of domestic animals must certainly go hand-in-hand with mass-improvement of the breeds. If indiscriminate raising of animals was unprofitable before the war, at the present time, with undreamt-of prices of fodder and labour, it is certainly a loss. The greatest obstacle in the way of a successful solution of the above problems is the shortage of progenitors suitable for stud and cattle- farms. Let us take for example horse-breeding. Before the war, the number of thorough-breds in Russia was not even 1 %, and for one stud-stallion, there were 600 mares. At present the difference between the demand and supply of valuable stallions can be expres.sed approxi- mately as 1 : 3000. But the number of mares served by one stallion in the pairing season is from 10-40, seldom higher, and on an average 25-30. E. I. IvANOPF 245 Consequently, in order to jirovide for all the available mares stud- stallions, it woidd be necessary either to increase the number of the latter at least a hundred times, or to increase the productivity of the present number as many times. The former, in view of the present economic conditions, cannot be realised even to the extent of 1 % of the requirements; the latter, though not fully, yet to a considerable degree can be reached by putting systematically into wide practice the method of artificial insemination which enables more than 300 mares to be inseminated during the pairing season by one stallion, instead of 25-30. The same applies to all other live stock. The aim of the present article is to show what has been achieved in Russia with this method and to indicate the possibilities connected with the practical side of this work. Under the term "artificial insemination" as applied to Mammals, and in particular domestic animals, is understood the introduction, by artificial means, of the semen of the male into the vagina of the female. The seminal cells or spermatozoa can be introduced either (1) in their natural medium, i.e. in the secretion of accessory sexual glands, or (2) in an artificial medium (physiological solution, Locke's fluid, serum, etc.). In either case the essential conditions of fertilization are the natural fundamental jirocesses and conditions: the maturity of sexual products, for examjale, their viability, etc. must remain unchanged. The only difEerence is that in artificial insemination the fate of the secreted male sexual cells is held in the hands of man, who can divide the collected material into parts and inject it into the female sexual organs by means of instruments (catheter, syringe). Thus, in the case of Mammals, strictly speaking, it is not a method of artificial fertilization, but of artificial insemination. The possibiUty of artificial impregnation by means of natural sperm was demonstrated long before my work by the famous Italian man of science, Spallanzani. The small number of carefully worked experiments on animals and lack of experience on the technical side for a long time prevented this method from gaining due importance in the practices of applied zoology. With regard to the second method of insemination, i.e. by means of an artificial medium, the very possibility of such a method was denied by some until my works appeared. Those who wish to get acquainted with the history and technical side of the subject will find detailed in- formation in these works and in the memoirs referred to at the end of this paper. ...... . . .. . , 240 Artijirial Insemination The practical iiu])()itaiice of artificial insemination of domestic animals is to be found in the possibility by means of this method to utilize the seminal fluid, secreted by the male when covering the female, for the purpose of inseminating a number of other females (10-20) who are "on heat"; to combat the barrenness of females caused by various mechanical obstructions (stenosis of the neck of the womb, deflexion of the neck of the womb, polypi, etc.); to cro.ss animals dirt'ering very greatly in size and weight; to cross various ty]ies of animals (horses with asses or zebras, cows witii bisons or aurochs, etc.); to utilize the repro- ductive capacity of a valuable male in case of fatal injury or even death of the latter resulting from causes of a non-infectious kind. (In the latter circumstance, the seminal fluid is collected from the excised .se.xual glands of the male, diluted with some solution beneficial to the life of the seminal cells, and injected like natural seminal fluid into the vagina of the female.) One of the great advantages of artificial insemination is to be found in the fact that it dispenses with the necessity of bringing a valuable male into close contact with an unknown female, as the semen can be obtained vnth the aid of a well-known female or one specially selected for the purpose. This circumstance is particularly important in areas where such diseases as dourine, glanders, etc. are met with. We must also point out that in artificial insemination when the presence of trypanosomes is suspected in the sperm, the possibility of making the seminal liquid free from infection without killing the sperma- tosomes, has in principle been proved. Finally, in artificial insemination the whole process takes place under the control of the microscope, which makes it possible to determine in every individual case before insemination the actiuil degree of mobility of the seminal cells and their number. This enables the breeder to follow and to determine the jjroductive abilities of the male before his stud career commences, and not after, as was usually tlie case in natural insemination. The greatest j)ractical importance of the method of artificial insemina- tion is to be found in the possibiUty of applying it for purposes of mass- raising of domestic animals and fullest utilization of particularly valuable males. In order to secure for this method wide practical aj)plication it was necessary to work out simple and safe technical means, to verify them on a sufficient number of animals, ascertain the number of possible inseminations from one "leap," the percentage of positive results, the strength, fecundity and working ability of the young, and convince one- E. I. IVANOFF 247 self that this method will not create an attitude of mistrust in the peasant population. Many years were devoted to these joroblems — beginning with my first experiments on the Dubrovski stud-farm in 1899. The practical work was carried on in the great majority of cases on horses. On cattle, chiefly on sheep, artificial insemination was practised to a fairly large extent, but usually with a scientific purpose. Experiments were also conducted on dogs, foxes, rabbits, birds and other animals. In the following, I shall submit data only from work on horses. In the history of the development of this work in Russia, two main stages must be noted, the first being the period of experimental pre- jmratory work on the problem under laboratory conditions (Institute of Experimental Medicine, Zoological Laboratory of the Academy of Science), and under conditions of practical life from 1899-1909 (on the special exj^erimental station of the Department of State Stud-Farming, in the Government of Orel, Livenski district, village of Mijnee-Dolgoe, and on the estate of Askania-Nova, in the Government of Taurida, formerly belonging to Faltzfein). Only after dealing with the fundamental problems and printing the results of the experiments, had the task of advocating this method for wide use been undertaken. The Government opened a special laboratory with a physiological section and a zoo- technical station, in Askania-Nova, attached to the Laboratory of the Veterinary Department, where from 1909 special theoretical and practical courses of study were pursued on the physiology and biology of insemina- tion and a body of specialists, mainly veterinary surgeons, was being trained for the practical a^jplication of this method on stud-farms and pairing stations. PreiJaratory experimental work was carried out apart from other animals on 579 horses, and during that period, together with problems of direct, practical interest (percentage of conceptions from artificial insemination, number of possible inseminations from one "leap," etc.) problems were experimentally investigated which have a more remote connection with stud-farm interests. As the manner of carrying out the exjjeriments and their results were made clear in my book Artificial Insemination of Domestic Animals, published also in German, I shall confine myself here to a very bare outHne. The above work shows that artificial insemination enables the number of mares inseminated by one stallion during the pairing season to be increased on an average ten times, and the percentage of foaling, if the work is done correctly and under conditions usual on pairing 24H Aiiifirldl IiiaciniiKttloii stations (stallion of undoui)te(l fertility, mares healthy, pairing season about three months, condition of mares ascertained and repeated in- semination effected) is higher than in natural coition with the very same stallions and reaches on an average 78 % . Single insemination, without repeated injections to follow it during the period of " heat," gives only 25 % of conceptions. Conception as a result of the mare being insemi- nated outside the period of "heat" is an exception. In cases of persistent barrenness, due to some anatomical irregularity in the construction of the mare's genital organs, artifical insemination is a very efficient means of bringing about conce])tiou. The technical side of artificial in.semination is sufficiently simple to be used under conditions prevalent on Russian pairing stations, to say nothing of stud-farms, and can be applied not only to horses but also to other domestic animals. Artificial insemination, when the necessary precautions are taken, removes the possibility of infection. Pregnancy and delivery are normal. The young in appearance, size, strength, capacity for work and fertility do not deviate from the normal, and such broods have produced a number of winners on the race- course (Dubrovski stud-farm), and in Askania-Nova over 50 horses, produced by artificial insemination, were bought for the cavalry and artillery at highest prices. Finally, it was proved beyond doubt that the peasants are quick to learn the advantages of such a method and that its wide use all over Russia is ultimately, therefore, fairly certain. With the establishment in 1909 of the above-mentioned laboratory (Physiological Section of the Laboratory of the Veterinary Department and Zootechnical Station in Askania-Nova) the aims of wliich, in addition to the investigation of scientific and practical problems of the physiology and biology of insemination, included also the popularization of the method of artificial insemination and aid for institutions and persons desirous of using this method — the possibility of acquiring fuller and wider knowledge of the method was afforded to specialists (veterinary surgeons and zootechnicians) and amateurs. A special course of lectures was given here on the physiology and biology of insemination, on problems of heredity, and practical work was done on artificial insemination of domestic animals. During the period from 1909 to 1919 over 400 persons, mainly veterinary surgeons, availed themselves of the opportunity of becoming familiar with artificial insemination. Some of these people have made practical use of this method. The work of the physiological section, in particular the work in connection with artificial insemination, attracted attention abroad, and a number of specialists (professors and surgeons) were sent to Russia E. T. IvANOFK 249 to gain acquaintance with this method {vide Review of the Activities of Phi/siological Section during 1909-1913, p. 22). • In 1914 the work was interrupted in consequence of the army's demand for veterinary surgeons, and many of the latter had no oppor- tunity of submitting the results of their work in 1913. The work done in Russia represents the first attempt to use artificial insemination in mass-raising of domestic animals, an attempt, as will be seen later, by no means perfect in regard to conditions; without systematic organization or strict compliance with definite instructions. The laboratory in my charge was a purely consultative institution. As was subsequently shown, that state of things made possible a great number of very material and injurious digressions from the technical method elaborated by me, and the work was often confined to one or two visits of the veterinary surgeon to a stud-farm, once or twice a week (in the course of one, two, or three weeks) and in the majority of cases was regarded as a side-issue. The data at our disposal refer only to the period 1909-1913 (data for the latter year are by no means complete) and were collected partly by means of a questionnaire, partly from reports of the Rural Councils. The questionnaire contained a number of questions, the answers to which should have given a clear idea of the conditions and results of the work. In deahng with the material of the questionnaire only those enquiry forms were made use of, the information of which was sufficiently full and left no suspicion as to the correctness of figures. The above material shows the general state of things to be as follows: during the period 1909-1913 inclusive, the method of artificial insemina- tion has been used in more than 30 governments of European Russia and in Siberia. , - . . ■ . •■ • Even in those cases where artificial insemination was introduced, not on private initiative, but as a measure of mass-improvement of the peasants' horses (governments of Kherspn, Ekaterinoslav, Taurida, Bessarabia), the organization and management of this work suffered from all kinds of irregularities; for example, at the head of affairs were often found persons without any or with second-hand knowledge of the subject, incompetent people; every worker could at his discretion intro- duce various changes into the organization or technical side of the work. Particularly injurious with regard to the percentage of conceptions was the idea that it is possible to carry out work successfully without regard to the absence of "heat" in the mare; also excessive simplification of the technical side, including in some cases the use, instead of index-cylinders 250 Artifieial Insemination aud pharmaceutical scales, of implements suitable only for home use. As artificial insemination of horses was included in the daily work of veterinary surgeons, often being done without any remuneration, only one day a week was often devoted to it, and sometimes the whole work did not go further than one or two visits of the surgeon. For anyone familiar with the conditions of pairing it is clear that work is impossible under such circumstances, as, apart from the method of insemination, they must lead to a considerable reduction of the percentage of concep- tions, owing to the impossibility of repeating the insemination on the fresh appearance of "heat" in the mare. In other cases, the lowering of the proportion of conceptions was due to injections into the womb being repeated every 7-9 days, wliicli was bound to cause an early abortion where conception had already set in. Further, at many stations for artificial insemination, stallions were used which were known to be un- suitable for this purpose, as in natural insemination the percentage of conceptions from them did not exceed 13 and in some cases was less than 4. It is true that, as has been ascertained, the percentage of conceptions from the same stallions with artificial insemination rose from 4 to 11, from 5 to 22, from 13 to 21, but the use of such specimens was bound to lower the general proportion of conception from artificial insemination. Finally, artificial insemination was performed mainly on mares very unreliable from the breeder's point of view, or known to be sterile. The work was carried on chiefly on Rural Councils' pairing stations, in villages tens, and sometimes hundreds, of miles distant from town or railway station. During the period 1909-1913 artificial insemination was performed, according to the data supplied by the questionnaire, on 6804 horses*. In 1909 on 3 stations 57 horses. 1910 „ 8 )) 288 1911- „ 31 J> 2285 1912 „ 41 j?> 3397 1913 „ 17 >? 777 The information for 1913, as pointed out above, is by no means complete. With regard to the normal condition of the young, freedom from infection of the sexual organs of the mares in the practice of artificial insemination, low proportion of abortions, etc., the data supplied by the ' The above figurea refer only to European Russia. E. I. IVANOFF -251 questionnaire are of a very uniform character and confirm all the con- clusions arrived at in the preliminary experimental work. Artificial insemination during the above short period succeeded in gaining the confidence of the ])opulation. In some cases the number of horses brought to the pairing station during one pairing season ran into hundreds. It has been ascertained that the number of mares fertilized during the pairing season by one stallion can average 300 and over when the method of artificial insemination is used. Sometimes the data of the question- naire sup])ly indications that the young born were of a superior type, or in any case perfectly sound. Only one case of deformity is recorded out of over 2000 foals born, which proijortion is quite normal. As regards the proportion of positive results, the figures, as one could expect from the foregoing, are by no means uniform. The percentage of conceptions varies between 4 and 907 and averages about 40. For 1909 the proportion of conceptions was 40 % „ 1910 „ „ „ 40-3 % 1911 35-5°, 0 • 1912 41-4 "/ 1913 4''-3 °/ When we look into that disparity in the results of the work of difierent people, we find that it was caused by those irregularities, which crept into the work, as the data of the questionnaire clearly show. For example, if we were to divide the material at our disposal into two sections, one under the heading "Artificial insemination performed on mares in period of 'heat,'" and the other "Artificial insemination per- formed regardless of presence or absence of 'heat,'" we should find that in the first group, where a number of other irregularities and deviations from proper conditions of work remained and only that factor of "in- semination without 'heat'" was excluded, the proportion of positive results equalled 49 % , while in the second group, where that factor remains, the percentage of positive results was 28. When we examine in both these groups the influence of the factor of duration of the pairing season, we find that when The duration of pairing season is The % of conceptions Less than one month for 1st group 38-6 ., 2nd „ 19 Over one month ,, 1st „ 54 „ 2nd „ 29 262 Artijirial Insemination With more careful organization of the work and reliable parents, the proportion of conceptions from artificial insemination reached in the work of my pupil-practitioners not only the figures shown above (78 %), but even exceeded them, rising to 87-88 and even 90 % . Table I. Artilicial inscminatiiiri took place under the following conditions: (I) the mares were *'on heat," (2) the pairing period for artificial insemination was lonyer than one month, (3) the complement of marcs was the usual one for a stud-farm, no mare being known to be sterile, (4) the stallions were kno^Ti to be reliable. a o c " ESj aS Eg S fS.2 J.S f2 f Year, Kovernment, district, station 0*^5 "S— o"^ o „ anil name of person in charge do« ox ds 02 a — of operations ^XS '^u ^.* ZS o^g5 1912. Don Army Province, Provalski Army Stud-Farm, vet. surgeon V.'Y. Kartasheif So 17 11 2 ()0-7 1913. Don Army Province, Provalski Army Stud-Farm, vet. surgeon V.V. Kartashcff 31 22 3 (5 88-0 1910. Govt, of Saratoff, dis- trict and town of Kuznetzk, pairing station, vet. surgeon N. A. Zhukoii 3(i 25 II 0 09-4 1911. (jiovt. of Saratoff, dis- trict and town of Kuznetzk, pairing station, vet. surgeon N. A. Zhukoff ... ... 27 17 10 0 62-9 1912. Govt, of Saratoff, dis- trict and town of Kuznetzk, pairing station. \et. surgeon N. A. Zhukoff ... ... 10 7 3 0 700 1912. Govt, of Kharkoff, dis- trict of Akhtyrsk, estate of Konig Bros., Trostinnetz, assistant-surgeon Gafner ... 32 28 4 0 87-5 1913. Govt, of Kharkotf, dis- trict of Aklityrsk, estate of Konig Bros., Trostinnetz, assistant-surgeon Gafner ... 43 39 4 0 90-7 Total ... ... ... 209 \na 40 8 771 It is clear from the above table that the proportion of conceptions in artificial insemination is usually higher than that in natural pairing of the same progenitors. The following table of results of artificial insemination of hor.ses on the zooteclmical station of Askania-Nova (1912) illustrates that fact still more clearly. E. I. IVANOFF 253 Table II. Results of artificial insemination of horses on the zootechnioal station of Askania-Nova in 1912 Results of natural insemination by the same stallions in the same year NvT of mares ' ' No. of mares sub iected to No. of "(, of o/ nf No. of subjected to Name of artificial in- concep- concep- concep- concep- natural in- stallion seminat ion tions tions tions tions semination Harry 11 9 81-8 40 2 5 Freiherr 7 4 57-1 .50 2 4 Irkutsk 11 11 100 — — — Wallenstein 4 4 100 50 4 8 Gadai-Zille 6 3 .'50 .33-3 6 Total 39 31 84-1) 43-4 10 23 Thus, the practical work of veterinary surgeons, carried out on some thousands of horses in various parts of Russia and generally under very primitive conditions, bears out all the fundamental conclusions of the preliminary investigations and shows the great usefulness and value of artificial insemination of domestic animals for the population. Reports and the material of the enquiry clearly indicate that where the work was conducted more or less succes.sfully, where the veterinary surgeon was able to devote to the artificial insemination of horses sufficient time and attention, there the number of horses inseminated durinc the pairing .season on one station reached sometimes 850, and the majority of these in such a case were peasants' horses. In the estimation of Rural Councils (Zemstvos) that work was gaining greater and greater importance, and in the year before the war meetings and conferences were being convened for the special purpose of discussing the problems of more efficient organization of the work of artificial insemination, in connection with mass-improvement and mass-raising of stocks of horses and other domestic animals; also for the purpose of preparing a definite plan of work and publishing special instructions binding for all those doing practical work in the subject (governments of Kherson, Taurida, Bessarabia, etc.). The interruption of the work, which was getting into stride and assuming large proportions, was due entirely to the war and revolution, and as soon as problems of organization in cattle-farming again came to the front, artificial insemination became a matter of prime importance, as shortage of procreators was the greatest obstacle in the way of all former endeavours. Already in 1919 at one of the All-Russian conferences of zootech- nicists of the Department of Cattle-farming of the People's Commissariat 254 Artificial 1 nsi ininatloii of Agriculture, aud later on, at the subsequent AU-Russian conferences on practical horse-breeding, the question of introducing the method of artificial insemination was considered to be of immediate imjjortance. However, lack of instruments, which were lost during the revolution, and the general economic and political -situation prevented work being commenced at once. The Collegium of the P. C. A. decided to open a "Central Experimental Station for investigating the problem of raising domestic animals," and one of its fundamental practical aims was the organization of the woik of artificial insemination. Ju 1921 the station received a great number of enquiries and re(]uests for instruments for artificial insemination of horses from the Don Province, Kuban Province, Turkestan, governments of Voionezh, Smolensk, and other places. In some places, like the Don Province, or government of Voronezh, it was proj)Osed to place the work of artificial insemination on a broad basis. In the Don Province a special laboratory has been opened ; in the govern- ment of Voronezh there are equipped stations, but the number of in- .struments and suitable horses is insufficient. Owing to the accumulation of imavoidable and unfavourable circumstances, the work of artificial insemination of horses could not be commenced in the spring of this year (1922), mainly in consequence of the impossibility of obtaining in ])roper time instruments ordered abroad. By spring 1923 the training of a staff of specialist-technicians should be completed and a number of stations opened, in the first place, for the artificial insemination of horses. With the practical devel()])ment of this method are clo.sely ccmnected also its economic possibilities. The price of good progenitors should rise and cattle and stud-farming should attract the attention of the popula- tion and of ca])italists who should find here a good investment. While formerly horse-breeding, for example, was in Russia by no means a ])rofitable business, but rather a necessity or a hobby, at the ])reseut time, when a stallion during the pairing season can bring a return several times greater than its cost and leave the owner a profit of some hundreds, and sometimes thousands, of gold roubles, the advantage of breeding a good stock is obvious'. In such circumstances there is a possibility of attracting to the work of mass-breeding and improving the stocks of domestic animals private enterprise, granting it one or another privilege and guarantee eliminating great risks. It would be possible to try to ' Kxample: expenses in connection witli equipment of station for artitioial insemination 200 gold roubles, cost of two .stallions, 1000 g.r., upkeep of horses for four uiontlis, iOO g.r., wages to groom for four months, 200 g.r., stabling, 100 g.r., total, 1700 g.r. Receipts: artificial insemination of 600 mares at 5 g.r. per mare; total, 3000 g.r. E. I. IVANOFP 255 create in Russia great liorse and cattle-ranches. In spite of the losses during the war and revolution, there are still in Russia millions of domestic animals. The females are preserved in greater numbers and the main shortage is in males. The latter could partly be brought from abroad and partly increased in number through artificial insemination. In any case Russia, with her enormous open spaces, her extraordinary variety of climate, soil, geographical surface, her wealth of still untouched zootechuical material and the millions of her population — chiefly peasant population — provides endless opportunities for the development of all kinds of animal-breeding on a large scale, from horse-breeding to, and including, the breeding of deer and silver foxes, in the case of which artificial insemination acquires particular importance, as foxes are monogamous. It must also be remembered that with the use of artificial insemination on domestic animals we not only gain material advantage, but also save time. Formerly, in order to create a local breed or even a considerable uniform herd, from one or another valuable progenitor, it required a period of several decades. Now, with artificial insemination, this can be. achieved in a much shorter period of time, and a surplus of progeny makes possible a more careful selection. Artificial insemination is bound in the near future to assume additional import.ance in the zootechuical practice in connection with the solution of two fundamental problems now being investigated, namely (1) in- creasing the number of females fertilized during the pairing season through artificial insemination by one male, and (2) preserving for some time and forwarding to distant places the seminal solution. The first 2>roblem has already been to some extent solved by my experiments and the observations of my pupils, veterinary surgeons, and for practical application it needs only to be verified on a number of animals. The second problem, theoretically quite reasonable, needs for its solution the investigation of a number of questions. The possibility of preserving alive the seminal ceils of Mammals outside the organism for many days and even weeks has been demon- strated by the work in our laboratory, as well as by other investigators ; this, however, is true only of seminal cells taken, with precautions of sterihzation, directly from the epididymes of the seminal gland. In this case it is sufficient to prevent them from drying up and to place them in a temperature of 1 or 2° C. But for practical purposes, it is most important to preserve the seminal cells not in the above state, but in the medium Journ. of Agric. Sci. xu 18 256 Artijicial Insemination of the secretions of the accessory sexual glands (natural sperm). In this medium there can be no question of sterilization and, what is particularly important, the seminal cells coming into contact with the secretions of the accessory sexual glands change their biological properties — acquiring a maximum of energy and dying in a comparatively short time (a few hours). Some methods of elucidating the nature of factors determining these properties are already mapped out and the solution of this problem, we are entitled to think, is only a question of time. There remain a few words to be said on the development of the technical side of the work. In view of certain inconveniences connected with the sterilization of the sponge collecting the sperm when tlie male covers the female, I have substituted for the 2 % solution of sodium carbonate a solution of spirits of wine (60-65 %) in which the sponge is kept about an hour, then thoroughly cleansed with sterilized water, and finally washed several times (being repeatedly squeezed out in the press) with a sterilizing physiological solution of kitclien salt, or a 10 % solution of refined sugar. This method of sterilization has been used in practical work by veterinary surgeons since 1911. Before this innovation was recommended for actual use, it underwent a number of suitable tests and investigations in the laboratory and on the experimental station in Askania-Nova which have shown its usefulness. LITERATURE. Heape. The Artificial Insomination of Mammals {Proc. Roy. Soc. 61, 1897). ■ The Artificial Insemination of Mares {Veterinarian, 1898). IvANOFF. De la F^condation Artificielle chez les Mammiferes {Arch, des Sciences Biol. 12, 1907, St Petersburg). De la Kcondation Art. clicz los Animaux domestiques {Arch, de Vet. Sci. 1910, St Petersburg). Die Kunstiiche Befnichiung der Haustiere, Hannover, 1912. Beschreibung von Ilybriden, etc. {Zeitschr. f. Induktive Abslammungs- und Vererbungslehre, 16, 1916). Marshall. The Physiology of Reprodiiction, 2nd ed., London, 1922. MabsH;\ll and Croslani). Sterility in Marcs, etc. {Journ. of Board of Agric. 24, 1918). Wolf. The Survival of Motility in Mammalian Spermatozoa {Journ. of Agric. Sci. 11, 1921). {Received June '21st, 1922.) THE EFFECT OF CHANGE OF TEMPEPvATURE ON THE BASAL METABOLISM OF SWINE. By J. W. CAPSTICK, O.B.E., M.A., Felloiv of Trinitij College, AND T. B. WOOD, C.B.E., M.A., F.R.S., Drapers Professor of Agriculture. {From the Animal Nutrition Institute, School of Agriculture, Cambridge.) Introduction. It has long been generally accepted that an animal requires more food in cold weather than in hot weather, that, in fact and within certain limits, an animal's food requirement increases as the temperature falls. The number of precise measurements however of the effect of change of temperature on food requirement is very small, and it was with the object of extending the knowledge of this important subject that the following investigation was undertaken. The writers' attention was directed to the subject by a paper by Armsby and Fries on " Net Energy Values and Starch Values " which appeared in this Journal in 1919'. In this paper Armsby and Fries point out that their Net Energy and Kellner's Starch Equivalents really measure the same thing, namely the amount of energy in feeding stuffs which is available to the animal for physiological purposes. They claim further however that this net energy only is available for maintenance, and that any excess of energy above this amount is converted at once into heat ivhich is of no value to the animal. The point will perhaps be made clearer by a concrete instance, and since figures on the subject are lacking in the case of swine, it may be permissible to supply them for a steer. Armsby states that the basal metabolism of a 1,000 lb. steer is about six therms or 6,000 calories per day. Most European authorities agree that 14 lbs. of average meadow hay supplies a maintenance ration for a ' 1 This .Journal, 9, 182. 18—2 258 Basal Metabolism of Swine 1,000 lb. steer. Now 14 lbs. of average meadow hay supplies about 11,500 calories of mctabolisable energy and about 6,000 calories of net energy. Armsby and Fries contend that only the 6,000 calories of net energy are of service to the animal, the remaining 5,500 calories being wasted. Many European writers, amongst them one of the authors (T. B. W.), have assumed that under certain conditions the whole 11,500 calories may be of service to the steer, the 6,000 calories of net energy sufficing for physiological purposes, which presumably will not vary greatly witli changes of temperature, and the balance of 5,500 calories .saving the oxidation of further material to meet the increa.sed energy required to maintain body temperature in unusually cold surroundings. It is significant to note that whilst Armsby's measure- ments were made in general in a calorimeter at about 65° F. or summer temperature, most European experiments on the maintenance of steers have been made in uiiheated stalls or sheds at winter temperatures of between 40° F. and 50° F. Such a difference in temperature, amounting to from 8 to 14" C. may possibly account for the difference of opinion between American and European investigators. It was with the hope of throwing some light on this point that the writers undertook the investigation described in the following pages. Unfortunately their calorimeter was not large enough to allow of experiments with a 1,000 lb. steer. The experiments were therefore carried out on a hog. The investigation also deals with another point. Previous investiga- tions on the effect of temperature on metabolism have shown that at ordinary temperatures an animal maintains its body temperature approximately constant by the rearrangement of its blood circulation. When the air temperature is low, the blood is deflecttid from the skin to the internal organs: the skin gets cold and loses less heat, but the internal temperature is maintained. When the air temperature is high much blood is sent through the skin where it loses much heat and cools the internal organs. This method of regulation is however insufficient to maintain a con- stant body temperature when the air temperature falls below a certain point which has been called the critical temperature. Below this point, a fall in the air temperature necessitates the oxidation of further material. The basal metabolism will therefore attain a minimum at the so-called critical point. Below this point it will increase by the heat value of the extra material oxidised. J. W. Capstick and T. B. Wood 259 Experimental. The experiments described below were carried out in the large calori- meter described by one of the writers (J. W. C.) in a previous number of this Journal'. As the method of working is fully described in that paper nothing need be said here as to the details of the manipulation. The exjjeriments were made on a Large White pedigree hog bred by Mr K. J. J. Mackenzie on the Cambridge University Farm. The hog was born on January 2(jth, 1921, and was castrated on April 1st. The experiments began at the end of November when the hog was 10 months old and continued to the following April. The food given to the hog was of the .same general character through- out the experiments and was gradually increased so as to be roughly proportional to the two-thirds power of his weight. As an indication of the nature and amount of his food, he received when his weiffht was o 300 lbs. the following daily ration in two meals : 2 lbs. sharps, 2 lbs. barley meal, 1^ lbs. bean meal, 1 lb. maize meal, 1 lb. fish meal. In the intervals between the fasting periods, which lasted from four to six days, he was kept in a small paddock in the open air with a shelter of wattle hurdles. In severe weather he was put in a sty in the labora- tory. As he stripped his paddock bare of vegetation very quickly he was given a little green food each day. The care of the hog was in the hands of Capt. J. S. Morgan who kept him in excellent condition throughout. The fact that the hog gained 157 lbs. in weight in the course of the experiments, which lasted 140 days, in spite of his being without food for about a quarter of that time, is sufficient indication that the fasts had no ill-effect on his health. The hog was put in the calorimeter about 9 a.m. and remained there in darkness and without food, usually for five days. Regular sup- plies of water were introduced from outside. During this period readings were taken whenever the galvanometer curve showed that he had been asleep for a sufficiently long time to get rid of the effects of muscular activity. From these readings the heat evolution was calculated. No analyses were made of the excreta nor was the respiratory exchange observed. After each fasting period he was given about a fortnight to recover 1 This Journal, 11. 1921, 408. 260 Banal Metabolism of Sivine before being put in the caloiimeter again. This proved to be long enough to keep him in good health. The heat given off by the hog was measured in the usual way by observing the rise of temperature of a stream of water circulating round the calorimeter, additions being made for the latent heat of the water vapour brought out in the ventilating air, for the sensible heat brought away by the ventilating air and for leakage of heat through the walls of the calorimeter. Details of the apparatus and the methods of measurement are given in the paper quoted above. It is sufficient to say here that the rise of temperature of the circulating water is measured by a thermoelectric couple which traces a continuous record on a Cambridge and Paul Thread Recording Galvanometer. This continuous record has proved to be a very valuable feature of the apparatus. The writers consider that it has enabled them to obtain more accurate measurements of the true Resting Metabolism than have been obtained hitherto. The curve shows at a glance whether the hog is asleep or nul, how long he has been asleep and whether his metabolism is rising or falling. The slightest movement of the hog is recorded at once and the nature of the hump on the curve often reveals its cause. An instance of this is given in Fig. 1. Moreover there are occasional lapses in the apparatus. An electric current may fail through a bad contact or other cause. The thermostat may strike work through a heater burning out, dirty mercury causing the relay to go out of action, etc. After some experience it has been found that all these have their ch.iracteristic effects on the curve, so that an observer watching the galvanometer can go at once to the source of the trouble and remove it before any harm is done. In the earlier experiments the hog was very regular in his habits. He seldom slept in the day-time but kept in continual motion, his metabolism being very irregular and mounting gradually higher until about 5.30 or 6 p.m. when he went to sleep, and usually did not stir antil about 6 o'clock on the following morning, except that at some time in the earlier part of the night he stood up to empty his bladder. Fig. 1 is the curve obtained on the second day of the experiment which began on January 5. It has been chosen as showing several of the characteristic features of these curves. The irregularity of the curve in the day-time is much the same as is found in most of the curves, except on the first day of the fast when the hog often slept a part of the day. The small hump at lO.oO p.m. is due to the hog's rising to empty Journal uf Agricultural ycienue, Vul. Xii. Part 3 Jun Utli, 1< )>2 ■ .- ■ *-^ -.- ^ \ ■" '"Xv.^ ■ ,- -■---. .■.■~-- ■-■' _ 'f 9 10 II Noon 1 2 3 4 5 6 7 8 between pp. 200^21) 1 Jan. Vth, \9-2-2 Mid I 2 3 4 5 6 7 9 10 J. W. Capstick and T. B. Wood 261 his bladder. Its shape is quite characteristic of urination. The warm urine falling on the floor of the calorimeter causes a rapid rise in the temperature of the outlet water followed by an etjually rapid fall as the urine cools. This fall is checked by the rise of metabolism due to the movement of the hog, which does not show itself so quickly as the rise due to the warm urine. The peak at 1.15 a.m. was caused by a small movement of the hog. It is obvious that it is impossible to take any measurements of the resting metabolism in the day-time or the earlier part of the night. When the hog has gone to sleep it takes many hours for the metabolism to sink to its correct resting value uncomplicated by the effects of muscular activity. The fall during the night is not entirely due to recovery from the day's activity. It is prolonged by the lag in the galva- nometer readings arising from the heat capacity of the calorimeter. This heat capacity, whilst not altering the total area of a hump, reduces its height and sjireads it over a greater time. Further it is known that the body temperature in man falls to a minimum in the early hours of the morning and it is possible that there is a similar fall in swine which would in itself cause a fall in the heat evolution during the night. The final conclusion from the study of these galvanometer curves is that it is only on rare occasions that observations of the resting meta- bolism free from the effect of muscular activity can be made at any other time than in the early hours of the morning. The galvanometer records have been taken continuously through the whole period of the fast. The readings of the various thermometers, etc. needed for calculating the heat evolution have in general not been taken during the day. Through the night they have been taken at hourly or occasionally at half-hourly intervals, unless the curve showed that the readings would be useless. At the end of each day the curve and the record of the readings were carefully studied to find the time or times at which the metabolism most nearly approached a steady minimum. It is not possible to be biased in making the selection for the calculation is so long and involves so many readings that it is quite impossible to foresee the result before the calculation is completed. Most frequently only one point was selected from the day's records — occasionally two or three — on a few occasions no point showed sufficient steadiness. The metabolism was in a few cases calculated at points within 24 hours of the hog's entering the calorimeter, but these early readings have less weight than those taken later. The hog never really settled down until the second night. 262 Banal Afefabolism of Strine Experiments have been made at a series of temperatures ranging from 10-3° C. to 23-7° C. It is not possible to state exactly what was the temperature of the hog's surroundings. Take for instance the experiment at what we have called 13'3° in Fig. 2. The water entered the circulating pipe at 13'3° and left it at 13".5°. The average temperature of the calorimeter body was therefore somewhere between these teniperatnre.s. The ventilating air entered the calorimeter at 12'4° and left at 14•4^ What then was the effective temperature of the calorimeter from the point of view of the hog's metabolism ? One can only make a guess. In order to have something definite the writers have adopted the temperature of the inlet water as defining the temperature of the hog's surroundings, fully realising that the actual effective temperature is probably a little different. The point is not one of any great consequence as the critical temperature is rather indefinite. It is not strictly a temperature but a set of circumstances of which temperature is the most important, for air circulation, humidity, etc. are not without effect. Table I. The time is given in hours since the last meal. In the 23'7° experiment the earlier readings were lost through a defect in the galvanometer circuit. Date Dec. 4 Date Dec. 19 Date Jan. 8 Wei" ;ht 216 lbs Weig ;ht 231 lbs Weight 252 lbs Temp. li-A" Temp. 16-9° Time Metabolism Temp. Time M 10-3° Time Metabolism etabolism 14 2-927 41 2-058 19 .3-188 43 2-139 67 1-869 42 2-374 66 1-926 90 1-702 66 2-227 93 1-852 114 1-725 90 2105 116 1-833 140 1-851 Date Feb. 5 Date Feb. 25 Date Mar. 22 Date Apr. 23 Weight 293 lbs Weight 306 lbs Weight 342 lbs Weight 373 lbs Temp. •20-4'' Temp. 13-6° Temp. 12-8° Temp. 23-7" 'ime Metabolism Time Metabolism Time Metabolism Time Metabolis 35 1-927 45 2-221 28 2-647 63 1-870 39 1-904 68 2-019 88 2-438 84 1-750 65 1-702 93 1-955 58 2-266 .14 1-.562 113 1-936 108 134 2-126 2-115 Table I shows the whole of the observations that have been made. At the head of each section is given the date on which the hog left the calorimeter, his weight at thf end of the fast before receiving his first J. W. Capstick and T. B. Wood 26^ meal and the temijerature of the inlet water. Below these data is given the calculated metabolism in kilogram-calories per minute at various numbers of hours from the hoy's last meal. 3 2 3-1 2 9 2-6 2-5 5 2 4 a S 2 3 2 2 2 1 20 1-9 IS 1-7 1-6 1-5 ■ "0 10 20 30 40 50 60 70 80 'JO 100 HO 120 '30 140 Hours since last meal Fig. 2. Fig. 2 shows the results in Table I plotted on a single diagram. The ordinates are calories per minute and the abscissae are hours since the last meal. The 12'8° curve is raised '3 cal. and the 23"7° curve is raised '5 cal. to keep them clear of the other curves. 264 Basal Metabolism of Swine It will be soeii that the points tall very well on a series of similar curves. The fact that the divergences of the points from the curves are so small atifoi'ds some evidence of the general accuracy of the measure- ments and supports the belief of the writers that they have obtained the true resting metabolism. The curves show a very close similarity to each other as regards their shape. This similarity formed the subject of a paper read by the writers before the Royal Society. The paper has been printed in the Proceedings of the Roijul Society' and nothing need be said on the matter here. The circumstances of the experiment at 23"7' made it impossible to get more than two satisfactory observations and these were both taken before the hog had (juite reached his basal metabolism. The basal meta- bolism was therefore obtained in this case by drawing a curve through the two observed points parallel to the remaining curves. This procedure seems to be justified by the conclusions reached in the paper read before the Royal Society. Inspection of the curves shows that in every case the terminal hori- zontal part, which gives the basal metabolism, is reached between 90 and 100 hours after the last meal. Tangl" states that the hogs with which he worked reached their basal metabolism in 72 hours. In the present experiments there was always a perceptible fall after 72 hours. From some experiments which the writers are cai'rying out at present it would seem that age has an etiect on the time at which the basal meta- bolism is reached, for a young hog weighing about 35 lbs. reaches his basal metabolism in about two days. The basal metabolism is found graphically from the curves in Fig. 2 and is tabulated in the fifth column of Table II below. Table II. Summary for Critical Temperature. Date liasal Age Corrected meta- Reduced Reduced correction basal 1921 Age Temp. Weight bolism to 300 lbs to 13-3" to 420 days metabolism Dec. 4 312 days 13-3° C. 216 lbs 1-840 2-291 2-291 -•391 1-900 ,, 19 327 ,, 16-9° C. 231 „ 1-715 2-057 — - -290 1-767 1922 .Jan. 8 347 ,, 10-3»C. 2.52 ,, 2-102 2-361 — - -195 2-166 Feb. 5 375 „ 20-4° C. 293 ,, 1-570 1-595 — - -092 1-.503 „ 2.5 395 „ 13-6° C. .306 „ 1-940 1-914 1-937 -■038 1-876 Mar. 22 420 „ 12-8°C. 342 ,, 2-120 1-943 1-905 0 1-943 Apr. 23 452 „ 23-7° C. 373 ,, 1-720 1-489 — -H-OlO 1-499 ' Froc. li.S. B, 94, 35. = Biol. Zeitsch. 44. J. W. Capstick and T. B. Wood 265 Before the measurements can be used for finding the critical tem- perature they must be corrected to a standard weight and a standard age. The standard weight chosen is 300 lbs., as this lies almost half-way between the extremes of weight. In making the correction it is assumed that the metabolism is proportional to the hog's surface area and that the area is proportional to the two-thirds power of the weight. The basal metabolisms in the sixth column of Table II are calculated from those in the fifth column on these assumptions. In order to provide data for a possible age correction observations of the basal metabolism were made in the neighbourhood of 13° C. on December 4, February 25 and March 22. These three were at 13-3°, 13"6° and 12'8° respectively. The difference is not great, but as they fall at a part of the range where the change of metabolism with tem- perature is considerable they should be corrected to the same temperature before they are used for finding the age correction. 2 4 — 2-2 V 20 ^^^^^^^ 1 8 " 16 - 1'4 1 1 1 1 1 1 1 1 300 320 340 360 380 400 420 440 460 Age in daj'S Fig. 3. This correction requires a knowledge of what it was the object of the experiments to find — namely, the relation between basal metabolism and temj^erature — and it is somewhat illogical to make such a correction at this stage. As however the correction is quite small it is not likely that any appreciable error will be made by assuming for this purpose that there is a linear relation connecting the metabolism and tempera- ture between 10-3° and 20-4°. This gives a fall of -077 calories per Centi- gi-ade degree, and using this value we have the corrected metabolisms shown in the seventh column of Table II. The final curve of Fig. 4 shows that a linear relation is not far from the truth. Fig. 3 shows the result of plotting the basal metabolism at 13-3° 266 Basal MefahoUsm of Sivine against the hog's age in days. It appears that at the end of the experiments age had almost ceased to have any effect. Wi' liave ne.xt to correct all the metabolisms in the sixth column of Table 11 to some one date. It is immaterial what date is chosen. The writers have selected March 22 when the hog was 420 days old. The corrections to be added or subtracted are given in the eighth column of Table II. In making these corrections it has been assumed that the effect of increasing age on the curve connecting metabolism and temperature is to cause it to move bodily downwards always keeping the same shape. It may be that it would be more correct to assume that the various points on the curve move downwards in the same proportion. 2 4 6 8 10 12 14 16 18 20 22 24 Degrees Centigrade Fig. 4. The writers see no adequate theoretical reason for preferring either of these methods to the other. For the present purpose it makes very little difference which method is used. The point might be worth in- vestigating experimentally, but it would be difficult to secure the necessary accuracy. We have finally the coi-rected values of the basal metabolism shown in the last column of Table II and plotted against temperature in Fig, 4. The points fall very well on the curve with the exception of that at 16"9° which is 6J percent, too high. The writers are unable to account for J. W. Capstick and T. B. Wood 267 this anomaly except on the grounds that the experiment at 16'9" was unsatisfactory throughout. The hog was scarcely ever really quiet throughout the whole fasting period with the result that the galvano- ■ meter curve was less regular than usual. It should be noted that errors due to the hog would almost certainly cause the observed metabolism to be too high, whilst experimental errors would be indifferently high and low. There was also an instrumental failure on the last day of the 16'9' experiment. One of the electrical heaters in the thermostat burnt out and there was a great disturbance of the curve before the fault could be remedied. The remaining points however are sufficiently consistent to enable the writers to state that the critical temperature is very near to 21° C. — remembering, however, what has been stated above, that the real average temperature of the hog's surroundings may be somewhat different. This conclusion agrees very well with that reached by Tangl who states that he found the critical temperature to be between 20° and 23°. The actual metabolism at the critical temperature is I'SO calories per minute for a 300 lb. hog or 2,160 calories per day. The exact relation between a hog's surf;\ce area and his weight is not known. Tangl accepts Voit's formula A = 9-02 IF". This gives 904 calories per day per square metre, which is near the value generally adopted for human beings. The values of the basal metabolism at different temperatures have a practical interest for pig breeders as they enable us to calculate the maintenance ration at various temperatures. Table III. Basal metabolism Basal metabolism Temp. of a 300 lb. hog Temp. of a 300 lb. hog 0 3-10 12 2-01 2 2-92 14 1-84 4 2-74 16 1-70 6 2-56 18 1-59 8 2-38 20 1-51 10 2-19 22 1-50 Table III gives the metabolism of a 300 lb. hog at intervals of 2° from 0° C. to 22° C. The metabolism between 10° and 20° is taken from the full curve in Fig. 4. Actual observations of the metabolism could not be made at very low temperatures as the temperature of the town supply water did not permit of the calorimeter being set to anything below 10°. In order to get an estimate of the metabolism below 10° it is therefore necessary to use the uncertain expedient of extrapolation. 268 Basal Metabolism of Swine As the observed curve is nearly a straight line below about 16" it has been continued backwards in a straight line to 0°, the dotted line being the part of the curve obtained by extrapolation. The values of the metabolism in Table III are taken from this extended curve. It may be presumed that in so far as the extrapolated values are in error, the error is on the side of their being too low since the observed part of the cui-ve is concave upwards. The}' will however provide a sufficient approximation for our present purpose. It will be seen that the maintenance ration at 0° is more than double that at 22°, and both these temperatures are not unfrequontly met with on farms in this countiy. Conclusions. The critical temperature of the hog under experiment was approxi- mately 21" C. At this temperature his basal metabolism was a minimum and amounted to 2,160 calories in 24 hours when he was 420 days old and weighed 300 lbs. This corresponds to 904 calories per day per square metre of body surface. As the temperature of his surrounding.s fell below 21° C, the basal metabolism increased at the rate of about 4 per cent, per degree Centi- grade, which corresponds to an increase of about 40 per cent, for a temperature difference of 10°C. which is commonly found between summer and winter conditions. Thus, if the same law holds in the case of a steer whose basal meta- bolism at 18° C. or summer temperature is 6,000 calories, his basal metabolism at 8° C. in an open yaid in winter would be 9,000 calories. The suggestion is that the increase of 3,000 cahjries is met b}' the utilisation of the thermic energy of the coarse fodder included in his ration. {Received June 2^th, 1922.) THE FUNGICIDAL PROPERTIES OF CERTAIN SPRAY-FLUIDS. III. By E. HORTON and E. S. SALMON, Research DepartiTient, S.E. Agric. College, Wye, Kent. As an aid to the elucidation of the problem of the exact fungicidal value of a mixture of lime-suljjhur and arsenate of lead — a matter of great importance to the practical fruit-grower — spraying experiments were carried out during 1921 with certain spray-fluids containing arsenic or lime-sulphur and its constituents. Method. In order to ascertain within narrow limits the fimgicidal value of any solution, it is obviously necessary to maintain as fixed a biological standard as possible. To ensure this, the fungus used in com- parative experiments should be in the same stage of development and, if it is a parasite, the host-plant used should also be " standardised " as far as possible — since it has been shown(2) that the same stage of a fungus may be more easily killed when on the older leaves of a plant than on the younger. In all the experiments described below, the fungus used was Sphaero- theca Huniuli (DC.) Burr., and the stage selected for spraying was the young, "powdery," conidial stage produced on young leaves, at the 3rd to 9th node, of rooted cuttings of hop-plants {Humulus Lupulus, Linn.) grown in an unheated greenhouse. To escape as far as possible variation on the side of the host-plant, with possible consequent effects on the vigour of the parasitic fungus, all the plants used were clone- plants, i.e., plants raised vegetatively by cuttings taken from one indi- vidual hop-plant. The general methods of sprajdng and of the examina- tion of the sprayed leaves, etc. have been described in previous articles (1, 2). In order to secure complete wetting of the fungus, calcium casein- ate (1 per cent.) was added to the solutions used. Where care is taken to obtain in this way strictly similar biological conditions of parasitic fungus and host-plant, it becomes possible, as shown below, to determine within narrow limits the fungicidal value of a solution. In all the experiments described below "powdery" conidial 270 Funf/icidal Properties of Certain Sprnij-Flidih. patches of : 20). Maintained approx. neutral 160 * 7-58 "{, of material dissolved. f ''"^S °o of material dissolved. Microscopical examination of the fractions obtained showed striking differences. The fine silt and silt obtained by the ordinary method con- tained considerable quantities of amorphous material. The same fractions obtained after preliminary oxidation were seen to consist entirely of crystaUine mineral matter. Similar fractions are obtained from raw clays and soils poor in organic matter. The use of ammonium persulphate proved inconvenient in practice and hydrogen j)en)xide was therefore tried as an oxidising agent. After .some prehniiiiary trials it was found that the most convenient method of oxidising was as follows. Ten grams of soil are weighed into a beaker of 600-700 c.c. capacity. Fifty c.c. of hydrogen peroxide (20 vols.) are added and the beaker placed on a boiling water bath. A vigorous reaction soon takes place with considerable frothing owing to the evolu- tion of oxygen. The contents of the beaker are stirred from time to time. After about 30 minutes the reaction dies down and a further 25 c.c. of peroxide are added, the froth adhering to the sides of the beaker being washed down with a small volume of water. After a further 15-20 minutes' heating the reaction is generally complete and frothing G. W. Robinson 289 ceases. In the case of soils with large j^roportions of organic matter, more peroxide will be needed. The beaker is then removed from the water bath and, after adding about 100 c.c. of water, boiled for about 15 minutes. A considerable oxidation of organic matter has now taken place and the oxidised material acquires the yellowish or light brown colour of a non-humous subsoil. The contents of the beaker remain approximately neutral so that no solution of mineral matter may be apprehended. In order to form some idea of the effect of this oxidation on the soil organic matter two soils were oxidised. After filtration and washing the residual material was dried and the loss on ignition determined. The filtrate and washings were evaporated to dryness and the amount of soluble organic matter determined. A separate experiment was also carried out in which the gaseous products were collected in normal sodium hydroxide and the amount of carbon dioxide estimated by double titration. The amount of organic matter completely oxidised to carbon dioxide was found by multiplying the weight of carbon found as carbon dioxide by 2, assuming as a first approximation that the organic matter contained 50 per cent. of carbon. The results obtained were as follows: Loss on ignition t).\idised Soluble Soil Loss on ignition after oxidation to T'O., org. matter Cili 2.5-4 8-1 o-,5 10-8 4T 101 40 1-9 3 7 It \vill be seen that the original organic matter is approximately accounted for by the unoxidised matter and the products of oxidation. A number of soils thus oxidised were submitted to mechanical analysis. The result of the treatment is to break down the compound particles very thoroughly and the separation of the fine gravel and coarse sand by means of the 100 mesh sieve is very easily effected with very little trituration. Generally speaking the sedimentations were carried out twice without the addition of ammonia and then as in the ordinary method. It may be added that the soluble products of oxidation would appear to be deflocculating in their action and quite considerable quantities of clay can be obtained without the use of ammonia. The following table shows the results obtained for a number of soils by the ordinary method including the usual preliminary treatment with HCl, and also by the hydrogen peroxide method, in which this acid treatment is omitted. 20- 290 Note on tlir JJcc/uaiical A/nili/t< of JJ mints Soils Ab. 4 B Ab. B S F 2 A F 2 B G 26 A C 3li A Soil Ord. H,0, Ord. Hj.Oj Ord. HjO, Ord. H,0, Ord. H.O, Ord. H,0, Fine gravel 3-8 3-3 30 2-8 — — •4 38 40 7-8 10-3 Coarae sandj Fine sand ( 29-5 29-5 .33-8 331 81 8-3 4-6 6-5 \4-8 )5-2 5-2 81 12-3 7-8 20-5 15.-. SUt 18-3 15*5 190 13-8 22-1 20-8 20-9 20-7 23-2 1 3-7 15-6 140 Fine silt 28(i 23-2 260 19-5 32-9 25-4 ,340 27-4 34-6 26-8 23-4 297 Clay 8-2 17-9 6-3 17-6 ^ -/ ' 1-3 201 283 V J 3-5 28-7 320 3-8 13-2 1 5-6 9.-. Moisture 2-3 3-2 3-3 2-6 Org. matter 100 90 10-9 5 •4 2, 5-4 10-3 Totals 100-7 101-7 US 4 97-1 97-G 97-2 96-8 95-6 104 1 99-7 98-1 99' Soil Ord. H,0, Ord. H,0, (ii.l. ILol Ord. H,0, Ord. H,0. Fine gravel 6-6 1-4 40 3-2 5-2 4-6 ■9 ■8 37 4-1 Coarse sand 2.1-7 27-9 20-5 380 22-5 24-9 23-6 25-3 11-8 12-2 Fine sand 27-1 20-4 29-7 210 270 18-7 26-8 21-4 32-8 18-6 Silt 10-8 10-5 14-8 10-5 15-7 12-5 160 12-0 20-0 17-5 Fine silt 10-7 17-7 10-1 17-2 12-5 19-0 13-2 15-7 131 19-0 Clay 3-8 6-8 V ■2 3-S 8-0 4-2 7-5 3-5 9-3 :!•:', 130 > Moisture 4 • 2 •>. 1 3-8 2 •6 Org. matter 8-1 10-9 8-7 9-6 110 Totals 980 980 98-0 98-0 97-9 980 97-4 97-9 98-3 97-0 It will be noticed that in every case there is an increase in the amount of clay obtained. Tn other words the effect of the oxidation has been to increase the degree of dispersion of the soil. That this effect was not an apparent increase in the clay owing to higher visco-sity of tJie aqueous solution of the oxidation products was shown by a viscosity determina- tion on a filtrate obtained from an oxidation. The ^^scosity was sensibly the same as for pure water and it may be assumed tliat tlie difference is due to an actual dispersion of complex particles which are not broken uy) in the ordinary method of dispersion. The hydrogen-peroxide dis- persion has of course been effected without preliminary acid treatment, in the case of the soils used. Possibly it might be desirable to include the acid treatment in the case of soils containing much carbonate. It is worthy of remark that the clay hquor obtained by the peroxide method is strikingly different from that obtained by the ordinary method. On agitation, it shows a satin-like effect owing to the reflection of hght from minute crystaUine particles. This effect is not shown by the clay liquor from soils in the ordinary method. Similar results are obtained with fine silt suspensions. It is also noticed that on flocculation of the clay hquor a smaller volume of flocculated material is obtained than in the ordinary method. (t. W. Robinson 291 No attemj)t lias been made to follow the chemical changes involved in the oxidation of soil organic matter by hydrogen peroxide. It may be mentioned that the soluble compounds formed appear to form a very suitable medium for the development of moulds and fungi as, when exposed to air, they quickly become covered with a scum on which growths appear. The products of oxidation would be well worthy of chemical examination. The results obtained particularly in the case of Welsh soils indicate that the figures obtained for clay by the ordinary method may be mis- leading and that there are considerable quantities of 'clay' which are wrongly grouped with the coarser fractions. In other words the ordinary method fails to secure the prime particle structure. This defect in the standard method may operate to some extent even in soils with smaller amounts of organic matter, for in every case larger figures were obtained for the clay after peroxide treatment. Treatment with peroxide offers a convenient means of removing organic matter from soil without altering the mineral portion. In these experiments it was not found possible to remove all the organic matter. Complete removal might be effected by successive treatment with peroxide. It would appear, however, that the humified organic matter is completely oxidised or rendered soluble for the dark colour is removed even from peat soils and the oxidised soil has the appearance of a raw subsoil. The unoxidised material in such cases appears to consist entirely of structural organic matter, a circumstance which suggests that only humified matter is attacked. {Received July Sth, 1922.) A BACTERIAL DISEASE OF TURNIP {BRASSICA NAPUS). By S. G. JONES, M.Sc. Lecturer in Botany, University College of Wales, Aberystwyth. (With Plate ill.) Of recent years a disease of root-crops known to farmers in Nortli Wales is one in which the heart or core of the root is converted into a soft putrid mass but which leaves the rind and mature foliage intact. The disease is prone to appear on land treated with lime as a pre%'entive against the roots being attacked by Plasmodiophora hrassicae or when the land has received a heavy dressing of nitrogenous fertiUzers. It is common knowledge that nitrogenous manures have a tendency to force the crop and so produce watery, sappy roots which easily fall prey to disease. Recently it was brought to the writer's notice by one of the County Organizers for Agriculture in Wales that some of the farmers in his area were strongly disinclined to lime the land for root-crops because this treatment was favourable to the appearance of "soft-rot." There is probably some modicum of truth in this contention, for as will be seen below, the writer in the investigation of the present disease found that the organism isolated refused to grow on any media which were not neutral or alkaline. The writer observed and investigated this disease on a crop of white turnips grown on the farm of the I'niversity College of North Wales and the land here had received a dressing of nitrate of soda. A casual glance at the crop did not show anything unusual, in fact judging from the amount of green foliage the crop looked very healthy. Closer examination however showed that the very young leaves at the centre of the crown had been destroyed thus forming a tiny wound into which one could push a probe without obstruction to the depth of some three or four inches. The fully expanded leaves very effectively concealed the wound and the disease seemed evidently con- fined to the internal tissues leaving the foliage and rind quite firm even up to time of harvesting the crop. Indeed the extent of damage was only fully revealed at the time of lifting when the harvester's knife S. a. Jones 293 lopped off the foliage only to discover tiiat the roots were bad. A diseased root cut in vertical section showed a Hask-shaped, soft, putrid core surrounded by a brown zone abutting on the healthv tissue (Fig. 1). The disease never made further progress into the rind nor into the lower part of the root and this would therefore account for the healthy turges- cent appearance of the fohage since the vascular tissue in the root and rind would still be functional. Such were the external features of most of the diseased jilants but in addition numerous cases were found where the entire apical bud had been destroyed thus forming a large wound only however to be concealed by no fewer than three, often five secondary crowns all bearing healthy luxuriant foliage (Fig. 2). There were also found numerous diseased roots with all foliage intact, but bearing deep cracks in the rind above ground; these were probably examples of "burst" roots a condition frequently found in sappy roots exposed to sudden weather changes. In all these types however the internal appear- ance of the diseased roots was precisely the same — the soft-rot was always confined to the core of the root and in addition there was always present the brown-coloured zone at the boundary of the diseased area. A white-rot of turnips similar to the one now described has been attri- buted by Potter to an organism Pseudomonas destructans (Potter). He states that plants attacked by this parasite "can be recognised by the drooping yellowish leaves, the older leaves being tlie first to show any indications of disease; they gradually flag and droop to the ground, at the same time becoming yellow and shrivelled in appearance. The leaves next in age gradually exhibit the same signs of premature decay and this proceeds until finally the young leaves at the growing point succumb; the entire rosette of leaves perishes, and the whole root becomes a soft, putrid mass, which eventually collapses etc." In the present disease, however, the external appearances of the diseased plants showed features very different from those accompanying the one described by Potter and these were deemed sufficiently striking to warrant further investi- gation. It has been a matter of extreme difficulty to detect the initial mode of attack of this disease on the plants in the field. The writer has not been able to investigate this important point and the evidence gathered from the farmers is most conflicting. One noticed that the turnip plants were to all appearances dead, with the leaves flat on the ground as if the cattle had been lying on them but that on a subsequent visit he had found the crop to all appearances fully recovered only to find however at the time of harvesting that the roots were bad. Another said that 294 A Bacterial Disease of Turnip (Brassica Napiiis) he had also seenjthe leaves drooping but that the seeming recovery of the fohage was due to new growth from secondary crowns. At the College farm the writer found intermixed in tlie drills along with diseased roots numerous healthy-looking plants possessing brown, dry, empty cavities with no sign of the pasty mass ever having been present; such plants invariably harboured slugs or their ova. These plants had, like the types of diseased roots already described, either the young foUage at the growing point destroyed or the foliage of the apical bud had entirely disappeared with the wound thus formed completely healed and surrounded by luxuriant fohage from secondary crow^ns. The Nature of the Disease. With a view to determine the nature of the pasty mass in the core, small quantities of it, taken from several affected plants were placed in a httle sterile water in a watch-glass. Microscopic examination of the turbid water showed isolated cells and cell-clusters floating in the liquid. There were no traces of protozoa such as are described by Priestley or of fungal hyphae. A small quantity of the liquid taken up on a sterile platinum looji and smeared on a cover-glass showed when stained a dense mass of bacteria. Cultures were then prepared for isolating the organisms. Methods. The sterihzation of apparatus and media was effected in the usual way. In the first instance nutrient gelatine acidified to 1 per cent, normal hydrochloric acid was employed but this medium w'as found unsuitable; then turnip-juice gelatine of natural acidity and neutralized was also used. Growth on these media, however, proved so inferior to that produced on beef-extract-peptone-gelatine neutrahzed with sodium hydrate that this was exclusively employed in the preparation of pure cultures. Precautions were taken to ensure its constant composition and uniformity of treatment in sterihzation. Diseased plants from which cultures were to be taken were thoroughly washed free from soil and placed for 15 minutes in a weak solution of corrosive subhmate. Without removing the chloride the- roots were cut open with a sterile knife from root-tip to crown, the root being inverted in order as far as possible to avoid contamination with foliage epiphytes. Petri-dish cultures were then carried out in the usual way and the several colonies transferred to agar-slants. The next step was to prepare sterile blocks of turnip for inoculation from the agar-streak cultures. By means of a platinum hook inocidations 8. a. Jones 295 were eflected from the agar-streaks into small excavations made at the centre of the blocks. After an incubation period of 15 to 24 hours at 20° C. most of the blocks showed a whitish-grey transparency around the inoculated part. In the more juicy blocks, the afiiected area, in 24 hours, would be about the size of a sixpence. After further incubation for 24 hours they had become completely diseased; during the next two or three days they assumed a yellowish hue and after a week a brown colour. The organisms causing the rot in the blocks had been derived from the agar-streaks which had been taken from round, whitish- grey hquefying colonies. Repeated cultures from field material showed that the colonies effective on the turnip blocks were of this type. Poured plates taken from successfidly inoculated blocks showed the colonies in crowded growth to be small, those situated below the surface of the gelatine being smaller. On a thinly sown jjlate the colonies were large, circular, with much hquefaction in the centre. Magnified under the low power of the microscope they showed a finely, closely fibrillated margin; within the margin was a narrow finely-granular zone, then a narrower denser granular band forming a very faint concentric circle within which was a wide zone gradually diminishing in a granular density towards the centre. The saucer-like depression contained massed bacterial debris floating in a thinly granular hquid. Buried colonies were small, round, fibrillated at the margin and homogeneously granular. The Organism .4nd its Flagell.\tion. After repeated cultures had been made in nutrient gelatine from the diseased blocks there remained no doubt that a pure culture had been obtained. Cover-sUp preparations of the organisms stained with carbol- fuchsin showed them to be rod-like in form and of varying length. Single specimens were short with rounded ends and pairs were frequently seen. Hanging-drop cultures in bouillon from several young agar-slants showed the organisms to be actively motile. After repeated attempts to stain the flagella, it was found effective to transfer the organism from agar- streaks of not more than 12 hours' growth into a number of small petri- capsules half-filled with sterile, filtered water kept at the same tem- perature as that in which the agar-cultures were incubated. The cover- sHps were placed in series of six in a large petri-dish bearing the corre- sponding number of the culture. A sterile platinum loop was then dipped into the small capsule and the drop quickly deposited on the cover-shp, six preparations being taken from the same capsule. These were air-dried •296 A Bacterial Disease of Turnip (Brassica Xapus) by iminediatelv placing the petri-dishes in an incubator at GO '. LoefHer's method of staining was first employed \vithout success, but with this method, the organisms showed uniform staining, except at one pole which was somewhat hyaline. When however the mordant was made sUghtly alkahne with caustic-soda (J c.c. of 1 per cent, alkali per 10 c.c. of tannin) most of the shdes after staining showed the presence of a single long polar Hagellum. Van Ermengen's stain, pre])ared according to the usual formula, also showed the presence of the cihum. No varia- tion in the number or .situation of tlie flagella was seen. Infection Experiment.s. The next step was to establisli the disease in healthy plants. It was decided to carry out infection of healthy plants by employing cultures from inoculated blocks. Preparations were accordingly made for inocu- lating a series of blocks from young cultures taken from the agar-streak tubes that had last been employed in the flagella staining. The latter process had entailed repeated failures and fresh attempts always involved the use of fresh subcultures. It was reckoned that the tubes now em- ployed for inoculating the turnip blocks contained cultures which had passed through some 20 generations over the medium. After the usual period of incubation, the inoculated blocks showed little or no signs of disease. Some of the blocks showed a slight browning around the inoculated part and made no further advance; others did not appear to have taken the disease at all. It seemed that the organism had either lost its virulence in its repeated passages through the medium, or that the turnip, through prolonged cultivation had passed over into a state of resistance to attack. It was therefore decided to submit sterile blocks of turnip to special treatment on the lines followed by Laurent for rendering resistant tubers sensitive to bacterial attack, by immersing them in alkaline solutions. Accordingly the blocks were first soaked for an hour, some in 0-25 per cent., some in 0-.5 per cent, and others in 1 per cent, caustic-soda solutions. They were then transferred to sterile test-tubes and inoculated from the same agar-streak cultures that had been employed on the supposed resistant blocks. After prolonged incuba- tion the parasite made no more progress on the blocks thus treated than on the control ones which had only been soaked in water previous to inoculation. It was therefore concluded that the organism had lost virulence by so many passages through the artificial medium. When, however, another series of turnip blocks was inoculated from an agar- S. G. Jones 297 culture derived directly from diseased material the usual signs of attack and rotting of the blocks took place. As above mentioned the farmers who had witnessed the attack in the fields said tiiat the disease had been preceded by a sudden collapse of the foliage. This reported phenomenon at once suggested infection of the leaves by way of the stomata or water-pores. Experiments were carried out on a number of plants in the following way. Uninjured leaves (attached to the plant) of varying ages were plunged into water in a series of petri-dishes to which had been added pure cultures of the organism from bouillon. The leaves were left immersed for an hour. In some plants they were first steriUzed with weak mercuric chloride which afterwards was removed by repeated plunging into sterile water; in others, no surface sterihzation was employed. The leaves of the control plants were also treated in the same way. These experiments failed to show any infection of the foUage. The writer is strongly of opinion that infection is preceded by mechanical injury through some such agency as leaf -cutting insects or slugs and the very earliest signs of tlisease in the very numerous cases seen in the field were the softening and water- soaked appearance of the young fohage. Further inocidation experi- ments were carried out by first .moistening the young foliage at the growing point and then depositing a pure culture from bouillon. The root was then covered over with a bell-jar plugged at the top with cotton-wool. These experiments invariably failed despite the greatest care being taken to keep the inoculated part moist. Accordingly the tender foliage at the growing point was pinched off with a sterile forceps and into the wound thus made a pure culture of the organism was de- posited, the wound being covered over with a piece of steriUzed cotton- wool, and the whole plant again covered over as before. This method of inoculation was always successful but the progress of the disease varied somewhat in different plants. The extent of attack was deter- mined from time to time by probing the cavity with a platinum wire. When a depth of some four inches had been reached the plants were cut open in vertical section. Those in which the pulp was watery showed the characteristic whitish-grey mass accompanied by the marginal brown discoloration. Other roots of a drier spongy texture showed a diseased core of a uniform brown colour. This difference of colour in the diseased parts suggested an idea that it might bear some relation to the water content of the cells and intercellular system or in other words to the extent of aeration of the tissues. During the earlier stages of the investi- gation it was noticed that when the inoculated turnip blocks in the test- 298 A Bacterid! DlHeaae of Turnip (Brassica Napus) tubes were examined at intervals, some were always found to be more advanced in disease than others. Those which were spongy in appearance were brown or yellowish-brown in colour but the more succulent blocks were always wliitish-f;rey. The difference of coloration in the diseased tissues may therefore be due to oxidation. Further proof of this was established by the appearance of the diseased blocks in the tubes after incubation periods of 24 hours, 48 hours, and 7 days. At first they showed a whitish-grey transparency around the inoculated part; later, they had become almost completely diseased but still whitish-grey, but after 7 days the blocks were considerably changed and distinctly brown in colour. The darkened colour-, however, was only seen in those portions of the blocks exposed to the air in the tubes; the basal portions immersed in a little water were still light in colour. Still further evidence proving the discoloration to be due to oxidation was obtained by the following experiment. Two conical flasks with side tubes, plugged and sterilized in the autoclave were partially filled with sterile blocks about a cubic centimetre in size, and a little sterile water added. Two rubber stoppers which had previously been boiled in mercuric chloride and dipped in sterile water carried a bent tube, reaching to within half-an- inch of the turnip blocks. Into each a small fragment of diseased turnip was inserted, tlie stoppers quickly replaced and sealed with paraffin wax, sterile rubber connections furnished with pinch-cocks were slipped over the side and the bent tubes. The flasks were connected each to three similar flasks containing a mixture of pyrogallic acid and potash. A current of air was then drawn through the series of flasks for 15 minutes. The rubber connections of the turnip-flasks were then closed with the pinch-cocks and the flasks detached. From one turnip-flask the pinch-cocks were removed, and the tube-ends closed with sterile cotton-plugs. Both flasks were then incubated at 20° C. for a week. After 24 hours the blocks around the diseased fragment of turnip intro- duced showed the usual whitish-grey transparency. At the end of the week the contents of the cotton- plugged flask had become brown in colour, whereas the blocks in the sealed flask, all diseased, still remained whitish-grey in colour, and showed no sign whatever of browning, nor did they show any change of colour even after several weeks. Inci- dentally the experiment shows the organism to be facultative anaerobic. Further evidence of this property is given below. As previously stated the discoloration in the field plants was con- fined to a distinct zone, between the healthy and diseased tissues. It is possible that the disintegration of the tissues following upon disease 8. a. Jones -ifUt would be attended by a collajjsing of the cells, so that access of air through the diseased core would be prevented and thus hinder atmo- spheric oxidation. At the time of infection in the field, the plants would be actively growing and drawing upon the minerals in the soil. It is conceivable that one or more of these substances would have an oxidizing efiect on the bacterial .secretions in the diseased tissues, causing a dis- coloration at the margin of the diseased area. E. F. Smith states that Laurent, experimenting with different fertihzers on potatoes upon plots which had been treated with sodium nitrate and sulphate of ammonia produced tubers, which when inocidated with an organism (beheved to be B. coli) gave a " black zone between the attacked and healthy tissues. As this stain was not noticed elsewhere, Laurent attributed it to the nitrogenous product formed by the bacteria at the expense of the tissues." Relation of Parasite to Ho.st. The relation of the organisms to the tissues of the host was first determined by examination of the diseased pulp of the field plant. A small quantity of the pulp mounted on a slide and gently pressed down with the cover-glass showed the cells to be completely isolated. The cell-walls were, however, intact and presented no appearance of being swollen. Numerous reticulated vascular strands were also seen with their walled prosenchymatous elements attached in places. The bacteria were generally seen outside the cells, but a few cells were so completely occupied by them that they appeared black amongst the contiguous imoccupied cells. Examination of the diseased pidp of inoculated plants showed exactly the same phenomena. Such densely occupied cells with the contiguous cells containing few or no bacteria were also seen in the microtomed sections of inoculated material. Field material was fixed in Flemming's solution and in Carnoy's acetic-alcohol. Small pieces of the diseased roots were cut to include the healthy tissue, the brown zone, and the loose \)\\\\). The shdes made from the material fixed in the Flemming mixture showed almost the complete loss of the soft pulp in the prolonged process of washing. The alcohol-fixed material was therefore employed. The pieces selected for paraffin embedding included all the tissues from the periphery to the soft pulp. Transverse and longitudinal sections were cut on the micro- tome to a thickness of 4fi and mounted on shdes thinly coated with white of egg. Considerable difficult}' was experienced in the differential staining of the tissues and the bacteria respectively. The organisms 300 A Bacterial Disease of Taniip (lirassicii Nai>iis) stained easily with tlie aiiilin dyes, particularly carbol-fuchsin, followed by differentiation in alcohol, showing the bacteria to advantage. Another method was to stain for 12 hours in HofFniann-blue saturated with picric acid, and after washing in 50 per cent, alcohol, to immerse for three minutes in carbol-fuchsin. This showed the tissues blue and the bacteria red, but the staining was not always uniform and hence the method was unreHable. After repeated attempts with various dyes satisfactory results were obtained by the use of Heidenhain's iron-alum- liaeraatoxylin followed by eosin in dove-oil. Tliis showed the bacteria black and the cell- walls light red. When the slides were exainincd under the Iow-jkiwci- of the micro- scope, the bacteria were seen to be apparently e.xclusively confined to the intercellular system of the medulla, and the disintegration of the tissues to be brought about by the dissolution of the middle lamella. Some cells in the same region were seen to be densely filled with bacteria, relatively few or none being present in contiguous cells, thus repeating the features already observed in the examination of the diseased pulp. Some of the xylem elements also seemed to be occupied by bacteria. The microscopic examination of microtomed sections derived from artificially inoculated roots revealed an exactly similar invasion of the tissues to that seen in the field material. Tlie bacteria abounded in the intercellular spares of the inner tissues; occasional parenchyma cells were densely filled with bacteria, whereas the peripheral tissues were intact. Separation ok thk Bacteria from their Products. The separation of the by-|)roducts from the bacteria was then attempted. A two-litre flask containing a little water and sterilized by intermittent steaming was half filled with sterile blocks of turnip, a diseased block from a test-tube dropped in, and the flask quickly plugged. After four days' incubation at 20° the whole of the contents had been reduced to a pulpy putrid mass with a highly offensive odour. This was pressed through a piece of coarse linen and the turbid filtrate passed through filter-paper. The grey-coloured liquid was then divided into two portions — one portion was passed through a Berkefeld filter which had been sterihzed for several hours in the hot-air oven and allowed to cool. The clear pale yellow filtrate was poured into a number of small, sterile test-tubes and corked with sterilized rubber bungs. To the other portion was added four times its bulk of 80 per cent, alcohol. After 24 hours a heavy flocculent precipitate had been formed. The super- 8. U. Jones 301 natant liquid was then siphoned out, the precipitate collected and washed with absolute alcohol. After careful drying in the incubator at 20° it was digested in 100 c.c. of sterile distilled water for three hours. The liquid was then passed through the Berkefeld filter. A clear pale yellow filtrate was again j^roduced and poured into a number of sterile tubes plugged with cotton-wool and protected by rubber caps. The action of the two filtrates upon living turnips was then tried in the following way. Thin sections of turnips were placed in small petri- capsules, into some was poured the filtrate obtained from the pulp, into others the filtrate from the watery digest of the alcohohc precipitate. Other capsules coiatained sections in sterile water. Microscopic examina- tion after 24 hours showed no differences between several preparations mounted and the control sections from the sterile water and examination again after several days showed no change in the appearance of the sections. Twelve tubes of the filtrates from both sources were employed upon the sterile turnip sections without any signs of loosening of the cells. The action of the filtrates upon a 1 per cent, starch solution was then tried, but the blue coloration after adding iodine still remained even after prolonged action. Tf an enzymic jiroduct capable of separating the parenchyma of the turnip was present in the diseased pulp, its power of action seemed to have been destroyed in the filter. A new series of experiments on similar lines was performed using a Muencke bacteria filter. This filtrate was again ineffective in loosening the tissues. The filtrates obtained from the watery digests of the alcohohc precipitates obtained from both bouillon and glucose-bouillon seemed to be quite inert in their action on the tissues. The use of the porcelain filters was then discontinued and the alternative method tried of treating the bacterial products with antiseptics. A solution of chloroform was em- ployed. The culture treated was the turbid filtrate obtained directly from the putrid mass formed from diseased turnip. The coarser portions of the pulp were removed as before by pressing through a cloth and merely filtering the liquid through filter-paper. An equal volume of the chloroform solution was then added to the filtrate and the flask vigorously shaken at frequent intervals. To determine the sterility of the product inoculations into tubes of nutrient gelatine were taken and poured into petri-dishes. At the same time thin sterile sections of turnip were treated with the product. After 24 hours" immersion the tissues had become disintegrated and further the poured plates were sterile. 302 .1 Bart I' r ltd Disease 0/ Turnip (Brassica Napns) Characters of the Organism. Habit. Causes a whitish-grey rot in turnip. Morphology. Bacteria l-3ju, to 3fi x -lofi to •9/x with a single polar flagellum. On sugar-rich media it forms long, unequally seg- mented filaments. Relation to Oxygen. Evidence for aerobism: Large surface colonies and rapid growth on the surface of the media, buried colonies small; gelatine stab funnel-shaped: rapid decay of turnip in the cotton-plugged flasks and tubes. Evidence for anaerobism: Slight growth on recently steamed agar-slopo tubes cultivated by Buchner'.s method for anaerobes; general turbidity in liquid media enclo.sed in Durham and Fermentation tubes; growth in gelatine stab after sealing mouth of track; decay of turnip blocks in a sealed flask. Growth on Gelatine Media. Surface colonies large, round, whitish- grey, margin fibrillated. Liquefies gelktin; stab-culture funnel-shaped in 24 hours, spreading to wall of tube in 48 hours, liquefaction nearly complete. Li the sealed track liquefaction not complete. Colonies on litmus-lactose and litmus-glucose-gelatine produce faint pink margin; on chalk-lacto.se and chalk-glucose-gelatine, hyaline ring; ])oor growth on acidified medium. Growth on Agar Media. Surface colonics round or sliglitly lobed, opalescent; sometimes spreading, filmy areas formed; deep colonies oval or spindle-shaped. Growth on agar-streak spreading and filmy. Good growth along agar-stab, spreading at surface. Colonies on litmus-lactose and litmus-glucose-agar produced jiink rnnigiii; or clialk-lacto.se and chalk-glucose-agar hyaline border. Growth i.v Bouillon. Rendeivs bouillon turbid, forming a heavy sediment; no pellicle; maximum turbidity in neutral and feebly alkahne broth; neutral reaction to litmus; with 2 per cent, glucose, lactose, and cane-sugar turbidity in open and closed ends of Durham and Fermenta- tion tubes, producing acid reaction without evolution of gas. In Pasteur's solution with cane-sugar turbiditv only in open end of tubes. Growth in Nitrate Solution. In Giltay and Aberson's culture fluid strong nitrite reaction after 24 hours with iodine-sulphuric-starch test, with formation of ammonia (Nessler's Test). Heavy sediment formed and film on surface of culture. Control tubes gave neither reaction. Growth in Nitrogen- Free Solution. No turbidity; after seven days, petri-cultures in gelatine inoculated from it produced very few colonies. Growth therefore feeble. 8. (r. Jones 303 1 a. u ^ eS o 3=3 ^ 3. W) ""^ X: X ai •9 » *o (0 ^ M j3 Tf 0.2 -p H O 3 zl-ti ^ -^^ ^ t-i £l -^ CD a •Geo ai "^ llJ 1— 1 a is o o t O oS -P Me^ 1- 33 o C^ . -d a "*^ a ■-E fl o rt -S 1> "3 ^ GC5 1 m c= o ti ^ ^r 3 (D 3 ■4J o a ft "^■S cS c8 2-C ■3 Hi a a.s .. u a & C8 o-s 9 £Pa E 5 ^. —J (3 tH .3 o T^ +J 4J _> c '^ >!•" O c s;2 go a « d o &, t4 ID __ ,^3 rf .3 fS o ^ fl" « (D O o 3 T3 P Q o Cj o o 3 ts 3 p o C-i o o te; •o^ '^ aj2 if5"5Pa ■ o a » ij-.i: S 6 = 5-^ o a o t< c 3 M'o'S S a i; fe t* ^ a> a i:^ i O.CJ<0 • c cs'S S 'C - a -T^ <:> ■ 0^ £ c o-^ a a 5 9.S 511: ■5 -a * >| .*:> .S a ca 1.3 = 1 CM ^ 0.-2 Spa c ^a '< a 3. 3 ca CCS 3^ t4 Q^ "o j2 a c3 o fe cq a5 .a-= S ■5 >>S oj a a a ^ 2 3 -j d .ii o N 1^. -^ 6i) ^ CD " "^ > cS^S OT m Lique tube only t. a a !§ d-d CO *0 trj t- — . • -: o ^ -■^ f^ .5 Cfi t>id = .2 t3Ph 131 a ""g^a ■^ -a & t-, d -J^ *tn - :S -1 d o a; " 03 ft _D d _, m iH , ■*" cr' '^ rt Ph ' ■^^•^ Si=-j ^1 T3 ^ U £ s ^J2, o ^ « O c^ P2 "o O' g ts ;2 i ' -== S 2 "S^s w 2 -a a * ■*" "B-TJ >> S £-= a, CT o .5 ^ c t^ c3 t. a d "> S o m ■— ' ^ 2.b^ 3 o rt =s^ >. S o ^- >>0.^ o a o cs d cs a » o o ^ '-3 ' ft ce ■^ o ^ -§1 Cli-S cs -a be a a ar. ^ "E 2 a d O 1-. ^ , •- 03 4) - O a -a-Z "^ rf d '" 'Z. <: a o ■-J 2 " ^ -- ■ O o ■- ca fO bij O "re Q s a 3 o >-. ti o t- _a 3_. o ji « § S-S a 5 ^ : ts cs a £ a; ^ a "r 3 cS 9 3^ ■ ^ 3 '5^ CD != >^° C6 ^ 3.2 o cS o 3 — 1-°|a • Mg2"« a o a S a o j:; =iHs S.a- i.'s a a ce a 3 "r 5 S >; -S ^ A-s c6 •" ^ j3 o tH o) d rt T^ d d o =i. 2 Q^;d t; ^ 'X' d (D . '^■- '—''-'> 1 q © ^ E^-d d. a +j +j oj H ^ Cfi Cfi *^ bD o CD a o 3 'S >> !?„'ca 5-a S 5 9 ■ III be a a Hi § =^ bD 3 -a a "CO . • .2 oj -a ; 3 -a 'a % o S o g' -^ _D ' — ' CO 5 a tl § a==^ as Qj ^ -^ c^ o bO o -o a ^ 3 a o o -a =3^ P q;) 03 5 ■g -a o o 9 _: & ^ a .z. cS o o' OJ ■ a S i; 3 SB S 2 a .tj — . CO * ca ^ .::^ Oj „-a X ^ a a --3 a O r- O ^ 3^.a_^ to i .s .s ° -3 :2 'a 1 i ^ Ph ca •la 2 3 £p.S ca '■/3 O 'j: a -a o Joum. of Agric. Sci. xn 21 304 .1 BdrtirldJ Disra.^r of Turin'/) (Brassica Napus) PARAisiTisM. Produces a soft -rot in swede, potato, carrot, radish and cabbage. No growth on beetroot. Ferments. Pectinase ; diastatic and peptonising ferments produced. Re.\ctiox of Bv-PROnrcTS. Diseased pulp neutral or feebly alkaline; bouillon neutral; acid produced in sugar media: production of ammonia in the vegetable media. COg evolved from diseased pulp. Reaction to Stains. Stains easily with various auilin dyes and by Gram's method. With Loeffler's flagella stain, one pole hyaline. rONCLUSION. While tlie organism under consideration has many characters in common with Pseudomonas campestris (Smith) and with Bacillus Oleracea (Harrison) it is clearly most nearly related to Pseudomonas destructans (Potter). The writer is of opinion that it is a varietal form of the latter. Of the differences mentioned in the table the most striking is the mode of attack of the disease while the root retains its shape, with the rind firm and the mature foliage healthy. This investigation was carried out in the Laboratory of Agricultural Botany, University College, Bangor. It was taken up at the instance of Prof. J. Lloyd Williams (formerly Adviser at Bangor) who had ob- served this disease on farms in North and South Carnarvonshire, but his enquiries in the other Counties failed to bring to light any other instances of its occurrence, although many cases were observed of other rots of swede and turnip. To him the writer wishes to express great indebtedness for guidance and kindly ad\'ice and also to Professor R. G. White by whose courtesy he had access to the ].,aboratory and the College Farm. LITERATURE. (1) Caeruthers and Smith. Journal of Botany, 39 (1901). (2) GnuNGHAM, C. T. Formation of Calcium Carbonate by bacteria in the soil. Journ. Agric. 8ci. 4. (3) Johnson and Adams. Bacterial Kot in Turnips and other Brassicas in Ireland. Economic Proc. of the Royal Dublin Society, 2, Xo. 1. (4) Potter, M. C. («) On a bacteiial disease of the Turnip (Br. Xapus). Proc. of Royal Hoc id y. 67. (b) On the parasiti^^m of Pseudontoiias deslructan.s. Proc. of Royal Society, 70. (c) A brief review of Bacteriological Research in Phytopathology. Science Progress, 5. (d) Bacterial diseases of Plants. Journ. of Agric. Sci. 4. (e) Brown Rot of Swedes. Journ. of Bd. of Agric. 10. ( / ) On a method of checking Parasitic Diseases in Plants. Journ. Agric. Sci. 3. JOURNAL OF AGRICULTURAL SCIENCE. Vol. XII. Part PLATE 111 -t -^,_ * , F.g. 2 e -'*««>-" * I Fig. 3 8. (i. Jones 305 (5) Pkiestley and Lechmere. A bacterial disease of Swedes. Joiirii. A(iric. Sri. 3. (6) Smith, Erwin F. Bacteria in Relation to Plant Diseases (Washington), 1, 2 and 3. (7) Taylor, T. H. Cabbage Top in Swedes. University of Leeds and The Yorkshire Council for Agricidlnral Education, No. 82. EXPLANATION OF PLATE IIL Fig. I. Diseased turnip with tlie young foliage of tlie growing point absent. Tiie plant had been left in the open and only removed at the time of winter ploughing. The rind had liecome hard, dry, and cracked. A small quantity of the shrunken pasty mass is shown. Fig. 2. Appearance of the crown of a plant with six secondary shoots. The aperture enclosed by them is surrounded also by leaf scars, which evidently belonged to the decayed apical bud. Fig. 3. Sections across the leaf-crowns of diseased plants. {Received June I'dlh, 1922. 21—2 A NEW METHOD FOR THE MECHAXTCAL ANALYSIS OF SOILS AND OTHER DISPERSIONS. By gilbert WOODING ROBINSON, M.A. Adviser in Agricultural Chemislry, University College of North Wales, Bangor. Introduction. Attention has been given of recent years to the possibility of devising methods of mechanical analysis which shall express the mechanical composition of a soil or clay by a continuous curve. The standard methods of mechanical analysis, as for example that followed in England, can of course be used to obtain such curves, but with the comparatively small number of fractious separated, very little detail can be inserted. Any multiplication in the number of fractions nnist ine\'itably be attended by an increase of experimental errors, and the problem must therefore be attacked by other methods. Od^n has devised an elegant method whereby the mechanical analysis of a soil or clay can be derived from an experimental curve showing the rate of accumulation of sediment from a column of material in sus- pension^. The weight of material is automatically registered from time to time and from the curve obtained a mass distribution curve is derived which gives a detailed representation of the composition of the material under examination. Wiegner'" has applied the theoretical principles of the Oden method in a very simple apparatus consisting essentially of a U-tube system in which a column of sedimenting suspension is balanced against a column of pure water. As the material falls below the point at which the water tube joins the fall tube, the specific gravity of the suspension decreases, resulting in a corresponding fall in the water column. The lieight of the water column is read from time to time and » Int. Mitt. Bodenkunde, 191"). 5, 2.57-311 : Koll. Zeit. 19U), 18, S.-J-^tS; Tram. Faraday Soc. 1922, 17. 327-348: Nefedof, ./. Exp. Lanilw. 1902. 3, 421-449, (nitlin.s a method similar in prinriple to that of Oden. but apparently purely empirical. • Lantlio. Vrrstirh.i SIfnf. 1918,91,41. G. W. Robinson 307 by appropriate calculation the experimental results can be thrown into the form of curves similar to those obtained by Oden's instrument. Both Oden's and Wiegner's methods have found apphcations in the study of dispersions 1. A serious drawback to Oden's apjaaratus is its expense. While it is of the highest value for the critical investigation of comparatively small numbers of samples, it can hardly come into use for routine work as the resources of an ordinary provincial institution would scarcely be equal to the outlay involved in setting up more than one such instru- ment and only comparatively small numbers of soils could be dealt with. Wiegner's apparatus is cheaper and could b}' modifications^ be made to give results of considerable accuracy. But here the time factor is serious. Using, as is necessary, a column about a metre long, several days would be required to obtain information as to the finer fractions which exert such an important effect on the properties of the soil. In the present work the writer has attempted to devise a method capable of giving more detailed data than the ordinary sedimentation methods and which, though not giving continuous curves, avoids the drawbacks of the Oden and Wiegner methods and can be used for standard mechanical analysis with great saving of time. Before discussing the new method of mechanical analysis a few remarks may be made on the expression of mechanical composition by means of curves. The simplest method of expressing the results of a mechanical analysis is to plot summation percentages* against particle sizes. Such a method is however almost useless in the case of highly dispersed substances, because in order to show the complete range the most characteristic particle sizes are cramped together near the «/-axis. A better distribution will be obtained by using the logarithms of particle sizes. The end of the curve corresponding to zero size is of course at — 00 but in practice a very manageable type of curve will be obtained^. Since the separation of particles of diameter smaller than -2 mm. is universally based on the principle of subsidence, whereby fractions are distinguished by their different setthng velocities in water, there are good reasons for using logarithms of setthng velocities instead of logarithms of particle sizes. In the ensuing treatment this method is used, the velocities being calculated in ordinary c.G.s. units. The same 1 Roll. Zeit. 1920,26. 100-121; ibul. 1920,26, 121^138. J. Landw. 1921, 69. 5-32. - As for instance in the method of reading the lieight of the water column. ' i.e. percentages of material of a given particle size or smaller. * Ct. Whittles, /. Agric. Sci. 1922. 12, 166-181. 308 Michdnlral A/iali/s!s of Soils ttiid otiter DispevslouK type of curve will be given and it will diispense with any assumption as to the size, shape and density of the settling particles. In this connexion, the probable effect of the gel coating which has been postulated for the finer particles of a soil or clay on the sedimen- tation of such material is worthy of consideration. It has been suggested that such coatings are not to be considered as uniform and discrete but as concentric shells of increasing degree of hydration. As an approximate basis of calculation let us assume that the emulsoid coating has the same density as the water in which the particles are suspended. What vnW be the effect of this coating on the settling velocity? If a be the radius of the falling particle and d the thickness of the gel coating or shell, then the velocity of fall of the particle in the absence of any coating is given by using the usual notation and assuming the density of water = 1, and the velocity of the coated particle by _2 gja + dy- where p^ is the mean density of the coated particle. weight of particle + shell volume |7ra»p + |7r[(a + df - a'] weight of particle + shell Now pi= — - -.. '^ volume fn{a+df a^p + [(a + df - o»] {a + df _ 2 g{a + df \a?p + [(a + rf)3-a'] _ ] lli.n ^^__. _ .^ - \a + df j ^2 flr a^(p-l) 9'ij' a+d ' v^a + d " Vi a ' a or i"o = V, a+d' It will be seen that the presence of a coating of appro.vimatcly the same density as the suspension medium will have a marked effect on the velocity of fall. The low velocities corresponding to the finer fractions G. W. Robinson :]09 will thus be conditioned not simply by the actual particle sizes but by the magnitude of the gel coating. The writer hopes to return to this point in a later paper. In the meantime it is suggested that the left hand portion of the curve in the case of clay soils may relate to such coated particles rather than to particles bounded by a sohd surface. By using log v instead of log particle size, the necessity for a decision as to the significance of these small velocities is postponed. It is of course obvious that the expression of the mechanical com- position of a soil or clay by a curve of this type offers a way out of the appalhng confusion created by the diversity of conventions used in different countries. Since practically all the methods in use are based on the principle of sedimentation, curves can easily be obtained if the settling velocities are known. Whether the principle is applicable to methods in which separation is effected by currents of water of varying velocity, as in the Hilgard method, the writer is unable to decide. As a first appro.ximation, it would appear that the method is applicable. We have assumed that the viscosity coefficient in the Stokes' equation is constant. With varying temperatures this is of course not the case. This difficulty can however easily be solved if a standard temperature be adopted, say, 15° C. By putting the viscosity coefficient of water at that temperature equal to unity and calculating the viscosities at other temperatures'^ with reference to this, the correction can be applied. Thus if results obtained at var3dng temperature are to be compared, it will be necessary to use log (v x specific viscosity) instead of log v. Theory of New Method. A fundamental assumption underlying all methods of mechanical analysis by sedimentation is that particles in a column settle inde- pendently of each other. That there are Hmits to this assumption is obvious. According to Oden this condition is fulfilled in suspensions of concentration not greater than 1 per cent.'^ Wiegner, on the other hand, brings evidence to show that concentrations of more than 5 per cent. can be used without serious inaccuracy, which may be due to the dangerous principle of compensating errors. Let us assume a suspension of soil or other granular material to consist of a number of fractions, a, b, c, etc., each uniform in itself, having hmiting velocities, i\, i\, v^, etc., respectively, and present in ' For the effect of temperature on the viscosity coefficient of water, see Hosking, Phil. Mat). 1907, 17. 509; ihid. 1909. 18, 260. - Int. Mia. Bodenlninde, 1915, 5, 276. 310 Mechanical Anah/'n's of Soils and other DinpevHlmis concentrations, A^, A^, A^. etc., respectively such that "ZA = C'^, the total concentration. Then if the fractions settle independently of each other, each fraction will behave as a separate column uniform in con- centration from top to bottom and we may represent the state of affairs at the beginning of sedimentation as in the upper diagram of Fig. 1, the relative amount or partial concentration of each fraction being represented graphically by the thickness of its column on the diagram. b { B Fig. 1. Diagrammatic representation of sedimentation. As sedimentation proceeds each column will fall bodily at its own appropriate velocity and the disposition after settling has proceeded for a certain time may be represented by the lower figure of the diagram. The black portion below the line CD will represent the amount of each fraction accumulated on the bottom of the sedimenting vessel, while the concentration of the suspension at any depth will be given by the total width of columns at that depth. Thus at depth d, the concen- tration will be equal to the sum of the partial concentrations of the ' Or Zi.-l + organic n\nftpr = C in the rase of nrdinarv Hiiila. Gr. W. Robinson 311 fractions, a to e, having velocities less than d/t. The ratio of the concen- tration at depth d after time t to the total concentration at the beginning of the experiment will thus give the proportion of material having velocities less than d/t. By determining the concentration for diilerent values of djt the data are obtained for a summation curve showing the relation between percentage of material and log setthng velocity. Experimental. The method used consists in allowing a soil suspension of known concentration to settle in a cyhndrical vessel and withdrawing samples for appropriate values of depth/time. By suitable choice of depth and time the concentration and hence the percentage of particles corresponding to any desired velocity can be obtained. Generally speaking a htre measuring cyhnder about 40 cm. in height and 6 cm. in diameter is used. There is of course no necessity to use a graduated vessel: any cyhnder of uniform cross section and suitable dimensions may be used. Samphng of the suspension is carried out by means of a 20 c.c. pipette passed through a cork or shive and adjusted so that when the cork rests on the top of the cyhnder the point of the pipette is at the desired depth below the surface of the hquid (see Fig. 2). The column having settled for the required time the pipette, previously adjusted for depth, is closed at the top with the finger, in order to avoid samphng the upper layers, and lowered very carefully till the cork rests on the top of the cyhnder. The finger is then removed and 20 c.c. of the suspension withdrawn. Every precaution is of course taken to avoid shaking or mixing the layers of the suspension at the point of sampling. With a column of the dimensions mentioned the withdrawal of 20 c.c. causes a fall in level of about 7 mm. This probably represents the extreme error in samphng. It is assumed that the 20 c.c. of suspension withdrawn represents the concentration at the point of the pipette. Probably the liquid comes mainly from above, but to some extent from below this point. A separate experiment with a column which had settled for several weeks and which had formed clearly defined strata, showed that it was possible to pipette to within 2 to 3 mm. of a stratum without disturbance. It will be shown later that an error of a few milhmetres in samphng involves a neghgible error in the final result. Careful manipu- lation is of course necessary in this operation. The 20 c.c. of suspension 21—5 Fig. 2. Method of sampling. 312 Mechanical Aufdi/sis of Soils and other Diapermnis is delivered into a fiat porcelain dish wliich lias been previously ignited and weighed. Dishes ordinarily used for the estimation of total solids in milk are convenient for the purpose. The sample is taken to dryness on the water bath and, if the estimation is to be made on unignited material, weighed after attaining constant weight. Ordinarily it is ignited in a muffle, an operation which only takes a few minutes at red heat, and weighed after coohng in a desiccator. From the weight of ignited material the concentration of the sample of suspension is calcu- lated. By sampling in such a way that successively smaller values of depth/time are used, the same suspension may be shaken up and sampled over and over again. The partial concentration of any fraction at a given depth is unaltered until the top of the fraction column has sunk below that depth, as will be seen by reference to Fig. 1. The removal of a sample of suspension does not therefore aiiect the concentration of the suspension with respect to fractions of smaller velocities. In an actual experiment a 2-5 per cent, suspension of a clay was prepared by shaking up 100 grams of powdered clay for 24 hours with 2 litres of water containing 100 c.c. of 1 per cent, sodium carbonate solution and making up finally to 4 litres. A litre cylinder was then filled to within 3-4 era. of the top with the well mixed suspension and after again shaking for a minute, allowed to stand for six minutes. A 20 c.c. sample was then withdrawn at 36 cm. depth as described above. After drying and ignition, the weight of ignited material was found to be -376 gram. Subtracting -005 gram for the amount of sodium car- bonate in 20 c.c, we have the nett weight of ignited material as -371 gram and the concentration of the suspension at the point sampled, 1-855 per cent. The original concentration being 2-5 per cent., we find that the concentration of the suspension at 36 cm. after six minutes is r855 •jr~^ X 100=74-2 per cent, of the original concentration. In other words 74-2 per cent, of the clay, reckoned as ignited material, has a settbng velocity less than -1 cm. /sec. Other determinations were made for successively smaller velocities and the residts are set out in Table I. For the sake of brevity the w-eight of ignited material is given after subtraction of the -005 gram of .sodium carbonate. In Fig. 3 cui-ves are shown for the clay of the experiment just described and for a few other ty})ical clays and soils. The vertical dotted lines, ^I, B, and t', are the ordinates corresponding to clay, fine silt, and silt, respectively, according to their settling velocities in the Enghsh method. In order to bring the fine sand, coarse sand and fiiie gravel G. W. Robinson 313 Table I. London Clay. 2-5 per cent, suspension in -025 per cent, sodium carbonate solution. Temperature 12-1G°. Ignited %of material Concen- original Time Deptli Velocity in 20 c.c. tration concen- sees. cms. cm./sec. log V gms. u /O tration 360 36 •1000 1^0000 0^371 b855 74-2 600 20 •0333 2^5227 0^358 1-790 71-6 600 6 ■0100 20000 0325 1-625 65-0 6000 20 •0033 3^5227 0^293 1-465 58-6 6000 6 •0010 3^0000 0252 1-260 .50-4 60000 20 ■O0033 4-.5227 0^220 1100 44^0 60000 6 •01 1010 4^0000 0^178 0-890 356 600000 20 ■000033 .^•5227 0135 0-675 270 600000 6 •000010 5^0000 0093 0^465 18-65 100 5 Log (' 3 1 1 3 Fig. 3. Summation Curves showing Composition of Typical Soils and Clays. 314 Mechanical Analysis of Soils and other Dispersions into the diagram, their velocity values were calculated on the assump- tion that they obey Stokes' law and that the diameter of particle repre- sented by the upper limit of the silt is -04 mm. The lines thus obtained are D, E, and F, respectively and are inserted for the sake of com- pleteness. By taking into account the coefficient of viscosity and using glycerine-water mixtures, these points on the di.stribution curve might be obtained by the above method. It is however simpler to use the sieve method to fill in the right hand portion of the curve. It may be added that the data of an ordinary mechanical analysis are given graphically by the difference between successive intercepts on the ordinates A, B, C, etc. For example the fine silt is given by the difference between the intercept on B and the intercept on A, the silt by the differ- ence between the intercept on C and the intercept on B and so on. By suitable choice of depths and times any required degree of detail can be secured in any part of the curve. If a large number of points are required it is convenient to prepare a large volume of suspension and work with a number of separate cyhnders. The details may be left to the convenience and ingenuity of individual operators. Possibly a more convenient, though scarcely less expensive, method of sampling may be devised. Any measurable physical property of dispersions which depends on concentration may be considered in this connexion. For very dilute suspensions it is possible that a nephelometric method might be devised. The method may find its best use as a substitute for the present standard method of mechanical analysis. The suggested procedure in this case is as follows. The air dried sample is treated with A'jb hydro- chloric acid exactly as in the ordinary method, using however 20 grams instead of 10 grams of soil. The fine gravel and coarse sand are separated in the ordinary way and the finer material passing the 100 mesh sieve is shaken with 600-700 c.c. of water and 50 c.c. of 10 per cent, ammonia in an end over end shaker for two to four hours, the longer period being necessary in the case of soils with much organic matter. After shaking, the suspension is made up to one Utre, which is equivalent to 2 per cent., reckoning on the original material. The following determinations are then made successively by the method described above. JtelJth Time Velocity Giving cm. hrs. min. cm./sec. percentivjie of (o) 30 0 5 01 silt + fine silt + clay (6) 12 0 20 001 tine silt + clay (c) (i 16 40 00001 clay or 7-2 20 0 ■ or 81) 24 0 (4. W. Robinson 315 Lastly to determine the tine sand, the suspension remaining after the last sampUng is poured away to about 200 c.c. without shaking up the sediment. The remaining suspension and sediment are then washed into a beaker and the fine sand determined by the ordinary method, using the 10 cm. and 100 seconds, or 7-5 cm. and 75 seconds sedimen- tation. The following figures are the results of an actual determination on a Pennant Grit soil from CTlamorganshire : By ordinary method Fiue gravel ... ... ^'6% Coarse sand ... ... 20-1 Moisture ... ... 4-2 Organic matter ... ... 10-2 Total 391 So that fine sand + silt + fine silt + clay should = 100 - 39'1 =60-9 (1). By new method, using 2 "(, suspension. For 5 mins. and 30 cm., weight of ignited material iu 20 c.c. = -112 gm. Therefore concentration =5 x -112 = -560 %. Therefore silt + fine silt + clay = '^-^^''-^-°^ = 28-0 % (2). Similarly for 20 mins. and 12 cm., ignited material = -065 gm. Concentration = -325 % and fine silt + clay = =16-25 % (3). Similarly for 20 hrs. and 7-2 cm. ignited material = -019 gm. Concentration =-095 % and clay = ," =4-75 % (4). Subtracting (4) from (3). fine sat = ll-5 %. Subtracting (3) from (2). silt =11-7.5 "o- Subtracting (2)from (1) fine sand = 32-9 %. By direct sedimentation at end of e.xperiment, tine sand = 32-5 %. Effect op Variations in Conditions of Working. In order to form some idea as to the latitude allowable in the con- ditions of working the following points were investigated. («) Effect of concentration. A number of determinations were made on a clay suspension of varying concentrations. Results are given for different times and depths. Depth 6 cm. Time 10 mins. Concentration Percentage of original of original suspension concentration 1 -50% 63-0 -625 61-0 1-00 620 1. 1 2-00 63-2 ns. ! 2-50 64-0 4-00 63-1 '5-00 65-7 316 Mechanical Analysis of Soils and other Dispersions i -50 530 Depth t3 cm. 1 100 51-5 Time 100 mins. 'i 200 52-2 1 400 490 „ .,1- (100 450 Depth lo cm. 2-00 44-2 T.mr 240 mms. ] g.^,, ^g.3 ,, ,, , f 100 37-5 Depth 7 cm. \ gg.^, T,.ne 27 hours | g.^^ ^,._ The variations in the results obtained with varyin<; concentrations are not serious when the nature of mechanical analysis and the nature of the material under examination is taken into account. In the standard Phighsh method the concentration used is about 2 per cent. In \'iew of the small weights to be dealt with in very dilute suspensions and the consequent magnification of weighing errors, it was decided to use 2 per cent, suspensions as a general rule and in the comparisons given later of the results by the new method with those by the old method, this concentration was used. (h) Diameter of column. The diameter of the column may be ex- pected to have some effect on the result obtained since, apart from any boundary effect, the fall in height due to the removal of liquid will be greater in narrow columns. The following experiment senses to test this point. Chiy suspension 25 %. Time 19-5 hrs. Dejjth 9 em. Diameter of cvlinder Weight of ignited material 7-7 cm. -102 gm. 60 -101 4-6 102 3-4 098 The last cylinder was an ordinary 250 c.c. measuring cylinder. It would appear that with J litre, litre and 2 litre cylinders consistent results may be obtained and that the diameter of the cylinder is im- material. In general, a litre or \ litre cylinder was used. (c) Equivalence of Depth/Time ratio. Theoretically the same con- centration should be obtained for different times and depths provided that the ratio depth/time, i.e. the hmiting velocity is constant. This point is investigated in the following experiments. Kaolin. 2-5 % suspension in -025 "„ sodium carbonate. Weight of ignited Depth Time material cm. hours gra. 5 5 -165 19-5 19-5 169 Clay. 2-5 % suspension in -025 % sodium carbonate. 4 1 -254 12 3 253 5 18-75 -178 18 68 177 G. W. Robinson :n7 {d) Errors Id Depth of Saniplijuj. Tlie effect of errors in depth of sampling may be best demonstrated by considering the nature of the vertical concentration gradients in a column of suspension after varying times. A series of curves showing the relation between depth and con- centration for different times can readily be derived from the summa- tion curve for the material under consideration. For example if it be known from the summation curve that 50 per cent, of the material has 100 2 0 2 Lot; V 20 30 40 im. L)c|itli Fig. 4. Concentratiiin Gradients at Different Ueptlis. a limiting velocity less than -01 cm. per second (log v = 2-0000), then at depth 10 cm. after 1000 seconds, the concentration will be 50 per cent, of the original concentration. A series of concentration gradients for a clay is shown in Fig. 4. AB is the summation curve and the curves I. II, III, IV, and V give the percentages of the original concen- tration at different depths for 10 sees., 1 min., 10 mins., 100 mins., and 318 Mechcuiical Anali/six of Soils and oflirr Dixpersioiu 1000 mins. respectively. The log f abscissae refer of course to the summation curve AB. and tlie depth abscissae to the concentration curves. It will he seen that the change in concentration witji depth after any given time is very gradual below the first few centimetres. Thus, after ten minutes, the concentration in the case illustrated only changes about 2-5 per cent, of the total concentration between 10 cm. and 20 cm. With longer times the gradient becomes rather steeper. Errors of the order of a few millimetres in depth of sampling have thus very little effect on the concentration obtained. This of course only holds so long as the material under experiment has a fairly smooth summation curve. With a material having an irregular type of curve, depth errors might be more serious. Similar considerations can be developed for the time and tem- perature error. The errors introduced by the above variations in working conditions though scarcely negligible are nevertheless not serious when the character of the material is considered and it may be doubted whether they are of significance from the point of view of the genetics and physical pro- perties of soils and clays. The Oden method is of course less exceptionable from the point of view of delicacy and must be used where critical data are required. The method described in the present paper could of course be made to give results strictly comparable among themselves. It is however desirable to have a method which can admit of some latitude to suit the convenience of individual worlvcrs and which has to that extent and within reasonable limits the character of an absolute method. Agreement with Re.sults obtained by the Standard Method. In view of the enormous number of results accumulated by the older method, it would be a doubtful advance to suggest a new method, even though more convenient and accurate, if the results obtained by it could not be used for comparison with the older results. A considerable number of determinations were therefore made by the new method and the results compared with those already obtained by the old method. In the following table the residts of this comparison are given. For con- venience of statement only three values are given, namely, clay, fine silt, and silt. The same method of dispersion was used throughout except in the case of certain clays, which contained practically no organic matter and were dispersed in -02.5 per cent, sodium carbonate. (1. W. Robinson :119 Table II. Comparison of Results hi/ Old and New Methods. Silt fcloil or Clay Pine silt A clay OlcT New 01(1 New A 43 A 9-5 10-8 26-8 26-8 A 55 A 10-5 101 21-7 21-7 F2B 29-0 28-7 340 34-0 D44A 3-5 5-9 381 38-0 G8 13-3 16-6 240 20-0 D 10 13-9 13-6 57-6 57-9 C34 3-9 4-0 13-6 13-5 Dll 2-7 3-8 14-6 14-4 D73 5-8 2-4 22-6 22-2 D49 20 2-2 12-2 9-3 Llangoed clay 43-4 48-4 30-7 30-8 Ruabon 25-8 23-2 (38-2 44-4 44-4 (26-4 London clay 400 37-2 (38-4 [24-2 22-6 2G-4 (27-4 (46-8 Kaolin 235 ^23-6 I23-8 48-4 46-4 (47-2 (110 121 245 28-7 Ab. 4 ^10-5 (l3-0 11-75 26-0 26-5 130 25-9 26-2 Old New 150 15-0 131 13-2 20-9 20-R 15-8 15-9 18-4 18-4 9-3 9-2 100 100 7-2 70 131 17-4 8-2 120 3-9 3-9 17-5 17-4 fl2-6 11-7 13-4 12-4 16-2 13-7 • 17'8 15-8 17-7 180 18-7 17-3 lS-5 17 3 The agreement between the two series of results is generally close and not unsatisfactory when it is considered that mechanical analysis by the older method is hable to considerable errors. On the average, putting the figures obtained by the old method as 100, the new method gives 102-7 for the clay, 100 for the fine silt and 99-6 for the silt. It would thus appear that there is no appreciable constant error in com- paring results by the two methods. In view of the large number of manipidations recj^uired in the old method, there are more occasions for error than in the new , and any serious disagreement is at least as likely to be due to errors in the former as in the latter method. It will be noticed that the most serious disagreements are in the case of light soils. To secure a perfect comparison it would be necessary to secure that the prehminary dispersion is exactly the same in both methods. In the standard method the material is repeatedly triturated after each pouring off. The method provisionally adopted for dispersion in the new method, namely a 2 to 4 hour shaking in an end over end shaker would appear to give a comparable degree of dispersion. The longer period is apparently necessary in the case of soils rich in organic matter. Dispersion by sodium carbonate^ gave good results with raw- clays but was unsatisfactory with soils containing much organic matter. 1 Cf. .Joseph and Martin, ./. Agnc. Sei. 1921, 11, 29,3-303. 320 Mechanical Aualiisia of Soils and other Dispersions Conclusion. The method of mechanical analysis above described has two recom- mendations. It is more expeditious and economical than the standard method. Granted that the new method gives reliable results, its adoption would remove one of the (gravest objections against the jiresent method, namely its laboriousness. With the new method it has been found possible in the writer's laboratory to carry out six mechanical analyses in a day. With pro])er organisation there should be no difficulty in carrying out 35-40 analyses in a week. Anyone familiar witli the routine of the older methods will realise that it would be impossible to deal witii such numbers single handed. There is the further consideration of economy in the use of beakers, filters, etc., to say nothing of the dis- tilled water, ammonia and hydrochloric acid required. On the other hand the suggested method requires careful manipulation. Working with 2 per cent, suspensions, an error of 1 mgm. in weighing corresponds to -25 per cent, in the result obtained. It is doubtful if the method would commend itself for teaching purposes. The new method mav be used within limits for obtaining continuous curves such as are obtained by the Oden and Wiegner methods. Pro- vided temperature conditions are controlled it is easily possible to carry the distribution curve for a soil or clay considerably beyond the limits of clay as defined by the standard method, and the constitution of the finer fractions can thus be investigated over ranges hitlierto scarcely explored^. An application may also be found in the investigation of changes in degree of dispersion consequent on manuring, cultivation and season. With regard to the errors of the method it may be remembered that the experiments recorded were carried out in ordinary measuring cylinders. These are rarely uniform in cross section. With perfectly cylindrical columns, no doubt, better results would be obtained. With regard to temperature effects, no attempt was made to secure rigid control of temperature conditions. As in the ordinary method, the settling took place in an ordinary laboratory \vith its unavoidable vicissitudes of temperature. Care was however taken to avoid the proximity of sources of heat. In certain cases where the settling columns were in the vicinity of radiating surfaces, unsatisfactory results were obtained. Generally speaking for the longer periods the columns were ' A considerable number of clays have, in fact, been followed by this method as far as log 0= 7, which appears to be near the lower limit. a. W. Robinson 321 put away in a large cupboard in a room without any burners or other sources of heat. The possibihty of convection currents was thus avoided as far as possible. By a suitable choice of depths and times the method could be readily adapted to systems of fractionation other than that in use in this country. In the American systems, where the numbers of fractions separated are large, the amount of work involved and the possibiUties of error are serious. The above method might be of use in such cases. By setting out the results as summation curves the mechanical analyses by any system of grading could be obtained by interjjolation. Summary. (1) The expression of mechanical composition by means of con- tinuous curves is discussed. It is suggested that a convenient repre- sentation will be obtained by showing summation percentage as a function of the logarithm of settling velocity. (2) The effect of a gel coating on the setthng velocity of a particle is examined and it is shown that a reduction in velocity takes place which is a simple function of the thickness of the gel coating. (3) A method is outhned by which the mechanical composition of a soil or clay is derived from determinations of the concentration of a settling suspension for different values of depth/time. (4) A shortened method for mechanical analysis is described which gives results in good agreement with results obtained by the present standard method. (5) The effect of various modifications in conditions of working is discussed. (6) The nature of the concentration gradients in a settling column of a suspension is examined. It is shown that below the first few centi- metres the change in concentration with depth is very gradual. (Received July 8th, 1922.) CAMBRIDGE AGRICULTURAL MONOGRAPHS General Editors: T. B. Wood and Sir E. J. Russell Basic Slags and Rock Phosphates. By George Scott Robertson, D.Sc. (Dunelm), F.I.C., with a preface by Sir Edward J. Russell, D.Sc. (Lond.), F.R.S. With 8 plates and a map. Small Royal 8vo. 14s net. Inorganic Plant Poisons and Stimulants. By Winifred E; Brenchley, D.Sc, F.L.S. Royal Svo. With 19 illustrations. 9s net. Plants Poisonous to Live Stock. By H. C. Long, B.Sc. (Edin.). Royal Svo. With frontispiece. 8s net. Cambridge University Press, Fetter Lane, London, E.C, 4: C. F. Clay, Manager CAMBRIDGE FARM INSTITUTE SERIES General Editors: T. B. WOOD, C.B.E., F.R.S. and E. J. RUSSELL, D.Sc, F.R.S. Crown Svo Fa rm Accounts. By C. S. Orwin, Hon. M.A. Director of the Institute for Research in Agricultural Economics in the University of Oxford. Large crown Svo. [Out of print Fungoid and Insect Pests of the Farm. By F. R. Petherbridge, M.A., Biological Adviser, School of Agriculture, Cambridge. Large crown Svo. With 54 illustrations. 5s 6d net A Student's Book on Soils and Manures. By E. J. Russell. Second edition, revised and enlarged. Large crown Svo. With 41 illus- trations. 8s net. Cambridge UniTersIty Press, Fetter Lane, London,, E.C. 4: C. P. Clay, Manager THE UNIVERSITY OF CHICAGO PRESS Agricultural Economics By EDWIN G. NOURSE. A volume comprising several hundred carefully edited selections with valuable original introductions, which make of the material a connected whole. This book follows the plan, successful in so many recent cases, of combining in one volume the virtues of the usual text and the collateral book of readings. Particularly strong is the emphasis upon the interrelations of the technique and the economies of agriculture. Most economists have made Uttle use of the data of agriculture in developing their exposition of economic principles. Chapter III of this book is a notable example of the attempt to found our agricultural economics upon a sound under- standing of the technical nature of the productive processes of agriculture. Many selections also show how changing economic or commercial con- ditions are modifying the practices of agriculture, and how at the same time technical considerations, such as the exhaustion of nitrogen from the soil or the growth of knowledge about plant or animal breeding, introduce new factors into the economic problem of agriculture. xxvi + 896 pages, Svo, cloth; 22s net. COMPANION VOLUME Outlines of Agricultural Economics 106 pages, paper; 4s net. The Cambridge University Press Agents for the British Empire (except Canada) London, Fetter Lane, E.C. 4 CONTENTS. (All Rights reterved.) PA6B 1. Deighton, Thomas. Some investigations on the electrical method of soil moisture determination. (With 6 Text-figures.) . 207 2. Woodman, Herbert Ernest. The chemistry of the strength of wheat flour .... * 231 3. IvANOFF, E. I. On the use of artificial insemination for zooteohnical purposes in Russia . ... . • -44 4. Capstick, J. W. and Wood, T. B. The effect of change of tempera- ture on the basal metabolism of swine. (With 4 Text-figures.) . 257 5. HoRTON, E. and Salmon, E. S. The fungicidal properties of certain spray-fluids. Ill 269 6. Collins, S. H. and Thomas, B. The sugars and albuminoids of oat straw .......... 280 7. Robinson, Gilbert Wooding. Note on the mechanical analysis of humus soils ......•••• 287 8. Jones, S. 6. A bacterial disease of turnip (Brassica najtm). (With Plate III.) 292 9. Robinson, Gilbert Wooding. A new method for the mechanical analysis of soils and other dispersions. (With 4 Text-figures.) 306 The Journal of Agricultural Science is issued in quarterly parts of about 100 pages, with plates and figures, four parts constituting a volume. Volumes I — XI (1905—21) now ready. In four parts, paper covers, 304. net per volume. Single numbers 10«. net each. Quotations can be given for Buckram binding cases and for binding subscribers' sets ; also for bound copies of back volumes. Papers for publication may be sent to Professor T. B. Wood, C.B.E., M.A., F.I.C., F.RS., School of Agriculture, Cambridge, or to the associate Editors. Other oommunicatious should be addressed to the University Press, Cambridge. Contributors will receive twenty-five copies of their papers free. A limited number of additional copies may be purchased if ordered when final proof is returned. In order to meet the increased cost of printing, contributors will be asked to pay a proportion of the cost of setting up tabular nfatter and of preparing blocks where these exceed a certain proportion of the whole communication. The subscription price, payable in advance, commencing with Vol. X is 30* net per volume (post free); single numbers 10». net. Subscriptions may be sent to any Bookseller, or to Mr C. F. Clay, Manager, Cambridge University Press, Fetter Lane, London, E.G. 4. The Cambridge University Press has appointed the University of Chicago Press agents for the sale of The Journal of Agricultural Science in the United States of America and Canada and has .■authorised them to charge the following prices : — Subscription, f€.60 net per volume ; single copies, ^i.ih net. FRmTED IN ENOIJIND AT TB£ CAMBBIDOE UNTVEBSTTY PBE6S Vol. XII. Part 4 October, 1©22 THE JOURNAL OF AGRICULTURAL SCIENCE EDITED FOR THE PLANT BREEDING AND ANIMAL NUTRITION RESEARCH INSTITUTES AT CAMBRIDGE, AND THE ROTHAMSTED RESEARCH INSTITUTES BY Professor R. H. BIFFEN, M.A., F.R.S., Cambridge Sir a. D. HALL, K.C.B., M.A., LL.D., F.R.S., Ministry of Agriculture, London F. H. A. MARSHALL, Sc.D., F.R.S., Cambridge Sir E. J. RUSSELL, D.Sc, F.R.S., Rothamsted Experimental Station, Harpenden Professor T. B. WOOD, C.B.E., M.A., F.LC, F.R.S., Cambridge IN CONSULTATION WITH B. C. ASTON, Department of Agriculture, Wellington, New Zealand Dr C. a. barber, CLE., School of Agriculture, Cambridge Professor B. T. P. BARKER, M.A., National Fruit and Cider Inititute, Long Ashton, Bristol W. BATESON, M.A., F.R.S., John Innes Horticultural Institute, Mcrton, Surrey J. R. CAMPBELL, B.Sc, Department of Agriculture, Dublin L B. POLE EVANS, Department of Agriculture, Pretoria, South Africa F. B. GUTHRIE, Department of Agriculture, Sydney, N.S.W. Professor J. HENDRICK, B.Sc, Marischal College, Aberdeen Sir T. H. MIDDLETON, K.B.E., C.B., M.A., The Development Commission, London Dr FRANK T. SHUTT, F.I.C, Experimental Farms, Ottawa, Canada Professor W. SOMERVILLE, M.A., D.Sc, School of Rural Economy, Oxford Dr a. C. TRUE, Department of Agriculture, Washington, D.C., U.S.A. Sir FRANCIS WATTS, K.C.M.G., Barbados Dr H. J. WHEELER, American Agricultural Chemical Co., Boston, Mass., U.S.A. Cambridge University Press C. F. Clay, Manager London : Fetter Lane, E.C. 4 also H. K. Lewis & Co., Ltd., 136, Cower St., London, W. C. t and WheldoN and Wesley, Ltd., 2-4, Arthur St., New Oxford St., London, W.C. 2 Chicago : The University of Chicago Press (Agents for the United States and Canada) Bombay, Calcutta, Madras: Macmillan and Co., Ltd. Tokyo : The Maruzen-Kabushiki-Kaisha 'Price Ten Shillings net [Issued 12 December 1922.] '♦♦HH* |»«»»«»»»^»»»»<>»m ti4»>««>»»».»»»»»»»»»»»»»»««»»>»t»»»»><»»»>»«»«»»»»»»»^»»»4»t»>«>»«t>«>>»»>>>«4*«»»»>»»««l»>«»>«*»l ll>^ laaaaaaaaanaaanaaanDanaanaanaannaannnDnaDDDGCDonnDna: Cambridge University Press British Grasses and their Employment in Agriculture. By S. F. Armstrong, B.A., F.L.S. With 175 illustrations. Demy 8vo. 105. 6d. net. Manuring for Higher Crop Production. By SirE. j. Russell, D.Sc, F.R.S. Second edition, revised and extended. Demy 8vo. With 17 text-figures, ^s. 6d. net. The Fertility of the Soil. By Sir E.J.Russell, D.Sc, F.R.S. With 9 illustrations. i6mo. Cloth, 35. net; lambskin, 45. net. Cambridge Manuals Series. A Course of Practical Chemistry for Agricultural Students. By H. A. D. Neville, M.A., F.I.C., and L. F. Newman, M.A., F.l.C. Demy 8vo. Vol. I, loj. 6d. net; Vol. II, Part I, 51. net. A Course of Practical Work in Agricultural Chem- istry for Senior Students. By T. B. Wood, c.b.e , MA. F.I.C, F.R.S., Drapers Professor of Agriculture in the University of Cambridge. Demy 8vo. Paper covers. 2s. net. Cattle andThe Future of Beef -Production inEngland. By K. J. J. Mackenzie, M.A., with a preface and chapter by F. H. A. Marshall, Sc.D., F.R.S. Demy 8vo. js. 6d. net. A Course of Practical Physiology for Agricultural Students. By J. Hammond, M.A., and E. T. Halnan, M.A. Demy 8vo. With blank pages interleaved. Cloth, 6s. 6d. net; paper covers, 45. 6d. net. Physiology of Farm Animals. By T. B. Wood, C.B.E., M.A.. F.I.C, F.R.S., and F. H. A. Marshall, Sc.D., F.R.S. Part I, General. By F. H. A. Marshall. Demy 8vo. i6s. net. Microscopic Analysis of Cattle-Foods. By T. N. Morris, B.A. With 54 text-figures. Crown 8vo. y. net. Insect Pests and Fungus Diseases of Fruit and Hops. A Complete Manual for Growers. By P. J. Fryer, F.I.C, F.C.S. Crown 8vo. With 24 plates in natural colours and 305 original photographs and diagrams. 455. net. Agriculture in the Tropics. An elementary treatise. By J. C Willis, M.A., Sc.D., F.R.S. Second edition, revised. Demy 8vo. With |^ 31 plates. 12s. 6d. net. Cambridge Biological Series. 1^ Rubber and Rubber Planting. By R. H. Lock, Sc.D. Crown Svo. With 10 plates and 22 figures. 6s. 6d. net. |l CAMBRIDGE UNIVERSITY PRESS FETTER LANE, LONDON, E.C.4: C. F. CLAY, MANAGER □aaaanaDaaaaaaaaDcnaaaanaanoaaaaaannaaaaaaaaaanacDnncxiaaGnaaxEEa ^iitn»t t i»«4i*t II tiiiTttt-fttttt-tT-H"M 1 1 tt tt t ttt »ttttttttttT-^ttt^t^-tt1-^-n-^ttt-tfttf1-Tl ft I ri i-t-tttttttttt t it t ti tt-tttttttttitt n mtiAt Volume Xil 0CT013EK, 1922 Part IV -LiSit.AKY NidW YOWC aOTAM tCAIL TEMPERATURE AND OTHER FACTORS AFFECTlNCi THE QUALITY OF SILAGE. By ARTHUR AMOS, M.A. AND THE LATE GWILYM WILLIAMS, M.A., N.D.A. (Schoul of Agriculture, Cambridge University.) Introduction. In 1883 Mr George Fry, F.L.S.^ described a series of observations which he had made upon silage. From these he drew the conclusion that if the conditions of silage making were such that the temperature exceeded 45° C. sweet brown silage resulted, but that if the temperature failed to rise above 40° C. then sour silage with a rather repulsive odour was produced. These results were obtained in the type of silos then commonly in use, which varied in depth generally between 12 and 18 feet, frequently had a considerable surface area and were filled comparatively slowly. In 1884 Dr Augustus Voelcker, F.R.S.- confirmed Mr Fry's ex- perience with silage in the regulation and maintenance of a proper temperature and mentions 125° F. (51-o° C.) as being the point below which sweet or hay fermentation does not take place. He further stated that sweet silage keeps only a short time on exposure to air, whereas sour silage may keep {J-9 months exposed to air. In 1886, Dr J. A. Voelcker^ described the making of sweet silage by ensuring the temperature of fermentation rising to 122° F. = 50° C, "the point which Mr George Fry considers must be reached to get sweet silage."' M. Goffart^, however, is quoted as follows: "My maize, my green rye, my fodders of every kind have scarcely changed colour after eight or ten months of ensilage."' From this it is obvious that the silage of M. Gofi'art, which had "scarcely changed colour" was very different material from the "sweet brown silage" advocated by Fry. Babcock and Russell'^ state that " the popular opinion that good silage can only be made with considerable heat is erroneous."" Good silage 1 Agricultural Gazette, Aug. 27th, 1883, Nov. 26th, 1883 and April Uth, 1884. cvj " Voelcker. Journal of the Eoijul Agricultural Society, \%M. ^ ' Voelcker. Il>id. 1886. ^ Silos for British Crops by the sub-editor of the Fivld. _^ ^ Babcock and Russell. Wisconsin Agricultural Experiment Station, I7th and 18th J " Annual Report, 1900. _., Journ. of Agric. Sci. xii 22 324 Tcnipenditn' qff'cctiii;/ the Qiiatif;/ of Silage was made by them in .small retainers at temperatures which did not exceed 80°F. = 27°C.; thus, as the authors say, "disproving Fry's theory, that a temperature of at least 120° F. was essential for good silage." Many American and other experimenters have obtained similar results, of which may be quoted those of Neidig'. In this case the maximum temperature recorded at the centre as distinct from the top surface of the silo was 91° F. = 33° C, and good silage resulted. When in 1917 the authors began to study silage-making in the American type of tower silo, agricultural opinion in England still re- tained the distinction between "sweet"' and "sour" silage which Fry had enunciated and did not realize the possibilities of other types of silage. For this reason, as well as for the fact that silage produced in the experimental silo at Cambridge varied very greatly not only from year to year, but also in different parts of the same silo, it became apparent that, before reliable feeding experiments with silage could be conducted, it was necessary firstly to define the different types of silage w hich were capable of being produced, and secondly to define the con- ditions under which each tyj)e could be produced. The observations in this paper are divided into two parts. The first part concerns those made upon silage produced by a large number of silage growers in the Eastern Counties and elsewluTe, from which a few characteristic examples have been described. The second part describes a more accurate series of observations made in the experimental silo at Cambridge. In the earlier years the late Mr G. Williams co-operated but his untimely death in 1920 prevented him from helping to complete the work. QUAIJTATIVE OBSERVATIONS MADE ON SILAOE PRODUCED BY FARMERS UNDER VARIOUS CONDITIONS. 1. On farms belonging to Mr .T. Thistleton Smith, Kakenham, Norfolk. Mr Thistleton Smith farms light to medium land, u])on which the silage crops generally stand well. The crop mi.xture consists of wheat, oats, and tares, which is allowed to become fairly mature before cutting (the oats being past the milk stage and the tares well seeded). If the crop is succulent it is allowed to wilt a few hours before being ensiled. ' Neidig. "Chemical Changes in .Silage ferniontation." lowii Agric. Kxp. Utation, Research Bulletin, No. Iti. X. Amos and G. Williams 325 The resulting silage over a period of years and in several silos has been of a yellowish-brown to brown colour with an acid though quite pleasant smell. The silage has been readily eaten by all classes of stock, which have invariably thriven upon it. This is much the most common type of silage now being produced in the Eastern Counties of England and appears to be universally produced in tower silos from a mature crop which is reasonably dry when ensiled. 2. In two silos belonging to Mr F. W. D. Robinson, Beccles, Suffolk, 1919-1920. The crops in this case were oats and tares grown upon hght land. They were cut in a medium condition of maturity, the oats being in milk and the tare pods full-grown in length but with immature seeds. The crops were ensiled immediately after cutting. The resulting silage possessed a green colour with a smell which was neither "sweet" nor "sour"; it can best be described as "fresh" and "fruity.'" Stock ate it greedily and throve upon it. It would seem highly probable that M. Goffart's " crops of every kind which had scarcely changed colour after 8 or 10 months of ensilage" must have been of this character. 3. On Mr Arnold Oliver's farm at Bures in Suffolk. In two successive years 1919 and 1920 the silos were filled with oat and tare crops cut in a medium condition of maturity, and under con- ditions very similar to those prevailing in Mr Robinson's silos. In each case the crop was ensiled immediately after cutting and produced a green silage with the same "fruity" smell observed with Mr Robinson's silage. In 1920 a maximum thermometer was inserted about the middle of the silo and this recorded 30'^ 0. 4. On General Adlercron's farm at Culverthorpe, near Grantham. In 1920 the silo was filled dTiring September and October from a late-sown oat and tare crop which was badly laid. The crop had grown to a great length and was semi-rotten close to the ground. Much rain fell during the ensiling process. The resulting silage was dark brown almost black in colour and possessed the most objectionably sour pungent smell. So tenaciously did this smell chng to anything touching it that the writer, who had occasion to handle some, was unable to get the taint from his hands for 36 hours. This silage was eaten by cattle but without rehsh. It is satisfactory to record that in the season 1921-22 beautiful silage has been produced on this farm by ensiling under conditions similar to those adopted by Mr J. Thistleton Smith. 326 Teinperatare a^'eeiiiKj the Qualifi/ of SUagc 5. Silage made by Capt. Nicoll of Alresford in 1920. Tlie crop was slightly overmature when cut, but not badly laid; during the greater part of the filling the crop was allowed to wilt after cutting and frequently got very wet with rain, but during the last two days of filhng the crop was ensiled directly after cutting. The top part coinciding with the dry period of filhng of the silo produced very good silage upon which the cattle throve, but the bottom coinciding with the wet period of filUng was poor silage and the cattle fell away whilst feeding upon it. In 1921, the silo was filled throughout with freshly cut material and the product was excellent: green in colour with the characteristic "fruity" smell. G. Silage made from immature crops on Mr Alfred Amos' farm at Wye, Kent. Maize was grown for several years in succession from 1899 to 1904 and ensiled in a tower silo, but the variety grown — American Horse Tooth — failed to ripen sufficiently for ideal silage purposes, rarely getting beyond the flowering stage. Under these conditions the silage was in- variably "sour"' with a pungent clinging smell. The cattle ate it, but not greedily and did not thrive greatly upon it. In 1921, a crop of winter oats which had grown very rankly and was hkely to be badly laid before harvest was cut oil between May 8th and 10th when a foot to 15 inches high, and put into a clamp silo after wilting for 24 hours. The crop was of course very immature. The re- sulting silage was of a greenish ohve colour with a most objectionable smell, similar to that described in General Adlercron's silage. The silage was fed to dairy cows in late summer, being scattered on the grass during the severest part of the drought in that year. The cows did not eat it greedily until the smell had partially blown away, but after lying in the sun for an hour the silage was readily eaten. The cows kept in good condition, milked well and no taint was noticeable in the milk. 7. Silage made in a stave silo in Sussex in 1920. The crop consisted of oats and tares in which a large proportion of charlock was growing. This was allowed to become very mature before cutting, so that the charlock had set seeds, which were almost ripe, and produced stems which were hard and woody. The crop after cutting was allowed to wilt 24 to 48 hours and was consequently very dry when ensiled. This fact combined with the woody character of the charlock stems prevented the chaffed crop being adequately packed by trampling, so that much air was included. A. Amos and G. Williams 827 When the silo was opened numerous tiny patches of mould were found throughout the whole depth of the silo and of necessity became mingled \vith the rest when thrown down for feeding. This silage was very dark brown almost black in colour, possessed a strong smell of ammonia and was musty. Cattle, when fed upon it, only ate it under compulsion unless they were able to pick out pieces uncontaminated with mould. SiLARE AT Cambridge, 1917-1921. The silo has been partially or completely filled each of the five years during the period, and in addition a silage stack was made in 1918. Careful records were kejDt of the crop as ensiled and of the silage as taken out. In many cases moisture content has been recorded by means of weighed samples enclosed in wire netting sample-bags. Temperature has been recorded by two methods. In the first a hollow iron gas-pipe was driven into the silage after the silo had been filled. A thermometer was then lowered to different levels in the silage, allowed to remain till it had taken up the temperature of the sur- rounding silage, quickly pulled out and the temperature read off. Readings were made at different depths at daily or longer intervals. The length of the gas pipe was never more than 8 ft. It is obvious, therefore, that the temperatures of the surface 8 ft. only could be ascertained by this plan. The method is open to the further criticism, that the silage is constantly settling; if, therefore, the tube is driven in 8 ft. from the surface one day> and the silage settles, the tip of the tube is no longer 8 ft. from the surface. In the observations recorded the tube was driven in to the full depth after trampling the silage at the time of the first reading and was not driven in further as the silage settled. The temperature readings on subsequent days were therefore taken at depths which corresponded approximately with the same layers of silage as those from which the temperature was taken on the first day. In the second method maximum thermometers were buried in the silo at more or less regular intervals as the filhng of the silo proceeded. These were carefully corked within short lengths of iron gas pipe to prevent breakage, and placed just beneath sample bags put in at the same time. The thermometers were recovered as the silage was used and the maximum temperatures recorded. In 1917 the silo was filled with a crop of oats and tares cut when fairly mature, the oats being well in milk and the tares with full-growTi pods and the seeds beginning to dent the pods. The crop, which was 328 Temperature atfecting the Quality of Silage lodged but not badly laid, was cut on July 16th and 17th in dull weather, a quarter of an inch of rain fell on each of July l~th and 18th and inter- fered with the coinmencement of filling on the hitter day. Filling had to be stopped on -hily 19th, when the silage cutter broke down. The percentage of moisture in the green crop as filled to the silo varied from 70-3 per cent, at the bottom when the crop though wilted contained some added rainwater, to only 64-6 per cent, at the top when the crop was wilted and dry. The following table gives a record of the daily temperature readings on the centigrade scale for ten days after filling, and subsequently at longer intervals of time. Table I. imi cwf. Tcmiieniture Tcinperaturc Temperature at 6 ins. at 2 ft. at 5 ft. Date "C. °C. "C. July 20 26 — — „ 21 49 33 27 22 60 34 29-5 ", 23 OSS 35-5 32 ,. 24 63 40 33 „ 25 6;">-;') 460 34-5 „ 26 62-.'> 47 35 „ 27 (>4r. 49-5 35 „ 29 63-5 48 36-5 „ 30 — 49-5 .36-5 Aug. 2 — 49-5 36-5 ,. 5 — — 37 Oct. 1 31-5 „ 23 — — 27 Nov. 3 — — 26-5 .. 1.5 — — 24-5 Except on July "2()tli. when the reading was made at 9 a.m., the thermometer readings were taken at or near 5 p.m. After July 29th in the case of the (i ins. depth and August .")th in the case of the 2 ft. depth, readings ceased to be taken because it was impossible to ascertain the corresponding depth, and moreover temperature changes were only those due to cooling. When the silo was opened on November r2th and subsequent days it was found that the level at which the (J in. temperatures had been taken consisted of s])oilt mouldy material from which much of the moisture had been driven out by the heat. The range of temperature therefore of 60° to 65° corresponded with moulding of the silage. The silage taken from the 2 ft. depth, where the temperature rose to 49° C, had a uniform dark brown colour with a characteristic ''sweet" pleasant smell similar to that of an overheated hay stack, and was evi- dently comparable to the '"sweet" silage described by Fry in earlier days. A. Amos and G. WilliajMS 329 The silage taken from tlie 5 ft. depth, where the maximum tem- perature did not exceed 37-5° C, was of a much paler brown colour with a strong, somewhat acid, flavour, similar to that described in Mr J. Thistletou Smith's silos. Two feet from the bottom of the silo where doubtless the temperature of fermentation was lower though records were not obtained, the colour was still brown but the smell was much more pungent and very un- pleasant, and similar to that described in General Adlercrons silo. The smell was most tenacious and when handled tainted the hands so that even washing with soap and water failed to remove the unpleasant smell for several hours. This silage in contrast with the previous two types was not relished by stock. In the light of later experience, it seems probable that the chief factor contributing to this condition was the rainfall upon July 17th and 1 8th, causing a certain amount of decomposition of the green crop in the field, and resulting in some rainwater being conveyed to the silo with this part of the crop. In 1918 the silo was filled at the bottom with rye and tares and with oats and tares at the top ; both crops were autumn sown. The rye and tares stood well whereas the oat and tare crop was somewhat but not badly laid. Cutting commenced on July 1st when the rye was rather old, the grain being full-grown but soft and the glumes dry; the tare seeds were denting the pods which were well developed; the oats were forward in milk. The crop was cut 24 to -18 hours in advance of filling. This continued from July "2nd to July 5th, when the silo was full. It was left over the week-end to settle and refilled on July 8th with oats and tares cut the same day. During the whole period of filling the weather was beautifully sunny and no rain fell. The following table gives a record of temperatures taken at or about 5 p.m. each day at first at daily and later at longer intervals. Table II. 1918 crop. Date July Temperature Temperature Temperature at 1 it. at 4 ft. at 8 ft. ;e °C. °C. °C. 9 37 32 45 10 41 34 42-5 11 47 34-5 41 12 47-5 35 40 13 47 35-5 39-75 14 47-5 35-5 39-5 16 49 36-25 38-75 18 46-5 35-75 38 21 45-75 35-5 37-5 330 Temperatnrc afeci'nuf the QuaHtji of Sildf/e The silo was opened on November J 0th wlieii it was found that the silage at 1 ft. deep, the level of the first set of temperature readings, was of dark hrown colour with a "sweet" pleasant smell in every way similar to that immediately below the to]) of the silo in the previous year although this part of the silo was filled with freshly cut crop. The maximum temperature recorded at this depth was 49° C. and the silage contained 72 per cent, of moisture when taken out. It is in fact almost invariably the case when a silo is filled with oats and tares or some similar crop that, after the mouldy surface is removed, a shallow layer of "sweet"' silage is found; this, however, in most cases rapidly gives place to silage of different character. At 4 ft. deep, where the temperature did not exceed 36-25° C, the character of the silage was of pnlnr brown colour and had a pleasant smelling though acid flavour. At 8 ft. deep the temperature records are higher than at 4 ft., and starting at the comparatively high figure of 4.')° C. on July 9th fall con- tinuously to .Tuly 21st. The explniuition of this apparent paradox is that the 8 ft. level di])])ed just below the top layer put into the silo on July Gth. This being easily accessible to air from July .")th to July 8th, during the interval of the filling of the silo, fermented readily and so reached a high temperature before the silo was refilled on the latter day. It is quite probable indeed that 4^i' V. was not the true maximum, for some cooling may have occurred before the thermometer was inserted on July 9th. The silage at this depth was similar to that at the 1 ft. level in that it was "sweet" with a dark brown colour, but the crop having been cut a couple of days before filling the silage was much drier. In this same silo five sample bags were put at regular intervals during filling, and below each bag a maximum thermometer was placed. Table III gives in the first column the number of the bag, in the second the condition of the crop when ensiled, in the third the per- centage of moisture in the green crop, in the fourth the maximum temperature, in the fifth the ])ercentage of Tuoisture in the silage, and in the last column tlu' type of silage produced. Tab le III. 1919 crop. % "f Maximum %o£ lo. of moisture temp. moisture Typo of bag Material green crop "C. silage silage .5 Oats and tares, no wiltin .■!.-) (>9I .. " .. 2 ,. 48 „ (i(>(i ;n . 1 Rye and tares, ,, 24 „ (w + 30 G9-8 »» »» A. Amos and G. Williams 331 Taking bag 5 first, because this was nearest to the top, and conse- quently taken out first, it was found to be situated within I ft. of the surface and so corresponded with the conditions discussed in relation to the silage 1 ft. deep in Table II. The sample contained a fair amount of moisture, nearly 72 per cent., and reached a maximum temperature of 47°-5 C. The silage produced was characteristically "sweet"' with pleasant smell and dark brown colour. The silage in bag 4, which reached a maximum temperature of 37° C, had an acid though pleasant smell. It was situated not far below the level where the break in filling the silo occurred and doubtless for this reason the temperature is above those in samples Nos. 3, 2 and 1. The silage in bags 3, 2 and I was in eacli case produced from a crop which had been consideral)ly wilted in dry weather and contained only about (15 per cent, of moisture when ensiled. The maximum temperatures were respectively 35° C, 31° C. and 30° C, and in each case a yellowish brown silage resulted with an acid, typically silage, smell. This was pleasant and by no means tenacious hke the smell of the silage from the bottom of the silo in the previous year. These samples seem to be typical of silage jiroduced in tower silos from fairly mature crops which are allowed to wilt under dry weather conditions before being ensiled. The silage is not "sweet" in the sense of the earlier writers and neither is it "sour" enough to be un])leasant. It is readily eaten by stock which thrive well upon it. In 1918 a silage stack was also made from a crop of spring-sown oats and tares. This was cut on .luly 20th, allowed to wilt for 24 hours, and built into a circular stack 12 ft. in diameter. The stack heated greatly, the maximum tem])erature in the bottom half of the stack — ascertained by the use of the same hollow gas pipe with thermometer previously described, but thrust horizontally into the stack — proved to be .58° C, whilst that of the top half rose to as much as 75° C. The whole of the stack was composed of sweet silage, for the most part dark brown, but in some jilaces, where the heat was greatest, almost black in colour. This silage was readily eaten by cattle, but the losses in fermenta- tion only, as ascertained from two sample bags, amounted to 19 and 21 jDer cent, of the dry weight respectively. So great a loss indicates that silage made at such high temperatures is uneconomical. In 1919 the silo was filled with a spring-sown oat and tare mixture. Cutting commenced on August 4th and was completed on August 5th. Filling was carried out on August 5th, 6th, 7th and 8th. The crop was IV. Mtl 1) rn>p. % of Maxiinum %of moisture temp. moisture 'J'y]ie of ;reen croji ° C. silago Hiluge 57-3 40 ()0() Sweet dark brown ()7-l .30 711 Acid lifrht „ 05-7 30 7 Mi •» »> •* G9-7 345 73-0 '» »» »» 70-1 31 73-8 »» t» ?• 72-(i 305 7.5-7 Sour dark ., '•V.Vl Teu^xraiinc <(i)(cfiii>j the QnulUij i)j' Sikaje luirly well lU'vuloped, the oats just pas.sing out of the milk stage aud the tares with full-grown pods and half-grown seeds. The crop stood up fairly well, but a slight shower of rain, amounting only to -01 in., fell on the crop on the evening of August 4th after the first days cutting. The rest of tiie filling period was fine though dull. In consequence of the rain on August 4th, the fodder in bags 1 to 5, Table IV, was slightly wetted by rain after cutting, but with the excep- tion of bags 1 and 2, was dry again before ensiling. Tabit No. of bag Material 7 Oats and tares, wilted 3 days f> .. .. .. 2 „ 5 ,, .. ., 3 „ 4 „ ., ., 2 ,. 3 „ M .. 2 „ 2 Oats and tares, wilted 1 day, but wet with rain 1 Oats and tares, not wilted, 79-8 24 83-4 „ „ „ wet with rain Table IV is compiled in exactly the same way as Table III. When the silo was opened it was again found that the topmo.st bag. No. 7, near the surface, contained "sweet" dark brown silage and this was associated with a maximum temperature of fermentation, taken just below the bag, of Hi" C. It is probable that the maximum temix'rature within the bag would have been a few degrees above this point. Bag 6, which had been cut in dry weather and allowed to wilt two days before ensiling reached a maximum temperature of 30° ('. and contained typical liglit brown acid silage with a pleasant smell. Bags 5, 4 aud 3 whicli had been slightly wetted with rain after cutting and then left long enough to dry oflt the rainwater before ensiling, produced silage very similar to that in bag (i. Bag 2, whicli contained a small amount of rainwater when ensiled, contained a slightly unpleasantly sour silage and bag I . wliich was next to the floor of the silo and was very wet when ensiled, contained the characteristic unpleasantly sour pungent brown silage. The thermo- meter indicated that the maximum temperature in this case was only 24° C, due partly to proximity to the floor of the silo and partly to the exclusion of air from the wet material containing as it did 80 per cent, of moisture. It must also be recorded that in the season 1919 the drain in the floor of the silo was intentionally closed so that no drainage was possible. This may have contributed to the souring of the bottom silage. A. Amos and G. Williams Table V. 1920 crop. "„ of .Maximum ''' (if No. of moisture temp. moisture Tyjje of bag Material green crop "V. silage silage 7 Oats and tares, no wilting (J9G 35-5 7!-.-) Acid light brown li t)ats, tares and beans, wilted (iS-1 31-5 70 Acid yellow- () hours lirown 5 Oats, tares and peas, wilted 0 hours (it) 32-5 70-4 Acid yellow- brown 4 Oats and tares, wilted 24 hrs. 70 33 69-7 Acid light brown 3 „ 48 ,. I14S 3()-2 02-8 ,, ,. o Oats and tares, cut 0 days before filUnt;. Soaked with rain 72-5 31-3 72-9 "Sour" dark brown 1 ., ., „ figi 305 71 « In 1920 the silo was filled with a cro]) of autumn sown oats and tares, except for two small quantities of material which contained in addition peas and beans respectively. The crop was well advanced when cut, the oats being just past the milk stage, the tare pods full grown in length, with seeds denting the pods. The pea seeds were full-grown in size though still soft, and the bean seeds were not quite full-grown. Cutting com- menced on July 3rd, but much rain fell during the next five days, so that filling was impossible till July 9th. The following Table VI gives the rainfall for the ])eriod. Table VI. Raiiifnll 1920. July 4 014 in. July S 0-44 in. July 12 nil 5 0-.'")(i ., 9 nil ,. 13 nil () OM „ „ 10 0(14 „ „ 14 005 in, 7 010 ,. ., 11 0-((5 ,, „ 15 nil Bag No. 7, which was situated 2 ft. 6 in. below the surface of the silage, was made from the oat and tare crop; this had been ensiled shortly after cutting. The temperature of fermentation was considerable, 35°-5 C, owing to its proximity to the surface and consequent access of air. This silage had a pleasant acid smell and was much relished by the stock. Bags Nos. C) and 5, containing beans and peas respectively, mixed with the oats and tares were in each case wilted only six hours, but the crops were dry and fairly mature before cutting, so that their moisture content as filled was low, 68-1 per cent, and 66 per cent, respectively. In each case good silage resulted characterised by a pale yellow-brown colour with a pleasant acid smell : this was greatly relished by the stock, the pea, oat and tare silage being particularly good. The silage in bag No. 4 was cut on July 12th and allowed to wilt 24 hours; that in bag No. 3 was cut on July 10th, wetted with slight 334 Trnipenitwe affect Iikj the Qiudiitj of Sihuje showers on tliat day and the morning of July 11th, but the afternoon was dry and sunny so that the crop when ensiled on July 12th was free from rainwater. Each of these produced light brown silage with an acid but not unpleasant smell. None the less it was not so good as that in Nos. 5 and (!. Bags Nos. 2 and 1 contained material which had been cuf on July .'kd, but owing to very wet weather remained in the field till July 9th before ensiling. A certain amount of decom])osition occurred in the field, and the crop was ensiled whilst it still contained some rainwater, the total moisture contents being 72-5 per cent, and f)9-l per cent, respectively. The silage from these bags was of the characteristically unpleasant "sour" variety, previously described; it possessed a dark brown colour and a pungent smell which clung to hands or clothes brought into contact with it; the maximum temperatures recorded with these samples amounted only to 30° C. to 31" C. Experimental cattle, when fed upon it, failed to thrive well. Table VTl. 1921 crof. "„ i)f Miixiiuiiin "„ nf Nci. of moisture teiii[) inoistiirc f.'^'Pf "f l>au Material green ero|) ■" ('. silage silage "-' Oats and tares, wilted 0 lirs. (i7-4 34-5 70-6 Green and "fruity" I ., , not wilted 7()-.") 24-.5 69-7 In 1921 the crop was ensiled in perfect weather, warm and dry, except that on the night previous to the first day's cutting -OC) in. of rain fell, so that the material in bag 1 contained a small amount of rain- water. The crop was not very mature, the tares being in full flower and the oats ju.st in milk. Wilting throughout the filling was reduced to a minimum, the croj) being cut only a short time before ensiling. Both bags, and in fact the whole of the silage, was of excellent quality. The silage had an olive colour with a tint of green throughout, though in some parts the green was more pronounced than in others. The smell was entirely dilTerent from that previously obtained; it had no suggestion of sourness, although it was made from sappy and rather immature material, nor had it any pungent odour, but it possessed a kind of "fruity" smell suggestive of pear drops and combined with this the smell of freshly cut lawn grass. When fed to stock it was ravenously eaten and under experimental conditions has given excellent feeding results^. It is to be remarked when such freshly cut material containing large quantities of moisture is ensiled that much juice may be expressed and ' The results of this feeding experiment will be puliliahed shortly. A. Amos and G. Williams '^V>l^ lost from vvoodeu silos. The juice may ruu to waste and after putrefaction produce unpleasant smells in the yard or it may be given to cattle or pigs as a liquid. Conclusions. 1. Many different types of silage may be produced from the same crop according to the conditions of ensiling; the characteristics of the resulting silage varying not only in physical and chemical but also in feeding properties. 2. The following types of silage have been differentiated and some of the conditions of their production ascertained. ((/) "Sweet" dark brown silage. This type of silage is produced when the temperature of fermentation rises above 45°-50° C. ; it has not been produced below 45° C. It is frequently produced in stack silage to which air has ready access, but not generally in tower silos except in a shallow layer (i in. to '2 ft. thick just below the mouldy surface to which air has ready access. This silage has a dark brown colour varying in intensity according to the temperature of fermentation and a sweet pleasant smell resembhng that of heated hay. It is readily eaten by stock, but has generally lost a considerable proportion of its food value through exces- sive heating. (b) Acid light-brown or yellow-brown silage is i>roduced in tower silos from crops which are moderately mature when cut and allowed to wilt for varying periods according to the initial dryness until the moisture content of the crop approximates 70 percent. The maximum tempera- ture of fermentation is generally between 30° and 37° C. This silage has a yellowish brown to brown colour with an acid though pleasant smell probably largely due to acetic acid, the yellowish types being generally the more pleasant. This type of silage is eaten greedily by stock, which thrive upon it and is to be commended. {(■) Green "fruity" silage is produced in tower silos from crops which are cut in the earlier stages of maturity, from the time of full flower till the seeds are half formed. The crop must also be ensiled soon after cutting. The temperature of fermentation is low and may vary from 22° C. to 34° C. This type has a green to olive green colour with a smell that is neither "sweet" nor "sour," but can best be described as "fresh" and "fruity." It is greedily eaten by stock, which thrive greatly upon it, and Woodman^ has recently shown that its digestible properties are very high. Green silage suffers from one practical disadvantage; large quantities 1 Woodman. This Journal, 12, Part II, April 1922. 336 TeiiijK ridnrc tijf'crfing the Quality of Sihuje of juice are liable to drain away from the silage and carry with them soluble food material, which if allowed to accumulate in the yard under- goes fermentation and ])r()duces a smelly mass of putrid material, if, however, the juice is collected cattle and pigs readily drink it. This problem is receiving further study. {d) Sour silage is j)roduced under at least two different sets of con- ditions. It may be ])roduced from an immature and succulent crop, or it may be produced from a crop which has been cut and then saturated by rain before ensiUng, especially when the crop has become laid and partially rotten at the base before cutting. Sour silage has a dark brown or olive brown colour witli a pungent and most un])l('asaiit smell ])ossibly due to butyric acid. Cattle will eat this, but not readily and do not generally thrive upon it. (e) Musty silage. One case only of tliis type had been recorded, in which an over-ripe crop containing much charlock was allowed to wilt and become over-dry before ensiling. A number of tiny mouldy centres were produced in which ammonia was generated. This silage has a dark bro\\ii almost black colour and has a musty ammoniacal smell. Stock refused to eat it or ate it only under com- 2)ulsiou. 3. The classiiication of silages given ai)ove is admittedly superficial, but it may serve to pave the way for the chemist, the plant phj^siologist or perhaps the bacteriologist to produce similar types of silage under well defined conditions and so make the distinctions more absolute. For the successful practice of ensilage in this country it is fundamental that the conditions by which each type of silage may be produced should be accurately known. 4. In the past persons conducting feeding experiments in this country have rarely attempted to define the type of silage used, and conflicting results have been obtained. When it is further noted that the quality of silage at different parts of a silo may be, and frequently is, fundamentally different, it is of the utmost importance that these facts should be recorded in future feeding experiments. Our acknowledgements are due to the Ministry of Agriculture who have financed this study, to Messrs English Bros, of Wisbech who generously i)resented the experimental silo and to Prof. Biffen and his farm managers Messrs TI. Y. Sliei-ingham and N. Langridge who have facilitated the ex])eriments u|ion tin- Plant Breeding Farm at Cambridge. (Received August 7th, I92"2.) AN INVESTIGATION INTO THE CHANGES WHICH OCCUR DURING THE ENSILAGE OF OATS AND TARES. By ARTHUR AMOS, M.A. AND HERBERT ERNEST WOODMAN, Ph.D., D.Sc. {School of AgricitUnrc, Vanibridijc U nircrsilij.) Introduction. The practice of ensilage has been rapidly gaining adherents in the British Isles since Mr Cxeorge Jaques began to advocate the oat and tare crop for this purpose in 1913. In view of this fact, it becomes increasingly important for the scientist to be in a position to supply the ])ractitioner with facts from which he may be able to calculate the probable economy of the system before investing the nece.ssary capital in a silo. One of us^ has already made such a tentative calculation based upon a some- what limited knowledge of those facts. It is the purpose of these experi- ments to amplify knowledge in one direction, namely, in that of the chemical changes which occur during the ensilage of oats and tares, so that one may know what are the losses and what the gains, if any, in chemical constituents during the process. One of us'^ has recently shown that it is possible to produce at least five distinct types of silage from the same crop. It is, therefore, equally important to be able to state definitely which of these occasion the least net loss during ensilage, and to endeavour by controlling the factors concerned to reduce to a minimum the losses in the making of those types of silage which appear most economical. The experiences gained with maize silage in America and elsewhere, useful though they be, do not bear directly upon the present problem of oat and tare silage, because the chemical composition of the two crops is very different. Prehminary work was begun in 19f8 when the late Mr Gwilym Williams assisted on the chemical side of the work, but in the earlier ' Ensilage by Aitliiir Amos, .louriial of Funiurs Clnb, March l'J20. - Amos and Williams, This Volume, p. 323. 3r{8 (}(il 1111(1 Tare SUdijc years progress was limited by the failure of some of the methods of investigatiou. In the first season high grade linen bags, such as are used by seedsmen in distributing seeds, were used as containers, but proved defective because the fabric perished as a result of the fer- mentation in the silo. Next fine-meshed galvanised wire netting was tried, but though in the first year the juices of the silage produced no effect upon this nuiterial, in subsequent years this was not the case and consequently accurate weighing of the contained sample became impossible. In the experiments now to be described, loosely woven jute bags, made of the same material as commonly employed for chaff bags, were used, and this has proved perfectly suitable. During this pre- liminary period certain progress was made in chemical methods which have been of use in the present investigation. It has, however, been thought undesirable to publish the earlier results in detail because of the uncertainties above mentioned. Saiii|)les of silage described in this ])a])er were made in two cases from the commercial silo used on tiie Plant Breeding Farm at Cambridge; in the remainder from three small wooden cylindrical silos, measuring 6 ft. high and 4 ft. in diameter, which have given highly satisfactory experimental results. In some measure experimental silos of this size constitute a new departure; they are free from the objection commonly used again.st bottles of silage, that the bulk- under ex])eriment is so small that any rise in temperature is immediately dissipated, and yet they enable the experimenter to vary at will one factor only, a condition which is difficult indeed to accomphsh by varying tiie jHisition of the sample bags in a large commercial silo. These small silos were constructed of I in. boards, tongued and grooved and kept tightly pressed together by three circular iron hoops which were capable of being easily tightened by suitable screws. The wood was of plain deal, tarred externally for purposes of preservation. The silos were placed in a shady .situation and the floor of each con- sisted of j)uddied gault clay. At the time of filling metal extensions i ft. high made of iron sheeting were fitted to the top of each silo. The sample bags, as previously stated, consisted of loosely woven jute, measuring 3 ft. (J in. long by '1 ft. in breadth; these contained about 2 cwt. of green fodder under average conditions of wetness. When filled with fodder these dimensions were considerably contracted, so that in no case were they within 6 ins. of the walls of the silo when placed in position. Immediately beneath each sam})le bag was placed a maximum A. A.MOS AXI) H. E. WooD.MAN oo5> thermometer encased within a piece of iron gas piping for protection. ]3abcock and Russell ^ have shown that such fermentations as produce increase of temperature in the silo take place very rapidly and that the maximum temperature is generally reached within four or five days. The use of a maximum thermometer therefore gives a useful indication of the vigour of the fermentation, jjrovided the heat is not rapidly lost by conduction. The results upon opening proved that good silage was capable of being made in these small silos, and in two out of the three cases the silage differed very slightly, if at all, from that produced commercially; more- over, the wa.stage due to spoilt material at the top compared favourably with that of commercial practice, and in the case of silage made from freshly-cut, succulent crop, the amount of spoilt nuiterial on top measured only i in. in thickness. Object of the Experime^t.s. The purpose of the present investigation was two-fold; firstly to determine the effect of varying the moisture content, both by the wilting of the crop after cutting and by allowing rain to fall upon the crop in the field, upon the quality of the silage produced, and so to ascertain how far this factor influences the producing of different types of silage; and secondly to ascertain the nature and magnitude of the chemical changes which occur when these different types of silage are made and to obtain a comparison of the fermentation losses involved in each case, as well as any benefits which may accrue from such fermentation. Further details of the chemical objects of the experiment are given later. Description of Crop. The crop utilised for the experiment was autumn-sown and was seeded at the rate of If bushels of tares and I5 bushels of grey winter oats per acre. In the dry summer of 1921 there was approximately equal growth of oats and tares at time of cutting, but no separation of the crop was made. The crop used for the samples was cut at recorded times between June 21st and 23rd, at which time the maturity of the crop was perhaps ideal for making hay, but would generally be considered slightly immature for silage. The oats were then well past flower, some being in milk, others having not quite reached this stage; the tares ^ Babcock and Russell, llth anrt lS//( Annual Report, Wisconsin AgriciiUnral Experi- ment illation. Joum. of Agrio. Soi. xii 23 340 Out and Tare Silcuje were ill full flower witli a few pods half an inch in length. The whole crop stood perfectly and was of excellent quality. The weather at the time of cutting formed part of the long drought in 1921, so that generally soil and crop were very dry, but -06 in. rain fell in the night previous to the commencement of operations on June 21st. June 21st was flull and tine, but June 22nd to June 26th were scorching days. During the night of June 2()th a shar]) thunder- storm fell, measuring -22 in. of rain, and about half this quantity next day. Samples Nos. 1, 2 and 3 were cut in the early morning of June 21st and were chafl'ed and ensiled without delay. Thus they contained some added rainwater, as Table 1 shows, and contained only 23-5 per cent, dry matter when ensiled. Samples Nos. 4 and 5 were cut at the same time, but allowed to wilt for 24 hours before ensihug, and lost moisture so that they contained 35-7 per cent, of dry matter. Sample No. G was cut at 8 a.m. on .June 23rd — a scorching day — and allowed to wilt six hours before ensiling, whilst Samples Nos. 7 and 8 were cut on June 23rd, dried by wilting in hot sun till June 2(5th, then wetted by the thunder- storm and linalh' ensiled on .June 27th, when they then contained 37-8 per cent, of dry matter, indicating that they must have been very dry before the rainfall. It is important to note that the crop used in Sami)les Nos. 7 and 8 showed no sign of rotting or putrefaction as a result of the rain, a condition that not infrequently occurs when a laid croj) of oats and tares becomes wet with rain after cutting. FlLLIN(i OF Sll.OS AND S.\iMPl,E BaG8. All the fodder was ehafi'ed with a I'apec silage cutter before ensiling. Before filling the sample bags and taking a sample of the original fodder, a suitable quantity of chaffed crop was shot down upon a smooth con- creted floor and well mixed by turning. From this well-mixed heap a 7 lb. biscuit tin of fodder was filled by taking small handfuls from different parts of the heap. The lid was closed at once and the sample sent down to the laboratory for analysis. The sample bags were im- mediately filled by taking portions at random from the same heap, tied U]), numbered and weighed. Sometimes more than one sample bag was filled from the same chaffed sample when rei|uired for immediate use in different silos or different parts of the same silo: in this way some economy was made in the number of control samples to be analysed and the resulting silage samples were more comparable. In filling the small silos, chaffed oats and tares were first put in until, when well trodden, the height of fodder in the silo was about 2 ft. ; A. Amos and H. E. Wuodjiax 341 then a nest was hollowed out into which was first put a short length of iron gas tubing containing a maximum thermometer. Immediately on top of the thermometer was placed a sample bag. Then more chaf1:'ed fodder was trodden in until the height of the fodder reached 4 ft., at which level a second sample bag was placed with maximum thermometer as before. More chaft'ed fodder was added and trampled until with the aid of the extensions previously described the toj) of the silage reached (j ft. 6 in. or 7 ft. in height. The silo was then left '24 hours to settle, after which about 6 in. of soil was thrown on top for the purpose of excluding air. Eventually the silage settled until at the time of opening the level of the tojj of the silage was about 5 ft. Two sample bags put into the big silo were provided with similar maximum thermometers, and in addition a 4 ft. length of stout string was tied at one end to the bag and at the other to a short thatching peg. The men filling the silo were instructed to keep pulling up the peg as long as possible, so that when the silage was being used the peg might come into view before the sample bag and so give warning of the proximity to the sample. Opening of the Silos. The small silos were opened at convenient times during November, by which time all fermentation had long since ceased and the tem- perature had for the most part cooled down nearly to that of the pre- vailing air temperature. Notes were made of the condition of successive layers in the silos as these were emptied, and so soon as a bag was exposed it was removed, cleaned and weighed. The bag was then quickly emptied, cleaned and weighed again; the difference between these two weights giving the net weight of contained silage. The silage, emptied from the bag, was then quickly mixed together, a sample taken from this, placed in a tin, and sent down to the laboratory for analysis. The two sample bags in the commercial silo were retrieved soon after the markers came into view and whilst the general level in the silo was still one to two feet above the bags. When silo 1 was opened on November 10th, it was found that there was no more than 4 in. of spoilt material at the top of the silo; then 14 in. of good silage was removed before the first bag (bag No. 2) was reached. This silage possessed a distinctly green colour and a pleasant odour with no suggestion of sourness. The aroma can best be described as "fruity."' Bag 1 was taken out on November 15th. It was situated 23—2 Kinall silu 1 C'OIIUH. Small silo 2 C'omiii. Small silo 3 silo A silo A \~ 2 3 ( 4 5 6 7 8 0 0 0 24 24 6 96 96 13 i;! 13 16-5 16-5 16-5 14-5 14-5 24-.-. 22r> 24-5 23-5 24-5 34-5 32-5 39-0 715-5:! 7(io3 76-53 64-28 64-28 67-40 62-24 62-24 7()l)5 75-56 69-31 06-17 66-25 70-07 63-46 {a) 02-06* {!)) 60-50 ■u-\ 262-4 262-9 334-0 323-6 287-8 293-8 294-0 2I4(i 227-3 232-9 306-5 294-3 267-0 264-2 2(5-9 842 (fdl 1111(1 Tare Siluijc about 2 ft. l)elo\v bag '1 aud possessed the .same pleasant ■'fruity' suiell but the colour, though still green, was not quite so good. Bag 3, which was taken from the same sample of crop as bags 1 and '1, was retrieved from the big silo on March 28th, and when opened was found to contain silage very similar to that taken from silo No. 1. It possessed an olive green colour and a^ pleasant, slightly "fruity"' smell. Table 1. No. of bai; Iiitorvrtl li<-t\vei!ii cutting aiul ensiling (hours) Initial t(MUiH'r-ature (° (.'.) .Maximum tompci-atui-e (" ('.) % moisture a.s ensiled % moisture after ensiling Dry matter as ensiled (ozs.) Dry matter after ensiling (ozs. ) % loss of ilry matter 13-2 13-4 11-4, 8-2 91 7-2 lO-o l(i-5 * (n) Unsjjoilt sample, (b) Spoilt sample. An examination of Table I shows that the conditions of silage in samples 1, 2 and -'5 are very similar, except in one respect. They were ensiled at the same initial and maximum temperatures, which latter was very low, the quality of silage was very similar, and the jiercentage loss of dry matter was rather high in all cases. The one condition which varied greatly. was the percentage of moisture in the final product; this in the two samj)les in the small silo remained practically the same as in the original crop, whereas in the big silo the percentage of moisture fell from TO-u per cent, at the beginning to 09 per cent, at the end, indicating that the superimposed weight of silage in the tall silo had caused some of the moisture to be expressed and this drained away. The drainage licjuid was not analysed, but similar drainage in other years has contained varying quantities, between 4 and 10 per cent., of soluble material and is therefore an im])ortant source of loss during ensilage whenever it occurs. As before mentioned, the percentage loss of dry matter in each of these three sam])les was high, namely, 13-2 ])er cent, and l."5-4 per cent, in the small silo, aud 11-4 per cent, in the large silo. There seems no very obvious reason to account for the greater loss of dry material from the small silo than from the similar sample in the large silo, especially in view of the fact that httle, if any, water drained A. Amos and H. E. Woodman 34:3 from the former whilst drainage must have been considerable from the latter. Silo No. 2 was opened on November 17th; it had been filled from a croij which had been cut 24 hours before ensiling, and was consequently much drier when ensiled, containing only (i4-3 per cent, of moisture. There were 8 or 9 in. of spoilt mouldy material on top, doubtless due to the drier crop being more difficult to press down by the soil covering for the exclusion of air. The silage in bags Nos. 4 and 5 contained in this silo was practically identical. It was olive brown in colour with no trace of the green colour found in bags 1, 2 and 3, and had no trace of the "fruity" smell associated with these. The smell, on the other hand, though pleasant, was distinctly acidic and was characteristic of much commercial silage made from similar crops under similar con- ditions. The rise in temperature as recorded by the maximum thermo- meters was again low, rising only from 16°-r) C. to 23°-5 C. and 24°-5 C. in each case. This is lower than the tem])erature commonly recorded with similar silage, for which the maximum temperature is normally about 30 " C. In both cases the moisture percentage increased during ensilage, a result occasioned by the disapjaearance of some of the dry matter during the process, and not by addition of water, for the silos were kept covered to protect from rainwater. Finally the loss of dry matter during ensilage in this case was con- siderably less than in the first three samples, a fact wliich suggests that drying the crop for 24 hours before ensiling is desirable. It should, however, be remembered that when this is done two other sources of loss in the field may be occasioned, namely, that due to respiration and to the breaking oft of the leaves as the crop dries. Bag () contained a sample which was allowed to dry for six hours only on a scorching hot day after cutting. It contained only 67-4 per cent, of moisture when ensiled, but this increased to 70 per cent, in the silo, partly by loss of dry matter during fermentation, and possibly also owing to absorption of moisture from other layers in the silo which were moister. In this case the maximum temperature rose to 34°-5 C, and yet the loss of dry matter was only 7-2. The silage produced imder these conditions had a pale brown colour with just a suggestion of green about it, and possessed a smell which was faintly fruity with little or no acidic smell. It was by no means typical silage and was characteristic only of a very small layer in the silo. Possibly this represented a transition stage between the green "fruity" silage below and the brown acidic silage about this layer in the silo. :>44 0(i( (iikI Tare Si/(t(/c Small silo No. '.i had been tilled from a part of the crop which was cut on June 23rd and left in the sun to dry during June 23rd, 24th and 2yth, by which time it was three parts made into hay. However, during the night of June 25th and 26th, a thunder.storm occurred, which pre- cipitated -22 in. of rain and wetted the croj) considerably. June 2Gth was dull and about •]() in. of rain fell. On June 27th a part of the crop was carted in early morning and ensiled in one of the small silos. This crop, in spite of the rain on the previous day, contained only l)2-2 ])er cent, of moisture when ensiled. The silo was opened on November 29th. when it was found that a considerable (l('])th, about 18 in., of spoilt mouldy material had to be removed from the top before the good silage was reached: even at this dejith the silage iie.xt tlie walls was mouldy and occasional spots of mould were found throughout the whole mass. This mouldiness was due in part probably to infection in the field as a consequence of the rain, and in part also to the fact that owing to the dryness of the material it could not be packed so tightly in the silo. When bag No. 8, situated in the top half of the silo, was removed it was found that the ma.ximum thermometer immediately below it had recorded a temperature of 39° C. J'robably the temperature in the bag itself would have been a few degrees higher. The contents of this bag were partly good and partly mouldy; approximately two-thirds were good. This had a yellow-brown colour and an aroma very similar to that of heated hay. It was in fact "sweet" silage. The loss of drv matter in fermentation was high, 16-5 per cent. This was due in part to the excessive heat of fermentation and in part to the acticm of the moulds present. Bag No. 7 at the bottom part of this silo contained a few small spots of nu)uld. It had reached a maximum temperature of 32^-.") C. It had a yellow-brown colour with a very slight indistinctive smell, with no suggestion of acid to the nose. It appeared different from any type of silage commonly met in practice. The loss of dry matter during fermen- tation amounted only to 10 per cent. P'rom experience gained from other sources it appears that the most desirable type of silage is the green "fruity" silage produced in bags 1 and 2 in the small silo, and bag 3 at the bottom of the large silo. This type of silage seems to be produced when a fresh green crop in an early Btage of maturity (soon after flowering) is ensiled, with little or no previous wilting, at a low temperature, at or about 2.5° C. The results of these experiments suggest that with this type a considerable loss of A. Amos and PI. E. Woodman 845 dry matter may result from the process, especially if much drainage of sap occurs. Bag No. 6 at the top of the large silo produced a silage bordering upon this type, and was ensiled under somewhat similar con- ditions as to freshness — it was cut only six hours before ensihng — but the maximum temperature recorded in this bag probably accounts for the alteration in type. The acid brown type of silage produced in bags 4 and 5 in the second small silo is characteristic of much commercially made silage, and is also a valuable type of silage. This is made generally from a moderately dry crop, dried either by wilting or by allowing the crop to matiire and so become drier before cutting. The analytical figures show a remarkably low loss of dry matter in these two samples, only 8-2 per cent, and 9-1 per cent. The small loss is certainly a point in favour of this type of silage. The silage in small silo No. 3 is not typical of commercial silage and therefore no important conclusions can be drawn except perhaps that rain washed fodder as well for its tendency to become mouldy, as for the washing of food material from it by rain in the field, cannot be expected to jiroduce first class silage. Methods of Analysis. The analysis of the green crop and silage samples was carried out in the following manner. The percentage of dry matter was determined by drying down representative samples of 200 grms. to constant weight in the steam oven. The estimation was carried out in duphcate and the dry residues were finely ground in a mill, allowed to air-dry for several days and then submitted to complete analysis. This involved deter- minations of moisture, crude, true and pepsin-HCl soluble protein, "amides," ether-extract, crude fibre, ash and nitrogen-free extractives. The results in every case were calculated to dry matter. When calcu- lating the amount of ether extract in the silage samples, it was necessary to make a correction for the volatile organic acids which are lost during the process of drying down. The analysis of the samples was left in the hands of Mr F. J. Aylett, to whom the writers would like to express their thanks for his careful work in this connection. In order to gain a further insight into the nature of the changes which occur during the ensilage of green forage, analyses were also carried out on aqueous extracts of the green crop and silage samples. Representative 200 grms. samples of the material were weighed out into .'MG Oat and Tare Silage wide-necked bottles of about 1200 c.c. capacity. To the material was then added 500 c.c. of distilled water (previously boiled and cooled) and the bottle was sto])|)ered with a rubber bung and vigorou.sly shaken in a shaking machine for four hours. The resulting dark brown coloured extract was filtered first through musHn. the residual material being well squeezed out. and then through a filter paper. 150 c.c. of the aqueous extract was then made up to 500 c.c. with alcohol; this caused the separation of a small amount of precipitate, which settled readily. The clear liquid, which was slightly yellow in colour, was then filtered oil' through a dry filter paper and submitted to analysis by means of the Foreman alcohol titration method. The amounts given in the following description of the method refer to the analy.sis of the silage extracts; the determinations on the green crop extracts were carried out on larger volumes, since the ingredients to be estimated were only present in relatively small amounts in such extracts. 10 c.c. of the alcoholic li(|uid were diluted with .50 c.c. alcohol and the solution was titrated with .V/10 NaOH in the presence of phenol- phthalein. When the neutral jjoint was reached, 12 c.c. of a ne\itral aqueous formalin solution were added, whereon the pink colour was dis- charged and the addition of one or two drops of iV/10 NaOH was necessary to restore neutrality. The total titre of ^V/IO NaOH gave a measure of the total acid grouj)s, free and combined, in the extract. It is im])ortant to remember that amino acids and amides of the asparagine type are also included in this measurement. The small amount of alkali re<]uisite to restore neutrality after the addition of formalin has been shown by Foreman to be an indication of the amount of dibasic amino acids and proline present in the solution. The titration method of Foreman also enabled the amounts of amino acids and volatile bases in the extracts to be determined. To 50 c.c. of the alcoholic liquid in a 500 c.c. distillation flask was added an amount of iV/10 NaOH sufficient to produce neutrality. The contents of the flask were then submitted to steam distillation, a vigorous current of COj-free steam being passed through for about five minutes. The bases with the alcohol were caught in 10 c.c. N jlO HCl which had ])reviously been pipetted into the receiver. The excess of acid in the contents of the receiver was determined by titration with iV/10 NaOH in the presence of alizarin and the amount of acid equivalent to the volatile bases estimated by dift'erence. The liquid remaining in the flask, which is alkaline after the distillation owing to the hydrolysis of the salts of A. Amos and H. E. Woodman :>47 amino acids, was cooled, diluted with distilled water and titrated with iV/10 HCl to phenolphthalein, the amount of free alkaH in the flask being equivalent to the amino acids in the original 50 c.c. of alcoholic liquid'. In order to obtain the amount of volatile organic acids in the alcoholic liquid. .10 c.c. were pipetted into a 500 c.c. distillation flask and an amount of N/iO RSO^ shghtly in excess of the equivalent to the volatile bases was added. The contents of the flask were then submitted to steam distillation, in every case 500 c.c. of distillate being collected. The acidity of the di.stillate was determined by titrating with N/U) NaOH to phenolphthalein. The results thus obtained, in conjunction with the moisture content of the material, enabled the amounts of the different ingredients in the original material to be calculated. The results were expressed in the following manner: e.cV Total aciflic groups, free and comiiinecl A Amino acifLs and aniide.s of asparagine type B Total organic acitis of lactie and ac-etie type .\-B Organic acids volatile in steam ( ' Non-volatile organic acids A -B-(.' Volatile bases I) Several advantages attach to the use of the above method of dealing with silage extracts which are not jjossessed by the titration methods previously in use. ( 1 ) The titration of total acidity is carried out in a liquid which is almost water clear, the light yellowish brown colour of the alcoholic liquid practically disappearing on further dilution with alcohol. The end point in the titration is much more satisfactory than that obtained when direct titration of diluted aqueous extracts is resorted to. (2) On treating the aqueous extract with alcohol as described above, proteins, albuminoses, etc. are precipitated. This treatment therefore removes certain classes of substances which might interfere with the determination of the simpler constituents. {?)) The determinations of the amino acids and volatile bases are carried out in one operation. A knowledge of the relative amounts of these con.stituents is of value in studying the degree of putrefaction which may have occurred in samples of spoilt silage^. (-1) Alcoholic silage extracts jtrepared in the manner described can be kept over long ])eriods without undergoing change. ' For a full account of the technique and significance of the Foreman method, the original publication should be consulted (Foreman. Biorh. J . 14, 4.51. 1920). - Foreman and Graham-Smith, J. Hygiene, IS, 109, 1917. ]4S Oat ami Tart Silage Table II. Showing clianges in contenl of dry matter, volatile and non- volatile organic acidx, amino acids and volatile baxcs undergone by the out and tare crop in the I'arious .silos per 1000 grnis. dry oats and tares. Silo I. Green oats and tares Bag 1 silage Bag 2 silage c.c. -V c.c. .V c.c. .V Volatile organic acids 9-0 299-0 194-9 Non-volatile organic acids .337-0 923-2 986-6 Amino acids 75-0 472-5 433 1 Volatile bases 17-0 116-4 108-3 Dry matter 1000 gms. 868-5 gms. 866-2 gms. Silo II. (ircen oats and tares c.c. .V Bag 4 silage c.c. .V Bap 5 silage e.e. .V Volatile organic acids Nonvolatile organic acids Amino acids Volatile bases 7-0 309-0 .-)5-() 19-0 268-9 601-1 444-2 89-0 284-7 534-8 493-9 90-0 ])rv maUer 1000 gms. 917- 009-5 gms. Sil( Volatile organic acids Non-volatile organic acids /\mino acids Volatile bases Dry matter Volatile organic acids Non-volatile organic acids Amino acids \'olatile bases Dry matter llrfoii rtiita Bag 7 silage c.c. X Bag 8 silage and tares Unspoilt c.c. .V S|X)ilt c.c. .V Total c.c. A' ti-0 243-0 43-0 230 1000 gins. 121-4 503-6 288-7 84-5 899-3 gms. 113-1 9-2 2()8- 1 69-8 174-1 69-8 63-5 .56-4 516-5gms. 318-5gm.s. 122-3 337-9 243-9 119-9 835-0 gm; Big silo. Green oats and tares Oat and tare silage Bags c.c. .V Bag~6 c.c. .V Bags c.c. X Bagfi c.c. A' 9-0 6-0 337-0 247-0 75-0 37-0 17-0 14-0 1000 gms. 1000 gms. 215-4 900-2 365-2 106-3 886-0 gms. 304-2 761-5 437-8 92-8 927-5 cms. A. Amos and H. E. Woodman 34:9 Table III. Condituenis of silage extract expressed as percentages of moisture-free silage. of bag 1 i; 3 4 5 6 7 8 0- /o o o. o/ o/ /o 0' /o () Unspoilt Spoilt t > o Volatile organic aeiils* 2-07 i-3r> 1-40 l-7(i 1-88 1-97 0-81 1-31 ()I7 Non-volatili- org. acidst (io3 7'22 6-11 311 2-51 5-17 2-85 2-48 — Amino acidsj 4-76 4-38 3(;i 4-24 4-75 413 2-81 2-95 1-92 Volatile bases]; 117 109 lUo 0 85 0-87 (••88 0-82 1-08 l-.j;") * r'ak'ulatecl as acetic acid. f Calculated as lactic acid (see table given later under "Changes suffered by Ether Extract"). .■f Calculated as crude protein. Comments on Tables II and III. The significance of the above data will be dealt with fully in the general discussion of the results. At this point, it is only desired to call attention to the following: 1. In all the bags containing normal silage, the non- volatile acids preponderated largely in amount over the volatile organic acids. '2. Where moulding of the silage occurred (as in bag 8), the volatile and non-volatile organic acids were not simply neutralised by the basic decomposition products of the jDrotein constituent, but were actually destroyed. This is in agreement with the observations of Dox and Neidig^ and the behaviour is further exemplified by the results of an analysis carried out on a samjde of mouldy silage from the top of silo II: CO. N Volatile organic acids — Non-volatile organic acids 2ri-9 Amino acids 1(1-7 Volatile bases UiO In this case, all the volatile organic acids and most of the non-volatile had disappeared as a result of mould activity. The meaning of the high ratio of volatile bases to amino acids will be discussed later. 3. The results for bag 3 silage are, as anticipated, quite normal, although this bag was allowed to remain in the big silo some mouths after the other bags had been removed. ' Dox and Ncidig, iJcNcri/r/i BiiUeliii. No. 10, Iowa Experiment Station. 350 Out aiul Tare Sihuje Table IV. Amounts of constituents of green oats and tares and oat and tare silage contained in bag I and bag '2 (sil-o I). Analysis of samples (i-alciilatod tu dry matter). Green oats Bag I Bag 2 and tares 0 silage O' silage o fVuile jirotein 121.-. 0 12-27 . o 12-01 Ether extract* :!r.!i 3-12 3-21 N-free e.xtractives oOGT 47-92 47-<>2 Crude fibre 25- 10 27-75 27-87 Ash 8-49 8-94 8-09 True protein 9-20 .1-03 5-7G '"Amides" 2-9.1 7-24 «-8.-) Pepsin-HCl soluble protei in 1002 943 9-90 I'rotein digestion eoeHicit ■nts (/// rilro) ,S2-.-. 7(i-9 78-5 * Not taking into account the volatile organic acids of the silage. Bagl Bag 2 \ f ^ C.reenoats Oat and Green oats Oat and % and tares tare silage increase and tares tare silage increase ()■/.. oz. nr loss oz. oz. or loss Moist material 1053-0 919-0 - 12-7 1118-0 930-0 - 16-8 Dry matter* 247-1 214-0 - 13-2 202-4 227-3 - 13-4 Ortianic matter* 226-1 19.5-S ^ 13-4 240-1 207-8 - 13-5 Crude protein .30-0 25-8 - 140 31-9 28-3 - 11-3 Ether extract* 8-9 10-9 + 22-5 9-4 10-2 + 8-5 N-free extractives 12.-.-2 100-8 - 19-5 1330 lOG-8 - 19-7 Crude lihre (J20 .58-3 - 5-9 05-8 62-5 - 5-0 Ash 210 18-8 - 10-5 22-3 19-5 - 12-5 Ti-ue protein 22-7 10-t) - 53-3 24-2 12-9 - 4(i-7 '■ -Amides" 7-3 15-2 ' 108-2 7-7 15-4 -1- 1000 Pepsin- KCl soluhio 24-8 19-S - 20-1 2«-3 22-2 - 15-6 prt)tein * -Mlowanee made for silage \'ohitiIe organic acitts as acetic acid- Amount of silage dry uuitler in bags calculated as residue after drying at 100° C: 210-3 oz. (bag 1) and 224-3 oz. (bag 2). A. Amu?; and H. E. Woudman 351 Table V. Amounts of con.slitiii'nts of green oats and tares and oat and fare silaye contained in bag i and bag 5 (silo II). Analysis of samples (calculated to dry matter) Green oats Bag 4 Bago and tares o silage o silage o Crude protein /o 11-37 12-81 (1 1309 Ether extract* 3-95 3-18 ?.ir) N-free extractives 49-93 48-62 47- 10 Crude fibre 2703 26-81 27-76 Ash 7-72 8-58 8-90 True protein 8-89 6-75 6-93 "Amides" 2-48 606 616 PepsinHCl soluble protein 9-35 9-97 10-57 Protein digestion coefiicients {in vitro} 82-2 77-8 80-7 * Not taking into account the volatile organic acids of the silage. Bag 4 Bag 5 ^^ A Green oats Oat and o Green oats Oat and ",' and tares tare silage increase and tares tare silage increase 07.. oz. or loss oz. oz. or loss Moist material 93.50 9060 - 3-1 906-0 872-0 - 3-7 Dry matter* 3.340 306-.) - 8-2 323-6 294-3 - 9-1 Organic matter* 308-3 280-7 - 8-9 298-6 26S-(i - 10-0 Crude protein 38-0 38-6 -f 1-6 368 37-8 + 2-7 Ether extract* 13-2 14-9 + 12-9 12-8 14-5 + 13-3 N-free extractives 166-8 14()-4 - 12-2 161-0 136-1 ^ 15-8 Crude fibre 90-3 80-8 - 10-5 87-4 80-2 - 8-2 Ash 2.5-7 25-8 + 0-4 25-0 ■25-7 + 2-8 True protein 29-7 20-3 - 31-7 28-S 20-0 ^ 30-6 "Amides" 8-3 18-3 + 120-5 8-0 17-8 + 122-5 Pepsin-HCl soluble 31-2 30-0 - 3-9 30-3 30-5 + 0-7 protein * Allowance made for silage volatile organic acids as acetic acid. Amount of silage dry matter in bags calculated as residue after drying at 100° C. : 301-3 oz. (bag 4) and 288-9 oz. (bag 5). 352 (hit mill Tan' Silaye Table VI. Amounts of consliliients of green oats and tares and cat and tare silage contained in bag 3 and bag 6 (big silo). Analysis of samples (calculated to dry matter) Ureen oats aad tares Oat and tare silage Bag 3 Bag (i Bags Bag G Crude protein 12-15 10(52 10-99 ll-5(i Ether extract* 3r>i) 3-38 3-37 3-05 N-free extractives 50()7 52-27 50-(i5 50-88 Crude fibre 2510 25-94 26-81 26-20 Ash 8-49 7-79 8-18 8-31 True protein 9-20 812 5-14 5-03 "Amides" 2-95 2-50 5-85 6-53 Pepsin-HCl soluble i>i'otein 1002 8-77 8-29 9-01 Protein digestion coefficients {in vilro) 82-5 82-6 75-4 77-9 * Mot taking volatile organic acids of silage into account. Bag 3 Bag 6 Green oats Oat and 0/ o Green oats Oat and \ O' o and tares tare silage increase and tares tare silage increase oz. oz. or loss oz. oz. or loss Moist material 11200 7.59-0 - 32- 1 SS3-(» 80-i() , 1-0 Dry matter* 262-9 232-9 -11-4 287-8 2()7-(l - 7-2 Organic matter* 240-6 214-1 -11-0 265-4 245-2 - 7-6 Crude ])rotein 31-9 25-2 -210 30-6 30-3 - 10 Ether extract* 9-5 11-0 -f 15-8 9-7 13-1 + 35-1 N-free extractives 1.33-2 116-3 -12-7 150-5 1.33-2 - 11-5 Crude fibre 6(i-0 61-6 - 6-7 74-6 68-6 - 8-0 Ash 22-3 18-8 - 1.5-7 22-4 21-8 - 2-7 True protein 24-2 11-8 -51-2 23-4 13-2 - 43-6 "Amides" 7-7 13-4 + 74-0 7-2 171 + 137-5 Pepsin- HCl soluble 26-3 19-0 -27-7 25-2 23-6 - 6-3 protein * Allowance made for silage volatile organic acids as acetic acid. Anmunt of silage dry matter in bags calculated as residue after drying at 100° C: 229-6 oz. (bag 3) and 261-8 oz. (bag 6). A. Amos and 11. E. Woodman 853 Table VII. Anmiuils of constiluenls vf i/rceii, uais and lares ami out and tare silage in bag 7 and bag S (silo III). Analysis of samples (calculated to chy matter) Ba" 8 .si la'' tJreen oats Bag 7 r ^ ^ and tares silage Unspoilt Spoilt 0/ /o o,' 0/ /O /o ,0 Crmle protein 11-97 13-34 12-68 15o5 Etlier extract* 3-87 2-81 2-70 2-88 N-free extracti vcs 45-43 44-49 40- 15 38-59 t'rudc tibre 3014 30- 10 29-82 31 -.32 Ash 8-59 9-20 8-05 11 (iO True protein 8-32 7-40 7-19 11-73 "Amides" 3 05 5-88 5-49 3-82 Pepsin-HCl sul ublc protein 1010 10-47 902 9-02 Protein digestion coefficients {in uitro) 84-4 78-5 75-9 01-9 * Not taking volatile ( Drganic acids of silage into account. Bag 7 Bag 8 A A Cireen Oat ^ f CJreen '' oats and (1 (J oats Un- "i, and tare increase and spoilt 8|joilt increase tares silage or tares silage silage Total or 07.. oz. less oz. oz. oz. oz. loss Moist material 778-(l 7230 7-1 780'0 4010 238-0 039-0 -18-1 Dry matter* -'93 8 204-2 1(1-0 294() 1521 93-8 245-9 - 1 0-5 Organic matter* :;(i8l> 240- 1 100 209:! 1391 .S2-9 2220 -170 Crude protein 3.-)-2 350 0-0 35 3 19-0 140 330 - 4-8 Ether extract* 11-4 9-4 17-0 11-4 (i-O 2-7 8-7 - 23-7 N-free extractives 133-5 1100 12-7 133-8 09-3 30-2 105-5 -211 Crude fibre 88-5 79- 1 100 88 '8 448 29-4 74-2 - 10-5 Ash 25-2 24- 1 4-4 25-3 130 10-9 23-9 - 5-0 True protein 24-5 190 - 20-0 24-5 10-8 11-0 21-8 -11-0 "Amides" 10-7 15-4 + 44-0 10-8 8'2 3I> 11-8 -i- 9-3 Pepsin-HCl soluble protei!i 29-7 27-4 - 7-7 29-8 14-5 9(1 23-5 -211 * Allowance made for silatre volatile organic acids as acetic acid. Am nurit of sil age dry matter in bag 7 calcndated as resi< lue after drying : It 100° C.=: 202-1 oz. In bag ; 8: 150-2 oz. (uns poilt) and 93-8 ozs. (spoilt). 0 7 8 -I-29-8 - 25-4 -92-3 + 5-0 - 5-2 -190 Discussion of Results. I. Moisture and dry matter ehanges daring ensilage. Numlier of bag 1 2 3 4 5 (_'hange of moisture content -101-5 -152-9 -331-0 -1-5 -4-7 (oz.) Percentage change of - 12-0 - 17-9 - 38-0 -0-2 -0'8 moisture content The above figure.s are of interest owing to the fact that certain of the soluble constituents of the silage are lo.st in the juice draining away from the silo. The losses in the silo are obviously compounded of two main factors: 1. Fermentation losses. 2. Losses in the juice draining away. Thus, where excessive drainage occurs, the percentage loss of dry matter may be high. 354 Old (Hid Tan SiL(((j< Such cases arc funiishod by bags 1 and '1 (silo 1) ami bag li (big silo). The reasons for the large losses of moisture from these bags are twofold : 1 . The green fodder carried an appreciable amount of superficial moisture owing to its having been rained on just previous to cutting. 2. The oats and tares were cut in an immature condition and were as a conse- quence exceedingly "sappy." No wilting was allowed to take place befoie filling the material into the bags and the moisture content was therefore mucli liigher than that of the material in bags 1, d, 7 and 8. It is clear that the relatively high losses of dry matter from bags ] and 2, as compared with those from bags 1 and "j, are to be attributed to the extra losses occasioned by drainage. It follows, therefore, that whilst green "fruity" silage of an excellent quality was obtained by preserving the unwilted, immature forage, yet the excessive drainage, consequent on "sappiness," led to a needlessly high percentage loss of dry matter. A study of the figures obtained for bags 4 and 5, which were filled with wilted oats and tares and from which little or no juice was lost by drainage, indicate that the actual fermentation losses need not exceed 8 i) per cent, of the original dry matter present. The large loss of moisture which occurred in bag 3 has already been referred to in an earlier section of the paper. The gain of moisture in bag G is explained by the fact tliat the bag was placed in tiiu big silo and was surrounded by material po.ssessing a higher moisture content. The infiltration of juice has to somi^ extent augmented tlu- amount of dry matter present in the bag, so that the net loss of dry matter as a result of fermentation is low, namely 7 per cent. It is of interest to note the high percentage loss of dry matter which occurred as a result of spoihng in bag 8, amounting to 16-5 per cent. II. Changes suffered 6// the nitroyvnuas constituents. It is obvious that during the conversion of the green crop into silage a profound change takes place in the character of the nitrogenous con- stituents, mainly resulting in the splitting up of a large proportion of the true protein into simple soluble ])roducts of the amino acid type. For example, only about i 1 i)er cent, of the crude protein in the bag I silage was in the form of true protein, whereas the true protein in the green oats and tares represented 7() per cent, of the total nitrogenous constituents. The figures in this connection for the material in the different bags are as follows: Xumbur of ban Amount of true j Green oats I protein exjiressed ; and tares ) as percentage of : (Jat and ] crude protein ) tare silage j los AND H. E. Woodman :J55 12 3 4 5 0 7 S Unspoilt Spoilt 75-7 7:)-7 7.5-7 7S-2 7S-2 7()-5 70U 7UU 70-U 410 45-7 4i;« 52-7 52-il 43-5 5(j-U 5li-7 75-4 The hydrolytic changes affecting the true protein constituent appear to proceed to the greatest extent during the ensilage of the moist un- wilted oats and tares. Thus in bags 1, 2 and 3, containing unwilted material, the change caused a disap])earance of roughly 50 per cent, of the true protein, this being associated with an increase in the amount of "amides" of about 100 per cent, in the case of bags 1 and 2 and 74 per cent, in the case of bag .3. A bigger increase in the amount of "amides" would undoubtedly have been registered had it not been for the loss of soluble nitrogenous constituents in the large volumes of juice draining away from these bags, since in bags 4 and 5, containing wilted oats and tares, and where little or no drainage occurred, a .splitting up of 30 per cent, of the true protein was accompanied by a 120 per cent, increase in the amount of "amides." The shghtly wilted material in bag 6, which was converted into silage possessing a high moisture content compared with that of the silage of bags 4 and .5 and which suffered no losses on account of drainage, suffered a large loss of true protein (44 per cent.) and the "amides" were augmented to the extent of 138 per cent. The material filled into bags 7 and fS had obviously been subject to change during the time it lay out in the field, since only 70 per cent, of the crude protein was in the form of true protein. The changes in the character of the nitrogenous constituents were not nearly so far reaching in these bags as with the material in the other bags. In bag 7, only a fifth of the true protein was hydrolysed and the "amides" were only increased by 44 per cent. In bag 8, where extensive spoihng occurred, only II per cent, of the true protein disappeared and the "amides" were only augmented to the extent of 9 per cent. In the actual spoilt portion of bag 8 silage, the true protein formed a larger proportion of the total crude protein than was the case in the original oats and tares placed in the bag, the rotting of the material having to a large extent used up the "amides." The above results indicate therefore that the ensiling of "sappy" unwilted forage leads to conditions which are favourable to the ex- tensive splitting up of true protein into "amides." From the nutritive standpoint, it is uncertain whether this can be considered advantageous. Jonrn. of Agric. Sci. xn 24 856 Oat and Tare Silage By combining the results obtained in the analysis of the dry matter of the saniph's with those obtained in the titration of the extracts, it is possible to ascertain in what form the "amide " fractions of the foodstuffs existed. Per 100 grins, dry matter (nitrogen expressed throughout as protein) Bag 1 Bag 2 Bag 3 Bag 4 Cireen (irecn Green Green crop Silage erop Silage crop Silage crop Silage o/ /o /o o/ o/ A) /o o/ /o 0/ o/ /o /o Amino acid.s, etc. oil 4-7() Oil 4-38 on 301 008 4-24 Volatile bases 002 1-17 002 109 ()()2 10.5 003 0-85 True amides 2-82 1-31 2-82 1-38 2-82 119 2-37 0-97 Ba J55 Bag 0 Ba g" Bag 8 Green (ireen Green Green Silage crop Silage crop Silage crop Silage crop , * ^ Unspoilt Spoilt o/ 0/ IV 0/ o/ 0/ O' O' o/ /o /o /o /o /o -o /o /o /O Amino acids, etc. 008 4-74 005 4-13 006 2-81 006 2-95 1-92 \'olatiU' l>ases 003 0-87 002 0-88 003 0-82 003 108 1-55 True amides 2-37 0-55 2-43 1-52 3-54 2-25 3-54 1-46 0-35 The amino acids (with amides of the asparagine type) and volatile bases are determined by titration of the extracts. The result obtained by subtracting the sum of these from the "amides"' as determined in the analysis of the dry matter of the samples has been designated "true amides"; i.e. amides which do not contain a free carboxyl grouj) and thus escape determination during titration of the extracts. It is of interest to note that amino acids and amides hke asparagine comprise only a very small fraction of the "amide" constituent of green oats and tares. This fact is in harmony with the conception that these substances represent stages in the synthesis of plant protein and there- fore exhibit no tendency to accumulate in the plant. The outstanding features of the silage "amide" figures are: 1. A big increase in the amount of amino acids, these forming, in the case of unspoilt samples, the bulk of the "amides" of the silage. 2. An increase in the amount of volatile bases. 3. A decrease in the amount of "true amides." The increase in the volatile bases is probably explained by the production of ammonia by a hydrolytic change affecting the "true amides." Thus, in bag 1, 1-5 percent, "true amides" disappeared and 1-15 per cent, volatile bases made their a])pearance. The figures for bag 8, where spoiling occurred, are excej)tionai. in the case of the spoilt portion of bag 8 silage, the amounts of amino acids and "true amides" were low, whereas the amount of volatile bases was relatively high. A. Amos and H. E. Woodman 357 Such features may be regarded as the chemical evidence of tipoihng, the volatile bases having made their appearance as a result of the destruction of amino acids. In the production of good silage, the main changes aflecting the nitrogenous constituents are probably brought about by proteolytic enzymes and an investigation of the material reveals a high ratio of amino acids to volatile bases. The rotting of silage is evidenced by a low ratio. The ratio in the case of bag 7 silage affords evidence of the slight degree of spoiling which actually did occur in this bag. A further ilhi.stration is seen in the figures already given for the titration of the extract of the spoilt silage samjjle from the top of silo II. Here the volatile bases are actually present in excess of the amino acids, the ratio of amino acids to bases being roughly 1 : 1-6. In bags I, 2 and 3, where large quantities of juice were lost by drainage, there were appreciable losses of crude protein, probably in the form of "amides" dissolved in the juice. In bags 4, 5, G and 7, where little or no juice was lost by drainage, only small changes in the crude protein constituent were recorded. The slight gain of protein in bags 4 and 5 arose probably from experimental error (difficulty of accurate samphng, etc.). It may safely be assumed that if silage can be made without excessive drainage from the silo, then the loss of crude protein will be small. A twofold problem awaits satisfactory solution if the practice of ensilage is to develop on a sound economical basis: 1. The prevention of excessive drainage from the silo. "2. The possible utilisation of silage juice in feeding. Both these questions are receiving attention at the present time. In every case the crude protein digestibility of the green oats and tares (as determined in vitro) suffered a slight depression as a result of ensilage. A study of the data points to the probability that the de- pression of protein digestibihty is greatest where large losses of the easily assimilated "amides" occur as a result of drainage. Where spoiling occurs, as in bag 8, the decrease in protein digestibility may be con- siderable. III. Changes suffered by the Ether Extract. In all the bags, with the exception of bags 7 and 8, the amount of ether soluble material was augmented as a result of the changes under- gone by the green forage during ensilage. The increase was very variable in amount and apparently bore no relation to the percentage loss of nitrogen-free extractives. Thus, in bag 4, where about 12 per cent, of the nitrogen-free extractives disappeared, there was a 13 per cent. 358 Oat and Tare Silarje increase in the ether extract, whereas in bag 6, an ahnost equal i)ei- centage destruction of nitrogen-free extractives was accompanied by a 35 per cent, increase in the amount of ether extract. The results obtained in this regard for bags 1 and 2 were curious. In both bags the percentage loss of nitrogen-free extractives was about 19 per cent.; in bag 1, however, the increase of ether extract was 23 per cent., the corresponding figure for bag 2 being only about 9 per cent. As bag 2 occupied the ujjper half of the small silo, it is ])ossible that a portion of the volatile organic acids escaped as a result of "heating," since analysis of the extracts showed that the bag 2 silage was iiiuch poorer in respect of volatile acids than was the bag 1 silage. In bags 7 and 8, where spoiling occurred, there was an appreciable decrease in the amount of ether extract during ensilage. It follows that where silage undergoes rotting, the process is attended by losses of ether soluble material. It has already been demonstrated that the organic acids are readily destroyed during the spoiling of silage by moulds. If the figures for the ether extracts be compared with those obtained in the titration of the organic acids in the extracts, certain interesting facts are disclosed which merit further investigation. The figures ob- tained for the green crop extracts show that the green oats and tares contained an appreciable amount of non-volatile material which titrated with iV/IO NaOH in the presence of phenolphthalein. The amount of volatile acidic constituent was negligible. The nature of the non-volatile acidic constituent was not ascertained. It was observed, however, that during titration of the green croj) extracts, a strong yellow colour de- veloped, which intensified as the neutral point was approached. It is probable tliat the oats and tares contain a constituent which reacts with the soda producing the colour change, and the soda used up in this reaction accounts for the high non-volatile acidic figure obtained for the green crop. As this phenomenon was noticed in an equal degree during the titration of the silage extracts, it follows that this constituent escapes destruction during ensilage, and thus it is not feasible to cal- culate the whole titration figure for non-volatile acidity in the silage in terms of lactic acid. It was further noted that the spoiling of silage by moulds occasioned the destruction of the constituent in question. The following table gives a comparison of the percentages of ether extract in the dried silage samples with th(> iion-voIatile acidity calculated as lactic acid, after subtracting from the uon- volatile acid titration figure the corresponding figure for the green oats and tares. A. Amos and H. E. Woodman 359 Per lUOgrma. dry oat and tare silage: Bags 1 2 3 4 0 G 7 S Unspoilt sample Non-volatUe acidity calculated 0-53 7-22 6-11 311 2-51 5-17 2'85 2-48 as lactic acid, % Ether extract in dry matter, o„ 3-12 3-21 3-37 3-18 3-15 305 2-81 2-70 It would be anticipated that the amount of ether extract in the silage after drying at 100° would show some correspondence with the amount of non-volatile acidity calculated as lactic acid. This was the case in the material of bags 4, 5 and 7 and in the unspoilt portions of bag 8. Wide discrepancies occur, however, in the figures for the silage in bags 1, 2, 3 and 6, warranting the assumption that a very different acidic fermentation has taken place in these bags from that which occurred in the remaining bags. It is clearly impossible to assume that the non-volatile acid in these bags was lactic acid wholly. In this con- nection, it is interesting to remember that bags 1, 2 and 3 contained unwilted material, and bag 6 slightly wilted material, and that the fermentative processes in these bags went on therefore under much moister conditions than obtained in the other bags. The question is worthy of further investigation, since excellent samples of silage were obtained from bags 1, 2 and 6. It is probably incorrect to assume that lactic acid is the only non-volatile organic acid which may arise as a result of fermentative changes in the silo. IV. Changes suffered bi/ the crude fibre constituent. In every case, a loss of crude fibre was recorded as a result of ensilage, the loss varying from about 5 per cent, in the case of the unwilted oats and tares to about 10 per cent, in the case of the wilted material. In the case of the spoilt sample in bag 8, the loss of crude fibre was as much as 16-5 per cent, of the original amount of fibrous constituent in the bag. The question arises as to the type of action in the silage process which results in the disappearance of crude fibre. The explanation is probably to be found in work carried out by Voelcker^, who found that the following change occurs when straw chaff is mixed with a small quantity of green rye or tares (1 ton straw chaft' to about 1 cwt. of rye or tares) in the spring or summer and the mixture is allowed to ferment until the autumn. The untreated straw contained 23-7 per cent. N.-free extractives and 54 per cent, crude fibre. The fermented straw contained 45-8 per cent. N.-free extractives and 34-5 per cent, crude fibre. 1 Voelcker, ./. of the R. Agric. Soc. of Eng. 7, 85, 1871. 360 Oat (iiul Tare Silaye It is tlius obvious that the fermentation to which the straw was submitted had the effect of converting an appreciable amount of the cellulose of the straw into a form which could be dissolved during the process of determining the percentage of crude fibre. The conversion of crude fibre into nitrogen-free extractives involved a distinct improve- ment in the feeding value of the straw. From the results of the present investigation, it appears probable that the cellulose of thi^ green oats and tares undergoes to some e.\tent a similar breakdown during ensilage, resulting in a gain of nitrogen-free extractives and a corres|)ondiiig decrease in the amount of crude fibre. Indeed, the nitrogen-free extractives so formed may conceivably undergo further change with the production of organic acids, although this may not hap])en if the green forage develops acidity rapidly in the early stages of storage in the silo. It should be pointed out that the oats and tares used in this in- vestigation were cut in an immature condition, so that the optimum conditions obtained for observing possible changes affecting the fibrous constituent. To what extent the cellulose of mature forage would be subject to such changes is uncertain, though Voelcker' s work indicates that similar changes would occur. The ])oint is under investigation. The clianges modifying the cellulose constituent during ensilage of fodder are of great importance from the jjoint of view of nutritive value. The gain is twofold: 1. Part of the fibre breaks down into nitrogen-free extractives (and, to some extent possibly, to organic acids), i. The residual crude fibre itself possesses a greater digestibility tlian the original crude fibre in the green oats and tares^. V. Changes Kiijjered bi/ the nitrogen -free extractives. In no case was the loss of nitrogen-free extractives less than 11 per cent, of this constituent originally present in the bags. The biggest losses were registered in bags 1 and 2, the diminution being roughly 19-5 per cent, in each bag. It is significant, in view of the remarks contained in the preceding paragraph, that the amount of crude fibre disappearing in these bags was much smaller than in the remaining bags. The results for bag 8 show that the rotting of a silage sam|)lc may lead to large losses of nitrogen-free extractives. VI. Changes suffered by inorganic constituents. Large losses of inorganic salts may occur during ensilage owing to the sohibli" portions escaping in the juice. Where little drainage occurs, ' VVoodmiui, J. Ayric. Hci. 12, 144, 1922. A. Amos and H. E. Woodman ;361 the loss of these constituents is not appreciable. These facts must be considered in conjunction with the importance attaching to the inorganic salt content of a foodstuff from the nutritional standpoint. If, for example, silage is to be utilised in feeding dairy cows, it is undesirable that the green crop should sustain large losses of inorganic salts during ensilage. Summary. Experiments have been described which had primarily as their object the investigation of the effect of varying moisture content in the green oat and tare crop on the type of silage produced from such forage. The magnitude of the changes affecting the constituents of the green crop under the different conditions of ensilage have also been detailed. The main conclusions are summarized below: (1) The ensihng of a fresh green crop in an early stage of maturity (soon after flowering) with Httle or no previous wilting, and with a fer- mentation temperature in the neighbourhood of 25° C., leads to condi- tions which favour the production of green "fruity" silage. The results of the experiments suggest that with this type, a considerable loss of dry matter may result from the process, especially if much drainage of sap occurs. The same conditions appear to be favourable to the ex- tensive splitting up of true protein into soluble nitrogenous products; more than 50 per cent, of the true protein of the green crop may be transformed into "amides," an appreciable pro]iortion of which may be lost, together with inorganic salts, in the drainage juice. Thus, though green "fruity" silage is much relished by stock and possesses excellent feeding value, yet its production may be accom- panied by substantial losses of crude protein and soluble salts. Loss by drainage should therefore be obviated. (2) The ensihng of a moderately dry crop, dried either by wilting or by allowing the crop to mature, produces conditions which favour the production of the acid brown type of silage. The production of this type of silage is accompanied by a relatively low loss of dry matter, the amount of juice drainage from the silo being very much smaller than that occurring during the production of green "fruity" silage. Approxi- mately .30 per cent, of the true protein of the green crop is spht up into "amides." (3) The ensiling of material which has undergone prolonged wilting and extensive rain-washing does not produce a good quality of silage, and the forage displays a tendency to become mouldy during the process. ;J62 Oat and Tan S'daye. Where moulding occurs, the volatile and non-volatile organic acids are not simply neutralised by basic products, but are actually destroyed. The process is accompanied by large losses of dry matter, the "amides," nitrogcn-frce extractives and ether extract being extensively destroyed. C'liemical evidence of the spoiling of silage by mould development is afforded by a study of the ratio of amino acids to volatile bases in the silage extract. In good silage, the ratio is high; in spoilt silage, the volatile bases may be present actually in excess of the amino acids. (I) In all the samples of normal silage investigated, the non-volatile organic acids were ])resent in good excess of the volatile organic acids. Evidence has been brought forward warranting the assumption that the acidic fermentation during the formation of green "fruity" silage is markedly different from that accom])anying the production of brown acidic silage, and it appears jirobable that lactic acid is not the only non-volatile organic acid which may arise as a result of the fermentative action in the silo. (.5) The increases in the amount of ether extractable material as a result of ensilage are very variable in the different experiments and bear no relation to the ])ercentage losses of nitrogen-free extractives. ((!) In every case, the crude protein digestibility of the green oats and tares (as determined in vitro) has been shown to suffer a slight depression during ensilage. (7) The outstanding features of the silage "amide" ligures as com- pared with the corresponding figures for tlie green croj) are: («) a large increase in the amount of amino-acids, these forming the bulk of the "amides'" of the silage; (b) an increase in the amount of volatile bases, the latter consisting probably of ammonia which has arisen us a result of hydrolytic changes affecting amides originally present in the green crop. (8) Results have been obtained which suggest tliat the cellulose of green oats and tares undergoes to some extent a breakdown during ensilage, resulting in a gain of nitrogen-free extractives and a corre- sponding decrease in the amount of crude fibre. Furthermore, as has been demonstrated in a [irevioiis communication, the crude fibre re- maining in the silage possesses a greater digestibility than that originally present in the green forage. {Received Aligns! 7l/i. I!)'2'2.') THE AVAILABILITY OF MINERAL PLANT FOOD. A MODIFICATION OF THE PRESENT HYPOTHESIS. By NORMAN M. COMBER. {Department of Agriculture, The Urdversity, Leeds.) The Present Hypothesis. Scientific soil researches have pussed through one fairly well defined phase. The important and pioneer work of empirical field trial has established the main physical and chemical requirements of crops and laid down the routine method of examining soil amendments. The researches of recent years have inaugurated a second phase : an enquiry into the constitution of the soil and into the mechanism of the growth of plants in soil. In connection with this second phase of soil investigations much progress has been made on the botanical side in developing knowledge of the mechanism of processes which go on inside the plant, and in the last decade views of the constitution of the soil have changed fundamentally. But the relation of the plant to the soil has received very scanty consideration in modern literature. Views on that subject are very much as they were after Daubeny's publication in 1845. Outstanding work has been done by Dyer^, by Cameron and Whitney-, and by HalP and his collaborators. There have been some changes of interpretation and some differences of opinion, but the fundamental assumption that the soil solution is the nutrient medium of the plant, that the state of solution is necessary to availability has remained almost unquestioned. The basis of practically all teaching on the chemistry of soil fertility and on the value of fertiUzers, is the behef that nutrient substances pass into solution in the soil water externally to the plant and subsequently diffuse into the root hairs. Much well-known discussion has arisen about the solvents concerned, particularly about the possible excretion of organic acids by the plant. ' Tran.-i. Chem. Sue. 189-t, 65. = U.S. Dept. Aijrk. Bureau vj Soils Bull 1903. No. 22. 3 Phil. Trans. 1913. 3G4 The Arailabili/i/ of Mineral Phmt luxnl The Inadequacy of the Present Hypothesis. There ;ire a number of facts wliieh are difficult to reconcile with this hypothesis. 1. The relatio)! of the composition of the soil solution to the mineral elements taken up, and the water transpired, by plants. In obvious and natural accordance with the view that the soil solution is the nutrient solution of the plant, many attempts have been made to express or dis- place this solution from the soil, and by its analysis to ascertain the essential chemical knowledjj;e of the fertility of the soil concerned. The results of these experiments show that the ratio of certain mineral ele- ments assimilated by plants to the water transpired is greater than the ratio of these elements to the water in which they are dissolved in the soil. Hall' has calculated that if the potash taken up by a clover C'rop is assumed to have been dissolved in the water transpired, the calculated solution is far more concentrated in respect of potash than solutions obtained from the soil. Raniiinn- makes a similar observation but arrives at a different conclusion. H;ill deduces that the dissolved material enters the plant from the soil solution at a greater rate than the water. Ramann is apparentlv unable to reconcile this view with the laws of diffusion and concludes that the i)lant takes material from the soil other than that which is in solution. 2. The absorption of iron by plants. Iron is a necessary element for plant growth. All plants take up a certain amount from the soil. The amount necessary is relatively small, but it is difficult to see how even this small amount can exist in sol id inn in soils containing a high percentage of (thalk. 3. The nvailabilil 1/ of phosphates. This is probably the greatest difficulty of all. Tlic value of various phosphatic substances to plants is not by any means in accordance with the present hypothesis. Basic slag and certain mineral phosphates are sparingly soluble and are yet valual)le phosphatic fertilizers. The evaluation of these substances entirely on a solubility basis has not been an un(iualified success. Their usefulness as phosphatic fertilizers is often comparable to that of a water soluble phosphate. Tlie phosphates of iron and aluminium are in a curious position in the literature. On t he one hand it is normally taught that these are insoluble » '/'At Soil. 19-20. - Lamlir. VirSlul. 1<)10, 88. N. M. doMHRR 365 and of little use to the plant, and that the formation of these phosphates in superphosphates diminishes the value of the material to the plant. On the other hand, direct ex])erimenti leaves no doubt that phosphates of iron and aluminium are very useful to the plant as sources of phos- phorus. There seems to be a prejudice, which prevails in spite of experi- mental evidence to the contrary, that sparing solubility must mean a low availabilitv. The e.xperimental fact is that sparint^ly soluble phosphates are often easilv available, and the inevitable conclusion is that solubility is not the dominatinji' condition of availability. A Modification of the Present Hypothesi.s. The hypothesis that all material entering the plant from the soil first becomes dissolved in the soil water and then diffuses into the root hair, arose on the basis of two fundamental beliefs. First, the physiological belief that only material in true solution can diffuse through the cell protoplasm and second, the belief that the soil is a simple system of particles moistened by a solution. Now tliese two fundamental concep- tions have become modified. Becaxise of the modifications and also because of the cbscrepancies already noted between the facts and the hypothesis, it is desirable that the hypothesis should be reconsidered. The absorption of colloids hi/ the plant. As a first point in the recon- sideration it may be noted that the teaching that only material in true solution can diffuse into the plant cells is not unquestionably true. There is some eviflence that less highly dispersed material can enter. Czapek- definitely states that colloids can enter the plant cells. "Die lebende Plasmahaut ist nicht nur fiir echte Losungen, sondern auch fiir kolloide Losungen durchlassig." He refers to the passage of dyestuffs and of fat emulsions into plant cells. Pfeffer^ records the diffusion of silicic acid into the cell sa]) of plants. There is also some recent experimental evidence in favour of the absorption of colloidal silica by plants. Jennings* describes a series of experiments in which wheat seedlings were grown in water solutions to which were added various absorbing substances. When colloidal silica gel was used there was a marked increase in growth and in drv matter, and it was clearly shown that the plant had absorbed silica. ' See Marais, Soil Set. 1922, 13, No. .") ;uui bibliography. " Biochemie der PJIanzeii, 1M3. ' Tlw Ptiysioloyij uf Plants. 1900. « Soil Set. 1919, 7. gill. (1 . it 0-2628 5-3 ()-4()()9 16-0 0-jt>35 31 10540 20-2 0-5743 2-8 1-1404 22-8 0-9l)8S 2-0 1 -4235 31 3 360 T/ir Arai/dhifili/ of Minerdl Plant FiukI The followini;; are some of Jennings' figures: Dry weights and silica contents of wheal seedlings grown in nutrient solutions and in nutrient solutions containing silica (Jennings). Dry weight Silii-a in of tops dry iiiatt<'r XT 1 - * 1 i- o- L -II- ( witliout silit-a Nutrient solution 8.) pts per million , ,. ■ , ,, ' ' I + silica gel 1 "„ .,-,, < without silica ( + silica gel 1 ",j -„„ (without silica .. ^>00 ,, -. .|. t I II I + silica gel I "„ |.N.,.. (without silica " " ] -f- silica gel 1 ",, Thus it appears that colloidal j^els can enter the plant, and entering the plant from the soil they may carry with them absorbed substances. The essential elements for plant growth. It is commonly held that there arc ton essential elements involved in the growth of plants. That tenet implies tliat all other elements universally found in plants grown in soil are there by accident. That this teaching has not been thoroughly acceptable is well known and from a studv of plants grown in soil it has been contended tliat sili(ton and a number of other elements are essential. The argument for ten essential elements and no more is, of course, the fact that apparently normal ])!ants can l)e grown in water solutions with- out the introduction of any detectalile tia('e of elements other than the well-known ten. That fact proves that only ten elements are required for the growtli of plants in water solutian. The application of the same con- clusion to plant growth in soil is entirely dependent upon the assumption that the mechanism of feeding is the same in the soil as in water solution. If the mechanism of feeding is different the chemical requirements may well be different too. If, for instance, silica acts as a kind of carrier of mineral substances to plants grown in soil, silicon will be an essential element for that purpose. By considering the possibihty of a mode of absorption from soil different from that which obtains in the case of plants grown in water solution, some explanation of the additional ele- ments always present in soil-grown plants may be found. The relation of the root hairs In the soil particles. Another point arises from the modified view- of the constitution of the soil. Whereas the soil system was formerly regarded as merely a system of moistened particles it is now regarded as a system of particles which are coimected with the external aiul free water by gel material. Some attention has been given to the bearing of this conception of the soil on transpiration and the N. M. CoMBKH 367 water content of plants '^. Its bearing on the mineral nutrition of plants still awaits consideration. Now it is a well-known Ijotanical fact that when the root hair comes into contact with a soil particle the outer la)'er of the root hair is trans- formed into a mucilage and an indissoluble attachment is made with the soil particle. One relation of the soil colloids to the plant is here apparent. The soil particle is coated with iiydrophilous colloid. The root hair be- comes coated with hydrophilous colloid. By the union of these colloids the plant and the soil particles become cemented together. Tlie particles so attached to the plant cannot be removed without damage to the root hairs. Underlying the present teaching is the conceptiou of the root hair "dipping into" the soil solution, and taking up its nutrient material in exactly the same way as it does from experimental water culture solutions. It seems, however, important to acknowledge the fact that by the union of their respective colloids the plant and the soil form one si/stcm and not I wo si/steni.? in mere contact, and to admit the possibility that the mechanism of the nutrition of plants grown in soil is not necessarily and entirely the same as the mechanism of the nutrition of plants grown in water solution. The physical possibilities of the union of the plant and the soil. The migration of ions in the colloidal complex by which the plant and soil are united has been discussed by Casale^. He argues that positive and negative colloids exist in the soil and that the charge on them is due to their throwing off anions and cations respectively. The negative colloids are dominant and absorb the positive colloids. The colloids of the plant, according to Casale, throw off hydrogen ions and are negatively charged, but less so than the soil colloids. Hence there is a potential difference between soil and plant and a migration of cations from the soil to the plant accordingly. Equilibrium is never reached, because the cations pass on into the plant by a similar electrical mechanism. Sufficient seems to be known about the different electrical charges on various parts of the plant cells to indicate that a migration of ions under difference of electrical potential probably plays an important part in plant nutrition and cannot be left out of account in a consideration of the availabihty problem. The chemical possibilities of the union of the plant and the soil. When the root hair and the soil particle are cemented together there is an obvious possibility of the cell sap of the root hair dissolving material 1 See Sludl, rmiix. Faraday Soc. 1922, 17. = Slaz. spa: agr. ital. 1921, 54; J.C.S. Abs. (i), 1922, 509. 368 Thr A railabi/ifi/ of Afineral Planf Food from the soil aggregate without any excretion of acid into the soil generally. The cell sap and the soil particles to which the cell is cemented are in such contact as to admit the direct dissolution from particular particles by the sap of the ])articulur cells attached thereto. Organic acids have a far greater solvent power for many mineral substances than minoral ;i(-ids have. Oxalic acid will dissolve much larger amounts of iron from the soil tlian hydrochloric acid. Calcium phosphate is much more soluble in citric, acetic, lactic and malic acids than in hydrochloric or nitric acids. But it must be noticed that the solution of mineral matter by the sap and nmciiagc of plant cells is not to be re- garded as essentially an acid dissolution. Many organic compounds besides acids will dissolve iron and aluminium compounds and phos- phates. Tricalcic phosphate, for example, is much more soluble in water containing starch, glue, sugars and many other organic bodies. This effect of organic matter on the solubility of phosphates and of compounds of iron and aluminium is very well known and in soil phenomena it must be far reaching. It affords a satisfactory explanation of the uptake of iron fi-oiu cluilky soils, for numy organic compounds have a high solvent power for iron over a wide range of both acid and alkaline reaction. The enormous amount of oxalate in lichens growing on limestone is not without significance in this connection. The availability of sparingly soluble phosphates becomes more in- telligible if it is supposed that the root hairs become cemented to the particles thus admitting the absorption of colloidal phosphate and the direcl dissolution of the particle by the organic solvents concerned. .\lso it is easy to see the connection between the root habit of wild white clover and its special response to the presence of these sparingly soluble phosphates. The available mineral phosphates have, according to the argument of a succeeding paper^, a hydrophilous surface with which the root hair can make its attachment. The unavailable phosphates appear not to have a colloidal surface. Ferric phosphate loses its colloidal properties on ignition and it also, according to Prianischnikow- becomes unavailable for the plant. Aluminium phosphate retains its colloidal properties after ignition and it remains available for the plant. There are, therefore, strong indications that colloidal properties in sparingly soluble mineral plant food are of first importance, perhaps enabhng the plant to absorb colloidal matter; certainly enabling the • ThLs \'oluine, p. .372. - liicd. CciUr. 1905, 34. N. M. Comber 3). .\ge Live weight "„ of No. Breed weeks lbs. skeleton 1. Windsor \, ,.,. 10 161 18-6 2. .. from same htUr „„ „, .„„ 4. Large White |, ..,, 9 24-3 ., •=■ Jirora same Mttcr .,., _. ,. 28 751 10-3 170 While some of the discrepancies between different authorities may be accounted for by the various methods adopted in cleaning the skeleton before weighing, yet on the whole the proportion of meat to bone varies considerably both with breed and with age. Material and Methods. The investigation was conducted in the same way as that previously described for cattle (7) and sheep (8) and consisted of a statistical treat- ment of the records of the Fat Stock Show held by the Smithfield Club at Islington from the year 1901 (when the classes for pigs were restarted) until the year 1913 inclusive. I am indebted to Mr E. J. Powell the Secretary of the Club who has kindly supplied me with these records. Two series of competitions for pigs exist (1) The Live Classes, and (2) The Carcase Classes. In the first Series — Lire Classes — a record is kept only of the age and gross weight of the animals. This .series is divided into classes for breeds and crosses of different ages, the sex not being specified, each pen con- sisting of two pigs. Classes also exist for ])orkers not exceeding 100 lbs. weight in which the age is not limited. In addition there are classes for single pigs of the different breeds under 12 months old. John Hammond 389 In the second series — Carcase Classes — wliicli were started in 1903, a record is kept not only of the live weight and age of the pig but also the weights of the carcase and pkick after the animal had been killed. Tliis series is divided only into classes for pigs of different weights; in some cases in addition there being an age limit as well. Each entry consists of a single pig only. The entries for this series are very small compared with the number of exhibits in the live classes especially in view of the fact that the carcase test is the ultimate object for which the pigs were bred. The weights of the animals and carcases have been carefully taken by the officials of the Club under the supervision of the Stewards. For the details of the methods of slaughter and dressing of the carcases I am indebted to jVIr Charles Bone who has been responsible for these and who has kindly furnished me with information on these points. Neither in the Live Classes nor in the Carcase Classes is there any specification of sex. Henry (9) found that boar pigs weighed slightly more than sows, but Danish (lO) experiments have shown very little difference between the sexes. The details given in the records of the Show have been treated statistically so as to give information on the points it was desired to investigate. Weights throughout have been given in lbs. and decimals of a pound, and ages in months and weeks. As the single pig classes would naturally attract the exceptional animals it was thought better to keep them separate. The fact that in the porker classes the weight was limited but not the age was not con- sidered to be of sufficient importance to warrant separate treatment and they have been grouped by their age; breeds maturing early would be exhibited in this class at a younger age than those of the late maturing breeds. In order to avoid confusion when discussing the results of the in- vestigation below, the following account gives the methods by which the various tables in the text have been compiled. Table I has been prepared from the records of the Show direct. As each pen consists of two pigs the total has been halved to obtain the weight of individual pigs. The divisions into age groups correspond more or less with classes at the Show, the bulk of the porkers falling into the "under 3 months" and "3-5 months" groups wliile the majority of the classes "not exceeding 9 months" and "not exceeding 12 months" fall into the groups " 8-9 months " and " 10-12 months " respectively. A few of the former constitute part of the " 6-7 months " group. The single pig 26—2 390 Groivth and Development of Breeds and Crosses of Pigs classes have been ke})t separate and are shown at tlic I'ud of the tabic Averages calculated from less than 10 individuals are shown in italic type. Table II has been calculated from Table 1 by dividing the weight by the number of weeks old (counting 4 weeks to a month). The weight at birth, owing to lack of data on this point, has been taken as zero. The numbers of individuals from which the results have been calculated are also given so that an estimate can be made of the reliability of the average. Averages calculated from less than 10 individuals are shown in italic type. Tabic III has been compiled from Table 1 by adding or subtracting the number of weeks growth (see Table II) reijuired to correct for age. Averages calculated from less than U) individuals are shown in italic type. The probable error of the mean has been calculated in several cases and is shown in Table 1 V. Table A. Probable error of mean (pen of two), lbs. Months old— 3 ;-) 7 9 11 Liu-go White Berkshire 3- 03 2-7t> 3-48 1-3I) 12-89 4-30 0-98 512 2-51 Table IV has been prepared from the records of the Carcase Classes direct. Results averaged from less than 5 individuals are shown in italic type. The parts into which the pig is divided on slaughter are as follows: Carcase. Includes the head, feet and skin (but not hair or claws). The skin varies in thickness in different types of ])igs(U). Henseler(i-) in a starved Bavarian belted sow of 79 lbs. found that it was li per cent, of the live weight and CoUn (13) that it was lO- 1 per cent, in a pig weighing 64 lbs. Within a breed the proportion would j)r()bably decrease with the increasing size of the animal. Pluck. Consists of the pluck proper — heart, liver, lungs, diaphragm, oesophagus and trachea — and in addition the caul fat. thus differing from the pluck of cattle and sheep. The total percentage is very similar to that found by Holm(i4) in Danish pigs of the same weight. " Unaccounted for.'' This is the difference between the sum of the two foregoing parts and the live weight; it consists mainly of stomach and intestines and their contents, but also includes blood, hair, spleen, pancreas and genitals as well as the weight lost on cooling. In order to give some idea of how the weight of the Pluck and " Un- accounted for" is distributed among the dift'erent organs the following Table B has been prepared: John Hammond 391 ■S3 I « I— I es EH F10X pooia (q) puB (») JO S'jua^uo,;) (q) sauijsaju];-, {ij) ip-eraoigj SAM![0 pUB JIBJJ Sni[0O^-) jBcjox ; sn3BqcIoB3() B3l[0BaX i uiSBjij({>n(j I .WAI'T '^ ■ i CO Q O 1.IB3H ;:; CO oi -r; CO CD O CC 05 y-j 'X -^ 'ri O O) •i: ■ J: CO -M ^ CO 1 ( I "^ c-. 6 6 6 1 1 1 O I I I CO CO o 6 ih CO 1 I U 'j''"v r>- I I I I I I I I 6- CO oi c J}"j » J= . -:.M TD > bt a; O O > X -* m m — (0 3 -M ■a ~ ID J; c ^ i I I I -t X CO oj oi -rj -Si.. Ti r^ — " ^ -H 01 oi J3 s c "5 =«i; > C 0) 1-1 3 <5 -i; S "J a a ^!CJ (•'rotcth and Dcrdujiitient of Breedf^ (iiul Cruises of P'ujs Tlie differences between authorities may be due partly to variations in method but are largely affected by differences in the size and age of the animals. The liver of the pig at birth is much larger in proportion (3 per cent, of liver) than it is at a year old (1 per cent, of liver). It will be noticed too that the proportion of the stomach and intestine contents increases from about 2-5 per cent, at birth to about 1 1 per cent, at 3 months old and then decreases again to about 2-5 per cent, at a year old. Table V has been calculated from Table lY in the same way that Table 11 has been jjrepared from Table I. Averages from less than 5 individuals are shown in italic type. Tlic four breeds from which the average at the base of tiie table has been calculated are as follows: Large White, Large Black, Berkshire and Middle White. The breed average only has been considered and not the number of individuals in each breed. Table. VI has been prepared from Table IV in the same way as Table 1 1 1 has been calculated from Table I. Averages from less than 5 indiNaduals are shown in italic type. The average at the base of the table refers to the same four breeds as in the preceding table. Table VII has been compiled from Table VI by calculating the weight of the carcase or part as a percentage of the live weight. Averages from less than 5 individuals are shown in italic type. The average at the base refers to the same four breeds as stated in Table V. Table VIII has been (uilculated from the averages of the four breeds given at the base of Table VI. The method used is described in the text (see p. 402). Table IX has been prepared from Table 111 by calculating the weights at various ages as percentages of the weights at 1 1 months old. The single pig classes at 9 months are shown as a percentage of the single pig classes at 11 months. Figures calculated from less than 10 individuals on either side are shown in italic type. Table X has been compiled from Table VI by calculating the weights at various ages as percentages of the weight at 11 months old. Figures calculated from less than 5 individuals on either side are shown in italic type. Table XI has been prepared from the Live Class records of tlie Show direct. All doubtful and second cross animals have been ehminated. Averages calculated from loss than 10 individuals are shown in italic type. Table XII has been calculated from Table XI in the same way that Table II has been calculated from Table 1. Figures from less than 10 individuals are shown in italic type. John Hammond 393 Table XIII has been compiled from Table XI in the same way that Table III has been prepared from Table I. The mean weights between the two breeds have been taken from Table III. Averages from less than 10 individuals are shown in italic type. Table XIV has been calculated from Table XIII in the same way that Table IX has been prepared from Table III. Figures calculated from less than 10 individuals on either side are shown in italic type. I'able X V has been prepared from the records of the Carcase Classes direct. All second and doubtful crosses have been eliminated. All figures are shown in italic type as in no case is the average calculated from more than 5 individuals. Table XVI has been compiled from Table XV by showing the per- centage of the carcase or part as a percentage of the live weight. All figures are given in itahc type as in no case is the average calculated from more than 5 individuals. Table XVII has been prepared from the records of the Live Classes of the Show after correction of each for age as described for Table III. Figures calculated from less than 10 individuals are shown in italic type. Table XVIII has been compiled from the records of the Carcase Classes after correction of each for age as described for Table III. Averages from less than 5 individuals are shown in italic type. Table XIX has been calculated from the records of the Live Classes of the Show after the weight of each pen of 2 pigs had been corrected for age as described in Table III. Table XX has been calculated from the records of the Carcase Classes of the Show after the weight of each animal had been corrected for age as described in Table III. Table XXI has been compiled from the records of the Carcase Classes of the Show by calculating for each animal the carcase and parts as a percentage of the live weight. Individuals were then classified into tliree equal groups — high, average and low — according to the live weight or percentage of the part they were grouped by. The averages for each of these groups were then calculated. Each different age group as shown in Table IV was treated separately in this way and then all the age groups were combined and their average is given in this table. The numbers of animals dealt with in each group were as follows: high 74, average 75, low 75. 394 Growth and Development of Breeds and Crosses of Pigs Results. The results obtained are considered in sections below under various headings which it was considered might have an eSect on growth and development — breed, age, early maturity, cross-breeding, selection, individual variation and correlation. In general, the results obtained from the Live Classes for actual gross weight have been considered first followed by the proportional development and weights of the various organs as obtained from the Carcase Classes. Breed. No satisfactory explanation has yet been offered why one animal or breed should produce meat more economically than another. Armsby and Fries (18) have shown that it does not depend on differences in digestive power; their experiments indicated that the well-bred animal had a lower maintenance requirement than a scrub, due to the hitter's more nervous disposition and greater restlessness. Feeding tests for the economy of gain in different breeds of pigs made at the Ontario College of Agriculture (if») have failed to show any uniform difference between breeds as so much depended on the strain used. Carlyle{20), however, who compared Berkshires with Razorbacks or semi- wild swine found that the latter required more feed per unit gained. Similar results were obtained by Diffloth(2i) when comparing Suffolk pigs with those of a fJerman breed. It is rather in the utilization than in the absorption of food that breeds differ; variations in utilization will be seen below. Little or nothing is known as to whether the underlying physiological differences in the metabolism of different breeds are due to variation in heat loss by the skin as shown by Wood and ]Iill(22) or to differences in the glycogenic coefficient of the blood controlled by the glands of internal secretion. Liihring(23) states that it is possible to distinguish between the different breeds of pigs by the protein differentiation method. Table I shows the average ages and weights of the different breeds of pigs as compiled from the records of the Live Classes. The following Table II shows these weights translated into lbs. per week gained since birth. The weight at birth has been neglected owing to the absence of data for the dift'orent British breeds. Carmichael and Rice(2i) found that the average birt h weights of breeds in America ranged from 2-25 lbs. for Duroc- Jerseys to 2-6 f lbs. for Berkshires but it was considerably affected by other factors such as litter size, sex, age of sow, etc. Meek (25) states that the birth weight varies from 1'.5 lbs. to 4-5 lbs. and Henr\M26) that the variation is from 1-3 lbs. to 3-1 lbs. John Hammond 395 p. CO o 05 S •« .^ '^' T ? "' T^ ■? '^ I I I 'O CO ^ I L ' o o a; lO -M ^ GO f-i to O-- CO 'M ■— I « CI O "I '^ "^O *^ _„H,— I — I— 1r-,P— 1 'l-^'-^ O CD CS -^ CO lO CO .£P S S K -o ^^ ^, o 01 J?! CO « CO ?C CO CO CO CO CO CO CO CO CO do 'X CO GO do do CO "Si "o o I ■C' r^ '>D '^o iXi >»«, ^H ' CO ^^ ^^ ^ I Ci --I O 1 ^^ O "--( '^ I li? --•- i I t- i >:. -•, "' I IS I g I I fC -+ I -+ ' I I " U^ I Ofi I, 3 I I ^ '-C ^ [ --i < ^1 '^f "-S '10 x ';» . Tl O O 'M O ilC 00 "^ GO '■o -f , I -I* cs rM o r- « ;?; 2 I I ^ t- -= IS rSi?: I I"* ., IS r |?^3 I I I "-" "'-* II 122 I I I Ci ? ^ ?>. •-, c^ '^ I I C: C^ ^. I -o '^ c: r- cc' •+ —- r- ■■•c c^ Ol -^ CO 'o CO -+ o OS -^ o O ^i o Cl 6i , f>-t Ol .■■■C CO o > t-- r<[ I o I 'i- r- (^ I -f 1 C^ I CO -t I I U ST;! I ii;:;! M OS O O s ii ;3 « g'Sa .« £f E*-S a 2 g g I ; J ^j M H ►-! Q -J3 CO ■ .t: J,- r ® ■S 2 is S <^ i^ -S £, o ■» 0^ S ^ "5 •fl — ^H ^< a -^ t- ^ 9 0) OS o =S 01 ^1 U h-1 h4 M H M S? 03 ■:a > £ 3 o S O en 390 (ji-oirth and Development of Breedx and Croftsea of Pigx T;il)lc 111. Comparatirc wrifjIitK of different breeds i)f piijs -Iba. Pen of 2 classes Single pig A classes Ago in months 3 r> 7 9 11 f ^ 9 11 Lincolnshire Curly Coatod 7J _ 341 404 510 465 483 Large Wliite ,S(i 107 — 390 481 421 490 Large Black — — 305 371 470 — 470 Somersetshire — 13S — 340 430 — 436 Tamworth — — 2.97 334 414 34!) 404 Berkshire... Sf) uu 21H 337 402 327 404 Middle White 77 118 210 325 384 — 385 SmaU Black — — 37!) 37/i Dorset — — — — — 23S 349 Small Wliitc — — 151 216 273 — — In Table I it will be seen that there is variation in the average age at which the difTerent breeds were exhibited so that the figures arc not strictly comparable. Corrections have therefore been made for this from the rate of growth shown in Table II and the comparative weights of breeds at the common ages of 3, 5, 7, 9 and 11 months are given in Table III and it is to this table that reference is made below. The breeds in Table HI are arranged in order of tlieir weight at 1 1 months old, the mature size of the breed ; the differences in the relative order at the other ages are due to variations in the rate of maturity and will be discussed below. The Lincolnshire Curly Coated (averaging 510 lbs.) is the heaviest breed with the Large Wliite next some 30 lbs. fighter while the Large Black follows closely weighing some 10 lbs. less. There is then a rather large drop of 50 lbs. to the Tamworth. with the Berkshire some 10 lbs. below and the Middle White about 20 lbs. fighter still. There is then a very big gap of some 1 10 lbs. to the now almost extinct Small White. A few animals described as the Somersetshire breed (probably Gloucester Old Spots) come midway between the Large Black and the Tamworth in size ; while a few called Small Blacks and Dorsets are sfightly fighter than the Middle Wliites. Much work on the comparative slaughter weight of pigs has been done by Hofman-Bang, Morkeberg and Lund (27) in Denmark and has been mainly directed to testing two breeds, the Large White and Native Danish and their crosses. These tests, which give data of gain in five weight per unit of food (consumed up to the age of approximately 200 days and 200 lbs. live weight, conclude with carcase weights at that age only and show consistent results in that the carcase weight of the Large White & John Hammond 397 is about 2 per cent, greater thau tliat of tlie Native Danish. The investiga- tion which was practical in nature was however limited to pigs of this age, so does not show differences due to age. Seniinler(2s) found that the carcase weights of Berkshires were on the average higher than those of German native breeds of the same type. Henry (29) quotes from Cuvier that the intestine length of the wild boar is 9 times the body length whereas that of the domestic boar is 13-5 times and he found himself that in fat hogs it was 21 times. No weights are given, however, and it does not follow that in<-reased length means in- creased weight; if the diameter were decreased the absorptive area would still be greater and the weight of contained foodstufi's the same or less. Henseler's(i2) data shows that a starved pig of 9 months old had an intestine length 18-1 times the body length, whereas in a well-fed pig of the same age and from the same litter the length was only 16-9 times; moreover, the intestine was 4-95 per cent, of the hve weight in the starved animal, whereas it was only 2-8 per cent, in the well-fed. The average weights of pigs killed in the carcase competitions are shown in Table IV which also gives tlie numbers exhibited and their average age. These figures have been translated into lbs. per week in- crease which are shown in Table V. The birth weight has been neglected as the weights of the different breeds at birth were not known. Owing to differences in age the weights of the different breeds given in Tal>le IV are not strictly comparable. Table VI has therefore been prepared to show all breeds calculated to common ages. If the live weights of animals exhibited in the "Carcase Classes" given in Table VI are compared with those shown in Table III for the "Live Classes" of the same age it will be seen that, as has been pointed out by Long (30), the live class animals are much heavier. With Berk- sliires the difference at 3 months is 12 lbs., while at 11 months it is 154 lbs. and again with Large Whites at 3 months the difference is 19 lbs., while at 11 months it is 156 lbs. The reason for this is probably that the "Live Class" animals are more heavily fattened. Evidence, which it is hoped to pubhsh shortly, has been accumulated on this point in sheep and it shows that the difference in weight and composition of the body of animals from the Live and Carcase Classes is one of fat alone. It would seem probable that this result would also apply to pigs and that the difference in weight between the two classes is due to extra fat in the Live Classes, wliich from the butchers" point of view is superfluous. It is worthy of note that the Tamworth, a breed which does not put on superfluous fat easily, has only a difference of 45 lbs. between "Carcase" 398 Groivth and Development of Breerh and Crosses oj Plys JOJ *o i-r; <>i OS #-* Ci 03 "^ -^ CO paiunoooimfj 'm « e^ co -^ •^ CO 00 C-l iM JfoniJ i ^, r>, - th »:' c ^ cs -M os«o.i«^ ii ic =^ 6 t^ ^ — ^ — ^i *>* 71 Ti 'M CC <^ CO fO O 'f5 OAIT ^ t^ ob -^ »c ■ ^ f^ — -^ ir:, ao ^l M "JJ 'M « O -M M C-i — C-) , ■'. -^ — -^ — — O e^ ® CO 00 ;o JOJ Ci CS N CC "O I o O 03 CO -f O <— > V5 ■^ o « t- "O ^-4 (M 1— ( -H »-, O 03 O lO »0 C^ 7J n 71 ^i J0| ^ 'v %' ^ ^ I p^unoooBUQ ^ 55 Js 15 ^ ' r- UD 71 ?D oc I >f.^nij 00 6 6 00 oc I CO X CO — ^ CO 7^ t-- CO Co .101 ^ 9 T' r '■? t-- 71 O CO 'f*N ponnu)Oi>i}!ij-| — oi7)i=.7", •"* l^ "* CO 'r% C5 X ^H t^ ;a oSBOJit' ) o li db 10 »o I'- Oi r- t>- f^ CO 10 1^ ^ » OAKf O) ^< S, 10 •* to l^ CI X ^ •- S J<: -=•3 " ■&5-S .a 4> ~ s p Si e s £.3 pq ■3 t; (. 1- a H g ^ O CC CC W PO <; cjob obcoob o o t^ ci 35 'M 0) 0 CO CO — -"^ s < r^ :c -c 1^ t^ O 05 :D ':0 71 CO *5; I* 1-0 -t Tf 'v^ O t^ O 00 00 •** ;^ ^ ITS *t" ^ CO C^ CO o c^05 cvj r- : : : d 0 . 0 ; Ji 3 0 ■"ij ^bm : ■ » eWh Whi Blac re . rth shii ;= 0 c ^ '^ ^ ■« if *c 2 ^ »-^ c « -0 »- t- ^ " a ::: rt aj 0 a .3 "►J -J 03Hm-I John Hajimond 399 roT o o CO Ci , a ^ ^on\^ ■^^'M CO CO ^ -i 6 o C' i s < 1 <"<. C; 'O '-^ t^ , osKoj^,") "* 9 -r r^ ^^ 1 O ^ --^ o t^ '^ -+ -t* QO CO '^ OAiq -^ -^ CO CO C: ;r lo i> 00 "-T ' ..o}55Sg-3 j po^unooouir^ :^ o -^ i -- J2 — ^ o; C^ '^ fO c o ' "^ o o o « o ' a r^' ffi- o -. i~ c:^ Tf 9SB0JT3,") '9 ^ ^^ "9 "T O lO O ^h lO ■X> ^ ^ ic CO CO ^■J c^ , 8AIT «> ■* "^ <:- tr ^ ' t^ CD l> '-■^ O ?=S CO JOI O GC X Ot J: pdJUllODOUlli-l 5 o — '^ -^ r-O ^-^ « *«* CO — O O =^ QO 1^ o Monij ^ c? ^ ^ ^ ' "1 o o o o o "^ s a^ £: t-- t^ to =c o fe5 1 '-^ Lo uo ;o -* CO ^ 1 ^ c: o CO t^ CO , «* oATT =p '^> =:' =? tr f CO CO X ^O I> JOT '-':) t_o o ^^ CO , CO Co IwjunoooiuiQ o 5 ^ .i, ^ 1 CO 2 Ti4 00 CO 0( tJ< , o 5 ■vjaiXT T — - '-J a: 'o — ' o dSDOJB;-) tr 9 T '^ T* ^ CO -f -»• CO -t e ■^ O O " Ol 'o — , aAiT 'r T V '■■^ 9 > -it t-O CO u^ LO -2 JOI C^ LO Ol , CO Co ■""l cr; =1 Ol -« cs 3 po)unoooviir-l q, ^ ^ ' -^ 'ii cS H -i= ^ CO -+ , CTl '^ s Honrj '7'?'^ \ '^'? o "^ cs O O ' c» cs a- 7 Ol 0> O , s>, -O asBOj-e,^ -o CO '^l CO CO Ji CO -* Tji ' -i, ko Co o: CD , ^ CO GAIT ? O T 'r -.' • -^ ^+j .-'-1 )'-\ :ry 1-^ o o CO CO 6 I CO o as JOl V . T -7" " ^ -^ O — ' c ( pjJUIlOJOUU^^ CO CO -* •+ : ■» O— CI ■ c o J a > CO Ol Ci Co dKLM.IU, ) >-^ CD - '-' "^ O I Ov[ r^l c •~H Tt^ t- 00 '■^ ot cr4 .10} Ci Tl CO c5 CO CO CO -)• ^ 00 c Ol CD c CO CO c OS CO c CD -^ c CO CO fi t^ Co C Obl.JJl.J CO CC — CO Ol Ol — ' Ol »-H _. : cr- — 01 cc, . 1 -^ O Cc CO 00 CO -CD ^ -tH 01 Ol Ol Ol 01 -'An i .«ft .loj ^ '^ r" ^ "^ ' pajunoooBun o'^ o! " ^ ?? ^•'"M CO cc o CO 6 'X O ^ UO -i< 1t;i' 1 II' "^ CO Ol CD Ko Ol J!>UJJl,J ,^ -* CO o> ^ X -T* -^ ^ -+ oArr o'l ra o "^i CD ■ ^ C5 t^ CO CO -^ ^ — H Ol -N OI JO} C5 ■T' O "-O CO pdjimuooBun 2 2 Jl S5 ?! I 'CL Tf O CD CO CO CO 'O CO ^ Ol dSBOJBJ ..•, o 10 lA o'l t^ GO CC i-X 'X _ Ol O C 01 ■'-Ml s ?^ ?; 55 i JOJ =C 9 Ol ^ ^ ^ pojtmoDO'BUQ ^ rt 21 ' £; "^1 " -*«-.' , i, o, _ asEojB,-) -:h CO o; I ,> < .-, - •, O -+ ^' ^ ^ '■^ T^ '?' I cc c^ Ol 1} o •= -i^ -^ £ 73 t- 1- a t- r-i o ;3 O cc c« ce .S O .tJ iJ .i! O J3 -Ch ^ CJ ~ CQ ;2^ boa ^J^o ^ ;^ S (S c3 o3 .S O 400 Growth and Development of Breeds and Crosses of Plcjs and " Live " Classes at 1 1 months old as compared with 154 lbs. and 156 lbs. for Berkshire and Large Whites respectively. At ') months old tlic Live Classes are only 16 per cent, heavier than the Carcase Clas.ses, yet at 11 months old this difference is increased to 62 per cent., the young animal being much more diflicult to overfatten than the old one. The carcase percentage for the different breeds is shown in Table VII ; the variations at different ages are discussed below (see Age) as also are the relative orders of the breeds at the younger ages (see Early Maturity). The Middle White has the highest carcase percentage and is followed by the Berkshire, Large White, Tamworth and Large Black in the order named. The proportion of pluck is in inverse proportion to the carcase percentage and so the order of breeds as regards this part is reversed. The carcase percentages are on the whole less than those of pigs exhibited at the International Live Stock Show at Chicago (3i) which vary from 83 to 88 per cent, but there a different type of pig — the lard pig — is required. Table VII. Proporlioii.s of or(jans in different bree(U of pigs — as percentage of the live weight. 3 months 5 month.s 7 months !t rnoiUhs 11 months t3 T3 Ts -a -0 ^ ^ <^ CJ 9 S .5 c S 3 ^ c 9 a Breed Q S 5^ O S &£ O S t)>2 O S D£ O S D^ Aliddle White ... 74-3 6-0 19-7 76-8 5-2 180 82-4 4-5 131 840 40 12-0 85-2 30 10-9 Berkshire ... 770 5-5 175 78-7 S-S lo-S 811 4-7 14-2 82-5 4-5 130 83-1 4-4 12-5 Large White ... 730 5-9 211 76-9 5-4 17-7 80-9 4-7 14-4 81-3 4-5 14-2 83-5 40 12-5 Tamworth ...— — — 70-1 6-1 23-H 76-3 5-2 18-5 80-4 4-7 14-9 84-8 4-0 11-2 Large Bhiek ... 72-9 5-8 21-3 73-9 GO 201 79-7 4-8 ir,-5 80-3 4-8 14-9 SO-7 51 14-2 Lincohishire Curly SO-9 4-3 14-8 — — — ^ __ ____ ___ Coated Average (4 breeds) 74-3 5-8 199 70(i 5 5 17-9 810 4-7 14-3 820 4-4 13(> 831 4-3 12i; Age. The effect of age on the rate of growth iu live weiglit in the different breeds, calculated as increase in lbs. per day since birth, is seen in Table II and V. The rates of growth of two breeds (Berkshire and Middle White) of which sufficient numbers have been available at all ages, have been averaged and are shown in Fig. 1 . The rate of growth in these two breeds falls from 3 to 5 months and then rises till it reaches a maximum at 9 months old when it falls again. The reason for the first drop between 3 and 5 months is to a very small extent caused by neglecting the birth weight which would naturally play John Hammond 401 a greater part m pii;i;s of smaller weight; this is not the main reason however which is probably due to the fact that the 3 mouths old pigs have not been weaned early but forced through, whereas the 5 months ones have often been weaned early and some may possibly have had a short store period. lOr » 7 - Fig. 1 5 7 9. n Jlonths old Rate of Growth — Live Weight — lbs. per week since birth. Average of i! breeds (Berkshire and Middle White). The maximum rate of growth in these two breeds appear to be made at 9 months old but if reference is made to Table II for the Live Classes and Table V for the Carcase Classes it will be seen that breeds vary greatly at which the maximum rate is attained: the Large White and Large Black attaining their maximum at 7 months. It is notable that although the pork type of pig (Middle White, Berkshire) appears to reach the maximum growth later than the bacon type (Large White and Large Black) yet the decline in rate of growth of the latter is not so great between the ages of 9-11 months as it is in the former. Since Uve weight growth is the sum of the growth of various organs and tissues in the body it does not necessarily follow that the rate of increase of meat is most at this time. The differences in rate of increase between the Carcase and Live Classes (shown in Fig. 1) which become greater as the pigs get older lead one to believe that the growth at this stage (9 montks) consists mainly of fat. 402 Gi'oirth and Dcrelopment of Breeds and Crosses of Pigs Till' variations in rate of growth in different breeds have yet to be analysed in terms of the components of the body ; the various systems of the body reach thoir maximum rate of growth at different stages and thus breed variations in the rate of growth may be explained. The figures given above for rate of growth may be criticized however because they show the rate from birth to that particular age ; thus they give the total result of growth to tlie particular age and not the rate at which growth is proceeding at the time. The rate of growth between certain ages has been calculated in another way — by subtraisting the weight at x months from the weight at // months and the age at x months from the age at // months and from this finding the rate of growth per week. This has been done for four breeds and the results are sjiown in Fig. 2 and in Table Vlll. From these it will be seen that, as before, the rate of growth falls between 3 and 5 months but attains its maximum (of approximately 1"2 lbs. per week) between 5 and 7 months and then falls again until between '.) and 11 months it is only about 3 lbs. per week. Data collected from various authorities for the growth in live weight of the pig are given in the following table : Groiolh in lbs. per week gain. Months old ... 1 2 3 4 5 G 7 Biicklcy(32) ... HarpciMiS) 0.stertag iintl Ziintz(35) 2-8 — — — — — — — — — — — The economy of gain in live weight however cannot altogether be measured by the rate of increase per week. J 1 enry and Morrison (36) found from data collected from the U.S. Experimental Stations that pigs of 78 lbs. (corresponding to 3 months old) put on 100 lbs. live weight from 400 lbs. food whereas pigs of 320 lbs. (corresponding to 9 months old) put on 100 lbs. live weight from .535 lbs. food; the composition of this gain in live weight as shown in Table Vlll however is very different. Morgan (37) has pointed out that man is larger than the rabbit because he grows for a longer period but the daily increases in weight are nearly the same. On the other hand, rabbits attain a larger size than guinea- pigs because they grow faster and not for a longer time. Meek (38) too has shown that although a donkey and a Large White pig attain much the same ultimate weight there is a great difference in the form of their growth curve. When one couipares the rate of growth of sheep (Murray (39)) with pigs it will be seen that the maximum rate of daily increase in John Hammond 40? weight occurs in the 2nd month in sheep, but not until about the 7th month in pigs. What factor decides the point of maximum daily growth is uncertain; it may be that the maximum comes where the falling curve of the velocity of growth (as measured in lbs. per 100 gained) cuts the ascending curve of increasing body weight^ — Robertson's (40) auto- catylic theory — or that the growth curve is in reality a combined curve 12 10 C3 2 - - Z./i'e - Carcase -Alimentary Canal ■Pluck 7to9 gtoll /^ye-Birth to3 3to5 5to7 months Fig. 2. Rate of Growth — lbs. per week — Carcass Classes. Weight at birth neglected. (Av. of 4 breeds.) Table VIII. Rates of growth and carcase percentages of pigs of the Carcase Class. Rate of growth — lbs. per week Percentage of the live weight Birtli Birth \ Age in months to 3 3—5 5—7 7—9 9—11 to 3 3—5 5—7 7—9 9—11 Live weight 5-84 4-74 11-96 6-35 3-07 — — — — — Carcase 4-34 3-82 10-27 5-50 2-90 74-3 SO-6 85-9 86-6 94-5 Pluck 0-34 0-24 0-45 0-21 010 5-8 51 3-8 33 3-2 Intestines, etc. 116 0-68 1-24 0-64 0-07 19-9 14-3 10-3 10-1 2-3 of the growth of different organs, each of which reaches its maximum at a different time; for example, the growth of lambs during the second month of life is above normal owing to the development of the rumen and its contents. Such an explanation would account for the differences 1 Thi.s depending on the relation hctween the weiglit of the individual young at birth to the weight of the mother which naturally falls as the number of young per birth increases. Journ. of Agric. Sci xn 27 404 Growth uml Development of Breed h and Crosses of Pigs in form of the growth curves of the pig and the donkey and these differ- ences are an exaggeration of what early maturity means in animals. In the guinea-pig Read (ii ) found that during uterine development cycles of growth took place and the same thing occurs in inan when growth is accelerated about the age of puberty (Robertson (42)). These facts indicate that it is impossible, in more than a very general sort of way, to construct growth formulae to fit the curve of growth of any animal. Carcase. Pluck. Intestines 84 6-3 22 Or 82 58 200i^ 80 5-3 180 78 4-8 160 76 4-3 140 74 3-8 120 3 5 7 9 11 months Fig. 3. Proportions of parts — pigs (4 breeds). Per cent, of live weight. The proportional development of the carcase, pluck and alimentary canal, etc. at the different ages is shown by the average of four breeds at the end of Table VII and is also given in diagrammatic form in Fig. 3. The carcase percentage shows a steady rise from 74-5 per cent, at 3 months old to 83 per cent, at 1 1 months old ; the greatest rate of rise occurs between 5 and 7 months and afterwards slows down somewhat. With this rise in carcase percentage the proportions of pluck and alimentary canal etc. fall rapidly between 3 and 7 months and after- wards more slowly; the alimentary canal etc. ("unaccounted for") on the whole falling more than the pluck, which is to be expected since the latter includes the caul fat. Semmler(43) found in pigs that the lungs of young animals are relatively larger than those of old pigs of the same breed. The actual rate of growth of the carcase, pluck and alimentary canal in lbs. per week is shown in Fig. 2 and also in Table VIII. From this John Hamiviond 405 table it will be seen that 100 lbs. of growth in live weight from birth to 3 months consists of 74-3 lbs. carcase, 5-8 lbs. pluck and 19-9 lbs. alimentary canal, etc. whereas the same weight put on between 9 and 11 months consists of 94-5 lbs. carcase, 3-2 lbs. pluck and only 2-3 lbs. aUmentary canal, etc. This difierence in the relative composition of the increase between animals of young and old ages should not be allowed to detract from the relative economy of meat production at early ages in general, but points to the conclusion that where ywfing animals are being killed for meat it is very essential to use an early maturing breed in which the age changes are hastened. Data have been collected from various authorities on the Carcase percentage of pigs of different live weights and are given in the table below : Live weifjht — lbs. Birth— 50 50—100 100—200 200—300 300—400 400—500 Robison<44) — 77-7 S3-4 870 88-2 88-2 Coburn(45) — 72 74 78 80 87 Smith(46) 74-5 — 70 79 SO 81 Tomhave<47) — — 80 84 — — It will be seen that there is a considerable divergence of opinion as to the dressed carcase percentage at different weights, which may be ac- counted for by differences in breeds and modes of dressing, but all agree that it increases with the rise in live weight. Whether this increase is a function of the age or of the hve weight should be made clear. Henseler(48) has shown that it is due to the latter; he took two pigs from the same Htter — one was fed well and the other starved, both being killed at the same age, but at very different live weights. The well-fed pig gave a carcase percentage of 87 per cent, and the starved one only 73 per cent. Under ' Correlation ' (below) it is shown that the carcase percentage within any one breed increases with size independent of age. That the carcase percentage will increase with age if the size is kept constant (by controlled feeding) would appear doubtful, since increased carcase percentage is brought about by the growth of muscles, fat and bones which attain their maximum rate of development late in life. When variations due to conformation and differential rates of growth of tissues in the different breeds are considered however it will be seen from Table VII that the carcase percentage may be independent of size : for example. Large Blacks of 5 months old weigh some 20 lbs. heavier than Middle Whites, but while the former have a carcase percentage of 74 per cent, the latter kill at 77 per cent. 27—2 40(5 Orojrf/i and DiKclopiiK nl of Breeds and Crosses of Pigs Earhj Maturiti/. In order to obtain an estimate of the rate of maturity of an animal it is necessary to consider three factors. Jn the first place maturity so far as live weight is concerned is the rate at which an animal attains its mature weight. This factor is seen in Table IX for the various breeds, the weights at different ages being shown as a percentage of the weight at 1 1 months. This table shows that the Middle White is the most early maturing breed with the Berkshire following closely behind it. Figures for the other breeds are somewhat fragmentary but indicate that the Large White and Tamworth come Axt while the Lincolnshire Curly Coated and Large Black are last in this respect. The second factor by which the rate of maturity may be measured is that of the carcase percentage at different ages. A breed which matures early forms bone, muscle, and fat early, and so the proportion of the t^arcase rises, that is, the age changes of low to high carcase percentage .uo hastened. Miiller(io) believes that the changes causing early maturity are controlled by the glands of internal secretion. Fig. 4 shows the carcase percentages of the breeds af different ages (see also Table VII). At 3 and 5 months old the Berk.shire has the highest carcase percentage followed by the Middle Wiite, Large White and Large Black in the order named. At 7 and 9 months the Middle AMiite overtakes the Berk- shire, but the other breeds remain in the same relative order, although they tend more nearly to approach one another. Another method of estimating the rate of maturity of different parts of the body is shown in Table X; the weights of parts at dift'erent ages are shown as a percentage of the weights of the parts at 11 months old. Considering the averages of the four breeds, whereas at 3 months old the alimentary canal etc. has made 40 per cent, of its ultimate growth the carcase has made only 23 per cent., the pluck 34 per cent, and the body as a whole 27 per cent. The two breeds in which the figures are significant (Berkshire and Large White) stand in the same relative order as given above. If the Live Weight figures in this table for "Carcase Classes" are compared with those given in Table IX for the "Live Classes" it will be seen that the ratio is higher in the Carcase Classes ; the reason has been pointed out above, namely, that the Carcase Class animals e.xhibited at the older ages are not so fat (and heavy) as those of the Live Classes and so the early maturity ratio is liighcr. For the same reason the ratio for Tamworths at 9 months in Table IX is probably rather high on account of the relatively small amount of fat this breed puts on at the older a--i 3 -^ p^ -i ^ »-, 'O , I GO "O-tGO'+'^'^'^'^^^O GO 1^1 t ^ "^ 1 I \Z \ \2t\ \ I i j= ^ T^ ffl 02 o 0-- I I I I « CO -fH "O I I I I !>„ C» P ^ ^ I I I MM ?: I I I I 6i Oi ^1 &0 X 3 2 ^ .a 5 f5 -a :g Si 2 5 ii S> > > ii j= 73 -S .t; « 'o 2 = - ■^ .« -"-^ -5 •- 112 hJ rt c6 ^ ^ 3s X c ^ X XhJ^ OJ ^ _-; M tJO-i*i c: -^ o til's MS C c3 ® m-1 ^ Cj ■ - „ O " — > — ^ — :3 .ti •■'■ CLj 1^ 0:1 Oj 0:1 p t-> ?*2 P^' ?*-' ?^ S c <» OCO:f-a:QCJ^Qj« O ifS Ol «^ >o ^ Ci cjoidii I |« |A I'-' I C5 '^ ao c-1 t^ -o '^ '^ '^ oca-— 'dsob'i'ioi^ S I lis I I I I i| I 1 1 1 1 1 i:is I s^s:: 1 1 1 1 igii I o .3 S ■*A IM 0) O "3*3*^ ^ ja -fi •« -a >-£^^ ■a 'o 60 £Q aj 4) & : 2 X • a X : ^ .t g .t) oO 3 05 Ci X J* 2 -^ ^'-^ U "^ tH t. t- 4; ^ 4j 03 aj MSCQhJM : : 3 X *^ - 0) .5 « J3 '^ -^ -a fc- b =" i^ o ^ c> fe to *0 a; a t. (H ^ HMhJJ .■§. _c 'o "2 ^ tX' a) o 03 S- if ►J ■« « rt-i X X 133 3 5 '^ -S — a) ■ - • a aJ 2 ■S X " J X -a -3 > t- _a a; =e « '^ tc d o rti — © cj i-j H « I c^ I r^ , CO , ^ Ci CO l^ll III M I I M ■* ^ -- t- I I CO c<» « -t I I , r-. <^ t^ *+< f^ oi ;c "^ ' PO *5< ^ >c (?i o LO ^ »o i^ "^ti c^i ^ CO '^ I I '^^^ 1 I I I I *^ CO CO O ?^ I t C-l CO CO <3i I I I IISI I I ■^ eo I I I I I I I l§ I I ■ r-j ^ . . ? rn ■ © * • ~ „ ID n O u c^ tH 0^ CI .g bo •O X ;^ X ,t^ X ^ QP -c ^ 73 7^ gi qj M g ^ "O -^ tjj^ d ^ "d t4 t4 ui » £ S ^ ^ ^ - -r .- •= - ^ ^ O O g (p c3 rt '^ M M t- — I © ri Ci 03 ™ e ^ '^ ^H M fcn H >-fi >i "^ r>n CO a)--;»sajta».3ie--:cecec8s £, "S 5 tT " _ — . . , HlMoQi^fi 414 Growth and Devdopmetit of Breeds and Crosses of Pigs Wellmann (65) has published results of comparative carcase tests made with the Lincoln and Manpalicza breeds and the crosses between them. He found that the live weight and carcase percentage of the cross-breds at 14 montLs was higher than that of Mangalicza's at 26 months old. Table XIV. Rate of maturity in live weight of cross-bred pigs — as percentage of weight at 11 months old. Pen of 2 classes Single pig Age in months 3 5 7 ^ 9 9 : Large White Large White X tJerkshire 25 30 — 92 89 21 22 — 79 78 i Large White X Large Blaek ... — 23 — — — 1 Large Black X Large White ... — 22 61 85 — Lincolnshire Curly Coated x Berkshire — — — 73 — Middle White x Large White 26 — — — — Selection. In order to study the progress made by breeders in altering the weights of pigs the records of the Show have been grouped in periods — the first froTn 1901 to 1906 both inclusive and the second from 1907 to 1913 both inclusive. The differences between the average weights of the two periods (Table XVII) would show the progress made. Whenever the numbers exhibited are sufficient to be reliable they show an increase in weight. In one or two cases, as at the younger ages of 5 and 7 months, there has been a decrease but the numbers from which the results have been calculated are small. The Large Black has shown the most notable increase in weight (Fig. 5) and it will be seen that the distribution curve has been shifted forward. This breed at 11 months old has made an average increase of 70 lbs. between the two periods; the Large White follows next with a rise of 50 lbs., the Middle Wliite with 30 lbs. increase and lastly the Berkshire with about 5 lbs. increase. Whether this effect is due to further fattening or is actual growth in meat it is difficult to say as no figures are available to give an estimate of the state of fatness of the pigs shown. However, as the increase is only at 9 and 11 months and not at 5 or 7 months there is a strong indication that it is due to state of fatness. In the Carcase Classes the Berkshire is the only breed which has been exhibited in sufficient numbers to make a comparison between the two periods po.ssible and the data for this breed is shown in Table XVIII. At all ages, with the exception of 9 months old there has been a decrease in the carcase percentage and an increase in the proportion of the John Hammond 415 33 '^' 1 1 >>? C ■-3 ^ ?3 ' s Z\ 1 f^'? 1 -o -c I I I I i A I I li':; I I I I I I I 1^ IS i 1^ I L s ^ I 3 i ^. I <5>Q! ■^ I |..>. 1 *= 'M ^n 1 ^' "O ij^ ^ MSa I I I I 05 • CO ^ I ii Ui I I •Bo ? ^ 'V: -s i* =«D fO 'iO *>» ^- Ci c^,. ^ i- ** Ci ^ '"-< Ol >-H c-i J cr^ I •-- i>, -^ »-K '"' r>t '^l 7j c^( '>i I I I I ) -rt c; i;* 'y I *? "? . ^i --^ '^'> I --^ I "M '^ »-l r-H '^ "^ '"H '^ ■^ f-t ^ 01 0^ K Berkshir ddle Whit Berkshire rge White erkshire inworth Berkshire rge Black Large Bla Large Wh liddle Wh , QO a o _-: ^ -^ ■ r- ^ O -< -t; O 1' i j;>.(ii-^ a>*S(U rfo)'^ rf.^ a 9 2 jji ODM a ^ bj:^ bd ao 0) bo ^ -111 boJai bC) bO q;, f_c Ih Eh t-i t- hrj '^S P. O) CO tf u t~> a 416 Growth and Developrtient of Breeds and Crosses of Figs 6C > :=3 OS a a » ^ o Oiks . 0» us *H 00 ^ Qo J^ I I I I o-i O •-S I (^ "^ 05 C^ ig oo «5«^"-i .°p I'je^ Qoco'oc«ooo=^ =o «o«c 1 00 - 0) 4) c "^ 1 1 •?=? 1 1 'T^ .s» d -w ^ -^ «5 1 1 ^ ^ C3 ■^ © •2, 1 ' — 1 .,. o 3 =? 1 1 CO "^ 1 1 «"* "? &4 i 1 til m <=. , oooo 1 C>» . . ^ «o . •? I ^ 'W lO -w 1 1 >o i •■C 1 M. s e J3 S COK5 1 1 "? © e» Ml Co , »^ sC => i fc =< O o t;?; ' 00 oo <^ oo e £^ ^ o 03 0) O ^^ ) . 2 1 1 C *^ ^ =o I^ '^ 1 1 s C ._S c -s ' -.C^l « ^ ^ ^ CC V 3 =? 1 1 'r-T 1 i>. U5 1 21 1 ■^ s '^ t^ s ? 1 •^ "^ 1 "^ 1 1 ^ r1 S t^ 1 >* rC. 1 tc CO (>. t^ ^c t^ O •- CO 0* 4- ^ 1^ l^^l X 1) d o H X X t6 X XX--. a ^ X .tic o 3 OS a 'o -S £ £ X 4> 4J -* H ■~ fl Oi DO ^ » -!> n 0} 2^ iP-^a ^ tiC-M 60 Sf) 4j bO T3 fc- U. I-. *- ^ Sr J'Ja John Hammond 417 alimentary canal. It is difficult to give an explanation of this fact unless it be due to changes in the systems of management and amount of exercise given. The same thing has occurred in sheep (66). 1 s 1 9 months old „ f __^__-T— ' 1 1 1 ' r 1 — -t- 200 240 280 320 360 400 Weight in Ihs 440 480 520 11 months old 290 330 370 410 450 Weight in Ibt 490 530 570 J^ 610 Fig. 5. Distribution curves of live weight in Large Black pigs. Light curve— 1901-6. Shaded curve— 1907-13. Table XVII. Changes in live weights of pigs from Period I (1901-6) to Period II (1907-13). Weight in lbs. Period 3 months 5 months 7 months 9 months 11 mon Middle White I II 77 125 109 236 195 324 327 369 398 Large White I II S6 161 105 z 359 395 447 499 Large Black I II — — 292 354 350 399 427 497 Berkshire I II 85 US 107 253 195 333 342 399 405 Numbers of pigs from which results calculated. Middle White I II 10 2 60 6 14 76 68 62 66 Large White I II 4 4 34 — 38 70 50 64 Large Black 1 II — z 20 8 34 44 56 66 Berkshire I II 16 124 10 22 128 118 120 148 418 Grrowth and Development of Breeds and Crosses ofPif/s Table XVIII. Changes in the proportional development of Berkshire ■pigs from Period I (1903-6) to Period II (1907-13). Percentage of live weight 3 munths 5 months 7 months 9 months 1 1 months 5 «■ « x" ^ 0) a> a^ V ■§ 1 J£ a '■3 1 ^ .3 0, s .a •J 1 ^ X M •1 •c Z ii S ^6 E 3 5§ g 3 5d ^ o 3 ■2^ «■ e 3 ^ d £ 6 04 c-2 S s s a- d s c5 3 111 (-1 o I 67-4 6:5 26- 1 78-2 5-3 16-5 82- 1 4-7 132 82-7 4-5 12-8 83-7 4-5 U-8 u 77-3 5-3 17-4 770 5-5 16-9 80-7 4-8 14-0 83- 1 4-5 12-4 82-8 4-3 12-9 Numbers from which calculated I 7 23 9 18 11 IT 20 3.5 23 51 27 Individual variation. All the figures given in this paper are averages of a number of pigs and there is a large amount of individual variation in weight apart from those due to causes which are known. Such things as strain, state of fatness, etc., influence the hve weight and all these come under the heading of individual variation. One of the most im- portant causes of variation is the amount of food contained in the aHmentary canal ; while it is not so important as in cattle and sheep where the contained food may be as much as 14 per cent, of the live weight yet it is a large factor. Table B shows that it may vary from 2 to 11 per cent, of the live weight ; while not all of this can be eliminated yet it is sufficient to cause considerable variation. As a measure of this variation in the various breeds and at different ages the standard deviation has been calculated and is shown in Table XIX; the coefficient of variability has been used as the measure of variation. The Large Black breed shows greater variation than the others and this is probably in part due to the increases in weight which have been made in this breed in the last few years (see Selection above). It is notable that the first crosses show rather less variation than the pure breeds. On the average of all breeds the coefficient of variability increases with age up to 7 months and then decreases again, variability being greatest when the rate of growth is at its maximum. This will be seen if the coefficients of variability at the base of Table XIX are compared with the rates of growth at the different ages given in Fig. 2. There is, however, more variation at the younger ages than at the older ones, as might be expected if improvements both by feeding and breeding were being made in the rate of early maturity. That variability is greatest when the rate of increase is at its maximum was first pointed out by King (67) for rats. Robertson's (68) results also show it to be true for mice John Hajimond 419 and men. Drununoud cl al. {*<■<) have found tliat the growth of the pig is affected by lack of vitaniinos in the diet while Elliot, Crichton and Orr(70) have shown that the more rapid growing pigs are the quickest to be affected by rickets. Table XIX. Variation in live weiglit of pigs (pens Standard deviation, lbs. oi2). Ago ill months Middle White Large White Large Blacli Berkshire Middle White ■ Berkshire Berkshire x Middle White Middle White Large Wliite Large Black Berkshire Middle Wliite Berkshire Berkshire x Middle White Average (5 breeds) ... 7-ti ll(i 7-3 U-2 17-5 •2ir, 13-2 lti-() 7 .33-7 49". 70-.5 S7-G Coefficient of variabilitji. •1(»7 -148 •USS -210 •141 •OW) •14'J ■113 •141 •()!(!• •130 •140 •25(i •1U2 •351 •3S7 •331 9 354 47'5 .5S^1 32-0 409 25-4 •109 •122 •157 •095 ■115 •litis •102 11 431 57^4 fi7^2 433 289 42-2 •112 ■119 •143 •108 ■069 ■097 ■101 Table XX. Variation in ■proportional development of Berkshire pigs. Age in mouths ... 3 5 Carcase Pluck Intestines, etc. Carcase Pluck Intestines, etc. 2^70 ■035 0^80 ■148 7 2-70 ■151 3 13 ■040 (•■68 •124 9 305 ■178 Carcase Pluck Intestines, etc. Carcase Pluck Intestines, etc. 339 ■042 051 •106 11 A 304 ■208 2^08 ■025 0^39 ■087 204 •155 Carcase Pluck Intestines, etc. 2^09 ■025 0^45 ■102 1-98 ■155 Standard deviation — lbs. Coefficient of variability Age in months Standard deviation — lbs. Coefficient of variability Age in months Standard deviation — lbs. Coefficient of variability The standard deviation and coefficient of variability of the pro- portional development of the parts of the body in the Berkshire breed at different ages are shown in Table XX. The carcase percentage like the live weight increases in variabihty up to 7 months and then decreases again, the variabihty being greatest at the period when the change in carcase percentage is most rapid (see Fig. 3). The percentage of pluck, on the other hand, steadily decreases in variability from 3 to 9 months ; this Journ. of Agric. Sci. xil 28 420 Growth and Development of Breeds and Crosses of Pigs is contrary to what occurs in sheep where variability in phick increases with Age but it should be remembered that with pigs the pluck includes the caul fat and that this in sheep decreases in variability with age. The variability in the proportion of alimentary canal etc. undergoes the same changes as the carcase. The alimentary canal has the greatest coefficient of variability, the pluck (including caul fat) comes next and the carcase has least. The causes underlying the variation in carcase percentage at any one age are probably mainly due to differences in fatness, early maturity and the amount of food contained in the alimentary canal; other factors however play a part. Sanborn (7i) and Henry (72) found that a ration low in protein affected the proportions of the body, and their results have been confirmed by Enmiett el al. (73). Hart and McCollum(74) have shown that a diet of maize alone reduced the rate of growth in pigs considerabh^ while Jackson and Stewart (75) have shown that rats subjected to periods of deficiency in diet were subnormal as regards skeletal and nmscular development, while the viscera were above normal in weight. ElUot, Crichton and Orr(7n) found that in pigs with rickets the head was abnormally large and Henseler(7()) found that by controlling the nutrition of pigs he could not only affect the hve weight but also the proportions of the body. Correlation. In order to determine the effect of weight, independent of age, on the proportions of the body and also the correlation existing between the proportions of the different parts of the body Table XXI has been prepared. This table shows that, independent of age, a liigh live weight is correlated with a large percentage of carcase and a small percentage of pluck and alimentary canal, etc.; that is within a breed the heavier a pig is for its age the higher will be the proportion of carcase to hve weight. This fact is consistently confirmed by the various groupings made. The remainder of the table shows that the proportions of the carcase vary inversely as the proportions of the pluck and aUmentary canal, this appearing as the result of every grouping. Similar results were obtained experimentally by Henseler(76) who took two pigs from the same litter, one was well fed and the other starved ; both were killed at 9 months old. The starved one weighed 58 lbs. and the well fed one 279 lbs., the former having a carcase percentage of 73 per cent., while the latter killed at 87 per cent. The main cause of the increased carcase percentage with age is the development of muscle and fat rather than great alterations in the size of the other organs of the body. John Hammond 421 The above conclusions seem to ])oint to the fact that a larij,e animal is a more economical producer of meat than a small one of the same age, especially when considered in conjunction with Voit's(77) results which confirmed Rubner's theory that the amount of energy required by a young animal when calculated in relation to unit of body surface is the same as that of an adult, the younger and smaller animals having a larger surface per kilo body weight than adult and larger animals. Table XXI. Correlafioii of live weight, and 'parts of the body in Berkshire -pigs. Grouped by Group Live weight lbs. Carcase 0/ /o Pluck o Intestmes. etc. % Live weiglit High Average Low- — 81-3 80-1 78-4 4-7 4-9 .5-2 14-0 15-0 10-4 °u carcase High Average Low 189-4 lti6-2 1.50-7 — 4-6 4-9 5-3 12-6 14-9 17-5 ".„ phick High Average Low 149-0 lG8-« 187-9 78-4 80-1 81-3 — 16-0 15-0 14-4 % intestines, etc. High Average Low 151-2 11)7 -9 186-.5 771) 80-2 82-() 4-9 4-8 — My thanks are due to Mr E. J. Powell, Secretary of the Smithfield Club, who has very kindly supplied me with records of the Shows, to Mr Charles Bone of the Central Meat Markets, London, who has given me information on the methods of slaughter and dressing adopted, and lastly to Dr F. H. A. Marshall, F.R.S. who has given me the benefit of his advice in the prej)aration of this paper. BIBLIOGRAPHY. (1) P. McCoNNELL. Notebook of Agricultural Facts and Figures, London, 1904. British Breeds of Live Stock, Ministry of Agricultiu-e and Fisheries, London, 1920. (2) Oermaii breeds of Live Stock, Deutsche Landwirts. Gesell. Berlin, 1912. (3) CoRNEViN, quoted from Miiller, Landimrtschaftliche Tierproduktionslehre, Berlin, 1900, 143. (4) LoNa. The Book of the Pig, London, 1906, 325. (5) Wei.lman. Laiidwirt. Jnhrb. \9H, iG. (G) TscHiEWiNSKY. Lnvdu: Vers. Stat. 1883, 29. 317. (7) Hammond, Jourii. Agr. Science, 1920, 10. (8) Ibid. 1921, 11. 422 Groii'tli and Develojmieiit of Bretfh and Cros!